Addiction and 

Choice:
Rethinking the Relationship
Addiction and Choice:
Rethinking
the Relationship
Edited by

Nick Heather
Northumbria University

Gabriel Segal
King’s College London

1
1
Great Clarendon Street, Oxford, OX2 6DP,
United Kingdom
Oxford University Press is a department of the University of Oxford.
It furthers the University’s objective of excellence in research, scholarship,
and education by publishing worldwide. Oxford is a registered trade mark of
Oxford University Press in the UK and in certain other countries
© Oxford University Press 2017
The moral rights of the authors‌have been asserted
First Edition published in 2017
Impression: 1
All rights reserved. No part of this publication may be reproduced, stored in
a retrieval system, or transmitted, in any form or by any means, without the
prior permission in writing of Oxford University Press, or as expressly permitted
by law, by licence or under terms agreed with the appropriate reprographics
rights organization. Enquiries concerning reproduction outside the scope of the
above should be sent to the Rights Department, Oxford University Press, at the
address above
You must not circulate this work in any other form
and you must impose this same condition on any acquirer
Published in the United States of America by Oxford University Press
198 Madison Avenue, New York, NY 10016, United States of America
British Library Cataloguing in Publication Data
Data available
Library of Congress Control Number: 2016943984
ISBN 978–​0–​19–​872722–​4
Printed and bound by
CPI Group (UK) Ltd, Croydon, CR0 4YY
Oxford University Press makes no representation, express or implied, that the
drug dosages in this book are correct. Readers must therefore always check
the product information and clinical procedures with the most up-​to-​date
published product information and data sheets provided by the manufacturers
and the most recent codes of conduct and safety regulations. The authors and
the publishers do not accept responsibility or legal liability for any errors in the
text or for the misuse or misapplication of material in this work. Except where
otherwise stated, drug dosages and recommendations are for the non-​pregnant
adult who is not breast-​feeding
Links to third party websites are provided by Oxford in good faith and
for information only. Oxford disclaims any responsibility for the materials
contained in any third party website referenced in this work.
 Editors’ preface

As its title suggests, the aim of this book is to rethink the relationship between addiction
and choice. All chapters address this aim in one way or another. But how is the relation-
ship between the two phenomena currently conceived? And why does it need rethinking?
A partial answer to the first of these questions is that views have tended to polarize
around two extremes: either typical addictive behavior involves choice in just the same way
as any other or it involves no choice at all. On the one hand, what may be called the pres-
cientific or “moral” view of addiction is that addictive behavior represents a free choice,
similar in kind to all the other autonomous choices people make every day. Although
some historians have discovered portrayals of “habitual drunkenness” involving some
restriction to autonomy dating back to the early seventeenth century (e.g. Warner 1994),
classic work by Harry Levine (1978) suggests that the “free-​choice-​like-​any-​other” view
was the common understanding of what we now call “addiction” until the late eighteenth
century and the emergence of the profound change in thinking about human behavior
known as positivism. If someone excessively consumed alcohol or some other substance
in a way that clearly did them harm, they nevertheless did so voluntarily and, presumably,
in full knowledge of the harm the behavior was causing them. As a consequence, they
should accept responsibility for their behavior and, when appropriate, disapproval and
public sanctions. This view of what we call addiction is, of course, still often found among
members of the general public today.
Surprisingly perhaps, this free-​choice view of addiction has been endorsed relatively
recently in scientific, scholarly, and professional opinion. For example, in an influential
book, John Booth Davies declared addiction to be a myth because “people who use drugs
(including so-​called addicts) do so for their own reasons, on purpose, because they like
it, and because they find no adequate reason for not doing so” (Davies 1997, p. xi, paren-
theses added).1 Similarly, psychologist Jeffrey A. Schaler, while recognizing that changing
one’s addictions may be hard, believes that “All the evidence we have supports the view
that drug addicts are conscious—​yes, even calculating—​responsible persons, in full com-
mand of their behavior” (Schaler 2000, p. 21, italics added). And lawyer David L. Wallace
(2013) argues that, for legal purposes, the conceptualization of addiction as pharmacolog-
ically compelled behavior is “bunkum” and that regular smoking is a free choice. While
the editors and many of the chapter authors of this volume would agree with Wallace
that it is unhelpful to see addictive behavior simply as pharmacologically compelled, and
would also agree with some of the criticisms of the disease concept of addiction made by

1 For a rebuttal of Davies’ arguments, see Heather and Segal (2013a, 2015).

vi Editors’ preface

Davies and by Schaler, it is the idea, as we shall see, that addiction represents completely
free, purely voluntary, and unconstrained choice that is the issue here. It should also be
noted that it is the free-​choice view that is misleadingly presented by some defenders of
the disease view of addiction as the only alternative to their view (e.g. Leshner 1997), thus
contributing to the unhelpful polarization of understandings of addiction of which we
shall shortly complain.
The diametrically opposite view is that addictive behavior, rather than being freely cho-
sen, is completely compelled and involuntary: addicts do not use because they choose to
but because they are compelled to. As implied above, the central problem in the study of
addiction is to explain why people repeatedly behave in ways they know are bad for them.
According to this view, the answer is that addicts experience cravings or urges that cannot
be resisted, no matter what.
An early and vivid illustration of this picture of addiction, quoted by chapter authors in
the current volume (see Holton and Berridge, Chapter 9; Ainslie, Chapter 13), came from
Benjamin Rush, the so-​called father of American psychiatry. Rush reported the words of
a patient who apparently told him: “Were a keg of rum in one corner of a room, and were
a cannon constantly discharging balls between me and it, I could not refrain from passing
before that cannon, in order to get at the rum” (Rush 1812, p. 266).2 Similar descriptions
of uncontrollable behavior by addicts can be found in more recent literature, especially
where addiction is held to be a chronic, relapsing brain disease. For example, Leshner
(2001) writes:
Over time the addict loses substantial control over his or her initially voluntary behavior, and it
becomes compulsive. For many people these behaviors are truly uncontrollable, just like the behav-
ioral expression of any other brain disease. Schizophrenics cannot control their hallucinations and
delusions. Parkinson’s patients cannot control their trembling.

(p. 3)

On this view, the entirely involuntary nature of addictive behavior is used to plead that
addicts should not be held responsible for their troublesome actions and should not be
blamed or punished for them, but rather given medical treatment, as for any other disease.

A plague on both your houses

Recently, there has been an increasing recognition by scientists and scholars interested in
addiction that restricting possible conceptions of it to either of these two extreme posi-
tions is very unhelpful and is, indeed, retarding progress on understanding the nature of
addiction and what could be done about it. Articles and chapters critical of this polariza-
tion are increasing in frequency and come from a range of professional disciplines and
scientific positions (e.g. Berridge and Robinson 2011; Henden, Chapter 7, this volume;

2 The source of this quotation is sometimes given as William James’ (1890) Principles of Psychology (Vol.
II, p. 543). Since the wording is identical, James must have borrowed the quotation from Rush’s (1812)
book, curiously without acknowledgment.
 Editors’ preface vii

Heather and Segal 2013b; Holton and Berridge 2013, Chapter  9, this volume; Hyman
2007; Levy 2011; Lewis, Chapter 10, this volume; Naqvi and Bechara, Chapter 12, this
volume; Segal 2013; Yaffe 2011).
Some uneasiness with this polarized landscape, albeit in different ways and for different
reasons, may be found in most chapters of this book.3 Authors would agree that the truth
lies somewhere between the two extremes described above and that choice, in some form
or another, plays a role in addiction. And most would also agree that when addicts choose
to use, their choice-making is constrained, impaired, or disordered in some way. In this
way, we may view addiction as being, or at least as involving, some kind of disorder of
choice.

But still plenty of room for disagreement

Accepting that addictive behavior is subject to processes of choice, however, still leaves
plenty of room for discussion about the theoretical and practical consequences of this
assumption and raises a number of crucial further questions. Some of these questions,
which we presented to prospective authors when we invited them to contribute to the
book, are as follows:
1 If addiction involves choice, does this make the concept of addiction meaningless?
If not, what meaning can “addiction” have if addicts choose their behavior? In other
words, what kind of concept of addiction is needed if the involvement of choice is
accepted?
2 Accepting that a concept of addiction embracing choice can be identified, does this
mean that it is misleading to describe addiction as a disease? If not, what kind of dis-
ease concept of addiction can be formulated in which the involvement of choice is
allowed?
3 What contribution can neuroscience make to an understanding of addiction once it is
seen as involving choice?
4 Depending on the perceived relationship between addiction and choice, what are
the implications of this for practical matters such as the prevention and treatment of
addictive disorders? How exactly does the conceptualization of choice affect the kinds
of preventive and treatment programs thought likely to be effective?
5 What are the implications of choice for notions of legal responsibility in relation to
addiction? How, if at all, should this affect the legal framework surrounding addictive
behavior?
6 How would acknowledgment of the involvement of choice affect the public under-
standing of addiction? In promoting the idea that addicts in some way choose their

3 An example of an exception to this is Flanagan’s Chapter 4. While he would certainly agree that addic-
tion, as he sees it, entails choice, Flanagan is concerned with those who choose willingly to engage in
addictive behavior and not those who do so unwillingly.
viii Editors’ preface

addictive behavior, how can the danger of a reversion to moralistic and punitive atti-
tudes among the general public be avoided?
We do not claim that these questions exhaust the possibilities for discussion once a role
for choice in addiction is accepted. We do regard them as reflecting some of the more
important issues that arise in this process of rethinking. We have also used these ques-
tions to structure the contents of the book (see below).
One particular demonstration of the point that the choice perspective leaves ample
opportunity for disagreement is that the editors themselves disagree about the broad
answer to the second question above—​whether it is still possible to see addiction as a dis-
ease. One of us (GS) believes that it is and that addiction is best seen as a disease of choice-​
making systems in the mind-​brain; the other (NH) believes that, to see addiction as a
disorder of choice4 summarizes many of the deficiencies of disease models of addiction
and is the key to an improved understanding and response to it. Differences between us
on this score are illustrated in the chapters we have contributed to the book (Chapters 1,
8, 20, 24, and 25). Rather than a limitation of the book, we see it as a strength that it can
encompass two very different takes on addiction within a more general framework of a
choice perspective.

Who is this book for?

The general aim of the book, then, is to explore the relationship between addiction and
choice and its implications for theory, practice, and policy. This aim is achieved by includ-
ing chapters from a range of scientists and scholars who have thought and written about
the relationship in question, have a view on it, and who therefore have answers, however
exploratory and tentative, to some or all of the questions posed above. As befits the range
of views on offer, the disciplinary backgrounds of these authors include philosophy, neu-
roscience, psychology, psychiatry, and law. We shall have more to say about this interdis-
ciplinary mix below.
While the focus of this book will be familiar to some scientists and academics, it is our
impression that it will be less well known to the majority of individuals working with
addictions in front-​line capacities as prevention policy-​makers, treatment providers, or,
indeed, addiction researchers. To many, the notion that addicts can be seen as choosing
their addictive behavior, however exactly that notion is understood, would be difficult at
first to accept, not to say highly controversial. We believe it is essential, if the understand-
ing and amelioration of addictive disorders is to progress, that the main body of workers
in the addictions field are informed of this new general perspective on addiction and
that an attempt is made to remove any misconceptions they may have about the role of
choice in addiction. To this end, we asked authors to ensure that their contributions to

4 The phrase “disorder of choice” is the subtitle of a recent and influential book by Gene Heyman (2009).
While NH regards this book as a very valuable contribution to the field, in his use of the phrase he is
not necessarily committed to everything in Heyman’s understanding of addiction.
 Editors’ preface ix

the book were clearly understandable by the great majority of workers in the addictions
field, as well as by the well-​informed general public. But as well as removing misconcep-
tions about addiction and choice, the book’s major task is to describe and develop ideas
about addiction as a disorder of choice that we believe have important implications for
improved scientific understanding and a more rational response to problems of addiction
in society at large.

Precedents in the literature

While the leading idea of the book is a radical departure from current mainstream public
understandings of addiction, this is by no means the first time that choice has been impli-
cated in addiction in the scientific and scholarly literature. Aside from any nineteenth-​
century precedents that might be mentioned (see, e.g., Valverde 1998), in more modern
times the work of George Ainslie (1975) and Howard Rachlin (1974) laid the founda-
tions for viewing addiction as a disorder of choice, at least in one tradition of theory and
research. While both these seminal articles were based on a behaviorist epistemology
and neither was focused on addiction per se, their discussions of “impulse control” and
“self-​control” led to what became known as the “behavioral economics” perspective on
addiction (see Vuchinich and Heather 2003). The merging of economics and behavio-
ral science immediately introduces the assumption that addicts are “consumers” making
choices among a range of available alternatives. The subsequent prolific writings of these
two pioneers, usually but not always more general in application than to addiction (e.g.
Ainslie 1992, 2000, 2001; Rachlin 1995, 1997, 2000), have had, and continue to have,
strong influences on addiction theory. We are privileged to have chapters from both of
them here (Chapters 13 and 14).
Also part of the behavioral economics perspective on addiction is Hernnstein and
Prelec’s (1992) theory. Their work has been taken forward by Gene Heyman (e.g. Heyman
2003)  and this has expanded to a more general theory of addiction (Heyman 2009), a
version of which, focusing on the implications for the public understanding of addiction,
Heyman and a colleague provide here (Chapter 21).
Another prominent contribution to the literature on addiction from a choice perspec-
tive is the rational addiction theory of Nobel Laureate, Gary Becker and a colleague
(Becker and Murphy 1988). This has proved influential, at least among economists, and
has been developed by others (e.g. Orphanides and Zervos 1995). For present purposes,
it is important to note that rational addiction theory is founded on the basic assumption
of classical microeconomic theory that human actions always seek to maximize utility
and that, therefore, choices made by addicts in respect of their addictions are rational
choices (for an exposition and critical review of rational addiction theory, see Skog 1999).
While any adequate discussion of rationality/​irrationality in human behavior is beyond
the scope of this Preface (but see, e.g., Stanovich 2004, ch. 6), choices attributed to addicts
in this book are almost invariably, in some sense, irrational choices. Becker and Murphy’s
theory and its modifications may be placed in the free-​choice category that represents one
side of the unhelpful dichotomy in views of addiction referred to above. It is also fair to
x Editors’ preface

say that, except possibly among economists, rational addiction theory is no longer found
credible as an explanation of addiction (for an excellent discussion and comparison with
other theories, see Yaffe 2002).
Not so long ago, just before the turn of the current century, one of us (NH) had a
manuscript rejected by a refereed journal on the ground that “weakness of will” was not a
scientifically acceptable term. One significant consequence of some of the work reviewed
here so far, especially the work of George Ainslie (e.g., Ainslie 1992, 2001), is that it has
made concepts like “will-​power,” and “weakness of will” respectable in scientific circles
(Heather 1994)—​or, rather, respectable again, because they were widely used in medical
science in the nineteenth century (Valverde 1998). This has enabled links to be strength-
ened between interest by psychiatrists and psychologists in weakness of will (or, in simi-
lar terms, impulsivity and impaired behavioral control) and philosophers’ time-​honored
preoccupation with akrasia and “incontinence” in human action (see Henden, Chapter 7;
Heather, Chapter 8; Segal, Chapter 24, this volume).
This link, in our view, has already proved invaluable in attempts to make sense of
addiction and promises increasingly to do so. Though concepts such as weakness of will
are always subject to technical definitions in the scientific literature, their use in explana-
tions of addiction gives them a reality and relevance to everyday life that was previously
absent, making accounts of addiction potentially more accessible to addicts themselves
and their relatives and friends. The concept of self-​control or self-​regulation has a longer
and more respectable history in modern psychology (e.g. Carver and Scheier 1998) and
with regard to addiction more specifically (e.g. Heather et al. 1991). More recently, how-
ever, thanks largely to the work of philosophers such as Richard Holton (1999, 2003) and
Alfred Mele (2001, 2012), psychiatrists and psychologists are beginning to see “weakness
of will” and “self-​control” as opposite sides of the same coin and as jointly relevant to a
full explanation of addiction. This tendency could be said to have reached its culmina-
tion in the work of Roy Baumeister and colleagues (e.g. Baumeister 2003; Baumeister
et al. 1994) whose research on “ego depletion” has direct and explicit relevance to weak-
ness of will and self-​control, and to related concepts such as “temptation,” in the theory
of addiction. We were fortunate to be able to persuade Baumeister and his colleagues to
contribute a chapter to this collection (see Chapter 16).
Another key development in the behavioral economics tradition was the publica-
tion of a book called Choice over Time, edited by George Loewenstein and Jon Elster
(1992). While not specifically concerned with addiction, this collection did include
Hernnstein and Prelec’s (1992) theory of addiction and other directly relevant chapters.
Of equal importance, the book provided economic, philosophical, psychological, and
political insights on aspects of human behavior over time—​impulsiveness, selfishness,
self-​command—​with profound implications for understanding addiction. A chapter by
Walter Mischel and colleagues (1992) gave an account of famous experiments conducted
at Stanford University in the 1960s on delay of gratification among preschool children,
work that many feel should be taken into account by any adequate theory of addiction and
recovery (for a more recent account, see Mischel 2014).
 Editors’ preface xi

Loewenstein and Elster’s (1992) book was published by the Russell Sage Foundation,
a long-​established foundation in the USA whose aim is to foster the development and
dissemination of knowledge about political, social, and economic problems. Under the
auspices of the Foundation, a group of scientists and scholars formed a Working Group
on Addiction that met annually from 1992 to 1997. The deliberations of this group cul-
minated in the publication in 1999 of two books edited by Norwegians, Jon Elster and
Ole-​Jørgen Skog (Elster 1999; Elster and Skog 1999). These two collections combined to
make a landmark event in the history of the choice view of addiction and contain many
chapters still widely quoted in today’s literature.
After a gap of about ten years, there has recently been another flurry of books attempting
to advance our understanding of addiction by analyzing and questioning basic concepts
and presenting new hypotheses of how its inherent mystery can be explained. We refer
to the collections brought together and edited by Don Ross and colleagues (2010), Jeffrey
Poland and George Graham (2011), and Neil Levy (2013). These three books address
somewhat different aspects of the topic and each has something unique to offer but they
can all be seen as advancing the radically revised view of addiction that concerns us in the
present volume. Also published recently and relevant to present concerns is a collection of
articles in Frontiers in Psychiatry edited by Hanna Pickard and colleagues (2015).
We do not claim that this brief history of theories of addiction as a disorder of choice is
complete. For one thing, there is work that is more broadly based but which includes the
choice perspective as one of its components—​for example, the role of choice at the begin-
ning of Robert West’s (2006, ch. 3) comprehensive theory of addiction. But we do believe
that most of the crucial developments have been covered and that, without this preceding
effort, the present volume would not have been possible.

An interdisciplinary collaboration
As shown by the three recent books referred to above and by the present volume, cur-
rent work on addiction from the choice perspective is very much an interdisciplinary
effort. With the rapid expansion of neuroscience over the last 20 years or so, collabora-
tions between psychology, psychiatry, and neuroscience have become common but it is
perhaps the addition of philosophy into this mix that is remarkable in more recent years.
Examples of active collaborations between philosophers and neuroscientists include two
chapters in this book (Pickard and Ahmed, Chapter 2; Holton and Berridge, Chapter 9).
Philosophy is continuous with science and can bring to the study of addiction essential
clarification, analysis, and development of basic concepts, as well as relating addiction to
long traditions of thought on human agency and autonomy and its limitations.
One absence from our book is work specifically by social scientists, though social aspects
of the choice perspective on addiction are, of course, mentioned in several chapters (see,
especially, Rachlin, Chapter 14). We see no reason why social scientists cannot and should
not make explicit contributions to the choice perspective and we encourage them here
to do so. The same might be said of political scientists and ethicists. Though they do not
xii Editors’ preface

feature in our book, they have contributed in the past (see, e.g., Collins 2010)  and we
hope they will make a greater contribution in future to discussions of the ethical and
political consequences of seeing addiction as a disorder of choice. More generally, we are
convinced that continued progress in understanding and responding to addiction will
require the combined efforts of people trained originally in different academic disciplines.

Structure and contents of the book

We made an early decision that it would be helpful to divide the chapters of the book
into separate sections, reflecting different areas of scholarship and science involved in the
disorder of choice perspective. However, an illustration of the interdisciplinary nature of
the topic described above is precisely that many chapters could be placed in more than
one section; they cover philosophy as well as neuroscience and/​or psychology, etc. Again,
in any one section, some chapters will have stated implications for treatment, prevention,
and public understanding. Thus the placement of chapters within sections is to some
extent arbitrary and the reader is asked to bear this in mind. We did, at least, obtain the
consent of chapter authors to the section in which their chapter was eventually placed.
Authors were asked to provide an abstract of their chapters, not exceeding 150 words,
plus between five and ten keywords. The latter will be used in the online version of the
book, mainly for online discoverability—​they ensure that when key points or words that
are relevant to the book/​chapter are searched for, the chapter and book are part of the
results. The abstracts also appear in the online version of the book to give a summary
of each chapter before it is opened. However, we decided that it would also be useful to
include abstracts in the printed book to give the busy reader a taste of what the chapter
contains. For this reason we saw it as redundant in this Preface to adopt the conventional
practice of mentioning each chapter and summarizing its contents. A few words on the
thinking behind the selection of individual chapters may nevertheless be of interest.
The introductory section contains only one chapter, addressing the issue of how addic-
tion can be defined. It is there asserted that articles and chapters on addiction often fail
to tell the reader at the outset what the term is thought to mean, thus leading to cross-​
purposes and impeding clear debate. Though we are aware that some authorities dislike
formal definitions because they restrict possibilities for exploration and change, we asked
authors to provide at least some indication at the start of their chapters of what kind of
thing they took addiction to be. In the opening chapter, one of us (NH) takes the oppor-
tunity to offer a tentative definition that he believes is useful in viewing addiction as a dis-
order of choice; the other (GS) discusses further the theoretical implications of different
uses of “addiction” and other key terms in his concluding Chapter 24.
Section II presents seven chapters on the philosophical foundations of the study of
addiction. With one exception, these chapters are written by professional philosophers
trained in the discipline (with in one case the addition of a neuroscientist colleague,
Chapter 2). The exception is Chapter 8 by a clinical psychologist with no formal train-
ing in philosophy (NH) but who has long been fascinated by philosophical analyses of
 Editors’ preface xiii

addiction and has had the chutzpah to offer his own reflections on the relevance of the
traditional philosophical concept of akrasia to understanding addiction.
Section III contains four chapters on the neuroscience of addiction. Two of these are co-​
authored by scientists whose names are associated with prominent neuroscientific theo-
ries of addiction—​Kent Berridge in Chapter 9 (Robinson and Berridge 1993, 2008) and
Rita Goldstein in Chapter 11 (Goldstein and Volkow 2002). Chapter 9 places Robinson
and Berridge’s incentive sensitization theory of addiction within a wider conceptual
framework, while Chapter 11 is concerned with important methodological issues in the
study of choice in addiction. Although other leading neuroscientists in the addictions
field have not contributed chapters, their work is mentioned in various places throughout
(e.g. Robbins and Everitt 1997; Hyman 2005; Koob and le Moal 2005).
In another chapter in Section III, Naqvi and Bechara (Chapter 12) give an account of
their highly original research findings that have attracted a great deal of interest in the
addictions field. The remaining chapter is by Marc Lewis (Chapter 10), a neuroscientist
whose autobiographical account of addiction (Lewis 2011), together with a more recent
book (Lewis 2015), has attracted much attention in the popular and professional media.
Lewis is unusual among neuroscientists in rejecting the idea that addiction is best seen
as a disease. However, it is a significant fact, and an indication of the breaking down of
boundaries between neuroscientific and other approaches to addiction, that all four chap-
ters in this section contain the word “choice” in their titles.
Section IV includes four chapters by psychologists and a psychiatrist (Ainslie). It com-
mences with chapters by the two great pioneers of a behavioral-​economic understand-
ing of addiction mentioned above, George Ainslie (Chapter  13) and Howard Rachlin
(Chapter 14). As also remarked above, these two authors continue to provide fresh and
profound insights into the essential nature of addiction. Another chapter here, by Miles
Cox and colleagues (Chapter 15), can be classified as belonging to a dual-​process account
of addiction (see Wiers et al. 2014), an organizing framework for understanding addic-
tion which we regard as of great significance for future theory and research (see Heather,
Chapter 25). The remaining chapter in this section is on “ego-​depletion theory” (Vonasch
et al., Chapter 16), already mentioned above as of paramount importance for the modern
understanding of addiction.
One advantage we claim for our book compared with other recent volumes (see above)
is that we have paid more attention to the implications of the disorder of choice per-
spective for the prevention and treatment of addictive disorders and for public health. In
Section V, we first asked someone who, rather than seeing addiction from a scientific or
academic stance, is at the coal-​face of helping people with addictive disorders to provide
a personal viewpoint on the choice perspective and how it relates to her day-​to-​day work
(Burgess, Chapter  17). Secondly, we have two chapters exploring the consequences of
behavioral economic theory for practice. Chapter 18 by Murphy and colleagues does so
for treatment or, more precisely, for the delivery of brief interventions (Heather 2003) for
alcohol problems informed by the enormously popular therapeutic approach of moti-
vational interviewing (Miller and Rollnick 2002). Chapter 19 by Tucker and colleagues
xiv Editors’ preface

discusses the role of choice-​based strategies in preventing and ameliorating addictive dis-
orders. Lastly in this section, Gabriel Segal in Chapter 20 articulates and defends a version
of the addiction-​as-​a-​disorder-​of-​choice view that is implicit in the foundational texts of
Alcoholics Anonymous and, based on that, proposes a scientific account of how 12-​step
programs work for the achievement and maintenance of abstinence.
Continuing the general theme of looking at the practical consequences of the disorder
of choice perspective, Section VI reviews implications for public understanding and for
legal responsibility for addictive behavior. In Chapter 21 Heyman and Mims expand on
the implications of Heyman’s (2009) theory of addiction for how addicts themselves and
the general public make sense of addiction. Next, in Chapter 22, Beth Burgess addresses
this issue from the perspective again of an addictions counselor. Last in this section but
by no means least in importance, in Chapter 23 Stephen Morse informs us how choice in
addiction is handled in Anglo-​American criminal law.
Finally, Section VII provides concluding chapters by each of the editors. It would have
been impossible, even if we had wished, to attempt to summarize the contents of the
chapters in this book. Instead, we offer a personal perspective on what seem to us the key
issues that arise and how the science of addiction should be taken forward in the light of
our conclusions.

A note on terminology
Throughout this Preface, in common with many of our chapter authors, we have used the
term “addicts” to refer to those who suffer from some form of addiction, however it might
be defined, to substances or activities. We are well aware that many find this usage objec-
tionable because it labels people rather than labelling the conditions they suffer from. Yes,
“addicts” are individual human beings whose lives cannot be summarized and dismissed
by a single term like “addiction.” Our only excuse is that “people who suffer from addic-
tions” is simply too cumbersome for repeated use in a book of this length. We merely ask
the reader to keep this apology in mind.

References
Ainslie, G. (1975). Specious reward: a behavioral theory of impulsiveness and impulse control.
Psychological Bulletin, 82, 463–​96.
Ainslie, G. (1992). Picoeconomics: The Strategic Interaction of Successive Motivational States within the
Person. Cambridge, UK: Cambridge University Press.
Ainslie, G. (2000). A research-​based theory of addictive motivation. Law and Philosophy, 19, 77–115.
Ainslie, G. (2001). Breakdown of Will. Cambridge, UK: Cambridge University Press.
Baumeister, R.F. (2003). Ego depletion and self-​regulation failure: a resource model of self-​control.
Alcoholism: Clinical and Experimental Research, 27, 1–4.
Baumeister, R.F., Heatherton, T.F., and Tice, D.M. (1994). Losing Control: Why People Fail at Self-​
regulation. San Diego CA: Academic Press.
Becker, G. and Murphy, G. (1988). A theory of rational addiction. Journal of Political Economy, 96,
675–​700.
 Editors’ preface xv

Berridge, K.C and Robinson, T.E. (2011). Drug addiction as incentive sensitization. In: J. Poland and G.
Graham (eds), Addiction and Responsibility. Cambridge, MA: MIT Press, pp. 21–​54.
Carver, C.S. and Scheier, M.F. (1998). On the Self-​Regulation of Behavior. Cambridge, UK: Cambridge
University Press.
Collins, P. (2010). Defining addiction and identifying public interest. In: D. Ross, H. Kincaid, D.
Spurrett, and P. Collins (eds), What is Addiction? Cambridge, MA: MIT Press, pp. 409–​13.
Davies, J.B. (1997). The Myth of Addiction (2nd edition). Reading, UK: Harwood Academic Press.
Elster, J. (ed.) (1999). Addiction: Entries and Exits. New York: Russell Sage Foundation.
Elster, J. and Skog, O.-​J. (eds) (1999). Getting Hooked: Rationality and Addiction. Cambridge,
UK: Cambridge University Press.
Goldstein, R.Z., and Volkow, N.D. (2002). Drug addiction and its underlying neurobiological
basis: neuroimaging evidence for the involvement of the frontal context. American Journal of
Psychiatry, 159, 1642–​52.
Heather, N. (1994). Weakness of will: a suitable topic for scientific study? (Editorial). Addiction
Research, 2, 135–​39.
Heather, N. (2003). Brief interventions. In: N. Heather and T. Stockwell (eds), Essential Handbook of
Treatment and Prevention of Alcohol Problems. Chichester, UK: John Wiley, pp. 117–​38.
Heather, N. and Segal, G. (2013a). Understanding addiction: Donald Davidson and the problem of
akrasia. Addiction Research and Theory, 21, 445–​52.
Heather, N, and Segal, G. (2013b). “A plague on both your houses”: a reply to John Davies (and to some
of his opponents). Addiction Research and Theory, 21, 455–​456.
Heather, N. and Segal, G. (2015). Is addiction a myth? Donald Davidson’s solution to the problem of
akrasia says not. International Journal of Alcohol and Drug Research, 4, 77–​83.
Heather, N., Miller, W.R., and Greeley, J. (eds) (1991). Self-​Control and the Addictive Behaviors. Sydney,
Australia: Maxwell Macmillan.
Herrnstein, R. and Prelec, D. (1992). A theory of addiction. In: G. Loewenstein and J. Elster (eds),
Choice Over Time. New York: Russell Sage Foundation, pp. 331–​60.
Heyman, G.M. (2003). Consumption dependent changes in reward value: a framework for understand-
ing addiction. In: R. Vuchinich and N. Heather (eds), Choice, Behavioral Economics and Addiction.
Oxford: Elsevier, pp. 95–​121.
Heyman, G.M. (2009). Addiction: A Disorder of Choice. Cambridge, MA: Harvard University Press.
Holton, R. (1999). Intention and weakness of will. Journal of Philosophy, 96, 241–​62.
Holton, R. (2003). How is strength of will possible? In: C. Tappolet and S. Stroud (eds), Weakness of Will
and Practical Irrationality. Oxford: Oxford University Press, pp. 39–​67.
Holton, R. and Berridge, K. (2013). Addiction between compulsion and choice. In: N. Levy
(ed.), Addiction and Self-​Control: Perspectives from Philosophy, Psychology and Neuroscience.
Oxford: Oxford University Press, pp. 239–​68.
Hyman, S.E. (2005). Addiction: a disease of learning and memory. American Journal of Psychiatry, 162,
1414–​22.
Hyman, S.E. (2007). The neurobiology of addiction: implications for voluntary control of behavior.
American Journal of Bioethics, 7, 8–​11.
James, W. (1890). Principles of Psychology (Volume II). New York: Holt.
Koob, G.F. and Le Moal, M. (2006). The Neurobiology of Addiction. New York: Academic Press.
Leshner, A.I. (1997). Addiction is a brain disease, and it matters. Science, 278, 45–​57.
Leshner, A.I. (2001). Addiction is a brain disease. Issues in Science and Technology, 17, 3.
Levine, H.G. (1978). The discovery of addiction: changing conceptions of habitual drunkenness in
America. Journal of Studies on Alcohol, 39, 143–​74.
xvi Editors’ preface

Levy, N. (2011). Addiction, responsibility, and ego depletion. In: J. Poland and G. Graham (eds),
Addiction and Responsibility. Cambridge, MA: MIT Press, pp. 89–​112.
Levy, N. (ed.) (2013). Addiction and Self-​Control: Perspectives from Philosophy, Psychology and
Neuroscience. Oxford: Oxford University Press.
Lewis, M. (2011). Memoirs of an Addicted Brain. New York: Public Affairs.
Lewis, M. (2015). The Biology of Desire: Why Addiction Is Not a Disease. New York: Public Affairs.
Loewenstein, G. and Elster, J. (eds) (1992). Choice Over Time. New York: Russell Sage Foundation.
Mele, A.R. (2001). Automomous Agents: From Self-​Control to Autonomy. Oxford: Oxford
University Press.
Mele, A.R. (2012). Backsliding: Understanding Weakness of Will. Oxford: Oxford University Press.
Miller, W. R. and Rollnick, S. (2002). Motivational Interviewing: Preparing People for Change
(2nd edition). New York: Guilford.
Mischel, W. (2014). The Marshmallow Test: Mastering Self-​Control. New York: Little Brown.
Mischel, W., Shoda, Y., and Rodriguez, M. (1992). Delay of gratification in children. In: G. Loewenstein
and J. Elster (eds), Choice Over Time. New York: Russell Sage Foundation, pp. 147–​66.
Orphanides, A. and Zervos, D. (1995). Rational addiction with learning and regret. Journal of Political
Economy, 103, 739–​58.
Pickard, H., Ahmed, S.H., and Foddy, B. (2015). Alternative models of addiction. Frontiers in
Psychiatry, 6. DOI: 10.3389/​fpsyt.2015.00020
Poland, J. and Graham, G. (eds) (2011). Addiction and Responsibility. Cambridge, MA: MIT Press.
Rachlin, H. (1974). Self-​control. Behaviorism, 2, 19–​74.
Rachlin, H. (1995). Self-​control: beyond commitment. Behavioral and Brain Sciences, 18, 109–​59.
Rachlin, H. (1997). Four teleological theories of addiction. Psychonomic Bulletin and Review, 4, 462–​73.
Rachlin, H. (2000). The Science of Self-​Control. Cambridge, MA: Harvard University Press.
Robbins, T.W. and Everitt, B.J. (1997). Drug addiction: bad habits add up. Nature, 398, 367–​70.
Robinson, T.E. and Berridge, K.C. (1993). The neural basis of drug craving: an incentive-​sensitization
theory of addiction. Brain Research Reviews, 18, 247–​91.
Robinson, T.E., and Berridge, K.C. (2008). The incentive sensitization theory of addiction: some cur-
rent issues. Philosophical Transactions of the Royal Society, 363, 3137–​46.
Ross, D., Kincaid, H., Spurrett, D., and Collins, P. (eds) (2010). What Is Addiction? Cambridge,
MA: MIT Press.
Rush, B. (1812). Medical Inquiries and Observations, Upon the Diseases of the Mind.
Philadelphia: Kimber and Richardson.
Schaler, J.A. (2000). Addiction Is a Choice. Chicago: Open Court Publishing.
Segal, G.M.A. (2013). Alcoholism, disease and insanity. Philosophy, Psychiatry and Psychology, 20,
297–​315.
Skog, O.-​J. (1999). Rationality, irrationality and addiction: notes on Becker’s and Murphy’s
theory of addiction. In: J. Elster and O.-​J. Skog (eds), Getting Hooked: Rationality and Addiction.
Cambridge: Cambridge University Press, pp. 173–​207.
Stanovich, K.E. (2004). The Robot’s Rebellion: Finding Meaning in the Age of Darwin.
Chicago: University of Chicago Press.
Valverde, M. (1998). Diseases of the Will: Alcohol and the Dilemmas of Freedom. Cambridge: Cambridge
University Press.
Vuchinich, R.E., and Heather, N. (eds). (2003). Choice, Behavioral Economics and Addiction.
Oxford: Elsevier Science.
 Editors’ preface xvii

Wallace, D.L. (2013). Addiction postulates and legal causation, or who’s in charge, person or brain?
Journal of the American Academy of Psychiatry and the Law, 41, 92–​97.
Warner, J. (1994). “Resolv'd to drink no more”: addiction as a preindustrial concept. Journal of Studies
on Alcohol, 55, 685–​91.
West, R. (2006). Theory of Addiction. Oxford: Blackwell.
Wiers, R.W., Field, M., and Stacy, A.W. (2014). Passion’s slave? Conscious and unconscious pro-
cesses in alcohol and drug abuse. In: K.J. Sher (ed.), Oxford Handbook of Substance Use Disorders.
Oxford: Oxford University Press; doi: 10.1093/oxfordhb/9780199381678.013.009.
Yaffe, G. (2002). Recent work on addiction and responsible agency. Philosophy and Public Affairs, 30,
178–​221.
Yaffe, G. (2011). Lowering the bar for addicts. In: J. Poland and G. Graham (eds), Addiction
and Responsibility. Cambridge, MA: MIT Press, pp. 113–​38.
 Acknowledgments

We wish to express our gratitude to our editors of OUP—​Martin Baum, Charlotte Green,
and Matthias Butler—​for their unfailing helpfulness, patience, and sagacity during the
preparation of this volume. We also thank most sincerely all the authors who agreed to
contribute: we hope they find the book a satisfactory vehicle for their ideas and hard work.
Nick Heather
Newcastle upon Tyne

Gabriel Segal
London

August 30, 2015
 Contents

List of Contributors  xxv

Section I Introduction
1 On defining addiction  3
Nick Heather

Section II  Philosophical foundations

2 How do you know you have a drug problem? The role of knowledge
of negative consequences in explaining drug choice in humans and rats  29
Hanna Pickard and Serge H. Ahmed
3 Addiction: the pleasures and perils of operant behavior  49
Bennett Foddy
4 Willing addicts? Drinkers, dandies, druggies, and other Dionysians  66
Owen Flanagan
5 Failures of rationality and self-​knowledge in addiction  82
Thomas Crowther
6 Normal and addictive desires  99
Patrick Butlin and David Papineau
7 Addiction, compulsion, and weakness of the will:
a dual-​process perspective  116
Edmund Henden
8 Addiction as a form of akrasia  133
Nick Heather

Section III  Perspectives from neuroscience

9 Compulsion and choice in addiction  153
Richard Holton and Kent Berridge
10 Choice in addiction: a neural tug of war between impulse and insight  171
Marc D. Lewis
11 Assessing drug choice in human addiction: costs, benefits, and findings
from current research paradigms  186
Scott J. Moeller and Rita Z. Goldstein
xxii Contents

12 The role of the insula in goal-​directed drug seeking and choice

in addiction  205
Nasir H. Naqvi and Antoine Bechara

Section IV  Perspectives from behavioral economics

and cognitive psychology
13 Palpating the elephant: current theories of addiction in light of hyperbolic
delay discounting  227
George Ainslie
14 Addiction as social choice  245
Howard Rachlin
15 Nonconscious motivational influences on cognitive processes
in addictive behaviors  259
W. Miles Cox, Eric Klinger, and Javad S. Fadardi
16 Self-​regulation, controlled processes, and the treatment of addiction  286
Andrew J. Vonasch, Heather M. Maranges, and Roy F. Baumeister

Section V  Implications for treatment, prevention,

and public health
17 The blindfold of addiction  307
Beth Burgess
18 Behavioral economics as a framework for brief motivational interventions
to reduce addictive behaviors  325
James G. Murphy, Ashley A. Dennhardt, and Ali M. Yurasek
19 Role of choice biases and choice architecture in behavioral economic
strategies to reduce addictive behaviors  346
Jalie A. Tucker, Susan D. Chandler, and JeeWon Cheong
20 How an addict’s power of choice is lost and can be regained  365
Gabriel Segal

Section VI  Implications for the public understanding

of addiction and for legal responsibility for addictive behavior
21 What addicts can teach us about addiction: a natural
history approach  385
Gene M. Heyman and Verna Mims
22 How a stigmatic structure enslaves addicts  409
Beth Burgess
23 Addiction, choice, and criminal law  426
Stephen J. Morse
 Contents xxiii

Section VII Conclusions
24 Ambiguous terms and false dichotomies  449
Gabriel Segal
25 Overview of addiction as a disorder of choice and future prospects  463
Nick Heather

Index  483
 List of Contributors

Serge H. Ahmed Javad S. Fadardi

Université Victor Segalen, Ferdowsi University of Mashhad,
Bordeaux, France Iran; and Bangor University, UK
George Ainslie Owen Flanagan
University of Cape Town, South Africa; Duke University, USA
Veterans Affairs Medical Center, Bennett Foddy
Coatesville, PA, USA New York University, USA
Roy F. Baumeister Rita Z. Goldstein
Florida State University, USA; Icahn School of Medicine
and King Abdulaziz University, at Mount Sinai, USA
Saudi Arabia
Nick Heather
Antoine Bechara Northumbria University, UK
University of Southern California, USA
Edmund Henden
Kent Berridge Oslo and Akershus University College
University of Michigan, USA of Applied Sciences, Norway
Beth Burgess Gene M. Heyman
Smyls Therapy, London, UK Boston College, USA
Patrick Butlin Richard Holton
University of Antwerp, Belgium; and University of Cambridge, UK
Hertford College, Oxford, UK
Eric Klinger
Susan D. Chandler University of Minnesota, Morris, USA
University of Florida, USA
Marc D. Lewis
JeeWon Cheong Radboud University, Netherlands
University of Florida, USA
Heather M. Maranges
W. Miles Cox Florida State University, USA
Bangor University, UK
Verna Mims
Thomas Crowther Harvard Extension School, USA
University of Warwick, UK Scott J. Moeller
Ashley A. Dennhardt Icahn School of Medicine
University of Memphis, USA at Mount Sinai, USA
xxvi List of Contributors

Stephen J. Morse Howard Rachlin

University of Pennsylvania, USA
James G. Murphy
University of Memphis, USA
Nasir H. Naqvi
Columbia University and New York State
Psychiatric Institute, USA
David Papineau
King’s College London, UK;
and Graduate Center,
City University of New York, USA
Hanna Pickard
University of Birmingham, UK
Stony Brook University, USA
Gabriel Segal
King’s College London, UK
Jalie A. Tucker
University of Florida, USA
Andrew J. Vonasch
Florida State University, USA
Ali M. Yurasek,
University of Memphis, USA
  1

Section I

Introduction
2
  3

Chapter 1

On defining addiction
Nick Heather

Abstract
The chapter discusses issues that arise in the attempt to define addiction
and, in doing so, reviews major positions in the literature on the essential
nature of addiction and how it shows itself. A specific but tentative
recommendation as to how it should most usefully be defined is developed:
“a person is addicted to a specified behavior if they have demonstrated
repeated and continuing failures to refrain from or radically reduce
the behavior despite prior resolutions to do so or if they would have
demonstrated such failures under different personal or environmental
circumstances.” This definition can be applied as a ‘litmus test” to
distinguish between activities that can reasonably be called addictions and
those that cannot. It is also argued to be a foundation for an emerging
theory of addictive behavior as a disorder of choice based on a dual process
account of human behavior.

1  What do we mean by addiction?

When reading literature on addiction, both in the popular media and in scientific publi-
cations, it is surprising how seldom authors actually tell us what they mean by the term,
whether by a formal definition, a rough characterization of how they see addiction, or
what they consider to be the “hallmark” (Skog 1999, p. 173) of addictive behavior and
experience. It is as though authors simply assume that, when speaking of addiction, eve-
rybody will know what they mean. This might have been understandable in the early days
of addiction studies in the 1950s or 1960s, when the word conjured up images in most
people’s minds of a dishevelled individual injecting heroin into a vein or desperately seek-
ing the means to do so. Most people would agree that such stereotypes should have been
abandoned long ago but there nevertheless remains a tendency for writers on addiction
to assume, for whatever reason, that the term needs no clarification.
Equally unhelpful is the practice of using different terms during the course of an
article or chapter to denote what the piece is about, but without trying to distinguish
between them. Thus an article could begin by discussing addiction but then describe the
results of an apparently relevant experiment in which “active substance abusers” were
compared with those who were not. Also disconcerting is the conflation still sometimes
4  

4 On defining addiction

encountered between, say, heroin users and heroin addicts, as if there were no mean-
ingful difference between use and addiction. Another source of confusion is the use of
“dependence” and “addiction,” as if they were synonymous terms but without explicitly
saying so. Following the decision by a WHO Expert Committee (1964) to repudiate
the distinction made in earlier WHO Reports between “addiction” and “habituation”
and replace them both by “drug dependence” on a particular drug type, dependence or
“dependency” have acquired a variety of meanings (see, e.g., Russell 1976; Fraser and
Gordon 1994). More recently, addiction has come back into fashion, notably in DSM-​5
(O’Brien 2011), but it is not always clear how it is intended to differ, if at all, from earlier
uses of “dependence.”
The truth is, of course, that there is a multitude of ways in which addiction could be
defined, depending on which aspect or aspects of this multifaceted phenomenon one
chooses to emphasize. It is astonishing now to recall that there was a time, not so long ago,
when use of cocaine (Van Dyke and Byck 1982) and of nicotine (Warburton 1989) was said
not to be potentially addictive, mainly because regular use did not conform to the primacy
of tolerance, withdrawal, and negative reinforcement in the opiate model of addiction
(Wikler 1965). Differences in the ways addiction can apparently be manifested are shown
by evidence on benzodiazepines (Licata and Rowlett 2008), marijuana/​cannabis (Budney
et al. 2007; Hall 2014), inhalants (Sakai et al. 2004), caffeine (Juliano and Griffiths 2004),
anabolic steroids (Midgley et al. 1999), sugar (Avena et al. 2008), chocolate (Collingwood
2006), water (Edelstein 1973), carrots (Kaplan 1996) and any other substance purportedly
showing unique features of addiction or dependence. Thus one reason for the diversity of
portrayals of addiction in the literature is that, despite common aspects, it seems to show
itself in different ways depending on the substance in question. Assuming we need a defini-
tion in order to clarify thinking about what is common to all the substance-​related behav-
iors that might be called addictive, we clearly need some kind of superordinate construct
that embraces different manifestations across different substances.

1.1  Behavioral addictions

If the task of defining addiction is made difficult by the different forms it takes for
different substances, it is made considerably harder as soon as we include behavioral
addictions (non-​substance or “process” addictions) in the mix. The concept of behavio-
ral addiction has been made officially scientifically respectable by the inclusion of gam-
bling disorder in DSM-​5, transferred from its place in “Impulse Control Disorders Not
Elsewhere Classified” in DSM-​IV, where it was called “pathological gambling.” A sign
of possible things to come is the inclusion of “Internet gaming disorder” in Section 3 of
DSM-​5, which is designed to stimulate further research (see Petry et al. 2014).
The claim that so-​called behavioral addictions should be considered genuine addic-
tions has often been made by attempting to show that their “symptoms” resemble those
of substance addictions, in terms of tolerance, withdrawal, salience, loss of control, and
other features of dependence syndromes listed in versions of the DSM or ICD diagnos-
tic systems (Potenza 2006; Griffiths 2011). Methods of measuring degrees of behavioral
  5

WHAT DO WE MEAN BY ADDICTION? 5

addictions are typically based on DSM criteria (e.g., Andreasson et al. 2012). This begs
the question whether the diagnostic classifications themselves are valid ways of defining
addiction (see below). Rather than rely on apparent similarities in the forms they take,
which may be superficial, it would be better to argue from first principles that putative
behavioral and substance addictions share a common property that suggests they belong
to the same general phenomenon.
Beyond gambling and gaming disorders, the behavioral addictions that have received
most attention in the literature so far, and might be thought good candidates for inclusion
in DSM in future, are various kinds of addiction to the Internet (Weinstein et al. 2014;
Griffiths et al. 2014), a range of addictions to food (Yau et al. 2014), various sexual activi-
ties (Rosenberg et al. 2014), love (Peele and Brodie 1975; Fisher 2014), shopping (Racine
et al. 2014), and exercise (Berczik et al. 2014). Also frequently mentioned in professional
and popular media are addiction to work (Griffiths 2011), smartphones (Pearson and
Hussain 2015), “kleptomania” (Grant et al. 2010), joy-​riding (Kellett and Gross 2006), and
pornography (Hilton and Watts 2011). But that by no means exhausts the list of excessive
activities that have been proposed as addictions. Among the more unusual are psychic
hotlines (Shepherd 2009), indoor tanning (Banerjee et al. 2015), street life and gang activ-
ity (Bergen-​Cico et al. 2014), “binge flying” (Cohen et al. 2011), approval (Dwoskin 2015),
muscle dysmorphia (Foster et  al. 2014), and Harry Potter books (Rudski et  al. 2009).
To show that not all such claims are necessarily frivolous, let us take an example from a
prestigious scientific journal (Cell: Impact Factor >33). Fell and colleagues (2014) review
evidence from experiments with rodents that UV light-​seeking behavior is a “primor-
dial addiction” mediated by “the hedonic action of β-​endorphin and anhedonic effects of
withdrawal” (p. 1527), thus contributing to the relentless rise of skin cancer incidence in
humans. At another extreme and extending the concept more widely,
Addiction in the modern world can be best understood as a compulsive lifestyle that people adopt
as a desperate substitute when they are dislocated from the myriad intimate ties between people
and groups—​from the family to the spiritual community—​that are essential for every person in
every type of society.
(Alexander 2001, p. 1)

Assuming we do not dismiss them out of hand, how are we to make sense of this prolif-
eration of claims for all kinds of addiction? Where do we draw the line between those
activities that are sensibly and usefully thought of as addictions and those that are not? We
clearly need some kind of litmus test to apply to these claims. I suggest such a test in this
chapter.
The aims of the chapter, then, are first to discuss some of the issues that arise in the
attempt to define addiction and secondly, in doing so, to review some of the major posi-
tions in the literature on the essential nature of addiction and the different ways in which
it shows itself, all of which have implications for how addiction could be defined. I shall
also, throughout the chapter, develop a specific but tentative recommendation as to how
addiction should most usefully be defined, beginning with an earlier attempt to do so
6  

6 On defining addiction

and taking into account actual and possible objections and resulting modifications to the
definition along the way.

2  What kind of definition is needed?

Given that, to make progress in understanding addiction, some kind of definition is
required, what kind of definition is this? Obviously, an adequate definition must specify
some common property of the phenomena that are to be included in the definition and
that distinguishes them from those phenomena that are excluded from it. Beyond that,
however, definitions can be of different kinds for different purposes.
One classification describes five kinds of definition with different functions: stipulative;
lexical; precising; theoretical; and persuasive (Britannica 2015; see also Sinnott-​Armsrong
and Pickard 2013). Of these, the one that comes closest to the kind needed here is the
theoretical definition. To propose a theoretical definition is to suggest a way of think-
ing about the thing defined that has theoretical consequences. In the present case, the
underlying theory is not a well-​established theory of addiction but one, it is argued here
and in other chapters in this book, that is presently emerging, viz, the theory of addiction
as a disorder of choice or decision-​making. Thus, in addition to its job of sorting out the
wheat from the chaff of alleged substance and behavioral addictions, the main motiva-
tion for advancing the definition of addiction that will be proposed in this chapter is that
it is argued to be the most useful for assisting the development of a scientific account of
addiction as a disorder of choice. It is an answer to the question, from the perspective of
such a theory, what is the property of so-​called addictive behavior that is in most need of
scientific enquiry and to the explanation of which scientific endeavors would most profit-
ably be directed. Where, among the many questions it is confronted with, should research
on addiction as a disorder of choice focus its attention?

3  Common usage and etymology

Although some would put it earlier (Nicholls 2009), the origins of the modern addic-
tion concept are usually placed in the late eighteenth and early nineteenth centuries, as
described by Harry Levine (1978) in his classic paper on “The discovery of addiction.”
This concept framed alcohol and, later, other addictions as “diseases of the will” (Valverde
1998). However, though the concept was widely understood, there was at first no specific
term for it and phrases like “a burning withering desire for drink” that was “overwhelm-
ing,” “overpowering,” and “irresistible” were used informally to define the condition
(Room et al. 2015). Such language is, of course, still used today to flesh out the disease
concept of addiction.
It was around the 1880s that “addiction” began to be used as short-​hand for this
nineteenth-​century disease concept, although a range of other terms, such as inebriety,
narcomania, and, particularly with reference to alcohol, dipsomania and alcoholism,
were also commonly employed in both professional and lay circles (Room et al. 2015).
Despite this terminological competition, for much of the twentieth century “addiction”
was common parlance for disorders specifically related to the use of psychoactive drugs.
  7

Addiction in the DSM and ICD 7

According to Room (2003), in modern society the addiction concept offers “a secular
equivalent for possession as an explanation of how a good person can behave badly, and
as an inner demon over which a hero can triumph” (p. 221).
However, “addiction” as a word in common usage existed long before the modern
concept but without the explanatory power Room discerns and without explicit ref-
erence to psychoactive drugs. Before the industrial revolution the word was used in
a way that is roughly equivalent to modern meanings of “committed,” “devoted,” or
“attached” (Room et al. 2015). An eighteenth-​century dandy, for example, was said to
be addicted to fashion. Nor were meanings always negative: some people were addicted
to vice but others to virtue and to God (Nicholls 2009, p. 64). A deeper but crucial dif-
ference between older and newer usage is that in the former “to addict” was a reflexive
verb, something that one did to oneself; although some loss of freedom was implied, it
was a willing renunciation of that freedom (Nicholls 2009, p. 64). In the newer version
there has been a shift from reflexive to passive verb and also from verb to noun. We now
speak of someone becoming addicted and as being an addict in a sense that our forebears
may not have understood.
Yet more recently, of course, over the last 20 or 30 years the addiction concept and
term have expanded, at least in English, to cover a very wide range of habitual behaviors.
Whether this is a consequence or a cause of the emergence of behavioral addictions as
topics of concern in the scientific and medical literature is unclear. The expansion has
also been viewed “as part of larger sociological trends of globalisation and the emer-
gence and diffusion of individualization, ‘risk society’ and new media formats across
societies and languages” (Room et  al. 2015, p.  11). Increasingly, one hears the word
applied to almost any activity that seems to be carried out in excess of conventional
norms. Groups of people, whole sectors of society, and even nation states are said to
be addicted in various ways, as when a financial expert might claim that “the UK is
addicted to credit cards.”
Going back to the origins of the word, addiction derives from the Latin verb addict (ad-​
“to, towards, at” + dīcō “say; declare”) and in Roman law meant “To deliver or hand over
formally (a person or thing) in accordance with a judicial decision” (English Language and
Usage Stack Exchange 2015). An addictio was thus a person who was enslaved through a
judicial procedure. This could happen when it was proved that a debtor lacked the means
to repay a debt, the justice being empowered to make the debtor a slave, thus “addicting”
the debtor to his or her creditor.
It is significant for present purposes that the term “addiction” derives from a Latin word
applied to slavery. How can we be sure that someone is genuinely enslaved? When they try
to break free from their bondage but are demonstrably unable to do so.

4  Addiction in the DSM and ICD

A straightforward solution to the general problem of definition that concerns us here is to
equate addiction or dependence (regarding them for the moment as synonymous) with
criteria laid out in the two major diagnostic classification systems for mental and behav-
ioral disorders in worldwide use today, the Diagnostic and Statistical Manual of Mental
8  

8 On defining addiction

Disorders (DSM), produced by the American Psychiatric Association but used routinely
in many countries outside the USA, and the International Classification of Diseases (ICD),
produced by the World Health Organisation (WHO). The DSM has been published in
successive versions since 1952 (Frances 2013)  and the ICD, after the formation of the
WHO, since 1948 (Room 1998). This solution would be a stipulative definition; for exam-
ple, researchers would say that, for the purposes of their study, addiction is defined as a
diagnosis of dependence according to DSM-​IV criteria. I shall focus on the DSM here but
similar considerations apply to the ICD.
In the DSM’s latest reincarnation, DSM-​5 (American Psychiatric Association 2013),
the main difference from earlier versions is the decision to combine items for depend-
ence and abuse, which had formed separate dimensions in DSM-​IV, in a single diagnosis,
substance use disorder. As mentioned above, another important change from previous
versions was the inclusion for the first time of a behavioral addiction, gambling disorder,
with the possibility of more to come. In DSM-​5, each specific substance is addressed as a
separate use disorder (e.g., alcohol use disorder, stimulant use disorder, etc.), but nearly
all substances are diagnosed based on the same overarching criteria. The 11 criteria used
for the diagnosis of alcohol use disorder are shown in Box 1.1.
For definitional purposes, an advantage of using DSM criteria is that it enables direct
comparisons to be made with other studies using the same criteria and with general pop-
ulation norms where they exist. Findings from clinical samples may be compared with
those from general population samples to see what differences appear. Such a definitional
tactic has been profitably employed, for example, in influential analyses of large-​scale
survey research by Gene Heyman (2009; 2013; see also Chapter  21, this volume). The
diagnostic items in DSM-​5 are based on decades of experience and careful observation
by skilled clinicians, backed up by extensive data from survey and other research, and
sophisticated psychometric analysis. They do, however, have major disadvantages as a
basis for definition, especially for the purpose of scientific enquiry into the fundamental
nature of addiction. Among many other criticisms, the DSM has been accused of: lacking
cross-​cultural applicability (Room 2006; Caetano 2011); confusing core features of addic-
tion by which it should be defined and the negative consequences arising from it which
do not belong in a definition (Babor 2011; Martin et al. 2014); “blind empiricism” in rely-
ing in DSM-​5 on statistical clustering techniques based on dubious measurement (Babor
2011); having no clear relationship to the need or demand for treatment (Room 2011);
and, in common with the ICD, leading to confusion among practitioners and other inter-
ested parties due to the “shifting kaleidoscope” of successive changes in terminology and
definitions (Room 1998). In more general terms, it has been alleged that the diagnoses in
the DSM do not represent real entities that exist in nature but arbitrary and temporary
categories of aspects of behavior and experience strongly influenced by the reimburse-
ment policies of insurance providers and the commercial interests of “Big Pharma”
(e.g., Kutchins and Kirk 1997).
More practically, given that the threshold for a diagnosis of substance use disorder in
DSM-​5 is the endorsement of two or more items from the list in Box 1.1 and ignoring the
  9

Addiction in the DSM and ICD 9

Box 1.1  DSM-​5 diagnostic criteria for alcohol use disorder

A problematic pattern of alcohol use leading to clinically significant impairment or
distress, as manifested by at least two of the following, occurring within a 12-​month
period:
1. Alcohol is often taken in larger amounts or over a longer period than was intended.
2. There is a persistent desire or unsuccessful attempts to cut down or control
alcohol use.
3. A great deal of time is spent in activities necessary to obtain alcohol, use alcohol,
or recover from its effects.
4. Craving, or a strong desire or urge to use alcohol.
5. Recurrent alcohol use resulting in a failure to fulfill major role obligations at work,
school, or home.
6. Continued alcohol use despite having persistent or recurrent social or interper-
sonal problems caused or exacerbated by the effects of alcohol.
7. Important social, occupational, or recreational activities are given up or reduced
because of alcohol use.
8. Recurrent alcohol use in situations in which it is physically hazardous.
9. Alcohol use is continued despite knowledge of a persistent or recurrent physical or
psychological problem that is likely to have been caused or exacerbated by alcohol.
10. Tolerance, as defined by either of the following:
a. A need for markedly increased amounts of alcohol to achieve intoxication or
desired effect.
b. A  markedly diminished effect with continued use of the same amount of
alcohol.
11. Withdrawal, as manifested by either of the following:
a. The characteristic alcohol withdrawal syndrome for alcohol (refer to criteria
A and B of the criteria set for alcohol withdrawal, pp. 499–​500).
b. Alcohol (or a closely related substance, such as a benzodiazepine) is taken to
relieve or avoid withdrawal symptoms.

Specify if:
In early remission: After full criteria for alcohol use disorder were previously met,
none of the criteria for alcohol use disorder have been met for at least 3 months but for
less than 12 months (with the exception that Criterion A4, “Craving, or a strong desire
or urge to use alcohol” may be met).
10  

10 On defining addiction

In sustained remission: After full criteria for alcohol use disorder were previously
met, none of the criteria for alcohol use disorder have been met at any time during a
period of 12 months or longer (with the exception that Criterion A4, “Craving, or a
strong desire or urge to use alcohol” may be met).

Specify if:
In a controlled environment: This additional specifier is used if the individual is in an
environment where access to alcohol is restricted.
Reprinted with permission from the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition,
(Copyright ©2013). American Psychiatric Association. All Rights Reserved.

distinction between Moderate and Severe, there are altogether 2036 possible combina-
tions of items that would qualify for a diagnosis (combinations of 2 from 11 + 3 from
11 + … 10 from 11 + 11 from 11). Thus the majority of individuals diagnosed as suffering
from substance use disorder will differ in the “symptoms” of the disorder they show and
some will have very little in common with each other. Indeed, especially in the Moderate
category, it would be quite possible for two individuals to receive the diagnosis without
having any symptoms whatever in common, hardly a satisfactory basis on which to form
a definition or for making progress in the science of addiction (Sinnott-​Armstrong and
Pickard 2013). Prospects for a useful definition would be greatly improved if just one of
the 11 criteria in Box 1.1 were given a special status for definitional purposes—​a kind of
primus inter pares. I suggest a candidate for this role in section 5.

5  A proposed definition of addiction

The criteria shown in Box 1.1 are all presumably features of what is popularly regarded as
addiction and can be seen as representing a current clinical consensus on the full range of
those features in the various forms addiction takes. The flexibility of the rule for making a
diagnosis of addiction (i.e., the presence of any two criteria from 11 in the past 12 months)
allows for the fact that addiction to different substances in different people in different
circumstances can be accommodated within the diagnostic framework. Outside the con-
fines of DSM-​5, it is safe to say that all, or nearly all, the individual criteria in Box 1.1 have
featured in formal definitions of addiction provided by respected authorities at one time
or another; on the other hand, nearly all could be criticized as definitional requirements
on one ground or another. But which of them could be picked as being common to all
forms of addiction, including behavioral addictions, and reasonably put forward as repre-
senting in some sense the essence or “hallmark” of addiction? The most likely candidates
are criteria 1 and 2 which both imply some kind of impaired control over the behavior in
question; impaired control or, originally, loss of control was the central, organizing idea in
descriptions of the first addiction concept, to alcohol, in the late eighteenth century, and
was then transferred to other psychoactive drug addictions and, more recently, behavioral
addictions (Heather 1991).
  11

A proposed definition of addiction 11

Some years ago I  advanced a definition of addiction as “a repeated failure to refrain

from drug use despite prior resolutions to do so” (Heather 1998). It is that definition,
with the extension of “drug use” to “a specified behavior,” that I now wish to defend or
modify where necessary. In order to accommodate the aim of controlled, moderate,
harm-​free, or less harmful use following addiction (Jaffe 2013), it would also be better
to replace “to refrain from …” by “to refrain from or radically reduce …” In addition, a
logical error is that the definition as it stands does not allow the possibility of success-
fully quitting drug use and recovery from addiction. Thus, after these adjustments, the
definition becomes:  “addiction is a repeated and continuing failure to refrain from or
radically reduce a specified behavior despite prior resolutions to do so.” This is similar to
criterion 2 in the DSM-​5 list in Box 1.1 except for the omission of a “persistent desire”
to cut down or control substance use. So why omit persistent desire from the definition?
It is conceivable that someone would persistently desire to radically change a behavior
without actually attempting to do so but, if so, the implication is that this person would
have very little confidence in her ability to change; she would be a type of sub-​clinical
addict described below (see section 5.3, “The counterfactual objection”, and Figure 1.2).
However, if evidence for the persistent desire is based only on the self-​report of the indi-
vidual concerned, it might reflect the functional advantages attached to the addiction
status described and studied by Davies and Baker (1987); on the other hand, actual failed
attempts to change behavior are, in principle at least, objective and verifiable. (Further, a
case of addiction could be described under certain assumptions that does not involve a
persistent desire to cut down or control use—​see section 5.3.)
Another objection might be that the definition makes no reference to harm. A frequently
encountered definition of addiction is the continuation of the activity despite the harm it is
causing, similar to criterion 6 in Box 1.1, or with the addition of an awareness of the harm
on the person’s part, criterion 9. Alternatively, a plausible definition of addiction would
be that it entails an awareness of harm caused by the activity together with unsuccessful
attempts to stop or reduce it, thus combining DSM-​5 criteria 2 and 9 in Box 1.1. So why
omit mention of harm? First, continuation despite an awareness of harm is arguably not
a sufficient condition for the attribution of addiction (see section 7, “A liberal account of
addiction”). Secondly, repeated attempts to stop or cut down must surely imply a recogni-
tion of harm. Why else would one repeatedly try to stop or cut down an activity except that
it results in some form of recognized harm? The reference to harm is therefore redundant.

5.1  The essence of addiction: a relapsing condition

It is commonly accepted that addiction is “a relapsing condition,” i.e., that it is relatively
easy to make an initial, radical change in addictive behavior—​quitting or drastically cut-
ting down smoking, illicit drug use, alcohol consumption, gambling, etc.—​but much
more difficult to maintain that change over time without reverting to self-​destructive
habits (e.g., Marlatt and Donovan 2007). This applies equally to individuals in treatment
and to those in the community who try to solve their problem without professional help;
in both cases many attempts at change may be needed before a lasting behavioral solution
12  

12 On defining addiction

is found (Prochaska et al. 1992), while some never find it before irreversible damage has
been done. This can be seen, I suggest, as the essence1 of addiction—​the struggle to change
a way of behaving that, implicitly, one knows to be harmful but cannot easily shake off.
From this viewpoint, the task for a theory of addiction is to explain why people continue
to behave in ways they recognize are bad for them and, more especially, why, after they
have resolved time and again to change their addictive behavior, they repeatedly fail to do
so. This also echoes the sense of slavery that seems to have been present in the use of the
word “addiction” since its earliest meanings.

5.2  A theoretical definition

The theoretical advantages of defining addiction in the way suggested can be illustrated by
looking at Figure 1.1, taken from Heather (1998). This depicts an explanatory framework
representing three groups of questions that a satisfactory theory of addiction must answer.
At the “lowest” level are questions concerning the way in which the brain and nervous sys-
tem adapt to repeated ingestion of certain drugs or to repeated enactments of certain behav-
iors. (For convenience, I will henceforth in this section speak of “drugs” and “drug users,”
etc. on the understanding that this is meant to include, where relevant, those engaged in so-​
called behavioral addictions.) This is the level of neuroadaptation or physical dependence
(Edwards et al. 1981). Also included here are pre-​existent differences in the neural response
to drugs, most probably due to genetic mechanisms, that require explanation at this level.
At the middle level in the framework are issues related to reports by users of abnormal
desires to use and feel the effects of psychoactive drugs. The more general term “desire”
is used in Figure 1.1 rather than “craving” because it has been suggested that the latter
should be reserved for reports of intense longing and not for “all dispositions … to use
the addictive substance” (Kozlowski and Wilkinson 1987, p. 31). Nevertheless, the desire
in question should be considered stronger than an ordinary, everyday inclination to do
something. Note that, whether or not it is necessary, a high degree of neuroadaptation
is not sufficient for drug desires to occur. The simplest illustration of this is the well-​
known phenomenon of hospital patients who have received frequent injections of opioids
to relieve pain without reporting any desire to continue using these drugs on discharge
from hospital (Lindesmith 1968; Højsted and Sjøgren 2007). More generally, there would
now be wide agreement among students of addiction that whether or not desire for drugs
is experienced and how intensely depends very much on set and setting (Zinberg 1984).
The third and “highest” explanatory level refers to the central problem in addiction
studies that, because of an awareness of the harm being caused, the habitual drug user
persistently resolves to give up or reduce drug use but repeatedly breaks this resolve and
returns to it.2 Figure 1.1 employs the Greek word “akrasia” to refer to this problem because
it is the term Aristotle used to describe a more general puzzle in human behavior which

1 More precisely, the “nominal essence” in John Locke’s sense (see Stanford Encyclopedia of Philosophy
2015). Thanks to Gabriel Segal for pointing this out.
2 This might seem to conflict with data presented by Gene Heyman regarding high rates of natural recov-
ery and “maturing out” in the common addictions (see Heyman 2009, 2013 and Heyman and Mims,
  13

A proposed definition of addiction 13

Other Determinants Other Determinants

“AKRASIA”

NOT
Other Determinants SUFFICIENT Other Determinants

DESIRE FOR DRUG OR ACTIVITY

NOT SUFFICIENT

NEUROADAPTION OR
PRE-EXISTENT
DIFFERENCES IN NEURAL
RESPONSE

Figure 1.1  Three levels of explanation for an adequate theory of addiction

Reproduced with permission from Heather, N. (1998). A conceptual framework for explaining drug addiction. Journal
of Psychopharmacology, 12, 3–7. Reproduced with permission of Sage. © 1998.

modern philosophers (e.g., Davidson 1980)  have simply called “weakness of will.” The
topic of akrasia will be returned to in this and in a subsequent chapter (see Chapter 8).
It is this “highest” level that is crucial to any satisfactory explanation of addiction and the
level that the lower levels of explanation subserve. No adequate theory of addiction can rest
at the lower levels without also attempting to explain why it is that drug users continue to
break resolutions not to return to using drugs. Moreover, an adequate theory of addiction
must explain, in addition to why addicts keep returning to drug use, why they sometimes
succeed, either temporarily or permanently, in refraining from it. In other words, if we had

Chapter 21, this volume). However, this difficulty is more apparent than real. Even though rates of natu-
ral recovery are much higher than would be expected from a “chronic brain disease” perspective and
though many addicts do mature out at relatively early ages, they may still experience great difficulty and
a number of failures before doing so. With regard to smoking, for example, Borland et al. (2012) found
that about 40% of smokers in four English-​speaking countries reported making a quit attempt each year
and over two attempts on average. Their data suggest that, by the age of 40, the average smoker may have
made over 40 quit attempts (see Borland 2014, p. 41). It is part of the folk wisdom on addiction that,
while many eventually succeed, smoking and other addictive behaviors are very difficult habits to break.
14  

14 On defining addiction

an adequate theory of failure of self-​control with regard to addictive behavior, we would

probably also have an explanation for how self-​control sometimes succeeds.
As with the relationship between the two lower levels, a high degree of drug desire may
or may not be necessary for akrasia to occur, but is certainly not sufficient for it to occur.
This is proved, as implied above, by the simple fact that people sometimes do succeed in
permanently abandoning repetitive drug use, with or without treatment or other help to
do so and despite occasionally experiencing intense desires to resume use. So other influ-
ences on the addict’s decisional processes in addition to degree of desire or craving must be
involved.3 Whatever precisely these other influences are, and whatever precisely the form
an explanation of addiction takes, the definition of addiction as a repeated and continuing
failure to refrain from or radically reduce a specified behavior despite prior resolutions to
do so directs attention to what, it is argued here, is the most crucial issue in any full and
adequate theory of addiction. In other words, if addiction is seen as a disorder of choice,
what is the choice that is disordered? Clearly, it is the choice to resume taking a drug after
one has resolved not to (and, possibly, when one currently judges it better not to—​see
Chapter 8). This puzzle may be implicit in conventional, medical accounts of addiction
but almost all theory and research on addiction so far, together with most definitions, have
been concerned with questions at the two “lower” levels in Figure 1.1 and not with issues
of choice, decision, akrasia, weakness of will, and self-​control that arise at the highest level.

5.3  The counterfactual objection

The late Ole-​Jørgen Skog objected to my proposed definition on counterfactual grounds
(see Skog 2003a, pp. 159–​64). He accepts that “persistent desire or unsuccessful attempts
to cut down” is the sine qua non for an attribution of addiction but argued that it is prob-
lematic to say that a person is not addicted unless they have struggled to change and failed.
Take someone who is a longstanding drug user but does not wish to quit and makes no
attempt to do so; in my 1998 article I call this person neuroadapted to the drug or physi-
cally dependent, but would withhold the label of addiction. But suppose, says Skog, this
person’s circumstances changed—​for example, if the cost of the habit rose, if the drug
taking in question became more stigmatized, or if their spouse threated to leave unless
they quit—​and they did now try to quit, they would most likely find it difficult and might
experience repeated failures. Skog asks whether it makes sense to say that the person was
not addicted under the first set of circumstances but addicted under the second, counter-
factual circumstances. To illustrate his argument further Skog imagines the five examples
of putative addiction summarized in Figure 1.2, and these will be discussed in turn.4
The naïve addict is illustrated by a long-​term heavy smoker in the 1950s before the
harmful effects of smoking were widely known. She is certainly neuroadapted to nicotine

3 More details of this argument and hence of the case for the definition of addiction being defended may
be found in Heather (1998).
4 Skog’s typology has interesting similarities to that described by Frankfurt (1971) and discussed by
Kennett (2013).
  15

A proposed definition of addiction 15

Feeling of enslavement?

No Yes

Ambivalence?
Naïve addict

No Yes

”Happy” addict Ever struggled?

No Yes

Subclinical addict Current struggle?

No Yes

Resigned addict Clinical addict

Low High
Degrees of dissonance
Figure 1.2  Classification scheme for cases fulfilling the counterfactual definition of addiction
From Skog, O.-J. Addiction: definitions and mechanisms. In: R.E. Vuchinich and N. Heather (eds), Choice, Behavioural
Economics and Addiction, pp. 157–75. Reproduced with permission from Elsevier. © 2003.

and often feels strong desires to smoke but has no good reason to quit, and makes no
attempt to do so. According to Skog, she is an addict because, if she did understand the
dangers of smoking and wanted to stop, she would find it difficult to succeed. However,
following Orford’s (1985) terminology, she is at present a consonant addict.
The “happy” addict is someone whose life may be in ruins as a consequence of a long
history of, say, heavy drinking and whose circumstances are such that he has no further
motivation for cutting down or quitting. Because of the harm he has experienced he is
ambivalent about drinking but, all things considered, believes he will be less unhappy if
he continues to drink than if he does not. He is therefore on balance a consonant addict.
The sub-​clinical addict is someone who knows that her drug use is harmful and should
be discontinued, and therefore desires to quit but has insufficient motivation to make the
attempt. Because such problems are relatively less severe, says Skog, this type is rarely
seen in the clinic. If her problems became more serious and caused her more distress,
she might well try to change her behavior but, once more, would find it difficult to do so.
Unlike the two previous examples, she is a dissonant addict.
The resigned addict has tried and failed, probably on many occasions in the past, to
quit drug use, still sees the habit as harmful, and would like to quit if he could but has
16  

16 On defining addiction

simply given up trying. This is obviously a dissonant addict but one who is unlikely to
seek professional help.
Lastly, the clinical addict is the type covered by my definition. He is still engaged in the
struggle to change behavior and may well have sought treatment or other assistance to
help him do so.5
How am I to respond to Skog’s counterfactual objection?6 There are two options: first,
despite some quibbles over detail, I could accept the thrust of his argument and modify
my definition accordingly; secondly, I could continue to insist that the term “addiction”
be reserved for the activity of struggling or having recently struggled to change drug
use (see Heather 2003); the counterfactual cases Skog describes would have to be called
something else, not addiction per se. I  now recognize that to take this second option
would be at odds with both lay and professional understandings of addiction, would
probably as a result lead to confusion about what addiction means, and would in any
case risk a descent into mere semantics. I  will therefore accept, reluctantly and with
some slight adjustments, Skog’s “counterfactual supplement” to my definition which
now becomes: “a person is addicted to a specified behavior if they have demonstrated
repeated and continuing failures to refrain from or radically reduce the behavior despite
prior resolutions to do so or if they would have demonstrated such failures under differ-
ent personal or environmental circumstances.” These changes notwithstanding, the defi-
nition is still focussed on the level of akrasia in Figure 1.1 and for much of the following
discussion in this chapter I will rely on the shorter definition without the counterfactual
supplement but with the understanding that I would concede if the counterfactual objec-
tion were made.

5.4  Denial and other kinds of self-​deception

Yet another possible objection is to argue that many people who should rightfully be
regarded as addicted fail to recognize the harm being done by their addictive behavior,
and therefore fail to try to address it, because they deceive themselves about it. These self-​
deceptions include the familiar denial, rationalizations, and other forms of “mental obses-
sion” and “insanity” described by Alcoholics Anonymous (1939) (see Chapters 2 and 20,
this volume). For this reason, so this argument would go, it makes no sense to exclude
such a person from the definition of addiction simply because their self-​deceptions pre-
vent them from trying to quit or cut down.
It could be countered that denial and resistance to change occur mainly in an inter-
personal context and are responsive to the style of the therapist (Miller et al. 2011) or,
more generally, the attitudes and conduct of other people in the person’s life. Accepting,

5 Besides implications for an understanding of addiction at the individual, clinical level, Skog’s taxon-
omy may also be relevant to responses to addiction at the population level. Although there will be no
space to consider these latter implications here, I thank Ron Borland for pointing this out.
6 For a fuller account of the exchange between Skog and myself, the reader is referred to the chapter by
Skog (2003a), my comments on his chapter (Heather 2003), and Skog’s (2003b) reply to my comments.
  17

Addiction as excessive appetite 17

however, that self-​deception in addiction is, at least to some extent, an intrapsychic phe-
nomenon, the solution is simple; these psychological factors are part of the personal cir-
cumstances applying to states of counterfactual addiction such that, if they were resolved
so that the harm of addictive behavior was recognized by the person and attempts to
change made, these attempts would probably be unsuccessful.

6  Addiction as excessive appetite

The work of Jim Orford has been mentioned above. His book, Excessive Appetites:
A Psychological View of Addictions (Orford 1985) was perhaps the first scholarly account to
advocate a more comprehensive model of addiction in which what subsequently became
known as behavioral addictions were given equal footing with substance addictions; in
this perspective, excessive forms of the use of drugs like heroin and cocaine, excessive
biological appetites for eating and sex, and non-​biological activities such as excessive
gambling were all seen as special cases of a more general and fundamental phenomenon
that Orford set out to describe and explain. The same approach was retained in a second,
equally influential edition of the book (Orford 2001).
There is much in Orford’s model that accords with the theoretical definition of addic-
tion proposed here. The central idea of a strong attachment that is subject to restraint,
ambivalence and conflict is clearly consistent with repeated resolutions and unsuccessful
attempts to quit or radically cut down an addictive behavior. Yet there are fundamental
differences too. The notion of a consonant excessive appetite in which, notwithstanding
how excessive the appetite appears to be, the person feels no need to change behavior
is at odds with my definition, except possibly where recourse is made to Skog’s (2003a)
concept of the “naïve addict” who would try to change if circumstances were different
(see above). More germanely, in the absence of explicit attempts to change or clear
reasons for believing that attempts to change would be made under different circum-
stances, even a dissonant excessive appetite might not be distinguishable from a strong
and unusual preference. Humans engage in many eccentric and statistically abnormal
pursuits, which have their upsides and downsides, their good and bad times, and their
beneficial and more or less harmful consequences, all of which can result in ambiva-
lence and conflict (stamp collecting, mountain climbing, and protracted TV watching,
among many other examples, spring to mind). Given that Orford wishes to replace con-
ventional but, in his view, misguided notions of addiction with his concept of “exces-
sive appetites,” the psychological and social processes he describes in the development
and amplification of an excessive appetite are not sufficient in themselves to justify the
attribution of addiction.
The key question is: who is to say whether or not an appetite for any sort of activity is
excessive? If it is a therapist, an academic expert on addiction, relatives and neighbors, or
even “society at large” who make(s) this determination, it must necessarily be made on
normative and thus contestable grounds. If it is made purely on statistical grounds, this
says nothing about the nature of the activity that is described as excessive. If made by the
18  

18 On defining addiction

person himself, this is at least prima facie evidence that it is something he might want to
do something about. With further objective evidence of motivation to change, in the form
of actual but unsuccessful attempts to do so, we would then be justified in seeing him as
addicted.

7  A liberal account of addiction

The above line of reasoning is also relevant to a “liberal” account of addiction, as provided
for example by Foddy and Savulescu (2007, 2010). On this account, addiction is charac-
terized, not as a brain disease or a moral condition, but as a particularly strong prefer-
ence, similar to appetitive preferences. Addictive desires are merely especially strong and
socially unacceptable desires that engage and activate the same basic and primal pleasure
pathways of the brain to generate pleasure.
The problem of addiction is the problem of the management of pleasure, not treatment of a dis-
ease. This differentiation becomes more and more important as the opportunities and sources of
pleasure increase. What place should pleasure have in our lives and how can we achieve that? The
moral opprobrium that attaches to addiction should be that which is appropriate for any activity
that harms others and should not be the result of disapproval of different ways of living or different
orderings of value. Much of the disrepute attached to addiction has been illiberal and the result of
one group, often the dominant political or religious group, applying their norms for personal living
to others, who share a different ideal of the good life.

(Foddy and Savulescu 2007, p. 32)

Leaving aside the issue of whether the treatment of a disease is the only alternative to
the management of pleasure, what this misses is that it is addicts themselves who often
complain that their behavior is out of their control and, in some cases, seek help, whether
lay or professional, to regain it. Foddy and Savulescu are right to insist that, although the
addict may not always like the results of her preference, she might rationally value the
pleasure of, say, heroin use above the value of her health or job. They are also correct that
the evidence shows drug users to have more control over their behavior than adherents
of “addiction as a chronic, relapsing brain disease” suppose. But a definition of addiction
implying that addicts might request treatment because their best efforts to control their
addictive behavior on their own have repeatedly failed cannot be accused of the illiberal-
ity they have in mind.

8  Addiction as hard-​to-​maintain behavior change

Ron Borland’s (2014) recent book, Understanding Hard to Maintain Behavior
Change: A Dual Process Approach, is highly relevant to the definitional issues under dis-
cussion here. Primarily a researcher on smoking, Borland presents a theory of how people
change, and can be helped to change, habitual unhealthy behaviors. This involves the
elaboration of a dual process account of behavior change, based on the work of Strack and
Deutsch (2004), which is an attempt to understand “the constraints on and the potential of
volitional attempts to change behaviour patterns that are under the moment-​to-​moment
  19

Addiction as akrasia 19

control of non-​volitional processes” (Borland 2014, p. 1). The theory describes the inter-
action and conflict between an Operational System (OS) in which bottom-​up processing
occurs automatically, mostly out of consciousness and reactive to the environment, and
an Executive System (ES) in which top-​down processing can be reflective, proactive, and
deliberative. Decisions to change made in the ES can only result in action via the OS
and only if they can generate greater affective force than those generated by bottom-​up
processes.
Hard-​to-​maintain (HTM) behavior change can be divided into hard-​to-​reduce or  –​
eliminate behaviors (HTR), to which the OS reacts positively but which are seen as
undesirable by the ES, and hard-​to-sustain (HTS) behaviors, where the reverse is true.
The former includes all the various forms of habitual drug use and other behaviors called
excessive appetites by Orford (1985, 2001), while the latter includes exercise regimes,
healthy eating patterns, sun-​ protective behaviors and safe-​sex practices. Sometimes
behavior change involves a combination of the two when the aim is to replace an HTR
with an HTS behavior.
The immediate relevance to the present discussion is that HTR behaviors are precisely
those which the research evidence shows are marked by persistent but unsuccessful
attempts to change and high rates of lapse and relapse (see Borland 2014, pp.  38–​44).
Even if overt attempts to change have not been made, people may be contemplating mak-
ing such attempts or may have aborted them before the attempt could be initiated, thus
conforming to a counterfactual addict as described by Skog (2003a). Seen in this broader
context, it makes less sense to ask which drugs or activities are “addictive” than to identify
addictive behavior with the occurrence of any behavior that is empirically difficult for
someone to resist or eliminate despite resolutions and attempts to do so.7 Thus the defini-
tion of addiction offered in this chapter can be seen as a foundation for the development
of a dual process theory of addictive behavior.

9  Addiction as akrasia
As mentioned above (p. 13), akrasia was Aristotle’s word for what can roughly be desig-
nated as acting intentionally against one’s better judgment and was a topic for discussion
and disagreement among the ancient Greek philosophers. It has continued to occupy the
minds of philosophers ever since and has spawned a huge philosophical literature. In
more modern terms the problem is often called that of “incontinence,” in the sense of
lacking control over one’s actions. Another, more familiar synonym is “weakness of will.”
The converse of akrasia for Aristotle and his contemporaries was enkrateia, roughly trans-
lated as self-​control and referring to action that is consistent with better judgment. The
philosophical problem is whether or not weakness of will is logically possible and, if so,
in what form, and how it can best be characterized and explained. The same explanations

7 A  radical implication of the idea that addictive behaviors are a subset of hard-​to-​reduce/​eliminate
behaviors is that the concept of addiction thereby becomes theoretically redundant.
20  

20 On defining addiction

will presumably apply to self-​control as the converse of weakness of will and to how one
can replace the other. It should be added that, in addition to philosophical discourse,
the same age-​old problem of understanding why humans behave in ways in which they,
in some sense, do not want to behave has also occupied the minds of psychologists and
psychiatrists since the foundations of those disciplines.
The thesis that addiction can profitably be seen as an extreme form of akrasia will be
defended in a later chapter of this book (see Chapter 8). Suffice it to say here that a view
of addiction as a form of akrasia (or of “weakness of will”—​see Chapter 8) is founded on
a definition of addiction as repeated failures to refrain from or radically reduce a speci-
fied behavior despite prior resolutions to do so. This definition is, at the same time, the
foundation for a theory of addiction as a disorder of choice or decision-​making based on
a dual process account of human behavior.

10  Conclusions
In case it has not been made sufficiently clear from the preceding contents of this chap-
ter, it is the advent of behavioral addictions that has complicated the task of defining the
essential nature of addiction and has demanded what must necessarily be an expanded
definition of it. In the absorbing book based on his heroic treatment of the street-​dwelling
inhabitants of Vancouver’s Downtown Eastside, whose health and welfare have been
nearly destroyed by their addictions to drugs, Gabor Maté (2012) includes a chapter enti-
tled, Takes One to Know One (Chapter 9, pp. 101–​115). This is an account of Maté’s long-
standing preoccupation with purchasing classical music CDs, a habit that cost him more
money and time than he could afford, involved frequent lying to his wife, and sometimes
led to neglect of his clinical duties. But is he right to call this habit an addiction, to com-
pare it with the travails and devastations of his patients, and to claim that his experience
gives him insight into their condition? The answer offered here is that, beyond the obses-
sion, the craving, the self-​loathing, and the harm, the only way we may meaningfully say
that the good physician was addicted in the same sense that his patients are is that he
resolved many times to stop buying CDs and tried but failed to do so.
The claim is that the definition offered is the only one that characterizes all those ways
of behaving that we are reasonably inclined to think of as addictions; other definitions,
especially those relying on continued use despite harm, are unable to distinguish between
addictions and preferences for excessive forms of behavior or are contestable on other
grounds. This does not mean that there are no important subdivisions within the class of
addictive behaviors as defined; for example, research may eventually make clear that some
kinds of addiction, presumably the main substance addictions, are the consequence of a
“dopamine hijacking process” and others, presumably the main behavioral addictions, are
not (Dill and Holton 2014). In either case, however, an adequate theory of addiction must
account for the “final common pathway” of addictive behavior, explain why attempts to
break free from it are repeatedly unsuccessful, and why and how they may eventually
meet with success. Thus the definition is a theoretical definition because it draws attention
to this theoretical requirement.
  21

Conclusions 21

Needless to say, many questions remain unanswered about this definition. In par-
ticular, I have not considered definitional differences between addiction and obsessive-​
compulsive disorder. Also missing is how the definition could be operationalized and how
degrees of addiction could be conceptualized and measured. These tasks will have to wait
for another occasion.
In this chapter, I have suggested a way of answering questions of the following kind: is Jill’s
curious and apparently harmful behavior an example of addiction? When Jack repeatedly
behaves in a certain way, is he showing addiction? The simple solution is, have they tried to
give up or radically change the behavior after resolving to do so and failed? This definition
is also argued to be a foundation for a theory of addiction as a disorder of choice based on
a dual-​process account of human behavior. I have tried here to refute actual and potential
objections to the validity of this definition or, where necessary, have made concessions to
some of those objections. I  have conceded that addiction may be said to exist in Skog’s
(2003a) counterfactual sense in which resolutions and failures to change behavior would
likely occur under different circumstances. But I recognize, of course, that other objections
may be possible that have not been considered here. This is why I regard the case for the
definition as tentative in nature and look forward to addressing other objections that may
be made to it.

Acknowledgments
My thanks to Ron Borland, Ed Day, Mark Jankowski, Jim Orford, Hanna Pickard, and
Gabriel Segal for valuable comments and advice on this chapter.

References
Alcoholics Anonymous (1939). Alcoholics Anonymous. New York: Alcoholics World Services Inc.
Alexander, B.K. (2001). The Roots of Addiction in Free Market Society. Vancouver, Canada: Canadian
Centre for Policy Alternatives.
American Psychiatric Association (2013). Diagnostic and Statistical Manual of Mental Disorders.
Washington DC: APA.
Andreasson, C.S., Griffiths, M.D., Hetland, J., and Pallesen S. (2012). Development of a work addiction
scale. Scandinavian Journal of Psychology, 53, 265–​72.
Avena, N.M., Rada, P., and Hoebel, B.G. (2008). Evidence for sugar addiction: behavioral and neurochemical
effects of intermittent, excessive sugar intake. Neuroscience and Biobehavioral Reviews, 32, 20–​39.
Babor, T.F. (2011). Substance, not semantics, is the issue: comments on the proposed addiction criteria
for DSM-​V. Addiction, 106, 870–​72.
Banerjee, S., Hay, J.L., and Greene, K. (2015). Indoor tanning addiction tendencies: role of positive
tanning beliefs, perceived vulnerability, and tanning risk knowledge. Addiction Research and Theory,
23, 156–​62.
Berczik, K., Griffiths, M.D., Szabo, A., Kurimay, T., Urban, R., and Demetrovics, Z. (2014). Exercise
addiction. In: K.P. Rosenberg and L.C. Feder (eds), Behavioral Addictions: Criteria, Evidence, and
Treatment. London: Academic Press, pp. 317–​42.
Bergen-​Cico, D.K., Haygood-​El, A., Jennings-​Bey, T.N., and Lane, S.D. (2014). Street addiction: a
proposed theoretical model for understanding the draw of street life and gang activity. Addiction
Research and Theory, 22, 15–​26.
22  

22 On defining addiction

Borland, R. (2014). Understanding Hard to Maintain Behaviour Change: A Dual Process Approach.

Chichester: John Wiley.
Borland, R., Parros, T.R., Yong, H-​H., Cummings, M., and Hyland, A. (2012). How much unsuccessful
quitting activity is going on among smokers? Data from the International Tobacco Control Four
Country cohort survey. Addiction, 107, 673–​82.
Britannica (2015). Philosophy Pages: Definition and Meaning. http://​www.philosophypages.com/​lg/​e05.
htm. Accessed 03/​04/​2015.
Budney, A.J., Roffman, R., Stephens, R.S., and Walker, D. (2007). Marijuana dependence and its
treatment. Addiction Science and Clinical Practice, 4, 4–​16.
Caetano, R. (2011). There is potential for cultural and social bias in DSM-​V. Addiction, 106, 885–​86.
Cohen, S.A., Higham, J.E.S., and Cavaliere, C.T. (2011). Binge flying, behavioural addiction and
climate change. Annals of Tourism Research, 38, 1070–​89.
Collingwood, J. (2006). Does chocolate addiction exist? Psychcentral.com, http://​psychcentral.com/​lib/​
does-​chocolate-​addiction-​exist/​000233. Accessed 30/​03/​2015.
Davidson, D. (1980). How is weakness of the will possible? In: D. Davidson (ed.), Essays on Actions and
Events. Oxford: Clarendon Press, pp. 21–​42.
Davies, J.B. and Baker, R. (1987). The impact of self-​presentation and interviewer bias effects on self-​
reported heroin use. British Journal of Addiction, 82, 907–​12.
Dill, B. and Holton, R. (2014). The addict in all of us. Frontiers in Psychiatry, 5 (Article 139),
doi: 10.3389/​fspyt.2014.00139.
Dwoskin, H. (2015). How to break free of your approval addiction. Huffpost Healthy Living. http://​
www.huffingtonpost.com/​hale-​dwoskin/​how-​to-​break-​free-​of-​your_​b_​6071692.html Accessed
01/​04/​2015.
Edelstein, E.L. (1973). A case of water dependence. British Journal of Addiction, 68: 365–​67.
Edwards, G., Arif, A., and Hodgson, R. (1981). Nomenclature and classification of drug and alcohol
related problems: a WHO memorandum. Bulletin of the World Health Organization, 59, 225–​42.
English Language and Usage Stack Exchange (2015). Etymology of addiction. http://​english.
stackexchange.com/​questions/​130423/​etymology-​of-​addict. Accessed 24/​03/​2015.
Fell, G.L., Robinson, K.C., Mao, J., Woolf, C.J., and Fisher, D.E. (2014). Skin β-​endorphin mediates
addiction to UV light. Cell, 157, 1527–​34.
Fisher, H.E. (2014). The tyranny of love: love addiction—​an anthropologist’s view. In: K.P. Rosenberg
and L.C. Feder (eds), Behavioral Addictions: Criteria, Evidence, and Treatment. London: Academic
Press, pp. 237–​66.
Foddy, B. and Savulescu, J. (2007). Addiction is not an affliction: addictive desires are merely pleasure-​
oriented desires. American Journal of Bioethics, 7, 29–​32.
Foddy, B., and Savulescu, J. (2010). A liberal account of addiction. Philosophy, Psychiatry and
Psychology, 17, 1–​22.
Foster, A.C., Shorter, G.W., and Griffiths, M.D. (2014). Muscle dysmorphia: could it be classified as an
addiction to body image? Journal of Behavioral Addictions, 4, 1–​5.
Frances, A. (2013). Saving Normal: An Insider’s Revolt Against Out-​of-​Control Psychiatric Diagnosis,
DSM-​5, Big Pharma, and the Medicalization of Ordinary Life. New York: William Morrow.
Frankfurt, H.G. (1971). Freedom of the will and the concept of a person. Journal of Philosophy,
68, 5–​20.
Fraser, N. and Gordon, L. (1994). A genealogy of dependency: tracing a keyword of the U.S. welfare
state. Signs, 19, 309–​36.
Grant, J.E., Odlaug, B.L., and Kim, S.W. (2010). Kleptomania: clinical characteristics and relationship
to substance use disorders. American Journal of Drug and Alcohol Abuse, 36, 291–​95.
Griffiths, M.D. (2011). Workaholism—​a 21st Century addiction. Psychologist, 24, 740–​44.
  23

Conclusions 23

Griffiths, M.D., Kuss, D.J., and Demetrovics, Z. (2014). Social networking addiction: an overview
of preliminary findings. In: K.P. Rosenberg and L.C. Feder (eds), Behavioral Addictions: Criteria,
Evidence, and Treatment. London: Academic Press, pp. 119–​42.
Hall, W. (2014). What has research over the past two decades revealed about the adverse health effects
of recreational cannabis use? Addiction, 110, 19–​35.
Heather, N. (1991). Impaired control over alcohol consumption. In: N. Heather, W.R. Miller,
and J. Greeley (eds), Self-​control and the Addictive Behaviours. Sydney, Australia:
Maxwell Macmillan, pp. 153–​79.
Heather, N. (1998). A conceptual framework for explaining drug addiction. Journal of
Psychopharmacology, 12, 3–​7.
Heather, N. (2003). Comments on Skog. In: R.E. Vuchinich and N. Heather (eds), Choice, Behavioural
Economics and Addiction. Kidlington, Oxford, UK: Elsevier, pp. 176–​81.
Heyman, G.M. (2009). Addiction: A Disorder of Choice. Cambridge, MA: Harvard University Press.
Heyman, G.M. (2013). Quitting drugs: quantitative and qualitative features. Annual Review of Clinical
Psychology, 9, 29–​59.
Hilton, D.L.Jr and Watts, C. (2011). Pornography addiction: a neuroscience perspective. Surgical
Neurology International, 2, 19, doi: 10.4103%2F2152-​7806.76977
Højsted, J., and Sjøgren, P (2007). Addiction to opioids in chronic pain patients: a literature review.
European Journal of Pain, 11, 490–​518.
Jaffe, A. (2013). Moderation in addiction treatment: a change is gonna come. Psychology Today, https://​
www.psychologytoday.com/​blog/​all-​about-​addiction/​201308/​moderation-​in-​addiction-​treatment-​
change-​is-​gonna-​come. Accessed 27/​04/​2015.
Juliano, L.M. and Griffiths, R.R. (2004). A critical review of caffeine withdrawal: empirical validation
of symptoms and signs, incidence, severity, and associated features. Psychopharmacology,
176, 1–​29.
Kaplan, R. (1996). Carrot addiction. Australian and New Zealand Journal of Psychiatry, 30, 698–​700.
Kellett, S. and Gross, H. (2006). Addicted to joyriding? An exploration of young offenders’ accounts
of their car crime. Psychology, Crime and Law, 12, 39–​59.
Kennett, J. (2013). Just say no? Addiction and elements of self-​control. In: N. Levy (ed.), Addiction and
Self-​control. Oxford UK: Oxford University Press, pp. 144–​64.
Kozlowski, L.T. and Wilkinson, D.A. (1987). Use and misuse of the concept of craving by alcohol,
tobacco and drug researchers. British Journal of Addiction, 82, 31–​36.
Kutchins, H. and Kirk, S.A. (1997), Making Us Crazy: DSM—​The Psychiatric Bible and the Creation
of Mental Disorders. London: Constable.
Levine, H.G. (1978). The discovery of addiction: changing conceptions of habitual drunkenness
in America. Journal of Studies on Alcohol, 39, 143–​74.
Licata, S.C. and Rowlett, J.K. (2008). Abuse and dependence liability of benzodiazepine-​type
drugs: GABA receptor modulation and beyond. Pharmacology, Biochemistry and Behavior,
90, 74–​89.
Lindesmith, A.R. (1968). Addiction and Opiates. Chicago: Aldine.
Marlatt, G.A. and Donovan, D.M. (eds) (2007). Relapse Prevention: Maintenance Strategies in the
Treatment of Addictive Behaviors (2nd ed.). New York NY: Guilford Press.
Martin, C.S., Langenbucher, J.W., Chung, T., and Sher, K.J. (2014). Truth or consequences in the
diagnosis of substance use disorders. Addiction, 109, 1773–​78.
Maté, G. (2012). In the Realm of Hungry Ghosts: Close Encounters with Addiction. Toronto: Vintage
Canada.
Midgley, S.J., Heather, N., and Davies, J.B. (1999). Dependence-​producing potential of anabolic-​
androgenic steroids. Addiction Research, 7, 539–​50.
24  

24 On defining addiction

Miller, W.R., Forcehimes, A., and Zweben, A. (2011). Treating Addictions: Guidelines for Professionals.
New York: Guilford.
Nicholls, J. (2009). The Politics of Alcohol: A History of the Drink Question in England. Manchester and
New York: Manchester University Press.
O’Brien, C. (2011). Addiction and dependence in DSM-​V. Addiction, 106, 866–​67.
Orford, J. (1985). Excessive Appetites: A Psychological View of Addictions. Chichester UK: Wiley.
Orford, J. (2001) Excessive Appetites: A Psychological View of Addictions (2nd ed.). Chichester
UK: Wiley.
Pearson, C. and Hussain, Z. (2015). Smartphone use, addiction, narcissism, and personality: a mixed
methods investigation. International Journal of Cyber Behavior, Psychology and Learning, 5, 7–​32.
Peele, S., and Brodie, A. (1975). Love and Addiction. New York: Taplinger.
Petry, N.M., Rehbein, F., Gentile, D.A., et al. (2014). An international consensus for assessing internet
gaming disorder using the new DSM-​5 approach. Addiction, 109, 1399–​406.
Potenza, N.M. (2006). Should addictive disorders include non-​substance-​related conditions? Addiction,
101 (Suppl. 1), 142–​51.
Prochaska, J.O., DiClemente, C.C., and Norcross, J.C. (1992). In search of how people
change: applications to addictive behaviour. American Psychologist, 47, 1102–​14.
Racine, E., Kahn, T., and Hollander, E. (2014). Compulsive buying disorder. In: K.P. Rosenberg and
L.C. Feder (eds), Behavioral Addictions: Criteria, Evidence, and Treatment. London: Academic Press,
pp. 285–​316.
Room, R. (1998). Alcohol and drug disorders in the International Classification of Diseases: a shifting
kaleidoscope. Drug and Alcohol Review, 17, 305–​17.
Room, R. (2003). The cultural framing of addiction. Janus Head, 6, 221–​34.
Room, R. (2006). Taking account of cultural and societal influences on substance use diagnoses and
criteria. Addiction, 101(Suppl. 1), 31–​39.
Room, R. (2011). Substance use disorders—​a conceptual and terminological muddle. Addiction, 106,
879–​81.
Room, R., Hellman, M., and Stenius, K. (2015). Addiction: the dance between concept and terms.
International Journal of Alcohol & Drug Research, 4, 27–​35.
Rosenberg, K.P., O’Connor, S., and Carnes, P.J. (2014). Sex addiction: an overview. In: K.P. Rosenberg
and L.C. Feder (eds), Behavioral Addictions: Criteria, Evidence, and Treatment. London: Academic
Press, pp. 215–​36.
Rudski, J.M., Segal, C., and Kallen, E. (2009). Harry Potter and the end of the road: parallels with
addiction. Addiction Research and Theory, 17, 260–​77.
Russell, M.A.H. (1976). What is dependence? In: G. Edwards, M.A.H. Russell, D. Hawks, and
M. MacCafferty (eds), Drugs and Drug Dependence. Lexington MA: Lexington Books, pp. 182–​87.
Sakai, J.T., Hall, S.K., Mikulich-​Gilbertson, S.K., and Crowley, T.J. (2004). Inhalant use, abuse, and
dependence among adolescent patients: commonly comorbid problems. Journal of the American
Academy of Child and Adolescent Psychiatry, 43, 1080–​88.
Shepherd, R.-​M. (2009). Dangerous consumptions beyond the grave: psychic hotline addiction for the
lonely hears and grieving souls. Addiction Research and Theory, 17, 278–​90.
Sinnott-​Armstrong, W. and Pickard, H. (2013). What is addiction? In: K.W.M. Fulford, et al. (eds),
Oxford Handbook of Philosophy and Psychiatry. Oxford UK: Oxford University Press. pp. 851–​64.
Skog, O.-​J. (1999). Rationality, irrationality and addiction: notes on Becker’s and Murphy’s theory of
addiction. In: J. Elster and O.-​J. Skog (eds), Getting Hooked: Rationality and Addiction. Cambridge,
UK: Cambridge University Press, pp. 173–​207.
  25

Conclusions 25

Skog, O.-​J. (2003a). Addiction: definitions and mechanisms. In: R.E. Vuchinich and N. Heather (eds),
Choice, Behavioural Economics and Addiction. Kidlington, Oxford, UK: Elsevier, pp. 157–​75.
Skog, O.-​J. (2003b). Reply to Heather. In: R.E. Vuchinich and N. Heather (eds), Choice, Behavioral
Economics and Addiction. Kidlington, Oxford, UK: Elsevier, p. 182.
Stanford Encylopedia of Philosophy (2105). Locke on real essence. http://​plato.stanford.edu/​entries/​
real-​essence/​ Accessed 17/​06/​2015.
Strack, F. and Deutsch, R. (2004). Reflective and impulsive determinants of social behaviour.
Personality and Social Psychology Bulletin, 8, 220–​47.
Valverde, M. (1998). Diseases of the Will: Alcohol and the Dilemmas of Freedom. Cambridge,
UK: Cambridge University Press.
Van Dyke, C. and Byck, R. (1982). Cocaine. Scientific American, 246, 128–​41.
Warburton, D.M. (1989). Is nicotine use an addiction? Psychologist, 4, 166–​70.
Weinstein, A., Feder, L.C., Rosenberg, K.P., and Dannon, P. (2014). Internet addiction
disorder: overview and controversies. In: K.P. Rosenberg and L.C. Feder (eds), Behavioral
Addictions: Criteria, Evidence, and Treatment. London: Academic Press, pp. 99–​118.
Wikler, A. (1965). Conditioning factors in opiate addiction and relapse. In: D.I. Wilner and
G.G. Kassenbaum (eds), Narcotics. New York: McGraw-​Hill, pp. 279–​85.
World Health Organization (1964). WHO Expert Committee on Addiction-​producing Drugs: 13th
Report. Geneva, Switzerland: WHO.
Yau, Y.H.C., Gottlieb, C.D., Krasna, L.C., and Potenza, M.N. (2014). Food addiction: evidence,
evaluation, and treatment. In: K.P. Rosenberg and L.C. Feder (eds), Behavioral Addictions: Criteria,
Evidence, and Treatment. London: Academic Press, pp. 143–​84.
Zinberg, N.E. (1984). Drug, Set and Setting: the Basis for Controlled Intoxicant Use. New Haven CT: Yale
University Press.
26  
  27

Section II

Philosophical foundations
28  
  29

Chapter 2

How do you know you have a drug

problem? The role of knowledge of
negative consequences in explaining
drug choice in humans and rats
Hanna Pickard
Serge H. Ahmed

Abstract
Choice models of addiction raise a stark puzzle. Why, if addicts are able to
control their consumption and choose to abstain in many circumstances, do
they routinely choose to continue to use despite negative consequences?
We delineate four options available to a choice theorist to explain this
puzzle with respect to human addicts; describe recent experiments with
addicted rats which determine the conditions under which they do and
do not choose alternative goods over cocaine and heroin respectively; and
consider how this animal research bears on our understanding of the nature
of addiction. We conclude by arguing for the importance to addiction
research of the popular notion of “denial” which, paradoxically, we can
begin to theorize by appeal to animal models: for the puzzle dissolves
if addicts do not know that the choice to use is the cause of negative
consequences, as rats, given their cognitive limitations, cannot.

1 Introduction
Addiction has harrowing consequences. It can destroy people’s bodies, causing disease
and death. It can damage people’s brains, causing cognitive impairment. It can cause and
exacerbate mental health problems, such as depression, anxiety, and psychosis. It ruins
relationships. People lose their families and friends, their jobs and social standing, their
sense of having a community to which they belong. It can be a source of terrible shame,
self-​hatred, and low self-​worth.
The orthodox conception of addiction is a chronic, relapsing neurobiological disease
characterized by compulsive drug use despite negative consequences such as these.1 This

1 We include alcohol, as well as all common illicit and pharmaceutical drugs of abuse, in the reference of
the term “drugs.”
30  

30 How do you know you have a drug problem?

conception is common to the American National Institute of Alcohol Abuse and Alcoholism,
the American National Institute of Drug Abuse, the American Medical Association, the
British Medical Association, and the World Health Organization, and reflected in the clinical
description and diagnostic criteria in the Diagnostic and Statistical Manual of Mental Disorders
Fifth Edition (DSM-​5) and the International Classification of Diseases and Health Problems
(ICD-​10). The prevalence and power of this orthodoxy can in part be explained by its capacity
to resolve the puzzle inherent in addiction. Common sense suggests that, if people know that
an action will bring about negative consequences, and they are able to avoid it, then they do.
We act, so far as we can, in our own best interests and the interests of others we care for. This
is a basic folk psychological rule of thumb for explaining and predicting human action. But
this is what addicts seem not to do. Although addiction has harrowing consequences, addicts
continue to use drugs.
Understanding addiction as a chronic, relapsing neurobiological disease characterized
by compulsive drug use resolves this puzzle. The explanation is that addiction “hijacks”
the brain, to use a common metaphor, thereby “nullifying any semblance of voluntary
choice” (Charland 2002, p.  41) so that addicts lose all control and cannot help taking
drugs, despite the consequences and against their best interests. As Carl Elliott describes
it, addicts “must go where the addiction leads [them], because the addiction holds the
leash” (Elliott 2002, p. 48). Hence the puzzle of why addicts use drugs despite negative
consequences is explained: if addicts could avoid using drugs, they would—​but they can’t,
so they don’t.
The orthodox conception of addiction is challenged by evidence indicating that, however
hard it is for addicts to control their use, and important for others to recognize this hardship,
addicts are not compelled to use but have choice over their consumption. There are multiple
sources of such evidence. Anecdotal and first-​person reports abound of addicts (includ-
ing those with a DSM-​based diagnosis of dependence) going “cold turkey” (Heyman 2009,
2013). Large-​scale epidemiological studies demonstrate that the majority of addicts “mature
out” without clinical intervention in their late twenties and early thirties, as the respon-
sibilities and opportunities of adulthood, such as parenthood and employment, increase
(for a review of the empirical data see Heyman 2009, Heyman and Mims, Chapter 21, this
volume; cf. Foddy and Savulescu 2006; Peele 1985; Pickard 2012). Rates of use are cost-​
sensitive: indeed, some addicts choose to undergo withdrawal in order to decrease toler-
ance, thereby reducing the cost of future use (Ainslie 2000). There is increasing evidence
that contingency management treatment improves abstinence and treatment compliance,
compared to standard forms of treatment such as counselling and cognitive-​behavioral
therapy, by offering a reward structure of alternative goods, such as modest monetary
incentives and small prizes, on condition that addicts produce clean urine samples (for a
review see Petry et al. 2011). Experimental studies show that, when offered a forced choice
between taking drugs or receiving money then and there in the laboratory setting, addicts
will frequently choose money over drugs when offered the choice (Hart et al. 2000; Hart
2013). Finally, since Bruce Alexander’s seminal experiment “Rat Park” first intimated that
something similar might be true of rats (Alexander et al. 1978, 1985), animal research on
  31

Explaining the choice to use drugs despite negative consequences 31

addiction has convincingly demonstrated that, although the majority of cocaine-​addicted

rats will escalate self-​administration, sometimes to the point of death, if offered no alterna-
tive goods, they will forgo cocaine and choose alternative goods, such as saccharin or same-​
sex snuggling, if available (Ahmed 2010; Zernig et al. 2013). Broadly speaking, the evidence
converges to indicate strongly that addicts are able to control their consumption and choose
otherwise in many circumstances. Drug use is responsive to incentives: addicted humans
and rats alike are able to forgo drugs when motivated to do so.
However, the emergence of a choice model of addiction as a rival to the orthodox con-
ception raises the puzzle in stark terms. Why, if addicts are able to control their consump-
tion and choose to abstain, do they routinely choose instead to continue to use despite
negative consequences?
Our aim in this chapter is to begin to answer this question. In section 2, we deline-
ate the options open to a choice theorist to explain the puzzle with respect to human
addicts. In section 3, we describe recent experiments with addicted rats which determine
the conditions under which they do and do not choose alternative goods over cocaine and
heroin respectively. Finally, in section 4, we bring sections 2 and 3 together and consider
how this animal research might bear on our understanding of the nature of addiction.
To anticipate, the overarching aim of the chapter is to argue for the importance of a bet-
ter philosophical and scientific understanding of the popular notion of “denial,” which,
paradoxically, we can begin to theorize by appeal to animal models. For, put simply, the
puzzle dissolves if addicts do not know that the choice to use is the cause of the harrowing
consequences, as rats, given their cognitive limitations, can never do.

2  Explaining the choice to use drugs despite

negative consequences
In asking why addicts use drugs despite negative consequences, it is paramount to
recognize that there is no puzzle at all with respect to why people use drugs in the
first place. Alongside factors such as peer pressure and cultural availability, drugs are
instrumental means to valuable ends, which people may want to secure and can easily
learn that drugs facilitate. Muller and Schumann (2011) delineate the following well-​
established functions for non-​addictive consumption:  (1)  improved social interac-
tion; (2) facilitated mating and sexual behavior; (3) improved cognitive performance;
(4)  facilitated recovery and coping with psychological stress; (5)  self-​medication for
mental problems; (6)  sensory curiosity—​expanded experiential horizon; and, finally,
(7)  euphoria and hedonia. Put crudely, drugs can be pleasurable—​they make us feel
good. And apart from their inherent pleasure, they help us do various things we may
want to do, like feel relaxed at a dinner party, have the confidence to flirt and dance,
or de-​stress after a hard day. Of course, the risk of addiction, alongside the possibility
of moral condemnation and legal sanction, may deter some people from using some
drugs, some of the time. But that does not affect the basic point that, broadly speaking,
drugs are means to multiple valuable ends.
32  

32 How do you know you have a drug problem?

Addiction occurs when use escalates and comes to dominate and adversely affect a per-
son’s life. Importantly, from a clinical perspective, as embodied both in the DSM-​5 and
the ICD-​10, there is no sharp divide between heavy and problematic use and diagnosis
with a disorder (cf. Wakefield and Schmitz 2014). In addition to the physiological symp-
toms of tolerance and withdrawal, the diagnostic criteria are polythetic and include: crav-
ings and a sense of compulsion to use; increased fixation and amount of time spent using
and recovering, in conjunction with the neglect of other pleasures and activities; struggle
to control levels of use despite efforts; serious negative consequences due to use with
respect to physical and mental health, relationships, and roles and responsibilities; and,
finally, persisting in using despite knowledge (DSM-​5) or clear evidence (ICD-​10) that it
is causing or exacerbating such problems (APA 2013; WHO 1992).
Our understanding of the acute and chronic effects of drugs on the brain is ever increas-
ing (Koob and Le Moal 1997; Koob and Volkow 2010). For instance, drugs directly affect
levels of synaptic dopamine as opposed to affecting them only indirectly via the normal
neural processes subserving learning and reward. Alongside other processes, such as incen-
tive sensitization, this may explain why drug cues become unusually salient and cause unu-
sually strong motivations to use (Redish et al. 2008; Montague et al. 2004; Robinson and
Berridge 2008). However, like most mental disorders, and as apparent in the diagnostic
criteria, addiction is not diagnosed via neurobiological mechanisms, but rather via psy-
chological states and behavioral patterns identified at the personal level that cause distress
and impair functioning (cf. Graham 2010; Stephens and Graham 2009; Levy 2013). This
is part of why, especially given the polythetic nature of the criteria, there can be no sharp
clinically derived line between heavy and problematic use and diagnosis with a disorder.
Consider, first, the obvious but important point that different people have different con-
ceptions of how they want to live, and, correspondingly, are distressed or disturbed by
different things. Although many, perhaps most, addicts feel shame, self-​hatred, and low
self-​worth about their addiction, there may be some addicts, within some sociocultural
contexts, who embrace their identity more willingly (cf. Flanagan 2013, Chapter 4, this
volume). If this is so, and if drugs are readily available, then the fact that a person craves
drugs, or that their life revolves around drugs, may not be experienced as a problem.
Correspondingly, there may be no attempt whatsoever to control use.
Relatedly, different life circumstances may protect people more or less well against
impaired functioning (for a review, see Martin et al. 2014). For example, it is well known
that addiction is associated with low socio-​economic status alongside other mental health
problems (Compton et al. 2007; Heyman 2009). But, in so far as addiction is diagnosed
via negative consequences, wealth, alongside other forms of privilege, may offer a protec-
tive factor (cf. Matthews 2014; Schmidt et al. 2010). For example, a wealthy mother who
drinks heavily but can afford a live-​in nanny to ensure her children are adequately cared
for is able to meet more of her role-​related responsibilities than a poor woman who drinks
equal amounts but whose children go hungry and miss school. The consequences are
more serious by shared social standards in the latter case than the former, and so too, as a
result, is the likelihood of a diagnosis.
  33

Explaining the choice to use drugs despite negative consequences 33

The implication that individual differences in conceptions of how to live and life cir-
cumstances can affect the likelihood of a diagnosis may give pause.2 Indeed, Martin et al.
(2014) have proposed that the negative consequences of use should be considered ancil-
lary rather than core features of addiction for this very reason, namely, that they introduce
significant individual and context specificity. However once we acknowledge that drug
use in itself is not indicative of any form of disorder, but rather offers instrumental means
to fulfilling valuable ends, the idea that negative consequences are fundamental to the
pathological nature of addiction becomes evident. For it is only when the costs exceed
the benefits, and yet use continues, that there is reason to believe that anything is wrong.3
And, whether or not the costs exceed the benefits will, evidently, depend on individual
differences in conceptions of how to live and life circumstances.
Why then, if addicts are able to control their consumption and choose to abstain, do
they routinely choose instead to continue to use in the face of harrowing consequences?
We suggest there are four possible forms of explanation.

2.1  Addiction and self-​harm

Some addicts may choose to continue to use, not only despite harrowing consequences,
but because of them. We noted above that the assumption that people act in their own
best interests is a basic folk psychological rule of thumb. But there are exceptions, particu-
larly with respect to people from underprivileged backgrounds characterized by child-
hood adversity and mistreatment, and who may struggle with a range of mental health
problems associated with addiction, especially personality disorders.4 People with such
complex needs, and for whom shame, self-​hatred, and low self-​worth are strong within
their sense of self, may deliberately and directly self-​harm5—​most notably through self-​
directed violence, such as cutting and burning, but also by other means, such as sexual and
other forms of risk-​taking behavior, overdosing, and, arguably, drug abuse quite generally.
For some addicts, they may not care about themselves enough to care about the negative

2 Note, however, that this arguably is a feature of most diagnoses of mental disorders. For discussion see
Graham (2010); Glover (2014); cf. too Murphy (2015).
3 Note for clarity that we do not claim this point refutes a disease model of addiction that identifies addic-
tion with neurobiological dysfunction. On such a model, the pathology is wholly neurobiological: what
is wrong or disordered is the functioning of the addict’s brain. Rather, we understand this point as high-
lighting a counter-​intuitive commitment of such a model, namely, that in identifying addiction with
neurobiological dysfunction, it entails that addiction can in principle occur in absence of any personal-​
level negative consequences or harm—​in other words, individuals may count as addicted, on this model,
even when costs in no way exceed benefits and there is no evidence of a problem (cf. Levy 2013; the
model will also, of course, need to specify what the brain dysfunction is, and why it should count as such,
without appeal to any negative consequences or harm). It is open to adherents of a disease model to
embrace this commitment, but it is, arguably, revisionary of the concept of disorder in general, and the
construct of addiction in particular. For discussion, see Stephens and Graham (2009); Murphy (2015).
4 For discussion see Maté (2009); Pickard and Pearce (2013).
5 For discussion see Hawton et al. (2012); Motz (2009); Pickard (2015).
34  

34 How do you know you have a drug problem?

consequences of their drug use—​indeed, they may, both consciously and unconsciously,
embrace these consequences, in keeping with their sense of self as a person who is bad or
worthless, and so deserving of them. In such cases, the costs of drug use in effect count as
benefits from their perspective. Fundamentally, the problem lies less in the choice to use,
than in the self-​destructive mindset that this choice serves.

2.2  Addiction and psycho-​socio-​economic context

Some addicts may choose to continue to use, notwithstanding the harrowing conse-
quences, because the benefits nonetheless outweigh the costs given a realistic apprecia-
tion of their circumstances and the options available (Pickard 2012). As noted above,
the majority of addicts “mature out” in the late twenties and early thirties. Those for
whom addiction remains a chronic problem are typically people from underprivileged
backgrounds who also suffer from co-​morbid mental disorders, particularly anxiety,
mood, and personality disorders, and who of course must equally face the stigma,
stress, and other problems associated with long-​term poor mental health (Compton
et al. 2007; Regier et al. 1990) and lack of psychosocial integration (Alexander 2008).
The “self-​medication” hypothesis has long been a staple of clinical understanding of
drug use (Khantzian 1985, 1997; cf. Muller and Schumann 2011). It is common knowl-
edge that drugs offer relief from psychological distress:  we “reach for the bottle” or
“drown our sorrows” when in need. For many chronic addicts, drugs may provide
a habitual and, in the short-​term, effective way of managing psychological distress,
caused by negative emotions alongside other symptoms and problems typically expe-
rienced by people with mental health problems living in impoverished circumstances.
Put crudely, drugs and alcohol offer a way of coping with stress, pain, and misery when
there is little possibility for genuine hope or improvement. For addicts in such circum-
stances, the cost of abstinence is likely to be very high, for the benefits of drug use are
many, and the alternative goods available are few. In such cases, the problem lies less
in the choice to use than in the psycho-​socio-​economic circumstances that cause suf-
fering and limit opportunities.

2.3  Addiction and temporal discounting

Some addicts may choose to continue to use, because at the moment of choice, they value
drugs more than they value a possible but uncertain future reward, such as improved
wellbeing with respect to health, relationships, or opportunities, which is consequent on
long-​term abstinence. The disposition to discount the future relative to the present is a
common feature of human psychology, standardly considered rational to the extent that,
adjusting for the relative value of the rewards, the present reward is certain while the
future reward is uncertain. But in addition, human discount curves are typically hyper-
bolic, so that as a reward nears in time, its expected value increases sharply, creating
shifts in preferences over time simply in response to current availability (Ainslie 2001;
cf. Heyman 2009). Addicts have steeply hyperbolic discount rates compared to the norm
(Bickel and Marsch 2001; Bickel et al. 2014). When the drug is within immediate reach,
  35

Explaining the choice to use drugs despite negative consequences 35

its value skyrockets and addicts may at that point in time prefer use to abstinence, even
if, when the drug is not within reach, they value it less than the possibility of improved
wellbeing consequent on long-​term abstinence and so, at that point in time, prefer absti-
nence to use.
Ambivalence is characteristic of addiction. Addicts typically report fluctuating desires
and resolutions, alongside vacillating hope and despair, which lends a sense of psycho-
logical reality to hyperbolic discounting models. Moreover, the success of contingency
management treatment testifies to the role of discounting in explaining drug choices. It
is remarkable that a small amount of money or a prize can provide sufficient incentive
for addicts to forgo drugs, when the harrowing consequences of their addiction do not.
However, the money or prize is directly and reliably available upon the delivery of a clean
urine sample, on a fixed thrice-​weekly schedule. There is no significant delay in gratifica-
tion, and there is no significant uncertainty as to delivery. In comparison, the rewards
consequent upon abstinence are not only temporally delayed, but also, for many addicts,
extremely uncertain.
Unlike contingency management treatment rewards, the good life does not spring
forth readymade simply because an addict quits. There may be long-​term physical and
mental health problems which cannot be fixed simply through forgoing drugs. Equally,
ruined relationships do not just snap back into shape, communities do not quickly forget,
and jobs that were lost are not automatically regained. For those addicts who come from
underprivileged backgrounds of poor opportunity, housing, education, and employment
opportunities do not simply materialize overnight. The creation of a life worth living
requires work, and, for many addicts, the cards are stacked against them even if they
kick their addiction.
Moreover, for addicts with complex needs, a “suicide option” may function to rational-
ize the discounting of any possible future reward consequent on a drug-​free life, given the
cost of abstinence in the present. The option of committing suicide can be very important
to people who live with long-​term psychological distress, because it offers an escape that
lies within their control if life becomes unbearable (Pickard 2015). In so far as drugs offer
relief from suffering, the cost of abstinence is very high unless and until alternative means
of coping are available: the person must bear not only withdrawal and other drug-​related
effects of abstinence, but also the psychological distress which the drugs were function-
ing to ameliorate. If an addict is committed to the option of committing suicide if life
becomes unbearable, then it is not obviously worth suffering through abstinence now,
because if it is unbearable they will take the option, ensuring that there is no possible
future reward for suffering in the present.
Discounting models can explain why addicts choose to use despite the harrowing con-
sequences, whether or not the rates of discounting indicate an impairment:  the expla-
nation depends simply on the fact that, at the moment of choice, present rewards are
preferred to future rewards. But it is an open question as to whether, given the life cir-
cumstances and options realistically available to many addicts, it is irrational to employ a
narrow temporal horizon.
36  

36 How do you know you have a drug problem?

2.4  Addiction and knowledge of negative consequences

These three options explain why addicts choose to continue to use drugs even though
they know that doing so has negative consequences. Lastly, some addicts may choose to
use because they do not know that use has such consequences.
To make this vivid, consider as an initial example nicotine addiction during the period
of US history prior to the discovery that smoking causes disease, when some people may
have guessed, but nobody knew, that cigarettes were bad for you. Indeed, some people
may have believed that cigarettes were good for you, due to misleading advertising and
government lobbying by the industry (Kruger 1996). Of course, during this period, smok-
ing in fact caused disease. But given that people enjoyed smoking, there is no puzzle
as to why they did it despite these negative health consequences, as they did not know
about them.
From a clinical perspective, smokers during this period could, of course, be diagnosed
with a disorder, in so far as they developed tolerance, suffered withdrawal, and experi-
enced cravings and a sense of compulsion. But, given the general availability and accept-
ance of smoking within US culture, together with the nature of the effects of nicotine, they
are unlikely to have neglected other pleasures and activities; or to have experienced any
non-​health-​related serious negative consequences due to tobacco use; and, as they were
unaware of health-​related consequences, they did not persist in using despite knowledge
or clear evidence that use was a problem, and would have been unlikely to have tried to
quit at all, and so therefore would have been unlikely to have tried and failed. From a
biological perspective, we can of course wonder how any species can have evolved and
survived with such a strong penchant for things that are bad for them as appears to be the
case with us.6 But if we adopt the perspective of cognitive science and our folk psycho-
logical conception of ourselves, addiction only appears to be a disorder or to indicate that
something is impaired or wrong with an individual if they know or should know, relative
to the cognitive capacities that are standard in the species and the information available
to them as an individual, that their drug use has harrowing consequences and yet persist
in using.7
So, if you’re an addict, how do you know that you have a drug problem? One way or
another, you have to discover that you do. The fact that one’s drug use is causing negative
consequences is not immediately manifest through introspection, but requires acquiring
causal knowledge.
There are at least two kinds of causal knowledge, typically acquired by two corre-
sponding routes. On the one hand, there are large-​scale generalizations, such as the
knowledge that smoking causes disease. Acquiring knowledge of large-​scale generali-
zations typically depends on equally large-​scale collective research efforts involving
data collection and hypothesis testing and confirmation. For example, the causal link

6 For an evolutionary perspective on addiction, see Nesse and Berridge (1997); Hagen et al. (2013).
7 The implications of this claim for the interpretation of animal models and the construct of addiction
are discussed in section 4.
  37

Explaining the choice to use drugs despite negative consequences 37

between smoking and disease was established by extensive longitudinal comparisons

of smoking versus non-​smoking populations, and confirmatory evidence from animal
models. Once such large-​scale generalizations are known in the research community,
they can become known to the public at large via the dissemination of expert testi-
mony through channels such as the media and public education initiatives. Crucially,
as individuals, we do not have the ability to acquire knowledge of these large-​scale
generalizations on our own, simply by considering our own experience, but must rely
on scientific discovery and its dissemination. But once disseminated, this knowledge
is then available for use in individual decision-​making. The large-​scale generalization
that smoking causes disease allows individuals to infer that, if they smoke heavily, then
they are at risk of smoking-​related disease.
On the other hand, there are small-​scale individual generalizations, pertaining espe-
cially to our actions and their outcomes. We can often acquire this knowledge on the
basis of our own experience alone. If we observe an association between two events—​such
as an action and an outcome—​we can test the possibility of a causal relation by inter-
vening and manipulating the hypothesized cause while monitoring the effect of doing
so. For example, although we cannot discover that smoking causes disease on our own,
we can potentially discover that, in our own case, smoking causes headache. We can do
this by first noticing the association and then testing the hypothesis by controlling our
actions: smoke a cigarette, then observe the effects; don’t smoke, then observe the effects.
Once this causal knowledge is acquired, it can be used in individual decision-​making,
allowing us to achieve outcomes by means of interventions such as our own actions. So,
armed with the knowledge that, in one’s own case, smoking causes headache, one can
choose not to smoke, so as to avoid headache.
Causal knowledge of the negative consequences of drug use can be difficult to acquire.
With respect to large-​scale generalizations such as health risks, individuals are not only
dependent on scientific discovery and dissemination, but must also be willing to trust
the testimony available, as well as to understand and assess probabilities and risk. With
respect to small-​scale individual generalizations, one’s experience may not offer clear con-
firmation. Given that the causal network of relations is likely complicated and thickly
interwoven, and drugs may well be contributory as opposed to single causes, interven-
tions and manipulations may not yield knowledge. As noted above, when addicts quit
using drugs, their problems, including those which were initially caused or exacerbated
by drug use, do not simply disappear. Indeed, things may get worse before they get better,
as they struggle to abstain and live without drugs. So an intervention—​forgoing drugs—​
may not produce the effect—​the disappearance of negative consequences of use—​which
would support the acquisition of knowledge of a causal relationship between them. To
take a well-​worn kind of example (cf. Segal 2013): if a person’s drinking is contributing
to ruining their marriage, but they are drinking in part because they are unhappy in their
marriage, then stopping drinking will neither immediately fix the marriage, nor corre-
spondingly offer clear evidence that drinking is a cause of the problem. In other words,
acquiring causal knowledge of the negative consequences of drug use should be seen as an
38  

38 How do you know you have a drug problem?

achievement. It can be hard to come by, and the evidence supporting it may be equivocal.
Perhaps it is no surprise, then, if it is also a form of knowledge particularly open to denial.
Denial has received surprisingly little attention in the philosophical and scientific lit-
erature.8 Within 12-​step programmes it is more central, understood as the unwarranted
belief that one can moderate use. Addicts may cling to such a belief, perhaps to avoid
facing complete abstinence. But there is only reason to moderate or give up using if it
is the cause of negative consequences. And denial can equally be directed at this causal
knowledge. Consider, as an example, the following first-​person report:
When I was about 8 I developed a keen desire to take drugs. This was in the 1960s, so drugs were all
over the media, movies, news reports etc. I started on solvents aged about 11, moved on to various
drugs at 14, took them until 20, when I switched to alcohol. I then drank too much more or less
every day. I did not until the age of 42 admit that I might be taking any serious risks at all with my
health. [But] my denial set in before I had even tried a drug or a drink. Rock stars would appear
on TV and say “Don’t do this! I tried it and it wrecked my life” sort of thing. Didn’t bother me. Yet
I was not a particularly foolhardy kid. I didn’t take abnormal risks in other areas. But I flew straight
into drugs, and into hard drugs. Not shooting up, but anything short of that, cocaine and speed at
15 years of age, without a thought of risk. So what was going on in me at 8, 9 years of age? There may
have been a very strong desire to escape reality … which motivated both taking drugs to escape and
denial itself … denial being an escape from reality.

(Anonymous addict, personal communication)

Denial pushes knowledge of the harrowing consequences of drug use to the side, thereby
serving to explain how addicts choose to continue to use. Even when there is clear evidence,
whether available through testimony or through experience, which should provide knowl-
edge and correspondingly incentive to abstain, addicts deny it, thereby removing the incen-
tive. With respect to large-​scale generalizations, they may distrust the testimony or, as in the
above first-​person report, deny that these apply in their case, suggesting belief in a sort of
personal exceptionalism, or, perhaps, omnipotence. With respect to small-​scale individual
generalizations, they may deny the causal consequences of drug use tout court: for example,
they deny the impact of drinking on their marriage, despite its obviousness to others.
The question this explanation raises, of course, is how it is possible to be in denial,
given the testimonial or experiential evidence available. How can addicts not know what
is staring them in the face? From a folk-​psychological perspective, the natural suggestion,
of course, is that they don’t want to know. They don’t want to know because they want
to continue to use drugs—​whether that is because of the strength of their motivation
or craving (cf. Robinson and Berridge 2008; Segal 2013); the pleasure of using (Foddy
and Savulescu 2006, 2010); the immediate functional value of drugs in their present cir-
cumstances (Pickard 2012); or some other reason. Perhaps, too, they don’t want to know
because of the shame such knowledge typically brings (Flanagan 2013). The key point

8 There are exceptions: Crowther (Chapter 5, this volume), Flanagan (2011) and Segal (2013, Chapter 20,
this volume); cf. too Moeller and Goldstein (2014, Chapter  11, this volume), who propose a self-​
awareness impairment in addiction which may connect to some forms of denial. The nature of denial
in addiction is further explored in Pickard (2016).
  39

IGNORANCE OF NEGATIVE CONSEQUENCES EXPLAINS PATTERNS OF DRUG CHOICES IN RATS 39

is that, according to this suggestion, denial is a drug-​related species of more standard

forms of epistemically irrational belief, such as self-​deception, motivated or biased belief,
or outright delusion. We do not at present have a clear and agreed understanding of any
of these phenomena individually, or the connections between them.9 Equally, there are
questions about whether people who suffer from these irrational beliefs at least some-
times do know, in some sense, or at some level, whatever truth they also at one and the
same time deny.10 The point we wish to emphasize here is that an understanding of this
broad set of phenomena may also throw light on addiction, in so far as denial is central to
why some addicts continue to use despite negative consequences.
As we saw above, lack of knowledge that drug use has negative consequences need not
indicate irrationality, if such knowledge is either straightforwardly not available or the
evidence presents a complicated picture. In cases of denial, however, it is epistemically
irrational. The evidence is overwhelming that drug use is destroying one’s life, and yet one
does not believe it. If this is right, then rather than a disease of compulsion, addiction may
be a disorder of cognition. What is impaired or wrong with at least some addicts is that, as
Segal (2013) has vividly put it, they are insane: given the evidence, they should know that
their drug use has harrowing consequences, but they persist in believing it doesn’t.
These four explanations are not mutually exclusive. Different explanations, combining
different factors, may be true of different addicts. But with this broad picture in view,
we now turn to recent studies on the nature of drug choice in addicted rats. Strikingly,
these studies in many ways mirror drug choice in humans. In particular, we suggest that
they help illuminate the role of ignorance of the negative consequences of addiction in
explaining patterns of drug choice that are both destructive and do not reflect the indi-
vidual’s actual preferences. In other words, they help illuminate the core and insufficiently
appreciated role played by denial.

3  Ignorance of negative consequences explains

self-destructive patternsof drug choices in rats
Rats are the most common non-​human animals used in experimental addiction research
(Ahmed 2012). They share with us important behavioral and neurobiological traits, includ-
ing liking and wanting many addictive substances (Nesse and Berridge 1997; Panksepp
and Biven 2012; Siegel 1989), but there are also important differences (Suddendorf 2013).
Notably, unlike humans, rats have a poorly developed prefrontal cortex—​a brain region
critically involved in foresight, planning, and decision-​making—​and limited ability to
know the consequences of their own choices (Ahmed 2005, 2008).
Obviously, knowledge of large-​scale generalizations is inaccessible to rats as a species.
However, less trivially, and in striking contrast to humans, rats are also limited in their

9 For discussion see Bayne and Fernandez (2009); Bortolotti (2010); Mele (2001); Pickard (2016).
10 Thisdescription may be especially apt when addicts who have been abstinent relapse. For discussion,
see Crowther (Chapter 5, this volume); Pickard (2016).
40  

40 How do you know you have a drug problem?

ability to learn small-​scale individual generalizations from their own experience about
the negative consequences of use. Above, we gave the example of knowing that smoking
causes headache, but we can easily multiply this example: there are countless situations
where non-​addicted human users abstain from taking a drug because they have learned
from experience that using brings unwanted consequences. For instance, occasional
alcohol drinkers may come to know from experience (and not simply via collective wis-
dom and testimony) that drinking excessively at night causes a hangover in the morning,
impacting on the capacity to fulfill social and occupational roles. Based on this knowl-
edge, they may opt not to drink excessively during the working week, thereby avoiding
these negative consequences. In contrast, rats are oblivious to the negative consequences
of drug choices. This ignorance explains why, in certain conditions, rats are vulnerable
to developing patterns of drug use that cause harm and even death. In other words, rats
provide a “proof-​of-​concept” that continued drug use despite harrowing consequences
can be explained by lack of causal knowledge.
Before proceeding, it is helpful to summarize what is known about the factors influ-
encing drug choices in rats. First, when provided with ready access to a drug for self-​
administration, and without access to alternative rewarding options, most rats will
self-​administer most drugs that can lead to addiction in humans, including cocaine,
heroin, or methamphetamine. They will also escalate their drug use if given sufficient
daily availability (Ahmed and Koob 1998)  and, as a result, eventually exhibit behavio-
ral changes resembling some of the diagnostic criteria of addiction (Ahmed 2012). For
instance, once their use of cocaine has escalated, most rats will expend more time, effort,
and cost to gain access to the drug (Ahmed 2012). Finally, post-​escalation, most rats will
also become more responsive to drug-​primed reinstatement of drug seeking after extinc-
tion, a widely studied model of drug craving (Shaham et al. 2003).
Second, even after escalation of cocaine use, rats nevertheless retain the ability to
quit or reduce intake when offered a valuable non-​drug alternative (Ahmed 2005,
2010, 2012). This is particularly well illustrated in a series of experiments where rats
were given a choice between pressing a lever to self-​administer cocaine or a different
lever to drink water sweetened with saccharin—​a potent, albeit non-​essential for sur-
vival, non-​drug option. An equal level of effort was required on both levers and choice
was either/​or to incite rats to express their preferences. Choice trials were also suffi-
ciently spaced in time to measure preferences while rats were not under the influence
of drugs. Faced with this choice, most rats quit or reduce their use of cocaine in favour
of sweet water (Ahmed 2005, 2010, 2012). Preference for sweet water is observed even
after escalation of cocaine use and development of a robust sensitization. Perhaps
more surprisingly, preference for sweet water cannot be surmounted by maximally
increasing the dose of cocaine. Sweet-​water preferring rats only shift their choice to
cocaine when the concentration of sweet water is very low, or when the effort required
to obtain it is very high. Overall, these findings demonstrate that, even after escalation
of use, rats retain the ability to forgo cocaine when offered an immediate and preferred
alternative pursuit.
  41

IGNORANCE OF NEGATIVE CONSEQUENCES EXPLAINS PATTERNS OF DRUG CHOICES IN RATS 41

However, third, at first glance, these findings conflict with other seminal research on
rats (Bozarth and Wise 1985; Dworkin et al. 1995; Fitch and Roberts 1993). In these stud-
ies, rats were given unlimited daily access to cocaine during several weeks in their home
cage-​environment, where they also had unlimited access to food and water. Most rats
consumed cocaine almost exclusively, and eventually to the point of death (presumably
caused by starvation, dehydration, and exhaustion).
There is therefore a puzzle with respect to how to resolve the findings of these two series
of experiments. In one series of experiments, rats quit cocaine for another reward that
had no immediate survival value; while in another series of experiments, rats continued
to take cocaine despite fatal consequences. It is by attempting to resolve this apparent
discrepancy that the importance of the fact that rats are oblivious of the negative conse-
quences of their own drug choices on other important (and indeed preferred) behaviors
becomes apparent.
Consider more fully a detail of the experimental procedure that was only briefly
alluded to above. In the studies where the rats quit cocaine for sweet water, choice trials
were sufficiently spaced in time for the direct anorexic effects of cocaine on motivation
to consume sweet reward to be avoided. In contrast, no such precaution was taken in
the studies where the rats continued to take cocaine to the exclusion of food and water.
Thus, in the latter series of studies, nothing prevented the anorexic effects of an initial
cocaine choice spilling over to subsequent choices and motivationally biasing the rats
against their normally preferred sweet reward, and in favour of more cocaine use. Put
differently, the first series of studies used a setting where choosing under the anorexic
influence of cocaine was prohibited, while the other series of studies used a setting that
allowed this influence.
This difference in choice setting can in principle explain the discrepancy in findings.
To directly confirm this hypothesis, the same rats were tested in these two different
choice settings (Vandaele et al. 2016). As expected, in the setting where choosing under
the anorexic influence of cocaine was prohibited, virtually all rats largely preferred
sweet water over cocaine, and quit cocaine. In contrast, when the same rats were shifted
to the setting where choosing under the anorexic influence of cocaine was possible,
their choices shifted from sweet reward almost exclusively to cocaine after only the
first few cocaine choices. This shift to exclusive cocaine choices was even observed in
hungry rats offered a nutritive sweet solution containing sugar. This shift was not due to
mere satiation for sweet water or to a direct behavioral competition or conflict between
operant responding for cocaine and operant responding for sweet water. Finally and
strikingly, note that this shift to exclusive drug use was not seen in rats tested in the
same choice setting but offered heroin which, unlike cocaine, has no anorexic effects
(Vandaele et al. 2016).
Thus, once rats make an initial cocaine choice in a setting where choosing under the
anorexic influence of cocaine is possible, it is as if they enter a locked-​in pattern of exclu-
sive drug choices from which they seem unable to escape, even after several hours. In such
a setting, the anorexic effects of a cocaine choice automatically spill over to subsequent
42  

42 How do you know you have a drug problem?

choices, creating a motivational bias against sweet water—​a pattern which is then end-
lessly repeated.
Importantly, rats are not only unable to escape from this locked-​in pattern once in it, but
are also unable to learn to avoid it through experience by refraining from making an initial
cocaine choice. With repeated testing, the pattern persists, despite the fact that it leads
to suppression of consumption of sweet reward, their normally preferred option. In fact,
the pattern persists even under optimal learning conditions. Briefly, rats were offered two
short periods of access to sweet water per session: directly before and after cocaine self-​
administration. In theory, by comparing sweet water consumption during these two peri-
ods of access, an animal should eventually learn to associate cocaine self-​administration
with suppression of sweet reward and thus refrain from initiating or at least reduce cocaine
self-​administration. However, this did not happen. Although cocaine self-​administration
repeatedly suppressed sweet water consumption during the second period of access, rats
nevertheless continued to self-​administer cocaine (Vandaele et al. 2016).
In sum, rats behave as if they are oblivious to the anorexic influence of their own cocaine
choices on other competing activities. The origin of this “irrational” behavior requires
further study, but presumably reflects rats’ limited ability to know the consequences of
their own choices: to foresee the negative influence that taking a drug now will have on
other competing, even preferred behaviors, in the near future.
To be clear, we do not deny that rats can learn to associate drug taking with some
immediate negative consequences. They clearly can (Krasnova et  al. 2014; Panlilio
et al. 2003). For instance, rats will eventually stop pressing a lever to self-​administer
a drug if immediately punished with a sufficiently high-​intensity painful event, such
as a foot-​shock. Similarly, we do not deny that rats can pursue a short-​term goal, tak-
ing the immediate consequences of their choices and anticipated events into account
(Dickinson and Balleine 1994). However, the time horizon in all known instances where
they display this capacity never exceeds a few seconds. This is too narrow a horizon
to encompass the influence of drug taking on other competing behaviors, which are
more significantly delayed in time (cf. Roberts 2002; Suddendorf and Corballis 2007).
Rats’ ignorance of the influence of drug taking on other competing behaviors makes
them vulnerable to harm themselves in certain choice settings. As explained, this occurs
when they are given unlimited access to cocaine for self-​administration in their home
cage environment, where they also have unlimited access to food and water (Bozarth
and Wise 1985; Dworkin et al. 1995; Fitch and Roberts 1993). Most rats self-​administer
cocaine to the exclusion of food and water, and eventually die within three weeks. The
only way to protect rats from developing this pattern of exclusive drug choices seems to
be through an outside intervention aimed at restructuring the choice setting to prevent
choices being made under the influence of the drug.
Rats’ behavior illustrates how a lack of knowledge can drive continued drug use despite
harrowing consequences. Because rats cannot know that drug use carries such conse-
quences, they do not even attempt to discontinue use. They continue to use despite the
fact that doing so means they forgo the reward which, given a forced choice, they prefer
  43

From rats to humans: the role of knowledge of negative consequences 43

to drugs—​and, even more, despite the fact that, given the biological importance of the
forgone reward to survival, they eventually die.

4  From rats to humans: the role of knowledge

of negative consequencesin understanding addiction
Rats cannot come to know on the basis of their experience that their drug use causes nega-
tive consequences, given their limited cognitive capacities. Human addicts, in contrast,
can know this when there is sufficient evidence—​whether testimonial or experiential.
Rats therefore represent what we might think of as a maximally idealized case of denial.
When in deep denial, we can understand human addicts as locked in to a pattern of drug
use despite negative consequences, similarly to rats. Hence one lesson from rat models of
addiction points, paradoxically, to the importance of capitalizing in developing treatment
on the human capacity that is precisely lacking in rats, namely, the capacity for knowledge
that drug use is causing harrowing consequences.
It is important to acknowledge that many effective forms of treatment do not capitalize
on this capacity, but rather aim to restructure human addicts’ choice settings, e.g. con-
tingency management treatment, or environmental restructuring to avoid drug-​related
triggers and cues. Arguably, we can also see social, education, and employment initiatives
designed to provide better life opportunities to addicts in a similar light as aiming to com-
pete with or crowd out drug choices. The value and importance of these various forms
of intervention is unquestionable. But humans also have an additional resource to help
overcome addiction: the capacity to know the negative consequences of drug use and to
use this knowledge in making drug choices. Some forms of intervention already capital-
ize on this. Public education initiatives aim to disseminate large-​scale general knowledge
of the risks surrounding drug use. Motivational interviewing, some forms of counselling,
and 12-​step programmes alike aim to help addicts recognize and face the small-​scale indi-
vidual consequences of drug use, to motivate change. Indeed, there is evidence that many
addicts who quit do so after an “awakening” process, which may be slow and erratic, but
which eventually culminates in recognition of the impact of drug use on their lives (cf.
Raab and Brown 2010). However, there is an urgent need to better understand the nature,
mechanisms, and ways to overcome the various kinds and degrees of denial present in
human addiction, to develop effective interventions targeting denial and appropriate to
different phases of the trajectory toward addiction. We should not cleave to the myth that
addicts must hit “rock bottom” before they are able to overcome denial.
However, a second lesson from rat models, equally paradoxical, is that it is unclear
whether or not rats can be addicted at all. Put otherwise, comparing rats and humans
reveals that there is an ambiguity inherent in the current construct of addiction.
The ambiguity concerns whether or not addiction involves drug use despite negative
consequences (as the orthodox conception of addiction suggests) or drug use despite
knowledge or sufficient evidence of negative consequences (as the diagnostic criteria
employed in the DSM-​5 and the ICD-​10 suggest). On the former understanding, rats
44  

44 How do you know you have a drug problem?

can be addicted, as evidenced in settings where they choose to use cocaine to the point of
death. However, if this is how to understand addiction, then it is unclear why addiction
is a disorder. For, as we stressed above, it is only if an individual knows or should know,
relative to the cognitive capacities that are standard in their species and the information
available to them as an individual, that their drug use has harrowing consequences and
yet persists in using, that there is reason to think anything is impaired or wrong with them.
If rats cannot know, given the cognitive limitations of their species, that the choice to take
cocaine will influence future choices, causing them to forgo rewards which they prefer
to cocaine and which are necessary to survival, then it is hard to argue that anything is
disordered or impaired in their choice. They both like and want cocaine, so they take it,
ignorant of the consequences of this choice due to species-​limitations, and hence through
no individual impairment due to drug use. On the other hand, on the latter understand-
ing, whereby addiction involves drug use despite knowledge or sufficient evidence of nega-
tive consequences, addiction is clearly a disorder, but rats cannot be addicted despite their
propensity to use drugs to the point of death. For, given their species limitations, they do
not use drugs despite knowledge or sufficient evidence of negative consequences, for they
have none.
We do not propose to answer this question about the construct of addiction here, but
rather simply to raise it as one of the new puzzles to emerge out of choice models of
addiction. The evidence is very strong that addiction involves choice, and we hope to
have made progress toward properly addressing the question of why, given the harrowing
consequences of drug use, addicts persist in using when they could instead abstain. The
first three explanations we sketched appeal to how addicts may represent and weigh the
costs and benefits of use; the fourth explanation appeals to denial and the point that, if
one doesn’t know that an action brings with it costs, then one will not weigh those costs
in choosing whether or not to act. According to this explanation, human addicts who
are in denial are epistemically irrational—​subject less to a disease of compulsion than a
disorder of cognition. They should know their drug use is having a harrowing impact, but
they do not. But to be epistemically irrational requires an individual to have the capacity
to be epistemically rational. Whether or not this is central to the construct of addiction is
an issue calling out for further research.11

References
Ahmed, S.H. and Koob, G.F. (1998). Transition from moderate to excessive drug intake: an elevation in
hedonic set point. Science, 282, 298–​300.
Ahmed, S.H. (2005). Imbalance between drug and non-​drug reward availability: a major risk factor for
addiction. European Journal Pharmacology, 526, 9–​20.
Ahmed, S.H. (2008). The origin of addictions by means of unnatural decision. Behavioral and Brain
Sciences, 31, 437–​38.

11 Weare grateful to Bennett Foddy, Cécile Fabre, Owen Flanagan, Nick Heather, Ian Phillips, and Gabe
Segal for comments and discussion.
  45

From rats to humans: the role of knowledge of negative consequences 45

Ahmed, S.H. (2010). Validation crisis in animal models of drug addiction: beyond non-​disordered drug
use toward drug addiction. Neuroscience Biobehavioral Review, 35, 172–​84.
Ahmed, S.H. (2012). The science of making drug-​addicted animals. Neuroscience, 211, 107–​25.
Ainslie, G. (2000). A research-​based theory of addictive motivation. Law and Philosophy, 19, 77–​115.
Ainslie, G. (2001). Breakdown of Will. New York: Cambridge University Press.
Alexander, B.K. (2008). The Globalization of Addiction. New York: Oxford University Press.
Alexander, B.K., Coambs, R.B., and Hadaway, P.F. (1978). The effect of housing and gender on
morphine self-​administration in rats. Psychopharmacology, 58(2), 175–​79.
Alexander, B.K., Peele, S., Hadaway, P.F., Morse, S.J., Brodsky, A., and Beyerstein, B.L. (1985). Adult,
infant, and animal addiction. In: S. Peele (ed.), The Meaning of Addiction, Lexington, MA: Lexington
Books, pp. 77–​96.
American Psychiatric Association (APA) (2013). Diagnostic and Statistical Manual of Mental Disorders
(5th ed.). Washingon DC: APA.
Bayne, T. and Fernandez, J. (eds) (2009). Delusion and Self-​Deception. New York: Psychology Press,
Taylor and Francis Group.
Bickel, W.K. and Marsch, L. A. (2001). Toward a behavioral economic understanding of drug
dependence: delay discounting processes. Addiction, 96(1), 73–​86.
Bickel, W.K., Koffarnus, M.N., Moody, L., Wilson, A.G. (2014). The behavioral-​and neuro-​economic
process of temporal discounting: a candidate behavioral marker ofaddiction. Neuropharmacology,
76, 518–​27.
Bortolotti, L. (2010). Delusions and Other Irrational Beliefs. Oxford: Oxford University Press.
Bozarth, M.A. and Wise, R.A. (1985).Toxicity associated with long-​term intravenous heroin and
cocaine self-​administration in the rat. JAMA, 254, 81–​83.
Charland, L. (2002). Cynthia’s dilemma: consenting to heroin prescription. American Journal of
Bioethics, 2(2), 37–​47.
Compton, W.M., Thomas, Y.F., Stinson, F.S., and Grant, B.F. (2007). Prevalence, correlates, disability,
comorbidity of DSM-​IV drug abuse and dependence in the United States: results from the national
epidemiologic survey on alcohol and related conditions. Archives of General Psychiatry, 64(5),
566–​76.
Dickinson, A. and Balleine, B.W. (1994). Motivational control of goal-​directed action. Animal Learning
and Behavior, 22, 1–​18.
Dworkin, S.I., Mirkis, S., and Smith, J.E. (1995). Response-​dependent versus response-​independent
presentation of cocaine: differences in the lethal effects of the drug. Psychopharmacology (Berl),
117, 262–​6.
Elliott, C. (2002). Who holds the leash? American Journal of Bioethics, 2(2), 48.
Fitch, T.E. and Roberts, D.C. (1993). The effects of dose and access restrictions on the periodicity of
cocaine self-​administration in the rat. Drug Alcohol and Dependence, 33, 119–​28.
Flanagan, O. (2011). What is it like to be an addict? In: J. Poland and G. Graham (eds), Addiction and
Responsibility. Cambridge, MA: MIT Press, pp. 269–​92.
Flanagan O. (2013). The shame of addiction. Frontiers in Psychiatry, 5, 120.
Foddy, B. and Savulescu, J. (2010). A liberal account of addiction. Philosophy, Psychiatry, and
Psychology, 17(1), 1–​22.
Foddy, B. and Savulescu, J. (2006). Addiction and autonomy: can addicted people consent to the
prescription of their drug of addiction? Bioethics, 20(1), 1–​15.
Glover, J. (2014). Alien Landscapes? Cambridge, MA: Harvard University Press.
Graham, G. (2010). The Disordered Mind. London: Routledge.
46  

46 How do you know you have a drug problem?

Hagen, E.H., Roulette, C.J., and Sullivan, R.J. (2013). Explaining human recreational use of
“pesticides”: the neurotoxin regulation model of substance use vs. the hijack model and implications
for age and sex differences in drug consumption. Frontiers in Psychiatry, 4, 142.
Hart, C.L. (2013). High Price. New York: Harper Collins.
Hart, C.L., Haney, M., Foltin, R.W., and Fischman, M.W. (2000). Alternative reinforcers
differentially modify cocaine self-​administration by humans. Behavioural Pharmacology,
11, 87–​91.
Hawton, K,. Saunders, K.E.A., and O’Connor, R.C. (2012). Self-​harm and suicide in adolescents.
Lancet, 379, 2373–​82.
Heyman, G. (2009). Addiction: A Disorder of Choice. Cambridge, MA: Harvard University Press.
Heyman, G. (2013). Quitting drugs: quantitative and qualitative features. Annual Review of Clinical
Psychology, 9, 29–​59.
Khantzian, E.J. (1985). The self-​medication hypothesis of addictive disorders: focus on heroin and
cocaine dependence. American Journal of Psychiatry, 142, 1259–​64.
Khantzian, E.J. (1997). The self-​medication hypothesis of substance use disorders: a reconsideration
and recent applications. Harvard Review of Psychiatry, 4(5), 231–​44.
Koob, G.F. and Le Moal, M. (1997). Drug abuse: hedonic homeostatic dysregulation. Science,
278, 52–​58.
Koob, G.F. and Volkow, N.D. (2010). Neurocircuitry of addiction. Neuropsychopharmacology, 35,
217–​38.
Krasnova, I.N., Marchant, N.J., Ladenheim, B., McCoy, M.T., Panlilio, L.V., Bossert, J. M., et al.
(2014). Incubation of methamphetamine and palatable food craving after punishment-​induced
abstinence. Neuropsychopharmacology, 39, 2008–​16.
Kruger, R. (1996) Ashes to Ashes: America’s Hundred-​Year Cigarette War, the Public Health and the
Unabashed Triumph of Philip Morris. New York: Vintage Books, Random House.
Levy, N. (2013). Addiction is not a brain disease (and it matters). Frontiers in Psychiatry, 4, 24.
doi:10.3389/​fpsyt.2013.00024.
Martin, C.S., Langenbucher, J.W., Chung, T., and Sher, K.J. (2014). Truth or consequences in the
diagnosis or substance use disorders. Addiction, 109, 1773–​78. doi:10.1111/​add.12615.
Maté, G. (2009). In the Realm of Hungry Ghosts: Close Encounters with Addiction. Toronto: Vintage
Canada.
Matthews, S. (2014). Addiction, competence, and coercion. Journal of Philosophical Research, 39,
199–234. doi: 10.5840/​jpr20147214.
Mele, A. (2001). Self-​Deception Unmasked. Princeton, NJ: Princeton University Press.
Moeller, S.J. and Goldstein, R.Z. (2014). Impaired self-​awareness in human addiction: deficient
attribution of personal relevance. Trends in Cognitive Science, 18(12), 635–​41.
Montague, P.R., Hyman, S.E., and Cohen, J.D. (2004). Computational roles for dopamine in behavioral
control. Nature, 431, 760–​67.
Motz, A. (ed.) (2009). Managing Self-​Harm. London: Routledge.
Muller, C.P. and Schumann, G. (2011). Drugs as instruments: a new framework for non-​addictive
psychoactive drug use. Behavioural and Brain Sciences, 34(6), 293–​310.
Murphy, D. (2015). Concepts of disease and health. In: E.N. Zalta (ed.), The Stanford Encyclopedia of
Philosophy. Spring. http://​plato.stanford.edu/​archives/​spr2015/​entries/​health-​disease/​. Accessed
03/​04/​2016.
Nesse, R.M. and Berridge, K.C. (1997). Psychoactive drug use in evolutionary perspective. Science,
278, 63–​66.
  47

From rats to humans: the role of knowledge of negative consequences 47

Panksepp, J. and Biven, L. (2012). The Archeology of Mind: Neuroevolutionary Origins of Human

Emotions. New York: W.W.Norton and Company.
Panlilio, L.V., Thorndike, E.B., and Schindler, C.W. (2003). Reinstatement of punishment-​
suppressed opioid self-​administration in rats: an alternative model of relapse to drug abuse.
Psychopharmacology (Berl), 168, 229–​35.
Peele, S. (1985). The Meaning of Addiction. Lanham, MD: Lexington Books.
Petry, N.M., Alessi, S.M. and Rash, C.J. (2011). Contingency management treatment of drug and
alcohol use disorders. In: J. Poland and G. Graham (eds), Addiction and Responsibility. Cambridge,
MA: MIT Press, pp. 225–​45.
Pickard, H. (2012). The purpose in chronic addiction. American Journal of Bioethics Neuroscience,
3(2), 30–​39.
Pickard, H. (2015). Choice, deliberation, violence: mental capacity and criminal responsibility in
personality disorder. International Journal of Law and Psychiatry, 14, 15–​24.
Pickard, H. (2016) Denial in addiction. Mind & Language, 31(3), 277–99.
Pickard, H. and Pearce, S. (2013). Addiction in context: philosophical lessons from a personality
disorder clinic. In: N. Levy (ed.), Addiction and Self-​Control: Perspectives from Philosophy,
Psychology, and Neuroscience. Oxford: Oxford University Press, pp. 165–​84.
Raab, D.M. and Brown, J. (2012). Writers On The Edge: 22 Writers Speak about Addiction and
Dependency. Ann Arbor: Modern History Press.
Redish, A.D., Jensen, S., and Johnson, A. (2008). A unified framework for addiction: vulnerabilities in
the decision process. Behavioural Brain Science, 31, 415–​37.
Regier, D.A., Farmer, M.E., Rae, D.S., Locke, B.Z., Keith, S.J., Judd, L., and Frederick, K.G.
(1990) Comorbidity of mental disorders with alcohol and other drug abuse. Results from the
epidemiological catchment area (ECA) study. JAMA, 264, 2511–​18.
Roberts, W.A. (2002). Are animals stuck in time? Psychological Bulletin, 128, 473–​89.
Robinson, T.E. and Berrdige, K.C. (2008). The incentive sensitization theory of addiction: some current
issues, Philosophical Transactions of the Royal Society of London, Series B, Biological Sciences, 363,
3137–​46.
Schmidt, L., Makela, P., Rehm, J., and Room, R. (2010). Alcohol: equity and social determinants.
In: I. Blas and A. Durup (eds), Equity, Social Determinants, and Public Health Programmes.
Geneva, Switzerland: World Health Organization, pp. 11–​29.
Segal, G. (2013). Alcoholism, disease, and insanity. Philosophy, Psychiatry, and Psychology, 20(4),
297–​315.
Shaham, Y., Shalev, U., Lu, L., De Wit, H., and Stewart, J. (2003). The reinstatement model
of drug relapse: history, methodology and major findings. Psychopharmacology (Berl),
168, 3–​20.
Shanks, D. (2004). Judging covariation and causation. In: D. Koehler and N. Harvey (eds), Blackwell
Handbook of Judgment and Decision Making. Oxford: Blackwell, pp. 220–​39.
Siegel, R. (1989). Intoxication: Life in Pursuit of Artificial Paradise. New York: E.P. Dutton.
Stephens, G.L and Graham, G. (2009). An addictive lesson: a case study in psychiatry as cognitive
neuroscience. In: L. Bortolotti and M. Broome (eds), Psychiatry as Cognitive Neuroscience.
Oxford: Oxford University Press, pp. 203–​20.
Suddendorf, T. (2013). The Gap: The Science of What Separates Us from Other Animals.
New York: Basic Books.
Suddendorf, T. and Corballis, M.C. (2007). The evolution of foresight: what is mental time travel, and is
it unique to humans? Behavioural Brain Science, 30, 299–​313.
48  

48 How do you know you have a drug problem?

Vandaele, Y., Cantin, L., Serre, F., Vouillac, C., and Ahmed, S.H. (2016). Choosing under the
influence: a drug-​specific mechanism by which the setting controls drug choices in rats.
Neuropsychopharmacology, 41(2), 646–​57.
Wakefield, J.C. and Schmitz, M.F. (2014). How many people have alcohol use disorders? Using the
harmful dysfunction analysis to reconcile prevalence estimates in two community surveys. Frontiers
in Psychiatry, 5, 10. doi:10.3389/​fpsyt.2014.00010.
World Health Organisation (WHO) (1992) ICD-​10 Classifications of Mental and Behavioral
Disorder: Clinical Descriptions and Diagnostic Guidelines. Geneva, Switzerland: World Health
Organization.
Zernig, G., Kummer, K.K., and Prast, J.M. (2013). Dyadic social interaction as an alternative reward to
cocaine. Front Psychiatry, 4, 100.
  49

Chapter 3

Addiction: The pleasures and perils

of operant behavior
Bennett Foddy

Abstract
This chapter compares the epidemiological and neuroscientific data on
different addictions and addiction-like syndromes, from drug addiction to
binge-eating disorders, gambling, and videogame addiction. It considers
the various neurological and behavioral differences that can seem to
differentiate these different behavioral syndromes, and it argues that these
differences are not essential to the underlying behavioral condition that
unifies various “addictive” behaviors. Based on these data, it is argued that
there is a hazard inherent in any rewarding operant behavior, no matter
how apparently benign: that we may become genuinely “addicted” to any
behavior that provides operant reward. With this in mind, addiction is rightly
seen as a possibility for any human being, not a product of the particular
pharmacological or technological properties of any one particular substance
or behavior.

1 Introduction
We use the terms “addict” and “addiction” in everyday speech over an extremely wide
range of domains. One may be addicted to alcohol or to cocaine, of course, but we also
speak of caffeine addiction, addiction to a television show, and addiction to eBay. We refer
to ourselves and others as workaholics, chocoholics, or shopaholics.
It is sometimes suggested that we use these everyday ascriptions of addiction much
more loosely than we do when we say that someone is an alcoholic, or that they are
addicted to heroin. I am not so sure. The scientific question of what addiction is—​and in
particular the question of whether it can genuinely arise around Internet use, or choco-
late, or work—​is not settled in the least; not among philosophers, nor psychiatrists, nor
neuroscientists (Foddy and Savulescu 2006, 2010; Foddy 2010, 2011; Gearhardt et  al.
2011; Sim at al. 2012).
In order to support a narrow definition of addiction, it would need to be the case that
there is something special about the cases of addiction that we all can agree are “genuine
addictions”—​such as heroin addiction, alcoholism, and perhaps problem gambling—​that
50  

50 Addiction: the pleasures and perils of operant behavior

differentiates them from “folk addictions” like chocoholism or television addiction. There
could be something unique about the natural history of genuine addictions, or they could
have unique behavioral effects, or essential, unique biological processes that undergird
them. If any of those things were true, then we could use those factors to distinguish
between genuine and folk addictions.
In this chapter, however, I will argue that the evidence points to the opposite conclu-
sion:  that the very broad folk usage of the term “addiction” is essentially correct, and
that the reasons usually given in favor of a more narrow conception are mistaken. Once
we clearly understand which processes and characteristics are essential in cases of drug
addiction, we will see that any voluntary, rewarding behavior can generate a disorder that
is behaviorally, biologically, and phenomenologically identical to a genuine addiction.

2  Must a “genuine” addiction involve suffering

or distress?
There are various “official” definitions of addiction that are used in psychiatric diagnosis,
neuroscience research, and public policy. The most influential of these official definitions
is provided in the psychiatric diagnostic manuals of the American Psychiatric Association
(APA), the DSM-​IV-​TR and the DSM-​5 (1994, 2013). In the DSM-​IV-​TR, the only “offi-
cial” addiction is termed “substance dependence,” while “pathological gambling,” “inter-
net gaming disorder” and “binge-​eating disorder” are defined as unrelated psychiatric
ailments. Every other addiction-​like behavior is omitted from the manual. The DSM-​5
groups gambling with drugs, now in a section labeled “addictions,” while food-​oriented
disorders are still defined as non-​addictive disorders, and other behaviors are relegated to
an appendix. So the official stance is that taking drugs and gambling are genuinely addic-
tive, while eating and other rewarding behaviors are not.
There is a range of reasons not to take the psychiatric definitions as the final word
on the matter, however. Most importantly, psychiatric disorders are always defined in
these manuals as conditions which cause significant distress or suffering. That is entirely
appropriate, since a psychiatrist has no business being involved with a patient who is not
in distress; who is happy and healthy being an addict. Psychiatric definitions of disorders
are intended to define who is in need of psychiatric help, not to define the disorder itself
at a functional level, partly because our functional understanding of psychiatric problems
often trails our ability to effectively treat them by decades.
Furthermore, it does not seem plausible that the distress or suffering that addictions so
often cause is an essential part of what it means to be addicted, since this distress is usu-
ally caused by—​or at least heavily contingent on—​the addict’s circumstances. A heroin
addict’s health suffers because the cost of her drug forces her to divert money from nutri-
tion or shelter toward drug use. And a problem gambler might become distressed when
family members blame him for gambling away their shared wealth (Sinnott-​Armstrong
and Pickard 2013). Meanwhile, the monetary cost of low-​stakes gambling will not cause
distress to a billionaire, and the criminalization of opiate drugs is clearly not a source of
distress for addicts who live in countries where those drugs are not illegal.
  51

What (if anything) makes addictive drugs unlike other addictions? 51

Distress over these consequences is also subjective and dependent on psychological

facts about the drug user that are wildly variable. In particular, addictive drugs are often
used to relieve pre-​existing psychiatric distress from other sources, as Pickard and Ahmed
point out in this volume (see Chapter 2). And it can be impossible to determine whether
these pre-​existing sources of suffering have been exacerbated by drug use or relieved by
it (Martin et al. 2014).
In our current sociological context, a “workaholic,” an “exercise addict,” or even a coffee
addict will face relatively few problems that are caused by their addictive choices, so even
if the underlying functional structure of workaholism or exercise addiction turned out to
be exactly like the structure of heroin addiction, we would expect it to cause much less
distress. Thus the psychiatric definitions of addiction will always be too narrow if we wish
to discuss the question of what a genuine addiction is and how it works, rather than the
questions of who needs treatment, or of what form the treatment should take.
So we return to the point posed at the beginning: it could be true that there are some
facts that distinguish drug addictions as “genuine,” whereas behavioral addictions are
not. And these facts could be neurological facts, psychological facts, or any other kind of
objective fact, but they cannot be facts about subjective suffering or distress caused by the
behavior in question.
An array of experimental and epidemiological data are available for every kind of addic-
tion. Although there is no unanimous consensus on how we should interpret these data in
each case, I think there are reasonable conclusions that may be drawn from a comparison
of the data between the various behaviors that we are generally said to be addicted to. We
may be able to find out what, if anything, makes a particular drug or behavior special—​
what makes it genuinely addictive when other behaviors or substances are not.
I will begin with the category of addictive drug use, since that is where the scientific
study of addiction began, it is a category of behaviors that everyone agrees are genuinely
addictive, and it is where the science is best developed and most widely discussed.

3  What (if anything) makes addictive drugs

unlike other addictions?
The science of addiction began with the study of alcoholism, but broadened as it became
apparent that there were strong similarities between alcoholics and chronic users of other
psychoactive drugs. Whether the drug in question is alcohol, heroin, or cocaine, there
is a litany of universal behavioral similarities: gradually increasing frequency of use and
dose, increasing urgency in drug seeking, failed attempts to limit or regulate use, feelings
of regret, deepening interpersonal problems stemming from chronic drug use, and so on.
In the mid twentieth century, Abraham Wikler’s theory of addiction became the most
widely accepted explanation of these behavioral symptoms. Wikler was puzzled by the
phenomenon of tolerance, in which a drug user experiences less pleasure from repeat-
edly ingesting the same dose of a drug; if a drug became less rewarding over time, why
not just stop taking it? His solution to the puzzle focused on the phenomenon of with-
drawal, in which the drugs of primary interest at that time—​alcohol and heroin—​produce
52  

52 Addiction: the pleasures and perils of operant behavior

uncomfortable physical symptoms when an addicted user stops taking them. In Wikler’s
(1961) view, an addict would keep taking a drug because he became conditioned to asso-
ciate drug use with the relief of his withdrawal symptoms.
This view was accepted by many until brain-​imaging techniques and rat studies began
to identify the pharmacological mechanisms by which these drugs induce pleasure, as well
as the effects that chronic drug use has on the structure and function of rat and human
brains. As new results arrived, a new orthodoxy emerged: according to most neuroscien-
tists writing today, drugs are uniquely addictive in that they “coopt neurotransmitter sig-
nals” by binding directly to the receptors in the brain in place of the natural, endogenous
neurotransmitters that mediate reward learning (Volkow et al. 2010). By binding directly
to the brain’s receptors, drugs have the power to alter the brain’s function and structure in
unique ways, or so the orthodox story goes.
In perhaps the most famous and widely replicated example of these experiments, the
brains of drug addicts have been shown to become less sensitive to neural “reward” sig-
nals. In a normal brain, rewarding experiences and behaviors elicit the release of the neu-
rotransmitter dopamine, which binds to D2 dopamine receptors in the brain’s reward
centres. The activation of these receptors is central to the reinforcement of behavior: for
example, rats whose D2 dopamine receptors are directly stimulated when performing a
given action will tend to perform that action more frequently (Nakajima 1989).
In people and rats addicted to cocaine, for example, fewer D2 dopamine receptors are
available for this kind of signaling, and so more cocaine must be taken for the same level
of reward (Volkow et al. 1993). The same finding holds true in the case of alcoholics and
heroin addicts (Hietala et  al. 1994; Wang et  al. 1997). The most popular hypothesis is
that these individuals gradually increase their dosage to compensate for the diminishing
availability of D2 dopamine receptors in their brains, thus creating the phenomenon of
“tolerance” and increasing dose.
Now, you might think, given this data, that drug use would be wildly more addictive
than putatively addictive non-​drug behaviors. Indeed, this view is held by many scientists
studying addiction and many clinicians who treat addicts, and if it were true it would
provide a solid basis for claiming that non-​drug behaviors are not truly addictive.
Unfortunately, we simply do not have a clear and settled scientific picture of what it
means for something to be more or less addictive in a neurochemical sense, or of exactly
how social and environmental factors interact with neurology to generate addictive
behaviors. For example, there is a large literature showing what seems surprising under
the neurological models: that cocaine addicts recover from their addictions more quickly
than tobacco or alcohol addicts (Heyman 2013). This may be partly explained by the rela-
tive ease of obtaining alcohol and tobacco, but not entirely: tobacco addiction remains
prevalent in prisons where its use is banned (Cropsey and Kristeller 2005). It seems the
amount of dopamine reward elicited by a drug does not predict how many people will
become addicted to it, or how long they will remain addicted to it, relative to other drugs.
There are other puzzles that ought to make us unsure about the neurological explana-
tions of addiction. For example, among outpatients who are given prescriptions of opiate
  53

What (if anything) makes addictive drugs unlike other addictions? 53

drugs like morphine to treat chronic pain, the rate of addiction is a mere 3% (Edlund et al.
2010). If drugs are uniquely addictive because of their neurochemical properties, and if
a chronic pain patient receives a daily dose of an addictive opiate, why isn’t the rate of
addiction higher? It seems reasonable to assume that it would be easy to forcibly addict
someone to a drug by injecting them with it against their will, allowing their brains to
develop tolerance, withdrawal, and any other relevant addictive adaptation. Indeed, this
is a scenario that is often presented in movies and books.
The evidence, however, suggests that this is not the case. One of the most important
experimental designs in studies of addiction using rats is called the “yoked control”: one
rat is rigged to receive an injection of cocaine (for example) every time it presses a button.
A second rat, the “yoked control,” has a button that does nothing—​instead, it receives the
same dose of cocaine every time the first rat presses the button. Only the first rat, which
administers its own cocaine, shows the dramatic increase in button pressing that we use
to model addiction (Bozarth and Wise 1981).
The same phenomenon is observable among inpatients who receive opiate painkillers
such as morphine while in the hospital (Lindesmith 1968; Zinberg 1984). Almost nobody
develops an addiction to these drugs in hospital, even if they were administered the drug
constantly during weeks or months of treatment, and even if they developed physical tol-
erance and withdrawal symptoms. By contrast, patients with chronic pain who are given
prescriptions of painkiller pills to voluntarily take at home do sometimes develop addic-
tions to their drugs (Korff et al. 2011). So it is clear that willful self-​administration is a
crucial part of the process of addiction; it is not enough to simply be given a substance
that activates one’s reward receptors.
To explain why that might be, we can refer to the psychological concept of operant con-
ditioning. In the process of operant conditioning, when an animal performs some willful
(“operant”) action and later receives some reward (or experiences some punishment),
the animal becomes more likely (or less) to repeat the rewarding behavior at a later time.
In the absence of willful behavior, there is nothing to be conditioned—​a patient in hos-
pital can learn to appreciate the sensation of her morphine drip, but without willfully
self-​administering it, she cannot learn to associate the pleasure with the heroin-​seeking
choices and behaviors that would constitute an addictive disorder.
The operant conditioning literature can also help to explain another puzzle of variation
between addictive drugs. We might wonder why cigarette smoking leads so frequently
to addictive use, whereas marijuana, which is typically administered in the same way
and has a more pronounced euphoric effect, does not. The most popular explanation for
this discrepancy is the “nicotine bolus hypothesis,” which points out that nicotine takes
effect within 15 seconds of inhaling cigarette smoke, while injected drugs take around 50
seconds, and marijuana smoke takes effect minutes after inhalation (Berridge et al. 2010;
Rose 1999). Along similar lines, smoked tobacco and cocaine seem to be far more addic-
tive than ingested tobacco and cocaine, even at the same dose, which suggests that the
speed with which a drug takes effect is a major determinant of its addictiveness (Samaha
and Robinson 2005).
54  

54 Addiction: the pleasures and perils of operant behavior

The function of operant conditioning in nature is that it helps animals to learn to repeat
those behaviors that yield rewards like food, warmth, and reproduction. Naturally, oper-
ant conditioning would not be useful to animals if the operant behavior did not cause the
reward, and so the brain needs some way to discern which rewarding outcomes were caused
from a given behavior. This needs to happen in simple animals as well as human beings, so
no complex model of causality is involved; instead, the brain associates the behavior with
whatever happened immediately afterwards (Redgrave and Gurney 2006). For example,
Wolfe showed in 1934 that rats would learn to associate successfully navigating a maze with
a food reward much less effectively if there was a delay before the food was issued.
This, then, is an important aspect of what makes any given drug-​taking behavior addic-
tive: it produces reward relatively quickly and reliably after the drug is taken.
Before I move on, let us take stock. Addictive drugs are claimed to be uniquely addictive
because of a number of their essential characteristics: they elicit intense reward by directly
or indirectly stimulating the D2 dopamine receptors in the brain’s reward pathways, even-
tually making those receptors less sensitive. However, a range of phenomena undermine
the simplistic claim that the unique chemical action of the drugs causes or constitutes
the addiction. First, drugs must be self-​administered to be addictive, allowing operant
conditioning to take place. And they must have a rapid onset of effect, strengthening the
conditioned association between self-​administration and reward.

4  Is food genuinely addictive?

When Nora Volkow (2010) says that addictive drugs “coopt neurotransmitter signals,” or
when Steve Hyman (2007a) says that addictive drugs “hijack the natural reward systems,”
they mean that the drugs are activating receptors in the brain that normally are activated in
response to certain “natural” rewards, such as food and sex. The reason we have these “natu-
ral reward” circuits in our brains is that we need to learn to pursue survival and reproductive
activities in order for the species to proliferate. When we eat food or drink water, have sex
or relieve our bladders, we receive a neural reward and this trains us (through operant con-
ditioning) to want to do these things again, at least in similar circumstances, in the future.
One of the implications of this is that intense “natural rewards” might be just as danger-
ous as drugs are, since they activate the same neural systems that drugs “hijack.” In this
section I will argue that food, and other natural rewards, can be considered addictive in
every significant respect—​and perhaps to the same degree—​as addictive drugs.
Of course, not every natural reward can be had to excess. For example, it would be
rather difficult to become addicted to voiding one’s bladder because once it is empty, that
is that. But food and sex rewards are examples of natural rewards that can be experienced
in excess of biological need, which means that they can produce neural reward in high,
frequent doses. To know if natural rewards like food can be genuinely addictive we need
to assess whether there really is an important fact about drugs that is not true about food
which explains why the regular action of natural rewards cannot produce an addiction by
the exact same mechanism as drugs. It could be a fact about the behavioral or appetitive
  55

Is food genuinely addictive? 55

patterns that drugs induce (versus natural rewards), or a neurochemical fact about the
amount of reward induced by drugs (versus natural rewards). And so on.
To make this kind of assessment, there are a number of things we can look at. First of
all, there is the behavioral taxonomy. People do develop a pattern of disordered eating that
is similar behaviorally to addictions: it is known as “Binge-​eating disorder” (BED), and its
clinical definition is modeled on the definition of drug addiction. To be diagnosed with
BED, one must eat a “larger amount of food than normal” at least twice per week, and feel
out of control while doing so, and affirm three out of five other criteria, such as eating until
uncomfortably full (APA 1994). As well as eating like they are addicted to food, binge eat-
ers crave food, they eat more than they intend do, they lie about using it, and so forth, in
essentially the same behavioral pattern expressed by drug addicts1 (Gold et al. 2003).
Binge-​eating disorder (defined and measured according to psychiatric guidelines) is far
more prevalent than addictions to every kind of drug except for cigarettes and alcohol,
but that is presumably only because everybody is exposed to food, and everyone is forced
by biological need to try it over and over again. It seems that around 3–​4 percent of all
United States citizens at any given time can be positively diagnosed with binge-​eating
disorder (Garfinkel et al. 1995; Striegel-​Moore 2000,2007,2009). To give a sense of scale,
in the 2011 National Survey on Drug Use and Health run by the US Government, 2.4 per-
cent of cocaine initiates matched criteria for cocaine dependence in the preceding year.2
Food rewards stimulate the same reward circuits in the brain as drugs do, and bingeing
on food has been shown to generate the same long-​term structural changes in the brain as
drugs of abuse in rat models (Avena et al. 2009). And it has the same effect of diminishing
the availability of D2 dopamine receptors in human beings (Wang et al. 2001).
As in the case of addictive drugs, we develop an increased appetite for food rewards by
over-​indulging, and we can lose our appetite by fasting. In one study, pregnant rats were
given either a normal diet or a fat-​rich diet. The pups in the high-​fat group had a much
larger appetite for fat after they were born (Chang et al. 2008). People who drink diet soft
drinks, which cannot be turned into fat, are significantly more likely to be overweight than
people who drink water, suggesting that the sweetness of a drink increases a person’s appe-
tite for sugar over time (Odegaard et al. 2011). And while binge-​eaters and obese people
have decreased dopamine receptor availability (like drug addicts), early evidence seems to
show the reverse effect among anorexics (Frank 2011). In short, at least one of the “natural”
appetites is subject to training and atrophy, just like drug-​oriented appetites are.
One common claim about drug addiction has been that the reason drugs are able to
“hijack” the brain’s pleasure circuits is that they are more pleasurable, and elicit more neu-
ral reward, than food. But this has been undermined by recent investigations into food
rewards. Salty and sweet foods, in particular, are extremely rewarding, with saccharin
sweeteners eliciting more neural reward than cocaine at a comparable dose (Lenoir et al.

1 We will leave bulimia aside, since it is heavily complicated by body-​image issues.

2 Morbidityfigures are freely available online at https://​nsduhweb.rti.org. The relevant code in the
SAMHDA database is DEPNDCOC.
56  

56 Addiction: the pleasures and perils of operant behavior

2007). Bingeing on sugar can even elicit a heroin-​like withdrawal syndrome, since sweet
taste sensations activate the same opioid receptors in the brain (Avena et al. 2007). So it
just isn’t the case that drugs are more potent.
We might think that drugs—​especially smoked or injected varieties—​would take effect
more quickly after administration than food, resulting in a stronger operant condition-
ing effect, just as tobacco cigarettes take effect faster than marijuana. But the chemicals
in food do not need to reach the brain to elicit this strong neural reward. Instead, it is the
taste of food that causes endorphins and dopamine to be released within the brain, and
this happens almost instantly when the food dissolves on a person’s tongue (Berridge
1996). If anything, food rewards are quicker to take effect, and thus more tightly paired
with eating than drugs are with drug-​taking behavior.
Finally, just as in the case of drugs, it seems eating must be voluntarily initiated if it is to
develop into an addictive behavior like binge-​eating disorder. In one experiment, if people
were commanded to eat a particular food every day, even if it was a food that they liked a
great deal, the reinforcing effect of the food was significantly reduced (Temple et al. 2008).
So binge-​eating disorder results in behaviors and problems for its sufferers that are
nearly identical to those experienced by drug addicts. It is more prevalent than cocaine
addiction is among cocaine initiates. Bingeing on food has exactly the same effect on the
brain as taking recreational doses of drugs. Food rewards take effect no less quickly than
drug rewards, and voluntary initiation of food intake is required if one is to develop dis-
ordered eating habits, just as it is required in the development of drug addiction.
On consideration of all these points, it is very difficult to mount the argument that
binge eaters are not literally addicted to food in the exact same way as drug addicts are
addicted to drugs. And although binge-​eating disorder is not defined as an addiction
in the DSM-​IV-​TR or DSM-​5, as I  have pointed out elsewhere, the consensus among
neuroscientists and psychologists is gradually shifting toward the view that food is liter-
ally addictive (Foddy 2010). Drug use cannot, after all, be the only genuine addictive
behavior—​it must be true that foods, at least, are also genuinely addictive.
Both drugs and food harness the natural reward system. We might think, then, that
drug addiction, binge-​eating disorder, and perhaps also other “natural reward addictions”
such as sex addictions, exhaust the category of truly addictive substances and behaviors.
Let us move to the next group to see if that is indeed a reasonable claim.

5  Gambling
The DSM-​IV-​TR has a definition of “pathological gambling,” which sounds a lot like gam-
bling addiction, but since the publication of the DSM-​5 it has been officially grouped with
the drug addictions (APA 1994, 2013). In the new revision, “gambling addiction” is the
only officially diagnosable type of addiction that does not involve drug use (Holden 2010).
The behaviors of problem gamblers closely mimic those of the users of addictive drugs.
Once again, their gambling escalates until it causes serious problems in their life. They lie
and steal to maintain their gambling habit. They gamble greater amounts of money, more
  57

Gambling 57

often, than they originally intended. And so on. But this presents an even greater puzzle
than food for the orthodox neuroscientific view of addiction. Gambling certainly involves
rewards—​in the form of monetary payouts—​but these are nothing like the natural rewards
involved in food addiction, and gamblers do not introduce any exogenous substance into
the brain that could activate (or “hijack”) their reward centers. We have not had money for
long enough in our evolutionary history to have developed “money receptors” in our brains.
In the neuroscientists’ conception of addiction, where only drugs can elicit “true” addic-
tions, this would absolutely rule out pathological gambling as a “real” addiction. And yet
the similarity in behavior and phenomenal experience between problem gamblers and
substance addicts poses a puzzle.
The most obvious, and most widely argued solution to the puzzle is to hypothesize
that human beings associate money with the natural rewards, since it allows us to obtain
those rewards. This hypothesis generally takes two forms (Lea and Webley 2006): in the
first, it is proposed that money is indirectly rewarding because it is literally a tool that
provides us with natural rewards. In the second, money becomes directly rewarding just
because we come to form a conditioned (Pavlovian) association between money and the
things we may exchange it for—​ultimately, when we are given money our brains react as
though we had been given a natural reward. Skinner (1953), for example, termed money
a “token reinforcer” in the sense that it was a “token” for natural rewards.
It is not settled whether either of these two theories is the correct one, and it could be
that both are true to some extent. However, we have some reason to suspect that problem
gamblers do not simply gamble as a means of obtaining money to pay for natural rewards.
In the first place, the delay between receiving money and buying a reward is often very
long, so the operant conditioning effect would be very weak if that were the only reward-
ing aspect of a gambling win. Secondly, obtaining money seems to be the least popular
motivation among problem gamblers, with “enjoyment” being the number one motiva-
tion among females, and “mood regulation” the reason given most often by males (Lloyd
et al. 2010a). Schull’s (2012) research suggests that people who are addicted to electronic
gambling machines most often refer to achieving a state of “flow” when explaining why
they gamble, and manufacturers take steps to maximize this cognitive/​emotional state
and minimize the frequency of monetary transactions. This suggests that the reason peo-
ple are motivated to gamble is that gambling itself provides a direct, immediate reward,
whether or not a monetary payout is received.
We certainly do see signs in the empirical evidence showing that gambling is function-
ing like an addiction to a natural reward. In the first place, there is the neuroscience: a
large number of studies now show the same decrease in D2 dopamine receptor activity
in the brains of gamblers that is seen in drug addicts and binge eaters (Comings et al.
1996; Zack and Poulos 2007). Genetic risk factors for drug addiction that implicate the
dopamine receptors (in a small subset of users) seem to pose a risk for pathological gam-
bling too (Comings et al. 1996). So the locus of gambling reward in the brain, and the
adaptations that occur in the brains of problem gamblers, are the same as they are in drug
addicts and binge eaters.
58  

58 Addiction: the pleasures and perils of operant behavior

The prevalence of problem gambling among gambling initiates seems to be close to the
levels of drug and food addiction too. A 2010 study based in the United Kingdom showed
that 59  percent of British men gambled in some way other than playing the National
Lottery, while 1.3 percent satisfied three of the DSM-​IV criteria for problem gambling
(Wardle et al. 2011). Therefore, 2.2 percent of British men who had ever gambled were (by
this broader criterion) problem gamblers at the time of the survey—​not at all far from the
prevalence of cocaine addiction among cocaine initiates.
Evidence increasingly suggests that the neurobiological effects of gambling are very
similar to those of drugs and binge eating (Leeman and Potenza 2011). So far so good for
the theory that people get addicted to gambling because it is a token for natural rewards.
Furthermore, if gambling addicts people using the same natural-​reward mechanism
involved in food and drug addictions, and if the prevalence and behavioral symptoms are
also very similar, it becomes harder to justify the exclusion of problem gambling from the
category of genuine addictions.
Before we move on, however, we might wonder exactly how gamblers are getting
rewarded often enough to become addicted, given that many forms of gambling (espe-
cially lotteries) only seldom pay out. An answer to this puzzle is suggested by a recent
neuroimaging study looking at the brains of gamblers using an electronic gaming
machine (“slot machine”) (Shao et al. 2013). As the gamblers became more involved in
a game, their brains gradually stopped responding to wins as reward states, and started
to process the mere “spinning” of the machine’s wheels as rewards. There are only two
reasonable ways to interpret the result, and they are not mutually exclusive: first, it is
possible that the anticipation of a win becomes the primary motivating reward of gam-
bling, as is predicted by Robinson and Berridge’s (1993, 2003) “incentive sensitization”
account of addictive choice (just as the anticipation of drug reward seems to become
more significant as people develop drug addictions). On that view, problem gamblers
become hypersensitized to the stimulus of the spinning wheels in a way which gradually
supplants the intermittent monetary reward. Second, it could be that the mere spinning
of the wheels and the chiming of the music is enough of a visual-​neurological reward on
its own merits.
Whether either or both of these hypotheses is correct, it is hard to see how either
wheel spinning or reward anticipation could be tokens for specifically natural rewards,
and indeed it is not intuitively obvious how a spinning wheel of fruit symbols could be
rewarding at all! But before we dismiss the reward value of spinning wheels, we should
consider the case of videogame addiction, where the rewards are frequently no stronger
than a spinning wheel and a series of electronic chimes.

6  Videogames
There has been a great deal of controversy over the question of whether videogame addic-
tion is a genuine addiction or nothing more than a bad habit. After some debate, in 2007
the APA decided to include only “internet gaming disorder” in the DSM-​5, and only as
an appendix, reasoning that there was insufficient scientific evidence to support grouping
  59

Videogames 59

videogames with addictive drugs and gambling machines, or for that matter with any
other psychiatric disorder (APA 2007). Since the APA made its ruling, further data has
emerged suggesting that videogame addiction is indeed a real phenomenon, although it
is still controversial whether it ought to be counted as a genuine addiction or merely an
unrelated behavioral disorder. This is certainly still an underdeveloped field of empirical
and theoretical study; most of the data collected so far focuses on children, despite the
fact that the average videogame player is now 37 years old (ESA 2011).
Once again, the same patterns of reward activation have been found in the brains
of videogame players as in gamblers and drug users (Han et  al. 2011; Ko et  al. 2009).
Videogame addicts develop the same reduction in the availability of their D2 dopamine
receptors (Weinstein 2010). People with chronic videogame problems seem to suffer from
the same psychiatric comorbidities as problem gamblers (Ha et al. 2006), and there have
even been cases where players’ children or even the players themselves have died of bodily
neglect (AAP 2012).
Yet videogames offer only the most ephemeral and indirect of rewards. If there is
something to the neuroscientists’ claim that drugs are uniquely or particularly addictive
because of their neurochemical mode of action, then we ought to expect that problem
videogame addiction is significantly different than drug addiction. Although the study of
this behavioral pattern is in its infancy, this does not appear to be the case.
The prevalence of problem videogame use is not yet clearly established in the epide-
miological literature, but the data we do now have is not that far out of step with the other
addictions. Using a tentative list of criteria based on the DSM-​IV protocol for diagnos-
ing pathological gambling, one group estimated the rate of videogame addiction among
players at 7.6–​9.9 percent, and showed that “pathological” gaming was a cause of some
social and psychological problems and not merely a symptom of an underlying psycho-
logical disorder (Gentile et al. 2011). A study of 3000 young Norwegian players found
that 2.6 percent showed signs of pathological play (Johansson and Götestam 2004). These
studies used preliminary measures and only involved young subjects—​many of whom
do not have any responsibilities that a game addiction could interfere with—​so they are
probably not representative or reliable results. However, they do suggest that the preva-
lence of videogame addiction is at least in the same ballpark as pathological gambling,
binge eating, and drug addiction, rather than being a rare, extremely abnormal pathology.
If videogame addiction is indeed the same phenomenon as drug addiction, problem
gambling, and binge-​eating disorder, it presents the biggest puzzle of all, since the behav-
iors are manifestly non-​consumptive, and cannot reasonably be said to provide natural
rewards or tokens for natural rewards. Hardly anyone ever gets to trade success at a vide-
ogame for a natural reward; indeed, playing videogames usually diminishes the likelihood
of obtaining natural rewards such as food, sex, or money. And, as in the case of gambling,
no substance is introduced to the body that can activate the reward receptors directly, as
drugs do. As Mark Griffiths (2007) puts it, “If videogame addiction actually exists, what
are people addicted to?”
The answer is probably multifactorial, since the design of videogames harnesses a
number of explicit and implicit reward mechanisms. To take one obvious example, many
60  

60 Addiction: the pleasures and perils of operant behavior

games employ social rewards, such as high-​score tables or status badges. There is data
showing that social approval generates the same pattern of neural reward as monetary
payments, thus “providing strong support for the idea of a ‘common neural currency’
of reward,” in the words of the researchers (Izuma et al. 2008). Other data shows that
rewards do not even need to be pleasurable in order to be motivating: for example, male
mice can become “addicted” to pressing a button which allows another male mouse
into their territory, suggesting that aggressive behavior can be rewarding (Couppis and
Kennedy 2008). Thus competition in videogames may be rewarding as well.
However, not every videogame has a high-​score table, and people have reported becom-
ing addicted even to solitary videogames which offer nothing but a particular pattern of
light and sound in response to player input, and sometimes (but not always) a numerical
representation of performance (the player’s “score”).
Yet if we cast our minds back to the material we have already covered here, it should
be easy enough to understand how a game can become genuinely addictive. The case of
electronic gambling machines tells us that something as simple as visual feedback (like a
spinning slot machine wheel) can become rewarding if it occurs after the player presses a
button or pulls a lever. And operant conditioning experiments demonstrate that reward
learning is strongest when there is a tight temporal pairing between action and reward.
Now, in videogames, the player’s input is nearly always followed immediately by visual
and auditory feedback—​the movement of a character, the firing of a gun, an explosion or
a pattern of flashing light—​so we should not be surprised to find a strong conditioning
effect even from these very mild videogame rewards.
If videogames are indeed addictive, it seems reasonable to form the hypothesis that
any willful behavior could be addictive, even when it offers only very small or indirect
rewards, so long as it is clear to the agent that she is causing the rewards through her own
actions. And it will be clear to the agent that she is causing the rewards just so long as they
happen very soon after the operant behavior takes place. Just as smoked drugs are more
addictive than ingested or injected drugs, highly reactive reward-​seeking behaviors will
be more addictive than behaviors with delayed rewards.
Videogames are unlike drugs and perhaps even food in one sense: they are specifically
designed to be maximally addictive. And they are unlike gambling machines in that their
perceived addictiveness is often touted as a selling point, a laudable quality for the best
games to have. If this were simply a casual, exaggerated use of the term “addiction” then
it would be easy to dismiss these things. But the evidence suggests that they are genuinely
addictive, and that the addicted players suffer from an addiction that was intentionally
engineered by the designer and joyfully appreciated by the user.

7  Every rewarding behavior can generate

a “genuine” addiction
The data I have reviewed here, spread over a range of disciplines and replicated in dozens
of experiments, present us with a conundrum. How can it be that heroin addicts, whose
  61

The pleasures and perils of operant reward 61

addiction seems to have been caused by repeatedly injecting a powerful narcotic, have the
same problems as gamblers and binge eaters? How can it be that the brains of videogame
players are activated in the same way as cocaine users? Surely a pattern of flashing light
and sound cannot be enough to “hijack” the brain’s reward system, as Hyman (2007b)
claims that drugs do.
The only plausible solution to this conundrum, in my view, is that each of these addictions
is fundamentally the same, and that neither drugs, nor binge eating, nor gambling really
belong to an exceptional category of brain-​altering behaviors. Although the behaviors are
different in each case, and the rewards are different, and the health risks are different, each
addiction involves a reward and voluntary control over the administration of that reward.
It is important to acknowledge that there are various factors that seem to differentiate
one addictive behavior from another in important ways. Chronic use of cocaine leads to
damage to the frontal lobe of the brain that governs judgment and impulse control, and
this surely helps to generate further cocaine use (Volkow et  al. 1993). Our food appe-
tites are partially governed by the satiety system, which generates feelings of “fullness,”
and some people overeat because they have disordered satiety functions (Ahima 2008).
Pathological gamblers engage in particular disordered thought patterns, such as falsely
believing that they can control the outcome of a dice roll (Lim et al. 2014). And person-
ality disorders are frequently involved in the generation and maintenance of addictive
substance use, eating, and gambling (Lloyd et al. 2010a; Woodside et al. 2001).
Each of these additional factors is important in explaining what the risk factors are for
a given addiction, why some behaviors are riskier than others, why a given addict has
become addicted, and how we should go about treating addicts in each case. But one of
the lessons of the cross-​disciplinary view I have adopted here is that these additional com-
plexities do not need to be present in order to generate the biological, psychological, and
behavioral phenomena that we would characterize as addiction in a significant number
of people. We get binge eaters who do not have leptin deficiencies or pituitary tumors, we
have compulsive gamblers who don’t have personality disorders, and we have addicted
cocaine users who have not been using long enough to damage the self-​control circuitry
in their brains. These risk factors are not causal prerequisites of addiction.
The factors that are common to all addictive behaviors are these: the behavior is (1) vol-
untarily repeated (2) a sufficient number of times, and (3) the reward is tightly paired to
the operant behavior.

8  The pleasures and perils of operant reward

In this chapter I have argued that there isn’t anything special about the rewards elicited by
drug use that would make drug addictions the only genuine species of addiction. I have
reviewed the evidence showing that drugs, food, gambling, and videogames all activate the
D2 dopamine receptors to produce neural reward and reinforcement of behavior. But the
same areas of the brain are activated even by rewards that are not widely considered to be
addictive: funny jokes, nice paintings and romantic love, to name just three (Bartels and
62  

62 Addiction: the pleasures and perils of operant behavior

Zeki 2004; Kawabata and Zeki 2004; Mobbs et al. 2003). It is not easy to pair rewards like
these to a convenient and instantaneous administration method (like swallowing a pill, or
pressing a button) which may explain the relative absence of joke addicts and painting junk-
ies. But there have been love addicts, and even addicts of very thin rewards, like carrots and
water (“psychogenic polydipsia”) (Ali et al. 2009; Kaplan 1996; Peele and Brodsky 1975).
There is even some emerging evidence that placebo pills provide neural reward (la Fuente-​
Fernández et al. 2004). The evidence suggests that there is nothing special or unique about
heroin rewards or gambling rewards that is required for the formation of addictions.
Whenever human beings have the opportunity to quickly and conveniently self-​
administer some rewarding experience this is going to generate problematic overuse in a
small percentage. Of course, not every rewarding behavior is equally addictive—​cigarette
addicts outnumber cocaine addicts, while cases of carrot or water addicts number in the
single digits. But particularly in cases when neural reward is tightly paired with an oper-
ant behavior, addictions will certainly arise.
If every rewarding behavior is a risk, this suggests that addiction is more fundamental
to human beings than is usually acknowledged. Pursuit of reward, after all, is not the
sole province of risk seekers and libertines. It is the basic motivation behind nearly every
human activity. This is why your child wants to press every button in an elevator. It is why
your co-​worker incessantly clicks the button on the end of his pen in and out. And it is
why you cannot stop pulling out eyebrows with tweezers, cracking your knuckles, or eat-
ing unsalted sunflower seeds.

References
Australian Associated Press (AAP) (2012). Taiwanese man dies after playing video games for 23 hours.
Herald Sun, Melbourne, February 3.
Ahima, R.S. (2008). Revisiting leptin’s role in obesity and weight loss. Journal of Clinical Investigation,
118, 2380–​83.
Ali, N., Imbriano, L.J., and Maesaka, J.K. (2009). The case: a 66-​year-​old male with hyponatremia.
Kidney International, 76, 233–​34.
American Psychiatric Association (APA) (1994). Diagnostic and Statistical Manual of Mental
Disorders: DSM-​IV (4th ed.). Washington, DC: APA.
American Psychiatric Association (2007). Statement of the American Psychiatric Association on
“Video Game Addiction.” psych.org. Arlington, VA: APA.
American Psychiatric Association (2013). Diagnostic and Statistical Manual of Mental Disorders (5th ed.).
Washington, DC: APA.
Avena, N.M., Rada, P., and Hoebel, B.G. (2007). Evidence for sugar addiction: behavioral and
neurochemical effects of intermittent, excessive sugar intake. Neuroscience & Biobehavioural
Reviews, 32(1), 20–​39
Avena, N.M., Rada, P., and Hoebel, B.G. (2009). Sugar and fat bingeing have notable differences in
addictive-​like behavior. Journal of Nutrition, 139, 623–​28.
Bartels, A. and Zeki, S. (2004). The neural correlates of maternal and romantic love. NeuroImage, 21,
1155–​66.
Berridge, K.C. (1996). Food reward: brain substrates of wanting and liking. Neuroscience and
Biobehavioral Reviews, 20, 1–​25.
  63

The pleasures and perils of operant reward 63

Berridge, M.S., Apana, S.M., Nagano, K.K., Berridge, C.E., Leisure, G.P., and Boswell, M.V. (2010).
Smoking produces rapid rise of [11C]nicotine in human brain. Psychopharmacology, 209, 383–​94.
Bozarth, M.A. and Wise, R.A. (1981). Intracranial self-​administration of morphine into the ventral
tegmental area in rats. Life Sciences, 28, 551–​55.
Chang, G.-​Q., Gaysinskaya, V., Karatayev, O., and Leibowitz, S.F. (2008). Maternal high-​fat diet
and fetal programming: increased proliferation of hypothalamic peptide-​producing neurons that
increase risk for overeating and obesity. Journal of Neuroscience, 28, 12107–​19.
Comings, D.E., Rosenthal, R.J., Lesieur, H.R., Rugle, L.J., Muhleman, D., Chiu, C., et al. (1996). A
study of the dopamine D2 receptor gene in pathological gambling. Pharmacogenetics and Genomics,
6, 223.
Couppis, M.H. and Kennedy, C.H. (2008). The rewarding effect of aggression is reduced by nucleus
accumbens dopamine receptor antagonism in mice. Psychopharmacologia, 197, 449–​56.
Cropsey, K.L. and Kristeller, J.L. (2005). The effects of a prison smoking ban on smoking behavior and
withdrawal symptoms. Addictive Behavior, 30, 589–​94.
Edlund, M.J., Martin, B.C., Fan, M.Y., Devries, A., Braden, J.B., and Sullivan, M.D. (2010). Risks for
opioid abuse and dependence among recipients of chronic opioid therapy: results from the TROUP
study. Drug and Alcohol Dependence, 112, 90–​98.
Electronic Software Association (ESA) (2011). 2011 Sales, demographic and usage data. theesa.com,
Essential facts about the computer and video game industry. Electronic Software Association.
Feldman, M.W. and Lewontin, R.C. (1975). The heritability hang-​up. Science, 190, 1163–​68.
Foddy, B. (2010). Addiction and its sciences-​philosophy. Addiction, 106, 25–​31.
Foddy, B. (2011). Addicted to food, hungry for drugs. Neuroethics, 4, 79–​89
Foddy, B. and Savulescu, J. (2006). Addiction and autonomy: can addicted people consent to the
prescription of their drug of addiction? Bioethics, 20, 1–​15.
Foddy, B. and Savulescu, J. (2010). A liberal account of addiction. Philosophy, Psychiatry and
Psychology, 17, 1–​22.
Frank, G.K.W. (2011). Brain circuitry models in eating disorders. Psychiatric Annals, 41, 526.
Garfinkel, P.E., Lin, E., Goering, P., Spegg, C., Goldbloom, D.S., Kennedy, S., Kaplan, A.S., et al.
(1995). Bulimia nervosa in a Canadian community sample: prevalence and comparison of
subgroups. American Journal of Psychiatry, 152, 1052–​58.
Gearhardt, A.N., Grilo, C.M., DiLeone, R.J., Brownell, K.D., and Potenza, M.N. (2011). Can food be
addictive? Public health and policy implications. Addiction, 106, 1208–​12.
Gentile, D.A., Choo, H., Liau, A., Sim, T., Li, D., Fung, D., and Khoo, A. (2011). Pathological video
game use among youths: a two-​year longitudinal study. Pediatrics, 127, e319–​29.
Gold, M., Frost-​Pineda, K., and Jacobs, W. (2003). Overeating, binge eating, and eating disorders as
addictions. Psychiatric Annals, 33, 117–​22.
Griffiths, M.D. (2007). Videogame addiction: further thoughts and observations. International Journal
of Mental Health and Addiction, 6, 182–​85.
Ha, J., Yoo, H., Cho, I., Chin, B., Shin, D., and Kim, J. (2006). Psychiatric comorbidity assessed in
Korean children and adolescents who screen positive for Internet addiction. Journal of Clinical
Psychiatry, 67, 821–​26.
Han, D. H., Bolo, N., Daniels, M. A., Arenella, L., Lyoo, I.K., and Renshaw, P.F. (2011). Brain activity
and desire for Internet video game play. Comprehensive Psychiatry, 52, 88–​95.
Heyman, G. (2013). Addiction and choice: theory and new data. Frontiers in Psychiatry, 4, 31.
Hietala, J., West, C., Syvälahti, E., Någren, K., Lehikoinen, P., Sonninen, P., and Ruotsalainen, U.
(1994). Striatal D2 dopamine receptor binding characteristics in vivo in patients with alcohol
dependence. Psychopharmacology, 116, 285–​90.
64  

64 Addiction: the pleasures and perils of operant behavior

Holden, C. (2010). Behavioral addictions debut in proposed DSM-​V. Science, 327, 935.

Hyman, S.E. (2007a). Addiction: a disease of learning and memory. Focus, 5, 220.
Hyman, S. (2007b). The neurobiology of addiction: implications for voluntary control of behavior.
American Journal of Bioethics, 7, 8–​11.
Izuma, K., Saito, D.N., and Sadato, N. (2008). Processing of social and monetary rewards in the human
striatum. Neuron, 58, 284–​94.
Johansson, A. and Götestam, K.G. (2004). Problems with computer games without monetary
reward: similarity to pathological gambling. Psychological Reports, 95, 641–​50.
Kaplan, R. (1996). Carrot addiction. Australian and New Zealand Journal of Psychiatry, 30, 698–​700.
Kawabata, H., and Zeki, S. (2004). Neural correlates of beauty. Journal of Neurophysiology, 91,
1699–​1705.
Ko, C.-​H., Liu, G.-​C., Hsiao, S., Yen, J.-​Y., Yang, M.-​J., Lin, W.-​C., et al. (2009). Brain activities
associated with gaming urge of online gaming addiction. Journal of Psychiatric Research, 43, 739–​47.
Korff, von M., Kolodny, A., Deyo, R.A., and Chou, R. (2011). Long-​term opioid therapy reconsidered.
Annals of Internal Medicine, 155, 325–​28.
la Fuente-​Fernández, R., de, Schulzer, M., and Stoessl, A.J. (2004). Placebo mechanisms and reward
circuitry: clues from Parkinson’s disease. Biological Psychiatry, 56, 67–​71.
Lea, S. and Webley, P. (2006). Money as tool, money as drug: the biological psychology of a strong
incentive. Behavioral and Brain Sciences, 29,161–​209
Leeman, R.F., and Potenza, M.N. (2011). Similarities and differences between pathological gambling
and substance use disorders: a focus on impulsivity and compulsivity. Psychopharmacology, 219,
469–​90.
Lenoir, M., Serre, F., Cantin, L., and Ahmed, S.H. (2007). Intense sweetness surpasses cocaine reward.
PLoS ONE, 2, e698.
Lim, M., Bowden-​Jones, H., and Rogers, R. (2014). Expressing gambling-​related cognitive biases in
motor behaviour: rolling dice to win prizes. Journal of Gambling Studies, 30, 625–​37.
Lindesmith, A.R. (1968). Addiction and Opiates. Chicago: Aldine.
Lloyd, J., Doll, H., Hawton, K., Dutton, W.H., Geddes, J.R., Goodwin, G.M., and Rogers, R.D. (2010a).
How psychological symptoms relate to different motivations for gambling: an online study of
internet gamblers. Biological Psychiatry, 68, 733–​40.
Martin, C.S., Langenbucher, J.W., Chung, T., and Sher, K.J. (2014). Truth or consequences in the
diagnosis of substance use disorders. Addiction, 109, 1773–​78.
Mobbs, D., Greicius, M.D., Abdel-​Azim, E., Menon, V., and Reiss, A.L. (2003). Humor modulates the
mesolimbic reward centers. Neuron, 40, 1041–​48.
Nakajima, S. (1989). Subtypes of dopamine receptors involved in the mechanism of reinforcement.
Neuroscience and Biobehavioral Reviews, 13, 123–​28.
Odegaard, A.O., Choh, A.C., Czerwinski, S.A., Towne, B., and Demerath, E.W. (2011). Sugar-​
sweetened and diet beverages in relation to visceral adipose tissue. Obesity, 20, 689–​91.
Peele, S. and Brodsky, A. (1975). Love and Addiction. Oxford: Taplinger.
Redgrave, P. and Gurney, K. (2006). The short-​latency dopamine signal: a role in discovering novel
actions? Nature Reviews Neuroscience, 7, 967–​75.
Robinson, T.E. and Berridge, K.C. (1993). The neural basis of drug craving: an incentive-​sensitization
theory of addiction. Brain Research Reviews, 18, 247–​91.
Robinson, T.E. and Berridge, K.C. (2003) Addiction. Annual Review of Psychology, 54, 25–​53.
Rose, J. (1999). Arterial nicotine kinetics during cigarette smoking and intravenous nicotine
administration: implications for addiction. Drug and Alcohol Dependence, 56, 99–​107.
Samaha, A.N. and Robinson, T.E. (2005). Why does the rapid delivery of drugs to the brain promote
addiction? Trends in Pharmacological Science, 26, 82–​87.
  65

The pleasures and perils of operant reward 65

Schull, N.D. (2012). Addiction by Design. Princeton, NJ: Princeton University Press

Shao, R., Wakeley, J., Behrens, T., and Rogers, R. (2013). Shifts in reinforcement signalling while playing
slot-​machines as a function of prior experience and impulsivity. Translational Psychiatry, 3, e213.
Sim, T., Gentile, D.A., Bricolo, F., Serpelloni, G., and Gulamoydeen, F. (2012). A conceptual review of
research on the pathological use of computers, video games, and the Internet. International Journal
of Mental Health and Addiction, 10, 9369.
Sinnott-​Armstrong, W. and Pickard, H. (2013). What is addiction?. In: B. Fulford et al. (eds), The
Oxford Handbook of Philosophy of Psychiatry, Oxford: Oxford University Press, pp. 851–​64.
Skinner, B.F. (1953). Science and Human Behavior. New York: Macmillan.
Striegel-​Moore, R. (2000). The epidemiology of binge eating. European Eating Disorders Review, 8,
344–​46.
Striegel-​Moore, R.H., and Bulik, C.M. (2007). Risk factors for eating disorders. American Psychologist,
62, 181–​98.
Striegel Moore, R.H., Rosselli, F., Perrin, N., DeBar, L., Wilson, G.T., May, A., and Kraemer, H.C.
(2009). Gender difference in the prevalence of eating disorder symptoms. International Journal of
Eating Disorders, 42, 471–​74.
Temple, J.L., Chappel, A., Shalik, J., Volcy, S., and Epstein, L.H. (2008). Daily consumption of
individual snack foods decreases their reinforcing value. Eating Behaviors, 9, 267–​76.
Volkow, N.D., Fowler, J.S., Wang, G.J., Hitzemann, R., Logan, J., Schlyer, D. J., et al. (1993). Decreased
dopamine D2 receptor availability is associated with reduced frontal metabolism in cocaine abusers.
Synapse, 14, 169–​77.
Volkow, N.D., Wang, G.-​J., Fowler, J.S., Tomasi, D., Telang, F., and Baler, R. (2010).
Addiction: decreased reward sensitivity and increased expectation sensitivity conspire to overwhelm
the brain’s control circuit. BioEssays, 32, 748–​55.
Wang, G.J., Volkow, N.D., Fowler, J.S., Logan, J., Abumrad, N.N., Hitzemann, R.J., et al. (1997).
Dopamine D 2 receptor availability in opiate-​dependent subjects before and after naloxone-​
precipitated withdrawal. Neuropsychopharmacology, 16, 174–​82.
Wang, G.J., Volkow, N.D., Logan, J., Pappas, N.R., Wong, C.T., Zhu, W., et al. (2001). Brain dopamine
and obesity. Lancet, 357, 354–​57.
Wardle, H., Griffiths, M.D., Orford, J., Moody, A., and Volberg, R. (2011). Gambling in Britain: a time
of change? Health implications from the British Gambling Prevalence Survey 2010. International
Journal of Mental Health and Addiction, 1–​5.
Weinstein, A.M. (2010). Computer and video game addiction-​a comparison between game users and
non-​game users. American Journal of Drug and Alcohol Abuse, 36, 268–​76.
Wikler, A. (1961). On the nature of addiction and habituation. Addiction, 57, 73–​79.
Wolfe, J.B. (1934). The effect of delayed reward upon learning in the white rat. Journal of Comparative
Psychology, 17, 1–​21.
Woodside, D., Garfinkel, P., Lin, E., and Goering, P. (2001). Comparisons of men with full or partial
eating disorders, men without eating disorders, and women with eating disorders in the community.
American Journal of Psychiatry, 58, 570–​74.
Zack, M. and Poulos, C.X. (2007). A D2 antagonist enhances the rewarding and priming effects of a
gambling episode in pathological gamblers. Neuropsychopharmacology, 32, 1678–​86.
Zinberg, N.E. (1984). Drug, Set and Setting: The Basis for Controlled Intoxicant Use. New Haven,
CT: Yale University Press.
66  

Chapter 4

Willing addicts? Drinkers, dandies,

druggies, and other Dionysians
Owen Flanagan

Abstract
A willing addict is one who reflectively endorses their addiction. Some say
that there are no willing addicts, only unwilling addicts who are trying to
stop but not succeeding, or resigned addicts who are so demoralized they
have stopped trying to stop. Willing addiction reminds us that addiction is a
phenomenon that involves a person-with-a-lifestyle-inside-an-ecology, that
there are good-making features of using even addictively, and that addiction
involves many of the same complex, negotiated features (experiential,
personal, interpersonal, structural, and cultural) of other person-level
lifestyle choices. Willing addiction shows that concepts such as choice,
voluntariness, and reflective endorsement have a place in the psychology and
phenomenology of addiction. There are implications for the psych-bio-politics
of addiction and for the implausible idea that addiction is a brain disorder.
Addiction is not a brain disorder, even though the brain is one promising site
for intervention aspects or features of addiction that involve brain.

1  Willing, unwilling, and resigned addicts

It is easy to understand how one might become an alcoholic or addict unwittingly. One
drinks or uses drugs for fun, for novel experiences, to relax, to ease the pain, and then one
gets hooked. It is puzzling to imagine someone becoming an addict on purpose, and it is
puzzling in a different way to imagine someone who finds himself in the grip due to inad-
vertence, or because he thought this was what he wanted, continuing to think it is what
he wants after he has experienced full-​on addiction. It is natural to think that addiction
is bad, feels bad, harms the body, the mind, as well as one’s relations, family, love, work,
and thus that it is never choice-​worthy. If there are people who rationally choose addic-
tion, either to become an addict or once having become one to stay one, they are willing
addicts. A willing addict prefers, all things considered, to be an addict. He has a settled
disposition to continue to use in the way he uses. He endorses using the way he uses.
Willing addicts, if there are any, are contrasted with two more familiar kinds of
addicts. Unwilling addicts are individuals who want to quit, try to quit, but so far
  67

WILLING, UNWILLING, AND RESIGNED ADDICTS 67

fail. Resigned addicts are former unwilling addicts. They are demoralized, and have
given up trying to quit (Flanagan 2011, 2013a, 2013b, 2013c; Kennett et  al. 2013).
Willing addicts are unfamiliar, but they are out there and have lessons to teach about
the nature of addiction and the complex but very real attractions of living in zones of
mind alteration.
There is some controversy about whether willing addicts are possible on grounds that it
cannot ever be rational to choose addiction. On this view, someone who seems like a will-
ing addict is better understood as an unwilling or resigned addict who is self-​deceived,
confused, rationalizing, or lying. I think that willing addiction is possible. Indeed, I think
it is actual.
In this chapter, I discuss the reasons some addicts choose addiction, mostly to stay in
addiction (rather than to become an addict) rather than “recovery.” I explore a set of cases
that cluster around loving mind-​altering experiences and the things, alcohol and other
drugs, that produce them, as well as possibly loving a certain lifestyle that involves using
and drug-​altered experiences. The inquiry has implications for our views about whether
addiction is always a disorder, and whether it is a physical, mental, or moral disorder,
all or none of these. Some will say that if there is such a thing as willing addiction, it
is not really addiction because the addict accepts, even endorses, her pattern of using
and accepts the consequences. That’s OK with me. This sort of response indicates that a
person’s own attitudes toward her use is part of what goes into determining whether the
normal negative normative force of saying someone is an addict is warranted. That this
is so is part of my overall point. My aim is mostly to assist in the project of helping us
think of addiction as a complex relational lifestyle phenomenon that is social and spectral.
There is something wrong with the picture that sees addiction as an individual problem
and addiction as a problem of the brain, where the brain has crossed a threshold between
using normally and using abnormally or abusing, between non-​addictive and addictive
use. Things are much more complicated than that and involve a variety of person level
attitudes, regimens, and choice patterns, as well as social attitudes and permissions.
There are three caveats about the claim that a willing addict prefers, all things consid-
ered, to be an addict, that they have a settled disposition to continue to use in the way they
use, and that they endorse the way they use. First, most settled dispositions, plans, and
projects, even most reflectively endorsed dispositions, plans, and projects, involve some
ambivalence. Good marriages and good careers involve moments of second-​guessing,
even occasional feelings of quitting or throwing in the towel. Second, when I say that a
willing addict endorses using the way he uses or prefers, all things considered, to continue
using the way he does, there is this complexity:  many addicts, willing, unwilling, and
resigned, play with dosing, and with types and mixtures of booze, prescription drugs, and
illegal drugs. Raikhel and Garriott (2013, pp.  28–​29) describe a program in Baltimore
where many addicts creatively mix heroin (illegal), Suboxone (a prescription drug that
helps moderate the effects of opiates), and OxyContin, a prescription opiate, which they
could technically, but usually do not, acquire through normal medical channels. Third,
I do not think that most addicts are willing addicts or that willing addicts choose to use
68  

68 Willing addicts?

only for pleasure (Flanagan 2011; Foddy and Savulescu 2010). Most addicts are unwill-
ing or resigned (Flanagan 2011, 2013a, 2013b, 2013c). People willingly engage in all sorts
of activities including addictive use of alcohol and many drugs for reasons other than
pleasure, for example, tuning up or down one’s personality traits, taming physical pain
or painful emotions (which is not remotely the same as seeking pleasure), to have novel
or meaningful experiences, to enable smooth social relations, and for reasons of group
solidarity.

2  Six examples of willing addicts

2.1 Morphine Mary
Imagine an elderly woman Mary, who is given six months to live, and is put on morphine
to control pain and allow her to be present to herself and her loved ones as she dies. She
(or her family) asks if this means she will become addicted. Suppose the hospice doctor is
a geriatric oncologist who knows about and cares about the latest in the medical literature
on addiction (the tandem is extremely unlikely). The doctor explains that recent DSMs
do not contain an entry for “addiction.” Ordinary concepts like “addiction” and “alcohol-
ism”—​even though they are still used every day in every way by doctors, nurses, psychia-
trists, psychologists, social workers, and ordinary people—​are replaced in DSM-​IV by
the terms “substance abuse” and “substance dependence,” where the first refers to harm or
impairment, and the second to physical dependence, determined by such things as nega-
tive effects of coming off without a titration regimen, as well as a host of familiar behav-
iors such as an inordinate amount of time using, reduction of other meaningful activities
in order to use, and so on.1 The hospice doctor tells Mary truthfully that she will become
physically dependent on the morphine and thus implies without saying so explicitly that
she will have a “substance dependency” in that sense. (DSM-​IV requires at least three of
its seven characteristics or symptoms for a diagnosis of the bad kind of substance depend-
ence; and Mary will suffer only one—​brain dependence—​and that one would reveal itself
to Mary and her loved ones only if she were to stop receiving morphine abruptly.)
If the hospice doctor knows about and is impressed by DSM-​5, she can say some even
more consoling things: the twin diagnoses of “substance abuse” and “substance depend-
ency” are replaced with a unified category “substance use disorder.” Thus, since Mary
will be under an expertly administered medical regimen, there will be no “disorder;” she
is not scoring, i.e. “abusing,” street drugs. The regimen will ease the pain, help with nor-
mal sleep; it will enable cognitive alertness so she can read, watch movies, enjoy music,

1 In DSM-​IV, “abuse” refers to cases like binge drinking or drunk driving or taking hallucinogens and
rock climbing (at the same time). Abuse can, indeed it often does, occur without dependency. And
dependency, as with prescription drugs or daily drinking, often occurs without abuse in the DSM-​IV
sense. DSM-​5 acknowledges the complexities in defining and distinguishing among abuse, depend-
ency, and addiction, not by providing more analytic distinctions among these kinds, but by opening
one very large bucket of “substance use disorder” into which all the multifarious phenomena are put.
  69

Six examples of willing addicts 69

and spend quality time with loved ones until the end. Of course, there will be physical
dependency in the sense that if there is a miracle and her terminal illness evaporates she
will need to be stepped down slowly. Otherwise, there would be major “Jonesing,” which
the doctor calls “withdrawal symptoms.” But, of course, that is not really in the cards. Thus
reassured, Mary says, “Fine.”
I say Mary is a willing addict in two ways: she chooses to become addicted to mor-
phine in one perfectly legitimate sense of addicted; she succeeds, and wants things to
stay this way until she dies. Others might say that Mary is willing but not an addict, or
that she is an addict but that it doesn’t matter (I agree with this last). In both these cases,
it is an extrinsic feature of Mary’s situation that is the key. If she isn’t an addict, the key
is the way the morphine is administered and the reasons it is administered. We should
assume, after all, that Mary’s brain will come to crave morphine and become marked by
morphine dependency in all the same ways as the brain of an individual who comes to
be addicted after major surgery or who scores morphine on the street does. Anyone who
thinks that addiction is a brain disease or disorder will need to say that Mary is an addict
(see, for example, the American National Institute on Drug Abuse definition:  NIDA
2014). The same situation applies if we say Mary is an addict but it doesn’t matter in the
sense that we are not going to judge her as such. Here it is nothing about the morphine
or what goes on in Mary’s brain morphine-​wise that makes it a case that doesn’t matter.
It is the goal of using the drug that takes it out of the domain of normatively disapproved
addiction.
The word “addict” is public domain. Even if DSM-​5 doesn’t use the term “addiction”
(or “alcoholic”), it is not going away. (DSM-​5 does use the term “addiction,” but only
for a certain variety of gambling disorder.) In itself this is not a big problem. In all the
ordinary senses or usages of terms like “addict,” “addiction,” and their suite, Mary has
decided to enter a regimen of drug use that will lead to pain relief, help with patterns
of sleep and alertness, enable social relations, that will cause her brain to have the
same character and profile as the brains of unregulated morphine addicts, and thus
to addiction in one ordinary, everyday sense of the term. She is a willing addict, but
perhaps, indeed probably, she is not an individual with a DSM-​5 style “substance use
disorder.” Why not? Again, the answer(s) is (are) not entirely about Mary since she
really is choosing to take an addictive drug and to become physically dependent on it,
and were her doses to become unavailable she would crave big time, and were she to
try to come off quickly, she would experience all the symptoms of opiate withdrawal.
The fact that she chooses addiction for a good reason, a really good set of reasons,
doesn’t change that. The fact that her relation with her drug of choice is well regulated
is a feature of the institutions of medicine, the FDA, the law, not about Morphine
Mary’s will or her brain.

2.2  Keith Richards

The lead guitarist of the Rolling Stones is a different kind of case. In his autobiography,
Life (2010) and many interviews, Richards confirms that he was a long-​time heroin user
70  

70 Willing addicts?

(plus a user of most other things). Nowadays, Richards sticks mostly to red wine and
pot, which he uses regularly. Being the lead guitarist in the Rolling Stones makes it fairly
easy to get well-​controlled, high-​grade heroin. If one is willing to pay enough, one can
have the drug cartel equivalent of FDA-​approved safe doses of junk. Richards’ goals in
using heroin were not like Mary’s. He was an experimentalist looking for fun, novelty,
and creative sources. “I’ve got to confess, I was very interested in what I could take and
what I could do. I looked upon the body as a laboratory—​I used to throw in this chemical
and then that one to see what would happen; I was intrigued by that.” He adds, “It’s like
what Churchill said about alcohol, ‘Believe me—​I’ve taken a lot more out of alcohol than
it’s ever taken out of me!’ And I kind of feel the same way about the dope and stuff. I got
something out of it.”
Imagine that, based on full information, we were to discover that Richards’ brain lights
up (or used to light up) in exactly the way the hijacked brain of an addict ought to light up.
So he is an addict in one sense. But one might also think that the concept of addict doesn’t
apply or, what is different, shouldn’t be used. The normal rules don’t apply. In Morphine
Mary’s case it is weird to talk about addiction because of the goals (pain relief) of her tak-
ing (getting addicted to) morphine. And she is dying. In Keith Richards’ case, well, the
normal rules don’t apply either. After all, he is a fuckin’ Rolling Stone and thus he is not
involved directly (he has agents who do the dirty work of procuring his drugs at a very
high organized crime level) in getting pharmaceutical-​grade heroin, or top-​of-​the-​line
drugs of whatever sort he wants, not in the seamy life of getting God-​knows-​what-​grade
heroin on the street like normal addicts do. If Keith Richard is, or was, an addict, he is a
willing one.

2.3 Drugstore cowboy
Here is a different kind of case: in the tragic-​comedy movie Drugstore Cowboy, based
on a true-​to-​life book of the same name by James Fogle, Bob Hughes, the junkie pro-
tagonist, ends up in a halfway house where he steps down from a regimen of “any
mind-​altering substance he can get his hands on” to a culturally acceptable, state-​
administered methadone maintenance program. At this point in the film, Bob no
longer robs drugstores and hospital pharmacies to fuel himself and his elective family
of fellow junkies. He ran into “bad luck” when Nadine, a member of his band of merry
jacked-​up folk, died from an overdose. That was when he decided he had “had it.” Bob’s
girlfriend, Diane, doesn’t see the death of Nadine as the same kind of “bad luck” that
Bob does, that is, as reason to go on the wagon or, what is different, to change their
felonious method of drug acquisition to a state-​sponsored kind of addiction. Diane
visits Bob in his halfway house and Bob suggests she join him there in “recovery.”
But Bob’s new lifestyle is a non-​starter for Diane. She loves dope, is still unreservedly
committed to “the whole lifestyle,” and seems not to comprehend why Bob’s is doing
what he is doing. In fact, she is already the “old lady” of Nadine’s junkie widower Rick,
formerly Bob’s sycophantic underling. As a parting gesture of abiding affection, Diane
gifts Bob a satchel filled with a potpourri of drugs. But Bob, now committed to the
  71

Six examples of willing addicts 71

methadone regimen, passes his bounty on to Father Joe, an elderly Catholic priest
whom he has known since childhood and who is still an active dope fiend. Father
Joe, who Bob thinks “ought to be a philosopher” for his vaguely depressive existential
pronouncements, judges Bob worthy of divine indulgences for sharing his bounty. Bob
sums up his own situation at this point: “I am a junkie. I love drugs. I love the whole
lifestyle. It just didn’t work out so well.”
Bob Hughes, the drugstore cowboy, is the case of someone who chooses to be an addict
as part of a whole lifestyle. Bob is a big loser from the readers’ or viewers’ perspective. But
it is not clear that he is a loser from his own perspective. He’s a junkie who loves drugs
and the lifestyle. Things didn’t work out. So he made some sort of all-​things-​considered
adjustment. Of course, he is not off dope. He is now a methadone addict, or, more politely,
he is physically dependent on methadone.

2.4  Drinkers who are not (yet) piss artists

“Drinkers” differ from junkies, both in terms of their drug of choice as well as typically in
terms of gender, social class, country of origin, and the execution of their habit. Like Keith
Richards, “drinkers” have a certain impunity from severe social criticism. But whereas
Richards’ impunity from normal categorization comes from being a rock n’ roll God and
super rich, the category of “drinkers” are recipients of a certain social permission to be as
they are for quirky historical reasons.
“Drinkers” might fit the typical brain profile of an alcoholic perfectly. But they are will-
ing participants in their own alcoholism. They have no trouble negotiating the relevant
lifestyle because liquor is legal and affordable for them. For them, but not for everyone
who uses as they do, there are structures in place, often audience permissions, even appre-
ciation, to use to excess. There are jobs that cooperate, no partners, or partners who are
OK with things. Being male, being English and working in a creative profession provide
the context for this class of willing addict.2
Christopher Hitchens was a “drinker.” In Hitch-​22: A Memoir, he writes:
Alcohol for me has been an aspect of my optimism: the mood caught by Charles Ryder in Brideshead
Revisited when he discourses on aspects of the Bacchic and the Dionysian and claims that he at least
chooses to drink “in the love of the moment, and the wish to prolong and enhance it.” I dare say that
some people have seen me the worse for wear in less charming ways, but I know that I have been
true to the original as well. The Commander [his father] was not a happy drinker. He didn’t actually
drink all that much, but he imbibed regularly and determinedly, and it was a reinforcement to his
pessimism and disappointment, both personal and political.

(Hitchens 2010, pp. 34–​35)

2 Qat (khat) is an amphetamine-​like substance commonly chewed on a daily basis by men in Yemen
(although women participate on weekends). The World Health Organization (WHO) considers qat
addictive. Suppose it is, then a lot of Yemeni men are addicts, willing addicts with social permission to
be addicts.
72  

72 Willing addicts?

His father’s attitude toward gin was, we are told, “a relatively devout one.” Later in the
memoir in a hilarious half-​dozen pages entitled “A short footnote on the grape and the
grain,” Hitch takes up the widespread view that he is a “drinker,” a prodigious drinker. He
does not deny this. But he says in his endearing, immodest way that his writing and public
speaking regimen would have been impossible were he what “the English so bluntly call
a ‘piss-​artist.’ ” Hitchens is well aware that “piss artistry” is “the professional deformation
of many writers and has ruined not a few. (I remember Kingsley Amis, no slouch, saying
that he could tell on what page of the novel that Paul Scott had reached for the bottle and
thrown caution to the winds.)” Hitch then explains:
I work at home, where there is indeed a bar-​room, and can suit myself. But I don’t. At about half
past midday, a decent slug of Mr. Walker’s amber restorative, cut with Perrier water (an ideal
delivery system) and no ice. At luncheon, perhaps a half bottle of red wine: not always more but
never less. Then back to the desk, and ready to repeat the treatment at the evening meal. No “after
dinner drinks”—​most especially nothing sweet and never, ever any brandy. “Nightcaps” depend
on how well the day went, but always the mixture as before. No mixing: no messing around with
a gin here and a vodka there. [A]‌lcohol can make other people less tedious, and food less bland,
and can help provide what the Greeks called entheos, or the slight buzz of inspiration when read-
ing or writing.

(Hitchens 2010, p. 351)

A bit later he writes:

“Hitch: making rules about drinking can be the sign of an alcoholic,” as Martin Amis once teasingly
said to me … Of course, watching the clock for the start-​time is probably a bad sign, but here are
some simple pieces of advice for the young. Don’t drink on an empty stomach: the main point of the
refreshment is the enhancement of food. Don’t drink if you have the blues: it is a junk cure. Drink
when you are in a good mood. Cheap booze is a false economy. It is not true that you shouldn’t
drink alone: these can be the happiest glasses you ever drain. Hangovers are another bad sign …
Avoid all narcotics: these make you more boring rather than less and are not designed—​as are the
grape and the grain—​to enliven company.

(Hitchens 2010, pp. 351–​52)

I admit that Hitchens’ case is determinate across a host of dimensions that make it hard to
know whether he counts as an alcoholic or not. He drinks a lot, loves drinking, loves the
effects, and has a wife who is OK with it. Friends (I know some) wondered and worried
about his prodigious capacity. If the brain and the body develop various kinds of depend-
encies through regular use, his body and brain have those kinds of dependencies. But he
is entirely on board with his drinking.

2.5  Drinking in Mecca

Lawrence Osborne’s The Wet and the Dry (2013) is another testimonial by a “drinker.”
Osborne is not quite a perfect example of a willing addict, but he is not unwilling or
resigned either, and he is, by his own standards, alcoholic. An experienced and distin-
guished travel writer, he decides on a journey to the Middle East to see if non-​drinking
Muslims have anything to teach, and wonders ambivalently whether new social atmos-
pherics and laws against drinking might change his behavior. They don’t.
  73

Six examples of willing addicts 73

Osborne is “a visiting alcoholic,” and insinuates at one point that he is “America’s great-
est vodka critic.” (He is, as “drinkers” tend to be, English, but has lived in NYC for a
long time.) It encourages Osborne when he comes upon an ancient Dionysian temple in
Lebanon and reminds himself, “Dionysianism was the most popular religion before the
arrival of Christianity and eventually Islam.”
He gets that drinking has its downside. There are the hangovers, the blackouts, and the
bad behavior. But he contrasts his kind of distinctively posh English drinking with mor-
bid and escapist drinking.
The Irish, I am bound to admit, have a particular penchant for this relation between intoxication
and morbidity. “I have absolutely no penchant,” Edgar Allan Poe once wrote, “in the stimulants in
which I sometimes so madly indulge.” It has not been in the pursuit of pleasure that I have periled
life and reputation and reason. It has been the desperate attempt to escape from torturing memo-
ries, from a sense of insupportable loneliness and a dread of some strange impending doom.

(Osborne 2013, pp. 165–​66)

The Plains Indians of the nineteenth century, Osborne reflects, became addicted in the
Irish way without an Irish psyche, and drink “destabilized them into madness, into altered
states of melancholic strangeness, and ultimately of course it destroyed them” (p. 167).
But Osborne is upbeat at least as far as his personal situation goes. The “drinker,” he
writes,
is a self-​critic, a connoisseur of his own altered states, he knows exactly how to tweak himself
upwards and downwards. He is an amateur alchemist when it comes to the drinks themselves. If
he is a writer, and wanted to explain himself to strangers, he would write a book called In Praise of
Intoxification. No one would invite him to explain his views in public.

(Osborne 2013, p. 64)

One might think here of great British actors like Richard Harris, Peter O’Toole, and
Richard Burton. Were they each perhaps a posh Bob Hughes? Harris and O’Toole
eventually went on the wagon, when “things didn’t work out,” but they never denied
that they loved it while it lasted. Were they willing addicts while it lasted? Probably.
Burton died young, probably related to his alcoholism. Was he a willing addict to the
end? Maybe.
Osborne explains how English culture allows willing addiction among people of certain
means and social class: “The English are very indulgent to episodes of alcoholic insanity.
They strike them as sympathetic, understandable, and a sign of being a real human being”
(Osborne 2013, p. 51). “The English relationship to drink is so deeply burned into my way
of being in the world that to write about drink is to simultaneously write about England,
a country I know almost nothing about since I have lived in New York close to twenty
years” (p. 58). Later he writes: “The drinker’s legendary unhappiness and frustration are
often exaggerated … The daily intoxication arises from an entire life’s experience, not
from an ‘illness’ that is supposed to be less mysterious” (p. 61).
Osborne is mostly able to regulate his drinking—​and has a certain sympathy for both
himself and the lifestyle of the English drinker. He is certainly not a resigned alcoholic.
74  

74 Willing addicts?

If he is still drinking, he fits the willing addict profile. Or he is at least, as they say, on the
spectrum.

2.6  True Detective

One last example: in the HBO show True Detective, Matthew McConaughey plays Rust
Cohle, an iconoclastic former Louisiana State police detective, who together with his
partner Martin Hart (played by Woody Harrelson), is under investigation for the irregu-
lar, possibly criminal, methods they used to solve a murder case 17 years earlier. Cohle is
retired and a drinker. He is not leading the most attractive life but he nonetheless seems
OK with it. We might think he must be resigned (how could anyone willingly live like he
does?). But he doesn’t come across as resigned, even if not 100 percent willing. Here is a
relevant piece of his interrogation:
Rust Cohle smokes, listens to his interviewer. Perhaps slightly buzzed, he speaks a little discursively
when he answers—​

COHLE: Yeah. Just married once. Came close another time. That was later on. Marty introduced
us. She broke it off. For the best. I gave her cause … I can be hard on people.

He puts his cigarette out, thinks, sips his beer. In an oddly confessional outpouring—​

COHLE (CONT’D): I can pick people apart. I don’t mean to. But I can be critical. I have problems
getting angry … I’ve had the thought before that I’m not good for other
people. Like it’s no good for them to be around me long. I wear them down.
They become unhappy. I don’t mean for that. But life works its way through
you … I can’t say the job made me this way. More like, me being this way
made me right for the job. I don’t know. I used to think about it a lot. Not so
much now. You reach an age and you understand who you are …

A beat as he stares into the camera, unable to fully hide the pathos in his eyes—​

COHLE (CONT’D): It’s better now. I live in a little room out in the country. Tend bar four nights a
week. In between I drink, and there’s nobody around to make me stop. Which
is good. I  understand who I  am. After all these years … There’s a victory
in that.

Cohle’s flat gaze carries no sense of achievement—​

This case is more puzzling than the previous cases, in large part because Russ Cohle
is buzzed during the interrogation and he seems alternately slightly amused, alienated,
pissed, disaffected, and removed. Is he therefore a resigned addict and not a willing one?
I don’t think so. Resigned addicts wish they could quit but can’t and have stopped mak-
ing the effort. Willing addicts want to keep using (although possibly on a better, less
debilitating dosing regimen). One might be locally or globally resigned. Imagine a per-
son who, like Morphine Mary, learns she is terminally ill. She is resigned to her fate. But
suppose she chooses to learn to play the piano in the time she has left. She is globally
resigned to her fate as one who is soon to die, but not locally resigned to the idea that
she should only be getting her affairs in order. On the other hand, a person might feel
  75

Social and cultural permissions 75

totally in charge of her life, but feel (locally) resigned, say, to her smoking habit. Cohle is
a smoker as well as a drinker. It seems pretty clear that cigarettes are addictive and that
some smokers are resigned addicts—​they have tried to quit, can’t, and have given up try-
ing. But others are willing addicts—​they get the risks, like to smoke, and don’t try to stop.
It is best overall to understand Cohle’s alcoholism as a variety of willing addiction. What he
says remember is this: “I live in a little room out in the country. Tend bar four nights a week.
In between I drink, and there’s nobody around to make me stop. Which is good. I understand
who I am. After all these years … There’s a victory in that.” If we take Cohle seriously he
understands who he is and sees that as a victory. He is best alone and drinking. There is some
embrace of this, some kind of reflective endorsement of himself as a loner and as a drinker.

3  Social and cultural permissions

There is evidence in what I have said so far that there are various kinds of cultural and
economic variation as well as differences in individual psychology that allow willing
addiction or an addictive lifestyle without the normal costs that lead to social and self-​
stigmatization. Being English, being rich, being an artist, and being male look like fea-
tures of a culture that lower the threshold for getting fussy about addiction, alcoholism,
and substance abuse disorders, as well as for the addict not experiencing the familiar
kinds of shame, depression, embarrassment, and even loss (of jobs, loved ones) that most
addicts suffer (Flanagan 2013b).

3.1  Artistic permission

Consider artistic permission: Kay Redfield Jamison, Touched with Fire: Manic Depressive
Illness and the Artistic Temperament (1993) is devoted to exploring the fact that “writers
and artists show a vastly disproportionate rate of manic-​depressive or depressive illness”
(p. 6). She claims that all the following fit the bill: Lord Byron, Percy Bysshe Shelley, Keats,
Schumann, Tennyson, van Gogh, Poe, Fitzgerald, Hemingway, Melville, Robert Lowell,
Anne Sexton, Delmore Schwartz, John Berryman, Graham Greene, John Cheever, and
William Styron.
There is also this link, which is important for present purposes: there are elevated rates
of alcohol and drug abuse among populations with bipolar disorder—​“60% have a history
of some kind of substance abuse or dependence” (p. 39); “conversely, there is a signifi-
cantly higher percentage of bipolar patients in populations of alcoholics and drug abus-
ers” (p. 37).
How are we to explain this apparent link between creativity, bipolar disorder and addic-
tion? One hypothesis that I am fond of is that creative genius is not best understood as
a property of individuals—​they are geniuses—​a gift such individuals have and others do
not have in themselves, in their psyches, but rather that genius is a relational property that
is partly constituted by the audience response to the artist’s expression. Artistic expres-
sion, especially well-​received artistic expression, makes or allows the audience to see bet-
ter what is there (think the French impressionists) or to go beyond some conventional
76  

76 Willing addicts?

boundary to what is different, tantalizing, interesting (think Stravinsky or Schumann).

So one idea is that bipolar people have unusual experiences more than most; they are less
inhibited than non-​bipolar people about expressing these unusual experiences; in manic
phases they have more energy to sustain these unusual expressions. High rates of alcohol
and drug abuse are explained by the triad of desires to self-​medicate, the disinhibition
that partly constitutes mania, and by social encouragement for their “creative” expres-
sions, which can be or can be thought (by them and by us) to be enhanced by drugs and/​
or alcohol.

3.2  Rainbow addicts: unwilling or resigned to willing?

The final example is not about posh folk or creative artists who have various kinds of
means and permissions to be willing addicts, but involves people who are not posh at
all. They are, or so it seems to me, recipients of a kind of compassionate social policy
that allows them a chance to move from being ashamed unwilling or resigned addicts to
something closer to a willing addict who is helped to find a way to be an addict and live in
a manner that is acceptable both subjectively and by the lights of the wider community.
The Rainbow Group is an NGO committed to helping people with such problems as
homelessness, psychiatric illnesses, drug use, and alcoholism. One program they run in
Amsterdam involves drinkers who are paid a salary and a beer stipend to pick up trash.
This is one of many techniques practiced across the world that involve contingency
management (Raikhel and Garriott 2013; Pickard and Ahmed, Chapter 2, this volume),
where the contingencies that are managed are managed by the addicts as well as by com-
munal, legal, and political institutions. The workday begins, as it often did in colonial
times at many workplaces in the Northeast of the US, with the workers being offered
(they always accept) two pints of beer. There are two more at lunch, and then two more at
closing time. Six pints of beer spread throughout the workday. The workers are also paid
a salary of 10 euros per day ($13), half a pack of rolling tobacco, and a free lunch. This
is more or less the alcoholic equivalent of a methadone maintenance program but with
work required. Workers call beer their medicine; some report continuing the slow drip
method of drinking after work, but also feeling renewed self-​esteem and self-​respect.
Once these men were (and would self-​describe as) alcoholics, losers, drinkers, piss art-
ists, drunks, and chronic relapsers. They once were unwilling addicts, trying and failing
to stop, or resigned addicts, feeling hopeless, demoralized, and no longer trying to stop.
Now one can imagine that they, like people on methadone maintenance programs, feel
like the term alcoholic or addict doesn’t apply or doesn’t quite apply. If this is right, then
what is different in these alcoholics is not that their brains don’t light up the way alco-
holic brains light up, or that they would not experience the phenomenology of craving
were they not to get dosed properly, but something about their attitude and that of (at
least some) others toward their using. The alcoholics accept, even embrace, that they
are alcoholic. They accept a regimen of partial moderation, really maintenance dosing,
but only during work hours, and they are able to make a living wage doing something
socially useful. This enhances self-​esteem and self-​respect. I  said earlier that resigned
  77

Is addiction a brain disorder? 77

addicts are typically former unwilling addicts. They have tried to quit, often with utmost
sincerity and great effort, but failed again and again. Eventually, they gave up trying to
quit. The Rainbow Group alcoholics might fit the bill of willing addicts who were for-
merly unwilling addicts and then, after that, hopeless resigned addicts. It is hard to know
if this is the best way to describe them and the situation. But something, actually several
things, individual and social, have improved.
The phenomenon is holistic; it takes particular individuals, in a particular culture,
with a certain set of attitudes (compassion; harm reduction), permissions (drink), and
demands (work), and an NGO that is able to implement a practical program. Would it be
best to manage contingencies so that willing addiction was more of a live option than it
is now by making addiction less costly to addicts by, for example, changing the monetary
cost and legal status of drugs and by promoting programs like the Rainbow Group’s Dutch
experiment? Probably.
Doing so would not mean full-​on advocacy of alcoholism or drug addiction as an excel-
lent way of living, a good option for most people. Even if we in liberal societies decide to
provide space for some addicts to be both with their addiction willingly and of service to
their community, there would still be all the remaining normative space for friends, fam-
ily, and employers to make clear even to willing alcoholics and addicts that there will be
prices to pay for choosing booze and drugs. Furthermore, for many, possibly most alco-
holics and addicts, drinking and drugging will eventually stop doing for them whatever
it once did for them; instead it will do something else. There will be psychic hell to pay,
and the alcoholic or addict will want to stop. Luckily, where there is such a will, there is
usually a way.

4  Is addiction a brain disorder?

What about the idea that addiction is ultimately a brain disease, that a common neuro-
logical disorder binds all addicts and is what ails them? One could in principle hold the
view that addiction is a brain disease and also accept my argument that there are some
addicts who are willing. The view would be that there is a spectrum of addiction types that
include willing addicts, unwilling addicts (addicts who are trying to stop but can’t yet),
resigned addicts (who have thrown in the towel on trying to stop), and former addicts
who, we will suppose, remain highly vulnerable. What makes them all addicts is the brain
disease (which in the case of the former addicts is, I guess, “in remission”).
While it is true that the view that addiction is a brain disease is compatible with my
argument that addiction comes in varieties, including the variety of willing addiction,
I resist the idea that addiction is a brain disease (Levy 2013; Hall et al. 2014). I claimed at
the beginning that an addict is a person and addiction is a lifestyle, so I prefer that addic-
tion does not turn out to be a brain disease because, well, a brain state (even a complex
dynamic, temporally extended one) is not a lifestyle.
Consider this: in AA (and in NA too), one hears that most everyone had some success in
stopping, but most had trouble staying stopped. We sometimes say “the off switch” is broken
78  

78 Willing addicts?

or works erratically. So if there is craving, the person uses, and then can’t exert normal will-
power to stop. This is not a bad metaphor. The idea that alcoholism, as well as other kinds of
substance addiction, involves a brain component and thus that certain aspects or features of
alcoholism are in the brain can be useful here. It is worth noting that AA says that alcoholism
is a disease; although interestingly the AA view is that alcoholism is a disease of body, mind,
and spirit, which I take to be closer to the lifestyle view than the brain view. Regarding the
brain, there are nowadays drugs that reduce the reinforcing effects of using drugs or alcohol.
There are drugs that will make one sick if one uses. There are drugs that reduce anxiety and
depression, which, depending on the alcoholic, make her vulnerable to relapse. There are
drugs for bipolar addicts who are most vulnerable to relapse during manic phases, and so
on. Are these drugs treating a brain disease? And if so, which one? The self-​esteem one? The
depression one? The mania one? The anxiety one? The stress one? Or the “off switch” one?
In my experience, it is usually the “off switch” one that those who believe that addiction
is a brain disorder will say is the problem and that needs to be treated, fixed, circum-
vented, worked around, or rewired. To be sure, the “off switch” problem might be the
major obstacle to doing what one wishes—​assuming one is an unwilling addict—​once
deep into addiction, namely, quitting. But the “off switch” problem is not the disease (or
condition) of addiction. If there is an “off switch” problem, and there commonly is, it is a
major source of difficulty in achieving the wish to stop. It is a major obstacle in the addict’s
way as they try to stop using. But the “off switch” problem is not their alcoholism, not
what their addiction is, or even where it is.
An “off switch problem” metaphorically names some set of actual problems in, we’ll
suppose, the mid-​brain with the dopaminergic system, the decoupling of the “wanting”
and “liking” systems—​I want/​need what I no longer overall like (Robinson and Berridge
2008),3 or the problem at the neural level that disables or blocks successfully executing
instructions to self that “This is no longer going well, cease and desist!” It is possible,
indeed I think it likely, that several aspects or features of addiction are quite literally in
the brain. The pathways that quickly reactivate craving after relapse in alcoholics are
likely largely in the brain, and the compromised systems that create the seemingly insur-
mountable problems in exerting willpower are almost certainly in the brain (Everitt and
Robbins 2013).
An addict, like everyone else, is a fully embodied being-​in-​the-​world, and every per-
ception, mood, emotion, thought, and action involves their particular nervous system.
But addiction is no more in the brain than walking or being married or having a job is in
the brain. Still, one can sometimes effectively leverage or reorient or help the addict man-
age to do what they wish, STOP, by leveraging some particular neural pathways. This is a
practical psychobiological fact, and has no ontological implications as far as what addic-
tion is or where it is. None.

3 It is a common mistake for non-​addicts to infer from the evidence that the “wanting” and “liking”
systems decouple in the brain, that alcoholics and addicts don’t find relief in and thus like, in the literal
sense, the first drink or dose. Thanks to Hanna Pickard for asking me to clarify this.
  79

Conclusion 79

Suppose one says: OK, maybe addiction is not in the brain, but surely its cause is
in the brain. There is this reply:  there are almost no mono-​causes in the world for
anything, and thus speaking of the cause of anything is almost always practical short-
hand, not metaphysically or scientifically precise. Becoming alcoholic or a drug addict
requires using a lot over time. The vodka, the heroin, or the cocaine are causes, as are
the histories and schedules of use. Genes make causal contributions. Depending on
the addict, idiosyncratic features, shyness or gregariousness, low or high self-​esteem,
depression or mania, being alone or with others, low stress or high stress, being happy
or sad, or a personality disorder, make causal contributions. Eventually, the drinks or
the drugs affect the “off switches” in many addicts’ brains, so that even if they want to
stop, they have trouble. This too—​an erratic or broken “off switch”—​is a causal contrib-
utor. But no single causal contributor is the cause. So addiction is not a brain disease,
nor is the cause in the brain. This is Logic 101. Neuroscience 101 too. That said, addicts
have brains and the brains of addicts are one of many features of the world that can
be leveraged in treating addiction. These include features of the socioeconomic sur-
round, culture and cultural history, the ends or goals of drug users, and the norms and
practices of communal, legal, medical, and therapeutic institutions. We live at a time of
the ascendency of “neurospeak,” so it matters that in the rush to find brain markers for
(certain kinds) of addiction and substance abuse we do not lose sight of other features
of these complex phenomena. Note finally, to emphasize the point, that I have not said
that addiction is not an affliction, a disease, or a disorder. I think it usually is subject to
such normative assessments, including in the typical case from the perspective of the
drinker or user himself—​but not always, as the case of willing addiction reveals. I have
denied that the affliction, disease, or disorder is best conceived as a brain affliction,
disease, or disorder. It isn’t, although some aspects of addiction, especially late-​stage
compulsivity, is likely due to neural systems that create major obstacles to reasserting
self-​control (Everitt and Robbins 2013).

5  Conclusion
In Addiction Trajectories (2013), Eugene Raikhel and William Garriott, leaders in the field
of the anthropology of addiction, write that thinking and speaking about addiction run
along three trajectories:
(1) the epistemic trajectories traced by categories and concepts of addiction as they change over time
and move across institutional domains; (2) the therapeutic trajectories of treatments as they move
through distinct cultural and organizational settings; and (3)  the experiential and experimental
trajectories of lives constituted through the terrains of addiction and subjectivity.

(Raikhel and Garriott 2013, p. 2)

These trajectories help explain how and why categories used to describe the phenomena—​
“addiction,” “alcoholic,” “addict,” “brain disease,” “disorder of choice,” “interpersonal
disorder,” “social problem”—​vary with treatment regimens, the variable statuses of the
sciences or research programs that endorse them, the city or nation state location of users,
80  

80 Willing addicts?

the socioeconomic status of users and the ends of users—​fun, pleasure, creativity, killing
mental or physical pain, and ennui.
The book reinforces in my mind the view that there is no natural kind that is addic-
tion; it is, at best, a complex sociocultural kind. One can stipulate, as some do, that
bona fide or “real” addiction to some substance has as its necessary and/​or sufficient
condition(s) some brain profile (“midbrain hijack”) or some phenomenological profile
(overwhelming desire to use and to stop using; shame; obsession and craving) or some
behavioral profile (certain frequency and amount of use; plan to stop using but inability
to do so; chronic relapsing). Such stipulation might be useful for certain purposes, for
example, creating priority lists for in-​patient rehabilitation facilities or making deter-
minations about who will benefit from what kind of prescription drug, social work
intervention or psychological therapy. But the phenomenon of addiction in real-​world
ecologies is not so tame that it can be corralled by a measure across a single dimension.
The phenomenon of willing addiction reminds us that there are different ways even
full-​on addicts can respond to their addiction, including reflective endorsement of it. The
conditions under which such endorsement or acceptance occur are, to a point, matters
of individual psychology or subjectivity, involving how much fun, pain relief, experien-
tial novelty, and meaningfulness altered states provide or have for a particular addict.
But willing addiction is not only an individual matter. It requires social cooperation,
various permissions (interpersonal, legal) that allow an individual to be as they choose.
Sometimes such permissions align in disturbing but not unfamiliar ways, with money
(lots), race (white), and gender (male). Should the impunities that allow for willing addic-
tion be extended to all rather than only to some and in ways that are not classist, racist,
and sexist? Unless one thinks that abuse, dependency, and addiction are always and eve-
rywhere bad, then one ought to answer “Yes.”
The phenomenon of willing addiction makes visible modes of being an addict (abuser,
dependent) that involve all-​things-​considered endorsement by the addict, and which we
who are not ourselves willing addicts can perhaps see as not an entirely awful lifestyle
for them. Possibly, in some cases we can even see the appeal, the charm of using mind-​
altering drugs, not just for them, but for us. This much can help with sympathetic under-
standing. Willing addiction also helps us see ways that some of the many contingencies
governing the ways we think of and treat addicts can be adjusted or modified. Many, per-
haps most, addicts will eventually want their way out of active addiction. Increasingly,
there are effective methods to help with this project of graduating to being a former
addict or alcoholic in recovery. For individuals who cannot find a way to becoming for-
mer addicts, practices like those of the Rainbow Group allow some unwilling or resigned
addicts to become willing self-​respecting addicts, which seems like a good thing.

Acknowledgments
Thanks to Hanna Pickard, Larisa Rachel Svirsky, Nick Heather, Gabriel Segal, Katherine
Rickus, Matthew Parrot, James South, Rachel Cooper, and Carwyn R. Jones for helpful
comments. This chapter is based on a paper delivered at the 2nd International Conference
  81

Conclusion 81

on Philosophy of Psychiatry, St Hilda’s College, Oxford, UK, July 2015 and at the 51st
Nobel Conference at Gustavus Adolphus College, St Peter, Minnesota, October 2015.

References
National Institute on Drug Abuse (NIDA) (2014). The science of drug abuse and addiction: the
basics. http://​www.drugabuse.gov/​publications/​media-​guide/​science-​drug-​abuse-​addiction-​basics.
Accessed 02/​05/​2015.
Everitt, B.J. and Robbins, T.W. (2013). From the ventral to the dorsal striatum: devolving views of their
role in drug addiction. Neuroscience and Biobehavioral Reviews 37(9 Pt A), 1946–​54.
Flanagan, O. (2011). What is it like to be an addict? In: J. Poland and G. Graham (eds), Addiction and
Responsibility. Cambridge, MA: MIT Press.
Flanagan, O. (2013a). Phenomenal authority: the epistemic authority of Alcoholics Anonymous. In: N.
Levy (ed.), Addiction and Self Control. Oxford: Oxford University Press, 67–​93.
Flanagan, O. (2013b). The shame of addiction. Frontiers of Psychiatry 4, 120.
Flanagan, O. (2013c). Identity and addiction: what alcoholic memoirs teach. In: K.W.M. Fulford et al.
(eds), The Oxford Handbook of Philosophy and Psychiatry. Oxford: Oxford University Press, 865–​88.
Foddy, B. and Savulescu, J. (2010). A liberal account of addiction. Philosophy, Psychiatry, and Psychology
17(1), 1–​22.
Hall, W., Carter, A., and Forlini, C. (2014). The brain disease model of addiction: is it supported by the
evidence and has it delivered on its promises? Lancet: Psychiatry, 2(1), 105–​10.
Hitchens, C. (2010). Hitch-​22: A Memoir. New York, Twelve.
Jamison, K.R. (1993). Touched with Fire: Manic-​Depressive Illness and the Artistic Temperament.
New York: Free Press.
Kennett, J., Matthews, S., and Snoek, A. (2013). Pleasure and addiction. Frontiers in Psychiatry, 4, 117.
Levy, N. (2013). Addiction is not a brain disease (and it matters). Frontiers in Psychiatry, 4, 24.
Osborne, L. (2013). The Wet and the Dry: A Drinker’s Journey. New York, Crown.
Raikhel, E.A. and Garriott, W. (2013). Addiction Trajectories. Durham and London: Duke
University Press.
Richards, K. and J. Fox (2010). Life. New York, Little, Brown and Co.
Robinson, T.E. and Berridge, K.C. (2008). The incentive sensitization theory of addiction: some current
issues. Philosophical Transactions of the Royal Society, Series B 363(1507), 3137–​46.
82

Chapter 5

Failures of rationality
and self-​knowledge in addiction
Thomas Crowther

Abstract
The focus of this chapter is the account of akrasia in alcoholism suggested
by the Alcoholics Anonymous literature. The chapter begins by sketching
out this account of akrasia. It then goes on to raise a number of questions
about this account. It attempts to resolve these questions in terms
of the idea that alcoholic relapse involves a disorder of mental state.
Various features of alcoholic action and thought in relapse are traced to
characteristic features of this state. The chapter develops the idea that such
states as “obsession with alcohol” are malformed varieties of the normal
condition of wakefulness, and that aspects of the disorder involved in
alcoholism can be explained in terms of this relation.

1 Introduction
According to a traditional characterization, akrasia is the phenomenon of freely or
intentionally acting against one’s knowledge, or belief, about what it is best for one to
do, and against what one so rationally desires to do.1 Akrasia appears to be essential to
addiction. Take Bill Wilson, the co-​founder of the Alcoholics Anonymous program of
recovery.2 Bill Wilson possessed abundant evidence of the damaging and tragic conse-
quences of his drinking; his alcoholic drinking had not only cost him numerous jobs
and a promising career, it had placed his marriage under unbearable strain and led him
to periods of hospitalization and psychiatric care. He was not unaware of this evidence.
He knew that the best course of action for him was to stop drinking, and he had a desire
not to drink that cohered with this belief. For a long period of time, he attempted to
stop drinking, seeking sources of relief both medical and religious for his problem. Yet

1 For similar characterizations of akrasia see Coope (2010), Lorenz (2014), Broadie (1991, ch. 5), Charles
(1984, ch. 3), and Davidson (1970). The locus classicus for the traditional notion of akrasia is Aristotle’s
Nicomachean Ethics, Book VII. See Aristotle (2002, pp.  385–​407) for detailed philosophical com-
mentary. A  different approach is developed in Holton (2009). For discussion of the implications of
Davidson’s approach for contemporary addiction science see Heather and Segal (2013).
2 For a brief history of Bill Wilson’s career as an active alcoholic see Alcoholics Anonymous ([1939] 2001,
ch. 1).
  83

INTRODUCTION 83

despite what he knew, and despite the many reasons he possessed for staying sober, he
continued to drink.3
The notion of akrasia is not a philosopher’s invention, nor a notion of narrowly academic
interest. Attempts to understand akrasia provide important aspects of the rationale for pro-
grams of treatment and recovery for addicts. The impact of akrasia on the life of an addict
and those around them is measured not only in the physical and financial consequences of
their addiction. In their akratic behavior, active addicts are hard to understand, and this is
an important sense in which addicts are lost to, or estranged from, their friends and families.
The topic of akrasia in addiction is the starting point of discussion here. The aims of
this chapter are modest, and the focus is narrow. There is much here that remains pro-
grammatic. The overall aim of the chapter is to use reflection on the problem of akrasia
in addiction to introduce into philosophical discussion of addiction the notion of a dis-
tinctive kind of “mental state” and to try to show what explanatory work such a notion
can do. I will here discuss not substance addiction in general, but only certain aspects of
alcoholism. And even then I will discuss only aspects of that condition as it is conceptual-
ized from the point of view of one approach to treatment and recovery.
Though the topic of the nature of alcoholism is a theme of this chapter, I make a few
starting assumptions. I  will be talking about alcoholism as the condition is found in
human beings, beings with the powers of rationality and for self-​consciousness. I assume
that alcoholism is a condition essential to which is a pattern of repeated drinking. The
pattern often begins in the alcoholic’s teenage years and then develops usually over a
long period of time. I will take it to be essential to alcoholic drinking that it is marked
by a degradation of control over the amount that is drunk. I assume that alcoholism is a
condition that has a history, and that the way that alcoholism manifests itself develops
over time with distinctive features in evidence at different stages in its progression.
Alcoholics Anonymous is a self-​help organization that was founded in 1935 and has grown
to become the pre-​eminent non-​professional source of support for the recovering alcoholic.
The question that is the point of entry for this chapter is: what account of akrasia, as a fea-
ture of alcoholism, does the literature of the Alcoholics Anonymous program offer? This
discussion needs to be prefaced with an important qualification. The AA literature does not
offer a formal statement of the problem of akrasia and an account of how to understand the
problem. The book was written to help alcoholics recover, not to contribute to philosophical
theory. Alcoholics Anonymous does not take any part in public or academic debates about
alcoholism, and the discussion of alcoholism in its literature is not to be understood as con-
stituting the endorsement of a position within such a public or academic debate.4

3 Bill Wilson had his last drink in 1934, as the result of a conversation during a chance encounter with
an old school friend. He remained continuously sober until his death in 1971. For more biographical
detail see Alcoholics Anonymous (1984).
4 Alcoholics Anonymous (1952, pp. 176–​79) provides a general characterization of Tradition Ten: that
“Alcoholics Anonymous has no opinion on outside issues, hence the A.A.  name ought never be
drawn into public controversy.” The main consequence of this stance is the neutrality of Alcoholics
Anonymous concerning matters of public health policy related to alcohol.
84  

84 Failures of rationality and self-knowledge in addiction

That being said, professional philosophers are at liberty to do as they please with this
literature. The core texts of the organization are available in their entirety, for free, on the
website of Alcoholics Anonymous, and non-​alcoholic members of the public are able to
purchase copies of these texts from the AA website. Suppose professional philosophers
were to treat what was advanced in the foundational texts of Alcoholics Anonymous as
an attempt to explain akrasia. What is that account? In section 3, I sketch in that account.
In section 4, I raise some questions about how elements of this account cohere. In the rest
of the chapter, I make some suggestions about how this coherence is to be understood—​
suggestions that draw on a relatively neglected topic in the philosophy of mind. Here the
approach is only drawn in outline. Questions of detail, and about the adequacy of this
approach to akrasia, will be pursued elsewhere.

2  Two initial responses

I want to determine some desiderata for the discussion that follows by considering briefly
two natural responses to explaining akrasia. One such line of thought would be the view
that alcoholic relapse involves a wholesale switch in attitudes and valuations; the relaps-
ing alcoholic no longer knows or believes that it is best to abstain, and no longer desires
to abstain. In their place are states such as a belief that it is unproblematic to take a drink
and desirable to do so, states that then motivate action in the normal way. A second line of
thought would be that the alcoholic’s drinking in relapse is not an instance of intentional
action at all, but something like an automatic bodily movement, akin to the narrowing
of the iris in response to changes in light. As such, it is not capable of being evaluated as
either rational or irrational.
Though I don’t have the space here to offer a fully satisfactory discussion of these lines of
thought, there is good reason to think that they ought to be resisted. The difficulty with the
first line of thought is how to make sense of what is involved in “losing” and then “regain-
ing” the relevant attitudes about alcohol, during the period of relapse and recovery from
relapse. The alcoholic who recovers consciousness after relapse does not need to reacquire
the knowledge that it is best for him to abstain from drinking or reacquire the desire not
to drink. The guilt, shame, and remorse that consume him on waking is a manifestation of
the presence of these states; and the best explanation for this is that those states persisted
during the period of relapse, as such states persist during sleep. The second line of thought
appears to involve an implausible characterization of relapse. Even if alcoholics do often fail
to control their drinking in various ways, it at least makes sense for an alcoholic to attempt
to control it, perhaps by exercising willpower in the face of temptation, or by exercising
intelligence in avoiding temptation entirely. The idea of an alcoholic attempting to exercise
control over his drinking makes sense in a way that it doesn’t make sense to think of an alco-
holic attempting to refrain from narrowing his iris in the face of changes in illumination.
Even if these approaches are wide of the mark, a good explanation of akrasia ought to
be able to account for why these options are prima facie attractive. With respect to the first
option, the alcoholic who relapses does appear to be acting in the light of at least some
  85

ALCOHOLISM ACCORDING TO ALCOHOLICS ANONYMOUS 85

positive evaluation of drinking, however inchoate that evaluation may be. With respect
to the second approach, even if it is not true that alcoholic drinking is simply a mind-
less reflex, rather than a full-​blooded intentional action, it seems right that the alcoholic
agent is “not fully present” qua agent of intentional action in those actions that constitute
relapse. A satisfactory approach to akrasia in addiction ought to be able to explain this.

3  Alcoholism according to Alcoholics Anonymous

The links between alcoholism and relapse suggest that the natural way into this question
is simply to look at some views about what alcoholism is. My focus here, as I have said, is
to be on what the AA literature says about alcoholism. The book Alcoholics Anonymous
describes alcoholism as a “three-​fold” illness, and these folds are characterized as “physi-
cal,” “mental,” and “spiritual.” This distinction between these three different elements of
alcoholism suggests the view that alcoholic relapse, and alcoholic drinking, is a complex
phenomenon, different aspects of which require different kinds of explanation. What fol-
lows is a summary of the different kinds of claims made in the prefaces, introduction and
pages 1–​164 of this text.

3.1  Bodily “abnormality”

Alcoholism involves a “physical allergy” to alcohol (Alcoholics Anonymous 2001,
p. xxviii).5 This is not an allergy in the sense in which one may be allergic to bee stings,
but in the sense of its being an “abnormal physical reaction” to alcohol as a stimulus.
When the alcoholic begins to drink, “a phenomenon of craving” is triggered that causes
him to continue drinking (Alcoholics Anonymous 2001, p. xxviii). This abnormality
of the body and of the body’s responses to alcohol is almost never found in the “aver-
age, or temperate” drinker (Alcoholics Anonymous 2001, p. xxviii). Once this physical
allergy is in place, alcoholics are not able to return to being non-​alcoholic drinkers
(Alcoholics Anonymous 2001, p.  xxx, pp.  30–​34). Because alcoholics in whom the
physical allergy has developed cannot return to being normal drinkers, the only last-
ing solution for those suffering from alcoholism is abstinence (Alcoholics Anonymous
2001, p. xxviii). The occurrence of a craving in this sense does not in itself explain the
onset of relapse. The phenomenon of craving is specifically understood as a craving for
more drink, once the alcoholic has started to drink. But the fact that the phenomenon
of craving develops on taking the first drink explains why an alcoholic, once he begins
drinking, finds that he has little control over the amount that he drinks or the rate
at which he drinks, and drinks more than he intended, desired, or wanted to drink.
According to AA, part of the explanation of akrasia, at least as it is a feature of the
development and prolongation of an episode of drinking, must make reference to this
component of the condition.

5 This section summarizes pp. xxv–​xxxii. The physical problem in alcoholics is described as an “abnor-
mality” on p. xxvi.
86  

86 Failures of rationality and self-knowledge in addiction

3.2  Mental “abnormality”

The typical alcoholic is aware of the painful consequences of his abnormal bouts of drink-
ing. In the middle of his hangover, and the shame he feels about last night’s behavior,
he decides to “swear off ” and “never touch another drop.”6 He may put in place various
strategies to prevent himself from drinking. But the time will come at which his patterns
of thinking around alcohol change. He will begin to engage in attempts to rationalize why
he lost control when he last drank, rationalization that sustains plans for further attempts
at controlled, moderate, and pleasurable drinking. At the times when he most needs it, his
awareness of the painful and destructive consequences of his drinking appears to desert
him: “The almost certain consequences of his drinking and the pain and suffering of a few
days or weeks will not crowd into the mind to deter him. If they do, they are immediately
replaced with threadbare excuses and spurious rationalizations. The alcoholic is without
mental defence against the first drink” (Alcoholics Anonymous 2001, p. 24). Not only do
alcoholics appear to exhibit “mental abnormality” in the sense of being “without mental
defence” when they are actively attempting to pursue strategies of relapse prevention,
but alcoholics, when not drunk, and while knowing the harmful and dangerous conse-
quences of their drinking, will also typically spend much of their time thinking about
drinking and planning their next drink. This thinking and planning, and the prospect of
drinking at a later time, is itself a kind of relief (Alcoholics Anonymous 2001, pp. 8–​9).
This is the alcoholic’s “mental obsession” with drinking alcohol. This mental obsession,
and the “strange mental blank spots” (Alcoholics Anonymous 2001, p. 42) that precede
episodes of drinking constitute “a subtle form of insanity” (Alcoholics Anonymous 2001,
p. 40). A central feature of these “strange mental blank spots” and the obsession with alco-
hol that grips him is that the alcoholic is all but oblivious to the facts about his situation,
and to their role in his drinking (Alcoholics Anonymous 2001, pp. 40–​42).

3.3  A “spiritual malady”

At the core of the alcoholic condition is a problem with how the alcoholic lives, and the
consequences of action that expresses his malformed attitudes to himself and to others.7
“Selfishness—​self-​centredness! This, we think, is the root of our troubles” (Alcoholics
Anonymous 2001, p. 62). Selfishness and self-​centredness appear to be character traits of
some kind. They describe a distinctive pattern of motivation. The alcoholic is only con-
cerned, at bottom, with acting in a way motivated by getting what he wants, and maximiz-
ing his own desire satisfaction. Satisfying other people, and acting in a way that satisfies
their desires, is an aim of the alcoholic’s behavior only insofar as it is a means to satisfy his
own desires. These patterns of motivation set the self-​centred alcoholic at odds with others
(Alcoholics Anonymous 2001, p. 62). As well as describing what appears to be a general pat-
tern of motivation, self-​centredness also characterizes a range of other traits, propensities,

6 This section summarizes material from Alcoholics Anonymous (2001, ch. 3).

7 This description of alcoholism is from Alcoholics Anonymous (2001, p. 64).
  87

ALCOHOLISM ACCORDING TO ALCOHOLICS ANONYMOUS 87

and emotional or affective conditions. The alcoholic is characteristically ruled by “fear”

(Alcoholics Anonymous 2001, pp.  67–​68), or “self-​centred fear—​primarily fear that we
would lose something we already possessed or would fail to get something we demanded”
(Alcoholics Anonymous 1952, p. 76) and, specifically, that he does not have, or will not,
for some reason, elicit the approval of others. The self-​centred alcoholic is “burned up with
resentment” (Alcoholics Anonymous, 2001, pp. 64–​65) at others, resentments the sources
of which are nearly always the result of the alcoholic’s own self-​centred action or inaction,
and often stem from the alcoholic’s perception of “hurt pride,” “fear,” and “self-​pity.”
The alcoholic’s self-​centredness, understood in all these different manifestations, causes
problems to pile up in the alcoholic’s life. The alcoholic becomes “restless, irritable and
discontented” (Alcoholics Anonymous 2001, pp. xxviii–​xxix). Alcohol begins to be rep-
resented as a route through which these problems in living can be ameliorated, escaped,
or even solved. As so represented, it comes to be “desired” or “wanted” by the alcoholic.
The alcoholic who relapses appears to have lost sight of the destructive consequences of
drinking, and so has become unable to stick to whatever resolution or commitment he
may have made to stay sober.
The Alcoholics Anonymous program of recovery involves the alcoholic’s complete absti-
nence from alcohol. At the heart of this program is the idea that, for the alcoholic to achieve
permanent and satisfying recovery from active addiction, it is the “spiritual malady” that
needs to be addressed (Alcoholics Anonymous 2001, p. 64). It would be a cause for despair
in the suffering alcoholic, and false, were complete removal of these habits, tendencies, and
patterns of thought and action claimed to be a condition of achieving sobriety. And the
AA literature doesn’t claim that it is. Rather, through taking part in AA life, the alcoholic
is able gradually to alter and adjust his outlook on life. He does this by engaging in a range
of activities that encourage this: working towards an honest and detailed examination of
his own behavior (Step Four) and arriving at a sober assessment of the consequences of
that behavior for himself and others with an AA friend or “sponsor” (Step Five); making
a sincere attempt to right wrongs done to others, where possible (Steps Eight and Nine);
engaging in “service” for the group, whether that be speaking in front of the group, setting
the chairs out, or making tea; keeping in regular contact with, and engaging honestly with,
fellow recovering alcoholics; engaging in regular self-​examination and meditation (Steps
Ten and Eleven); and, most importantly, helping newly recovering alcoholics to stay sober
(Step Twelve). This process of re-​engagement with sober life is described as “finding a
power by which we could live” and finding “a Power greater than ourselves.” Step Three
asks the alcoholic to turn his will and his life “over to the care of God as we understood
Him.” Alcoholics Anonymous is not a religious organization and neither is it denomina-
tional. The spirituality involved here is minimal in its commitments. To solve their prob-
lems, recovering alcoholics need to discover a source of strength by which they can live,
which is something more than their own, failed, strategies for achieving sobriety.
Let’s refocus on the puzzle of akrasia with which we began. Suppose for the moment
that the species of akratic action that is the focus of discussion is that which accompanies
the onset of drinking. And suppose for the moment that the triggering of a phenomenon
88  

88 Failures of rationality and self-knowledge in addiction

of a craving cannot explain this event (being something that occurs after the first drink
has been taken). Then the conception of akrasia in alcoholism suggested by the kind of
account offered here is that such action is the manifestation of a kind of mental disorder.
This mental disorder manifests itself in the way that an alcoholic, prior to relapse, relates
both to what he knows about the past consequences of his drinking and also to how he
currently evaluates taking a drink. The alcoholic fails to bring these past consequences
to mind in the way that is needed to preserve his sobriety. And he is subject to desires or
wants for alcohol that motivate his drinking, motivations that make his drinking explica-
ble as an action and not a mere reflex or piece of involuntary behavior, and that mark his
drinking out as an attempt to gain relief from problems in living. These problems of liv-
ing have their source in patterns of “self-​centred” thought and action, which may or may
not involve drinking. These patterns of thought and action themselves we can also think
of as the manifestation of some more basic form of ethical maladjustment of the subject,
concerning how the agent conceives of himself, and how he relates to other agents.

4  Some questions about the Alcoholics

Anonymous account
I want to develop some suggestions about how to understand aspects of this approach by
raising a number of questions that might be prompted by the discussion so far.

4.1 Question 1
A central part of the explanation of relapse is that the alcoholic suffers a specific kind of inca-
pacitation with respect to his preserved knowledge about what he did and how he behaved
when he last drank. But what is disordered—​“insane” even—​about incapacitation? A rational
agent who is asleep, and in dreamless sleep, is incapacitated, in a very specific way, with respect
to the full range of his mental powers, including the capacity to draw on preserved knowledge.
But such an agent isn’t disordered in virtue of being so incapacitated. So what exactly is the link
between the idea of incapacitation and the kind of “disorder” that’s at issue here?

4.2 Question 2
In some parts of Alcoholics Anonymous, it is suggested that the alcoholic has no idea why
he relapses. Here is an example from Chapter 2:
If you ask him why he started on that last bender, the chances are he will offer you any one of a
hundred alibis. Sometimes these excuses have a certain plausibility, but none of them really makes
sense in the light of the havoc an alcoholic’s drinking bout creates … Once in a while, he may tell
the truth. And the truth, strange to say, is usually that he has no more idea why he took that drink
than you have. Some drinkers have excuses with which they are satisfied part of the time. But in
their hearts they really do not know why they do it.

(Alcoholics Anonymous 2001, p. 23)

One of the distinguishing features of intentional action appears to be that the agent who
is engaged in such action knows what the action is intended to bring about without
  89

Defects of the wakeful state in alcoholism 89

observation or inference, simply in virtue of being the author of that action.8 When I am
engaged in the intentional action of walking to the shop to buy milk I know that that is
what I am doing just by being the one doing the walking. But if the alcoholic who relapses
does not know why he drank then that might suggest that his drinking—​according to this
approach—​is not an instance of intentional action at all, but some other occurrence that
involves bodily movement: a piece of reflex or involuntary behavior, say. So how does the
idea that the alcoholic is ignorant of why he drinks, when he relapses, cohere with the idea
that his drinking is a case of intentional action?

4.3  Questions 3 and 4

In the account sketched out above, a claim was made that alcoholics use alcohol in an
attempt to deal with emotional difficulties and other problems with living. This moti-
vation for drinking is expressed in many of the recovery narratives that make up the
“Personal Stories” sections of Alcoholics Anonymous. For example:
Bewilderment, fear and resentment moved into my life. And yet my ability to lie outwardly and to
kid myself inwardly grew with each drink I took. Indeed, I had to drink now to live, to cope with
the demands of everyday existence. When I encountered disappointments or frustrations—​as I did
more and more frequently—​my solution was to drink.

(Alcoholics Anonymous 2001, p. 555)

This prompts two further questions:

How does the idea that the alcoholic uses alcohol as a way to solve problems with living
cohere with the idea that alcoholics do not really know why they drink?
Given the facts about what drinking does to an alcoholic, facts of which the alcoholic is
aware, the idea that the alcoholic thinks that the solution to disappointments, frustra-
tions, or other problems is for him to drink seems so extraordinary that it is simply
unclear why we ought to attribute such motivation to him. Is this really what motivates
the drinking that constitutes relapse, or just a spurious piece of post-​hoc rationalization;
an “excuse” that “doesn’t really make sense in the light of the havoc that an alcoholic’s
drinking bout creates?” If this is the motivation, how is it that alcohol is supposed to
help? What kind of solution is it that the alcoholic uses alcohol in an attempt to achieve?

5  Defects of the wakeful state in alcoholism

5.1  Addressing question 1
The link between the notion of disorder and incapacitation can be illuminated by a par-
ticular notion of a “mental state” and by the notion of what is proper to that state. The
inaccessibility of knowledge about past attempts to drink safely or moderately is proper
to the state of sleep in a way that it is not proper to the state that obtains during relapse.

8 For important discussions of this feature of intentional action see Anscombe (1957), Moran (2001),
and the papers collected in Roessler and Eilan (2003).
90  

90 Failures of rationality and self-knowledge in addiction

There is much that is not well understood about the details of the functional role of the
state of sleep in the life of an animal. But the outline of the role is clear enough; sleep is a
condition in which the animal is alive, but at rest. In this state the animal is able to recover
from fatigue, and so through sleeping it is able to continue to function effectively as an
animal during the time it is awake. What enables it to fulfill this function is that in sleep
the animal “switches off,” where to be switched off is for the animal to be incapacitated
with respect to the powers or capacities that it possesses.9 The facts about how animals
rest may differ. Some animals may shut off in different ways from others.10 However, sleep
in a mature human being is a state in which it is shut off with respect to what appears to
be the full range of the capacities it possesses as a rational animal; its capacities for physi-
cal action, sense perception, intentional mental action, and memory, amongst others. The
unavailability of capacities in the state of sleep is not an instance of mental disorder. So the
unavailability or inaccessibility of memory in sleep is not an instance of disorder.
But the unavailability of capacities when the agent is in a state of wakeful consciousness
is disordered. The nature of the state of wakeful consciousness is a question that has been
largely neglected by philosophers of mind and action.11 But at least to a first approxima-
tion, wakeful consciousness is that state in which the animal is “switched on”; the waking
animal is in the position to be at work being an animal. Its being capacitated in this way
normally entails that it is capacitated with respect to each of the capacities it possesses as
an animal.12 What it is for an animal to be awake depends on what capacities it possesses.
In the case of rational animals—​those creatures with the capacity for rationality and self-​
conscious thought—​being awake is being in the state in which it is in the position to be
at work being a rational animal. Then for a rational animal to be awake, but to be inca-
pacitated with respect to some of the capacities that it possesses as a rational animal, will
prevent it from being in the position to be properly at work as the kind of thing that it is.
States such as this will be varieties of the wakeful condition but which constitute forms of
failure given the nature of that condition. Take, for instance, the state of shock or trauma
that can come to obtain after witnessing extremely painful, disturbing, or dangerous epi-
sodes. In a state of shock, an animal is awake, but temporarily incapable of accessing
its powers of intentional mental action and for self-​initiated bodily movement; forms of
incapacitation that issue in confusion and disorientation. Even if the state of shock may
have a biological function, it is nevertheless a disordered state. Shock is a state that inher-
its the aims of the wakeful condition, given that it is a type of wakeful condition. But it
fails to fulfill the explanatory goals of the wakeful state; it is a state in which in which the

9 An account along these lines is presented in sections 1 and 2 of Aristotle’s De Somno. For introduc-
tory material on the contemporary research on sleep see Hobson (2002) and the essays in Smith et al.
(2008).
10 For interesting discussion of “unihemispheric slow-​wave sleep” in animals see Rattenborg et al. (1999).

11 Anexception is the discussion of wakeful consciousness in O’Shaughnessy (2000), part  1. See also
Crowther (forthcoming).
12 For interesting discussion and argument for a view of this kind see Aristotle De Somno sections 1 and 2.
  91

Defects of the wakeful state in alcoholism 91

animal is not properly in the position to be at work being an animal, because it involves
incapacity of the relevant kinds.
While we have a notion of “being obsessed with something” according to which it is a long-​
term character trait, we can also think of the state that precedes relapse in the alcoholic as a
state of obsession with alcohol understood along the lines just proposed. The alcoholic who is
obsessed with alcohol in this sense is in a kind of wakeful state that is deformed or degraded in
various ways. Central to this is the incapacity to bring to mind in the right way the memory of
the suffering and humiliation of previous relapses. A kind of knowledge that is available to the
drinker in the normal non-​degraded state of wakefulness is here not capable of being brought
into play in determining what to do. Characteristic also of the state is the alcoholic’s preoc-
cupation with thoughts about drinking, with the construction of plans for drinking, as well as
immediately “reading” his environment in terms of a series of opportunities for, or obstacles
to, drinking. This conception of obsession suggests answers to the other questions raised.

5.2  Addressing question 2

Suppose that the intentions with which an agent acts are his reasons for acting. Then we
may respond to the question posed in section 4.2 with the following thought: the alco-
holic in the state of obsession has reasons for action, but they are extremely minimal.
These reasons may be no more than that it seemed to the alcoholic that taking a drink is
what ought to be done, or that taking a drink is what the situation calls for. So the alco-
holic’s drinking can be understood as intentional. The alcoholic drinks with the intention
of doing what, in this particular situation in which a drink is offered, is to be done, or
what the situation calls for. The event that occurs when the alcoholic drinks can then be
distinguished from a mindless or reflexive bodily movement.
In the passage that prompted this question, the claim was that the alcoholic “does not
really know why” he started out on that last relapse. We can respond that there is some
minimal knowledge that the alcoholic has of why he drank, knowledge that he might
express by citing the trivial and minimal reasons for drinking just discussed. This is con-
sistent with two further truths. The first is that there is some deeper and more informative
explanation of why he drank of which the alcoholic is ignorant. The second is that if the
alcoholic is asked why he took that drink, given his drinking history, then giving such
minimal reasons for action will indeed seem like a particularly inadequate, evasive and
uninformative answer.13 But that doesn’t entail that they weren’t his reasons.
A natural question at this point is why things strike the alcoholic in this way; why it
seems to him that drinking was what was to be done. The peculiarity of these reasons, in
the light of the actual consequences of his drinking, is to be explained by features of his
mental state. In a state of obsession with alcohol, the world presents itself to the alcoholic
in a distinctive way. In a state of obsession with alcohol, and when the opportunity to
drink arises, taking a drink presents itself to the alcoholic as what is there and then, in

13 That this response is so inadequate may be why such an alcoholic may be happy to concede, under
interrogation, that he has no idea at all about why he drank.
92  

92 Failures of rationality and self-knowledge in addiction

those circumstances, to be done, or to be what the situation calls for. The alcoholic acts on
reasons for drinking that he has, given his state of obsession with alcohol.

5.3  Addressing question 3

The idea that alcoholics drink to deal with frustrations, disappointments and problems
with living coheres easily enough with the idea that alcoholics do not really know why
they drink. Perhaps there are occasions on which alcoholics deliberately and knowingly
drink in order to deal with problems. But the more natural way to understand such
testimony, as was quoted in section 4.3, is as an articulation of a pattern of motivation
for alcoholic drinking of which the agent was unaware when drinking. That alcoholics
can be ignorant of the sources of motivation for their action is no more problematic,
on the face of it, than any similar claim about the existence of unconscious motivations
for action.
This issue about awareness of reasons for action reveals another way in which the
notion of a mental state may be mobilized in explanation. When they are in the state that
precedes and accompanies relapse, alcoholics lack self-​awareness or self-​knowledge of
various kinds. And it is plausible that at least some such absences of self-​awareness can
be traced to various respects in which the state of obsession differs from the normal and
properly formed state of wakeful consciousness. One who is in the state of non-​degraded
wakefulness normally knows, when he is in that state, why he is doing what it is that he
is then doing. And he normally knows what mental state he is in at that time. But at the
time they drink, alcoholics do not have an awareness of the real reasons for their drinking.
They are, perhaps, aware of the minimal reasons for action. But they are not aware that
their drinking is, for example, aimed at “coping with the demands of everyday existence.”
Further, what makes the state of obsession so dangerous for alcoholics is that they are
oblivious to their being in this state. In the state of obsession, alcoholics, I will say, lack
“occurrent self-​knowledge” of various kinds.
Another characteristic failure of self-​awareness is worth distinguishing from this. An
alcoholic may have acquired (let us just assume) knowledge about his condition. He
knows that he has a physical allergy to alcohol, that at times he has “no defence against
the first drink,” and that his absence of such defence is associated with a condition of
obsession in which certain bizarre forms of motivation prevail. Nevertheless, he may
drink, and he may drink as a result of becoming obsessed with alcohol. Such an alcoholic
didn’t lose this knowledge about himself. He doesn’t have to reacquire it when he wakes
up after a relapse. But in the state of obsession this knowledge wasn’t accessible to him
in the way that it is in a normal state of wakeful consciousness, where memory makes it
available to the rational agent in the normal way. So, where an obsessed alcoholic does
possess knowledge of his alcoholic condition of the “dispositional” kind, incapacitation
with respect to memory renders this kind of self-​knowledge inaccessible to him.

5.4  Addressing question 4

Perhaps there are many different ways in which the alcoholic may attempt to solve his
emotional difficulties by drinking. I want to explore just one, which relates to the way
  93

Another defect of the normal wakeful condition 93

that the alcoholic, when in a state of obsession, represents his drinking. Consider the
descriptions that recovered alcoholics offer of their early experiences of drinking. “A few
years later, in junior high school, a few friends and I got a bottle of rum from a bootleg-
ger. I got really drunk and it was great. I had a feeling of complete freedom” (Alcoholics
Anonymous 2001, p. 495). “Drinking released me from the suffocating fear, the feelings
of inadequacy, and the nagging voices at the back of my head that told me I would never
measure up. All of those things melted away when I  drank” (Alcoholics Anonymous
2001, p. 310). “(I) do remember this much: When I was drinking, I was okay. I under-
stood. Everything made sense. I  could dance, talk, and enjoy being in my own skin.
It was as if I had been an unfinished jigsaw puzzle with one piece missing; as soon as
I took a drink, the last piece instantly and effortlessly snapped into place” (Alcoholics
Anonymous 2001, p. 320).
The content of these descriptions of drinking experiences plays a part in explaining the
alcoholic’s motivation for drinking. When in the state of obsession with alcohol, alcohol-
ics don’t represent their drinking as a route to loss of control and to physical and emo-
tional damage. They represent drinking as a mode of self-​realization, as a route through
which they can be the way they are supposed to be. Again, it can be helpful to think of
this form of motivation in terms of the notion of a state of consciousness. The alcoholic
attempts to use alcohol as a solution to his problems because he represents alcohol, in the
obsessive condition, as a way of changing the prevailing psychological condition to one in
which he is in a non-​defective and normal wakeful state of consciousness—​that psycho-
logically healthy state in which he is in a position to exercise those rational and affective
capacities that make him the kind of being that he is. That includes being in the position
to “think straight” about things, to interact sanely and normally with other people, and
being in a position in which he is not assailed by self-​doubt, fearful thoughts, and other
troubling occurrent emotions. The past-​tense content of these descriptions offered by
recovering alcoholics is also relevant to the point at issue here. Alcoholics in an occurrent
state of obsession with alcohol represent their drinking as being capable of putting them
into a normal psychologically healthy wakeful state because they recollect, or appear to
recollect, that alcohol had these liberating effects on them on certain occasions of drink-
ing in the past. “The idea that somehow, someday, he will control and enjoy his drinking is
the great obsession of every abnormal drinker” (Alcoholics Anonymous 2001, p. 30) and
is, at least in part, the idea that some day the alcoholic will control and enjoy his drinking
in the way that he takes it that he used to.

6  Another defect of the normal wakeful condition

In the light of the idea of a distinction between defective and non-​defective states of
wakefulness, two kinds of error might be distinguished in this reconstructed motivation.
One kind is obvious. For the chronic alcoholic, drinking does not have the desired effect.
Drinking in relapse, for the chronic alcoholic subject to an obsession with alcohol, does
not bring about a return to a normal non-​degraded condition of consciousness. It brings
only more of the same. And this is surely part of what is distinctive of chronic alcoholic
94  

94 Failures of rationality and self-knowledge in addiction

drinking. The chronic alcoholic, when drinking, remains obsessed with drinking. His
excessive preoccupation with drinking now manifests itself in extending indefinitely
the episode of drinking that is under way. He formulates plans about how to extend the
episode of drinking, and then tries to execute the steps of this plan, all the while revis-
ing these plans to achieve the same end in the face of new information he continues
to receive. He continues to read his immediate perceptual environment in terms of his
obsession, now in terms of the opportunities for prolonging the episode of drinking that
is under way and in terms of obstacles to prolonging it. His way of drinking may involve
special vigilance to time, to how quickly or slowly those with whom he drinks, if any, are
drinking, and also to how considerations about time affect the material possibilities for
him to get more alcohol. His overriding fear is that there will not be enough alcohol to
continue the episode of drinking or that something will interfere to bring it to an end.
The possibility that a further, more fundamental kind of error is also involved in the
chronic alcoholic’s motivation for drinking points to an explanatory role for another kind
of defective variety of wakefulness. Suppose that the drinking of an alcoholic never, in
fact, fitted the representation of drinking that figured in the relevant form of apparent
recall, in which drinking is represented as a route to a normal wakeful state, a state in
which the alcoholic “has complete freedom” and in which “the missing piece of the jigsaw
is in place.” If that were so then the motivation for drinking, as it figures in explaining
drinking in the state of obsession, involves a basic cognitive error. But why might this be
true? What is the reason for thinking that an alcoholic’s drinking, even in the early phases
of his drinking, was never drinking that initiated and sustained a healthy wakeful state of
consciousness? Is it really plausible to claim that from the very first drink, the alcoholic
drank in a way that manifested obsession with alcohol in the way that it has been under-
stood here? Doesn’t this also entail that we must refuse to take at face value those aspects
of personal testimony that describe the spontaneity and freedom from care and worry
that accompanied the alcoholic’s very early drinking career?
The literature that has been the focus of our attention here describes the fundamental
problem of an alcoholic as “self-​centredness.” “Being self-​centred” (or “self-​obsessed”)
might pick out a personality characteristic, a set of personality characteristics, or habits
of thought and action. These characteristics are things that rational agents can possess
in dreamless sleep. But self-​centredness might also be understood as an instance of the
kind of state that has been at issue here—​a defective instance of the wakeful condition
that involves forms of incapacity that we might describe as broadly ethical, and patterns
of thought and motivation that we might describe as ethically disordered or maladjusted.
Very roughly, we might think of this state as involving incapacity with respect to those
capacities that embody our ongoing sensitivity to, and practical acknowledgment of, the
ethical demands made upon us, as rational agents, by other agents like ourselves. We
might also think of it as involving incapacity with respect to the capacity to be sensi-
tive to the demands that we ourselves also make on others as the rational beings we are,
and toward what is rightly owed to us. These incapacities are revealed in the range of
thoughts and feelings that strike the one who is so incapacitated in reflection, in patterns
  95

Conclusion 95

of practical deliberation in which the demands of others do not figure fundamentally,

and in which action is only motivated by the agent’s own desires. But they might also be
manifested, crucially, in a form of perceptual incapacity, an incapacity to experience our
immediate perceptual environment as making characteristic sets of demands on us, as
the rational beings that we are. Being self-​centred in this sense, I emphasize, would not
be to be psychopathic or sociopathic. Being disordered in such senses appears to involve
an absence of the capacities themselves; in individuals of such a kind these capacities
never developed properly at all or have been destroyed through physical damage. Self-​
centredness, as I understand it here, is a condition in which capacities that the agent pos-
sesses are inaccessible in the circumstances, given the mental state.
In order to establish the claim that the alcoholic’s motivation for drinking involves a
cognitive error about the effects of alcohol on him, it need not be argued that from his
initial experiments with drinking he was in a state of obsession with alcohol. For it may
nevertheless be true of him, and it may be what marks out the early drinking experiences
of alcoholics, that they take place in the state of “self-​centredness” or “self-​obsession.”
Developing an understanding of exactly what such a state is, and properly substantiating
the thesis that a state such as this is distinctive of alcoholic drinking, even during the early
stages of alcoholic drinking, requires more than I can do here. But the basic idea need be
nothing more than the following. When alcoholics drink, one of the things that makes
their drinking characteristically alcoholic is that it takes place against the backdrop of
such a condition: an insensitivity to the demands that others make on them as the things
that they are and an inability to entertain, as a basic reason for doing anything, what other
people may want of them. Take any of the episodes of drinking from early in the careers
of alcoholics, which are described in the pages of Alcoholics Anonymous, where those epi-
sodes are ones which would be later characterized in terms of “things making sense” or
“having a feeling of freedom” or a “freedom from fear.” Now imagine that halfway through
any such episode, as it actually occurred at the time of the episode of drinking, someone
the alcoholic narrator held very dear had requested them to stop drinking right away,
because there is something that they (the intervener) wanted to do which was incompat-
ible with their (the drinker) continuing to drink. I predict the following: that while the
alcoholic may eventually have acceded to such a request, it would have been met with
more or less extreme irritation and reluctance, perhaps with hostility. The content of the
request would not immediately strike the alcoholic as something that ought to be done
and which in the situation was required. The request would be experienced as something
to be avoided, resisted, or shrugged off by any available means.

7  Conclusion
Understanding the kind of disorder that is crucial to sustaining alcoholic drinking as a
deformed state of wakeful consciousness has consequences for how to conceive aspects of
recovery from alcoholism. I earlier briefly described recovery from alcoholism, according
to the programme of Alcoholics Anonymous, as involving action geared around adjusting
the alcoholic’s attitudes and outlook on life, action informed by the alcoholic’s establishing
96  

96 Failures of rationality and self-knowledge in addiction

contact with “a God” of his or her own “conception,” or a “Higher Power.” Were one to
read these claims in terms of the suggestions made here then one might be led to propose
that the fruit of recovery is the reclamation, on a consistent basis, of a way of living that is
characterized by the obtaining, during the waking hours, and on a reasonably consistent
basis, a non-​disordered state of wakeful consciousness.
Given the account suggested here, that is something that involves, at least, capacita-
tion with respect to a range of capacities for rational thought and rational action, where
such action is also informed by, and embodies, a stored set of attitudes that encode the
agent’s conception of what is worth caring about, at the core of which is the idea that other
rational agents and their desires make demands on us that we ought to acknowledge. As
I have been trying to emphasize, the alcoholic’s problem is not a problem with respect to
what the agent knows or believes. Alcoholics coming around after binges, filled with guilt,
shame, and remorse, do not have to reacquire the knowledge that they are not capable of
drinking safely, or that it is best for them not to drink at all. That is something that they
knew all along. Neither need it be understood as a straightforward defect with respect to
what the agent really values or cares about. The same alcoholic, returning to conscious-
ness after relapse, does not need to relearn that it is worth caring for one’s family and
friends, or worth valuing what other people want as well as what they themselves want.
Rather, what is needed is for a state to obtain in which this knowledge is able to make itself
manifest in the way that it is supposed to in rational agents who have it, and for a state to
obtain in which these standing evaluations make themselves manifest in action in a way
that they are supposed to.
One might think of one way that Alcoholics Anonymous functions as the provision of
a scaffold for supporting the state of non-​deformed wakeful consciousness, or indeed for
redeveloping it. Alcoholics Anonymous does this by providing environments in which
recovering alcoholics are able to do things with one another and for one another. Some
of this might involve taking stock of the consequences of past behavior and arriving at
a full and balanced picture of the facts about their alcoholism with a sponsor. Or it may
involve making tea. Either way, through doing these things, alcoholics are able to spend
time living in a non-​deformed state of consciousness. Spending time living in this state,
and gradually feeling at home in it, strengthens the alcoholic’s dispositions to inhabit the
normal wakeful state. But whatever exactly these dispositions might be, to possess these
dispositions is not what recovery from alcoholism itself consists in. The alcoholic may
possess these dispositions while he sleeps, or indeed, in relapse. Similarly, however, so
also may an alcoholic who is “recovered” by the lights of the kind of account sketched
here possess a range of long-​term dispositions and traits that distinguish him from non-​
alcoholics. The recovered alcoholic is not able to drink safely. The incapacity of his body
to respond to the drinking of alcohol remains. Alcoholics who relapse after long periods
of sobriety develop cravings soon after taking a drink, or exhibit the response to alcohol
that characterized their previous drinking in the space of one or two further episodes of
drinking. So also the recovered alcoholic is disposed to re-​enter the state of occurrent
self-​obsession or self-​centredness; the kind of disordered state of wakeful consciousness
  97

Conclusion 97

that it has been suggested here is the soil in which the potential for alcoholic drinking
grows. This is why the recovered alcoholic may find that he continues to need to engage
with the supportive environment and activities that he finds in Alcoholics Anonymous
on an ongoing basis. This is not because the recovered alcoholic becomes “addicted” to
Alcoholics Anonymous, or because he has become “obsessed with” attendance at meet-
ings. It is because the dispositions that the recovered alcoholic possesses, as an alcoholic,
for self-​obsession and self-​centredness, have as their natural manifestation the erosion of
the structure of the non-​disordered wakeful condition.
The suggestions I have made here about akrasia in the chronic alcoholic are qualified.
First, I do not present these ideas as suggestions about every instance of akrasia. Second,
I have only here been talking about psychological or cognitive aspects of chronic alcohol-
ism, and the coherence of a range of psychological claims associated with the perspective
of a particular program of treatment and recovery. A full understanding of this condi-
tion would need also to address the nature of the “phenomenon of craving,” something
about which I have been neutral here.14 The suggestions are also programmatic in obvious
respects. There remain significant questions about the precise nature of the various kinds
of conditions of obsession mentioned here and how they differ from wakeful conscious-
ness. And there is the further question of whether a properly developed account of this
kind can provide a satisfactory explanation of the relevant psychological features of akra-
sia in alcoholism. That is something that I do not have the space to undertake here. But
I hope to have made at least some kind of case for the view that the notion of a mental
state that involves a specific kind of departure from the state of wakeful consciousness
may be a fruitful way to understand important aspects of alcoholism.

Acknowledgments
Thanks to Nick Heather and Gabriel Segal for helpful comments on an earlier draft.
Thanks also to Hemdat Lerman, Naomi Eilan, Bekah Sparrow, Guy Longworth, and
Matthew Soteriou for helpful discussion.

References
Alcoholics Anonymous (1952). The Twelve Steps and Twelve Traditions. New York: Alcoholics
Anonymous World Services.
Alcoholics Anonymous (1984). “Pass It On.” New York: Alcoholics Anonymous World Services.
Alcoholics Anonymous ([1939] 2001). Alcoholics Anonymous (4th ed.). New York: Alcoholics
Anonymous World Services.
Anscombe, G.E.M. (1957). Intention. Oxford: Blackwell.
Aristotle (2002). Nicomachean Ethics. Translated by Christopher Rowe with commentary by Sarah
Broadie. Oxford: Oxford University Press.
Berridge, K. (2001). Reward learning: reinforcement, incentives, and expectations. In: D. Medin (ed.),
The Psychology of Learning and Motivation: Advances in Research and Theory, 40, pp. 223–​78.

14 For relevant work on this aspect of alcoholism see Berridge (2001), and Robinson and Berridge (1993).
98  

98 Failures of rationality and self-knowledge in addiction

Broadie, S. (1991). Ethics with Aristotle. Oxford: Oxford University Press.

Charles, D. (1984). Aristotle’s Philosophy of Action. London: Duckworth.
Coope, U. (2010). Aristotle. In: T. O’Connor and C. Sandis (eds), A Companion to the Philosophy of
Action. Oxford: Wiley-​Blackwell, pp. 439–​46.
Crowther, T. (forthcoming). Experience, dreaming and the phenomenology of wakeful consciousness.
In: F. Dorsch, F. Macpherson and M. Nide-​Rumelin (eds), Phenomenal Presence. Oxford: Oxford
University Press.
Davidson, D. (1970). How is weakness of the will possible? In: J. Feinberg (ed.), Moral Concepts.
Oxford: Oxford University Press.
Heather, N. and Segal, G. (2013). Understanding addiction: Donald Davidson and the problem of
akrasia. Addiction Research and Theory, 21, 445–​52.
Hobson, J.A. (2002). Dreaming: An Introduction to the Science of Sleep. Oxford: Oxford University Press.
Holton, R. (2009). Willing, Wanting, Waiting. Oxford: Oxford University Press.
Lorenz, H. (2014). Aristotle’s analysis of akratic action. In: R. Polansky (ed.), The Cambridge Companion
to Aristotle’s Nicomachean Ethics. Cambridge: Cambridge University Press, pp. 242–​62.
Moran. R. (2001). Authority and Estrangement. Princeton: Princeton University Press.
O’Shaughnessy, B. (2000). Consciousness and the World. Oxford: Oxford University Press.
Rattenborg, N.C., Lima, S.L., and Amlaner, C.J. (1999). Half-​awake to the risk of predation. Nature,
397, 397–​98.
Robinson, T.E. and Berridge, K.C. (1993). The neural basis of drug craving: an incentive-​sensitization
theory of addiction. Brain Research Reviews, 18, 247–​91.
Roessler, J. and Eilan, N. (eds) (2003). Agency and Self-​Awareness. Oxford: Oxford University Press.
Smith, H.R, Comella, C.L., and Hogl, B. (2008). Sleep Medicine. Cambridge: Cambridge
University Press.
  99

Chapter 6

Normal and addictive desires

Patrick Butlin
David Papineau

Abstract
This chapter aims to understand drug addiction in the context of an overall
theory of human behavior. It describes three systems that contribute to
determining our choices and actions and are of particular relevance for
understanding addiction. These are the habit system, the desire system,
and the planning system. Locating theories of addiction in a framework
for understanding normal behavior is an important prerequisite to
answering moral and practical questions concerning addiction, and also
makes it easier to compare these theories. This chapter also argues that
the standing desire theory is the most elegant theory of drug addiction
currently available and explains how this theory relates to other prominent
proposals. Finally, it outlines a major challenge to the standing desire
theory, deriving from the ongoing debate concerning the function of
dopamine.

1 Introduction
Our aim in this chapter is to understand drug addiction within the context of an overall
theory of human behavior.
In our view, behavior is to be understood as influenced by a number of different
control systems. We shall focus in particular on what we will call the habit system, the
desire system, and the planning system. These three systems are of crucial importance
for determining almost all of our choices, addictive and otherwise. It seems likely that
the first two of these systems are shared with other animals, but that the planning
system is peculiar to humans. By describing these three systems, we aim to provide a
framework that will clarify the mechanisms of addiction and the loss of control they
involve.
In this context, one possible theory of the mechanism of drug addiction stands out as
offering a particularly simple and elegant explanation of how addicts feel and behave.
According to this theory, which we call the standing desire theory, addictive drugs cause
long-​term changes to the desire system, which in turn cause addicts to experience
100

100 Normal and addictive desires

abnormally strong occurrent desires to take drugs.1 These occurrent desires can be strong
enough to prevent addicts from acting on the intentions formed by their planning sys-
tems. We will compare this theory to some other prominent proposals, and assess its
current prospects.
From any perspective, a distinctive feature of addiction is that addicts are characteristi-
cally unsuccessful at sticking to long-​term intentions to abstain.
Such undermining of intentions is not, of course, peculiar to addiction. It occurs
whenever somebody forms a resolution and then later gives in to temptation. To this
extent, we view addictive behavior as similar to normal behavior. Just like normal
agents, addicts are perfectly capable of forming rational intentions, and in particular
intentions about indulging or refraining from their addictions. And again, just like
normal agents, they sometimes fail to stick to these intentions under the influence
of contrary later impulses.
What distinguishes addicts, according to the standing desire theory, is that their
occurrent desires to indulge their addiction are often abnormally strong. Addicts are
subject to unusual influences that accentuate their cravings and hamper their plans to
abstain.
The standing desire theory, like other modern accounts of drug addiction, attributes
these unusual influences to the role of dopamine. The ingestion of addictive substances
results in unusually high levels of dopamine in the brain, and this leads to a standing
disposition to experience strong desires for those substances in future, in ways discussed
in detail below.
As far as a definition of “addiction” goes, we shall restrict this term in what fol-
lows to the state of being disposed to experience abnormally strong cravings, due
to abnormal effects involving dopamine. A less restrictive definition would allow the
term also to encompass abnormal cravings that are due to unusual effects not involv-
ing dopamine. It is a moot point whether there are any other such non-​dopaminergic
exceptional causes for abnormal cravings. In any case, we shall not consider any such
phenomena in this chapter. Given this, it will be more convenient to stick to the more
restrictive definition and understand “addiction” as implying “dopamine-​caused” in
what follows.
If some condition involves abnormal desires, whether caused by dopamine or some-
thing else, does that make it a disease? We shall not spend any time on this question.
Indeed it is not obvious that it is a good question. There is no agreed understanding of the
term “disease,” and in our view it is doubtful that it tracks any natural kind, as opposed to
some conventional category tied to the historical idiosyncracies of the medical profession.
Nor is there any agreement on the moral or practical consequences of something being a

1 Standing desires are (roughly) persisting mental states in virtue of which we prefer or value cer-
tain outcomes, and occurrent desires are manifestations of these states at particular times that play a
direct role in motivating us to pursue the outcomes concerned. We discuss this distinction in detail in
section 2.4.
  101

Systems for behavioral control 101

disease. It is unclear, to say the least, whether or not having a disease means sufferers are
not responsible for their actions, are not to blame for their condition, should be treated
rather than punished, and so on.
Of course, issues of responsibility, blame, and punishment are real and pressing, and
particularly so with respect to addicts. But they are best addressed directly, without a
detour into the issue of disease. We can simply ask straight off about the responsibility,
blame, and punishment of addicts, without also worrying about whether addicts are ill
or not.
While these practical and moral questions are not themselves within the scope of this
chapter, we take our arguments to be directly relevant to them. We may not need to know
whether addiction is a disease to know how to deal with it. But we will be hard pressed
to work out the right way to respond to addicts if we don’t have a proper understanding
of the mechanisms behind their plight, and specifically of how they relate to the normal
mechanisms of action control. To this we now turn.

2  Systems for behavioral control

2.1  Three systems
The three systems for behavioral2 control which we will describe are the habit system, the
desire system, and the planning system. The habit system is likely to exist in many animals;
the desire system has been extensively studied in rats and mice, and can therefore be
expected to exist in most mammals; and the planning system may well be present only in
humans, at least in its full form. Although these three systems are of particular interest,
they are not responsible for all human behavior; one type of exception is behavior caused
by reflexes, and there may be others.
We will argue that humans use all three of the systems we shall describe. How exactly
these systems interact is a complex question which will be of relevance below. At first pass,
we can suppose that, in cases where the systems conflict, the desire system will dominate
the habit system, and that the planning system will dominate both the others. But as we
shall see, this domination is not absolute in either case.

2.2  The habit system

In the habit system, behavior is controlled by learnt associations between stimuli and
responses. So the resulting patterns of behavior, or habits, are sometimes called S-​R associ-
ations, and are the products of S-​R learning. Stimuli are features of the person’s or animal’s
circumstances, and responses are behaviors that they perform in those circumstances. S-​R

2 Philosophers are much concerned with definitions of such terms as “action” and “intentional action.”
We shall bypass these issues, and simply use the neutral term “behavior” as a general term for all forms
of purposive motor activity (that is, motor activity that can be explained teleologically in terms of its
furthering some end). So our talk of “behavior” should not be thought to exclude items normally called
“actions.”
102  

102 Normal and addictive desires

learning takes place when stimuli are followed by responses, which are in turn followed
by either rewards or punishments that between them amount to a level of reward different
from what is expected in that situation. Habits get stronger when responses are followed
by more reward than expected, and weaker when they are followed by less reward than
expected. What makes a situation rewarding varies between species and between indi-
viduals, but typical rewards include getting food or sex.
Experiments on rodents suggest that internal states, including states of physiological
need like hunger or thirst, influence the habit system by acting as stimuli. Niv and col-
leagues (2006) found that rats that had been trained when hungry to press a lever for
sugar solution reduced responding when they were thirsty but not hungry. Since sugar
solution is good for both hunger and thirst, these results suggest that habits are triggered
specifically by those physiological needs that were active when they were acquired.
One crucial question about the habit system is how it interacts with the other systems
of behavior control. On one picture of the habit system, it only affects human behavior
when we let it—​when we are performing familiar routines, and don’t concentrate on what
we are doing. If this view is correct, then it seems that addiction could not be explained by
drug-​taking habits, because these habits would be easy to resist. In addition, addicts per-
form complex sequences of apparently pre-​planned behaviors in order to source drugs,
which again would seem hard to explain if addictive behavior were solely habitual.

2.3  The desire system

The desire system is defined by responsiveness to information about the values of out-
comes. The existence of the desire system in rodents was established in the 1980s, primar-
ily by outcome devaluation experiments, which produce results that are hard to explain if
we assume that rats are only capable of habitual control (Adams and Dickinson 1981). In
these experiments, rats are first given the opportunity to press a lever, and receive a food
reward when they do so—​peanuts, for instance. They remain in this environment for long
enough to learn to press the lever. Then the rats are taken away from the cage with the
lever, and some of them undergo “outcome devaluation”: they are fed with peanuts, and
then given a lithium chloride injection that makes them sick. The other rats in the control
group receive both the food and the injection, but at different times. When the rats are
returned to the cage with the lever, those that went through the outcome devaluation
press the lever less than those in the control group, even though no rewards are delivered
to either group. It seems that the habit system cannot explain this result, because the only
difference in the experiences of the rats in the two groups concerns the outcome—​it does
not involve them being presented with the stimuli, or performing the responses, that are
relevant to the behavior that changes. Apparently, the rats were storing information both
about the value of the outcome—​getting peanuts—​and about the relationship between
the action and the outcome—​between pressing the lever and getting peanuts.
Another finding of outcome devaluation experiments is that they don’t work if the rats
are given too much initial training on the lever; in this case, they do not reduce respond-
ing as a result of outcome devaluation, instead behaving as though their behavior were
  103

Systems for behavioral control 103

controlled by a habit (Adams 1981). So rats are thought to use both the habit system and
the desire system in parallel, with habits being formed and modified relatively slowly.
Further evidence for this claim comes from studies by Yin and colleagues (2004, 2005,
2006), who selectively destroyed parts of the striatum in rats, and thus identified distinct
brain areas that are apparently responsible for habitual and desire-​system control. The
idea that rats use both systems, and that outcome devaluation is a good way to distinguish
them, is now central to an extensive research programme on behavior control and reward
learning (Balleine and O’Doherty 2010).
Unlike the habit system, the desire system uses states that track the values of outcomes
to control behavior. It combines these with its learned information about the relationships
between behaviors and rewards to calculate which available behavior is expected to bring
about maximum reward. In effect, the choice of behavior depends both on how strongly
desired outcomes are, and on how likely experience has shown those outcomes to be conse-
quent on given behaviors. To this extent, the desire system works something like the econo-
mists’ model of a utility maximizer, selecting the behavior with maximum expected utility.
Both the habit system and the desire system seem to exist in humans as well as in rats.
One striking piece of evidence for this is that outcome devaluation experiments work in a
similar way with human participants, including the fact that overtraining leads to habitual
behavior and the loss of sensitivity to devaluation (Tricomi et al. 2009; for a very similar
experiment on rats, see Balleine and Dickinson 1998). In a neat demonstration of this latter
point, researchers found that people who regularly ate popcorn at the cinema would eat
roughly equal amounts whether the popcorn was fresh or stale, but that those who ate it
less frequently were more discriminating (Neal et al. 2011). Further, imaging studies have
found that the brain areas in humans involved in the two systems correspond to those that
the lesion studies of Yin and colleagues found in rats (Balleine and O’Doherty 2010).
It is worth noting that desires seem to be involved, not only in determining how we
behave, but also in generating the reward signals responsible for the formation of new
habits and desires.3 Evidence for this claim comes from the phenomenon of secondary
reinforcement, in which behaviors are learnt when they lead to outcomes that animals
have been trained to find rewarding, such as lights and tones, rather than outcomes they
might be expected to find rewarding regardless of past experience, such as food (Skinner
1938; Hull 1943). This is an important point, because it helps to explain how humans
come to have the wide range of—​sometimes unlikely—​desires that they do; the idea
would be that these desires can all, in principle, be connected to “primary rewards” like
food or apparent social advancement via long chains of associations.

2.4  Standing and occurrent desires

It is important to distinguish between the roles of standing and occurent desires in control-
ling behavior. We can think of standing desires as relatively stable features of our characters

3 This idea is central to the theory of desire proposed by Schroeder (2004).

104  

104 Normal and addictive desires

that persist over time; occurrent desires, by contrast, are the active manifestations by means
of which these stable features make direct contributions to the control of behavior.
Some obvious features of our behavior provide ample initial reason for making this
distinction. On the one hand, if we did not have standing desires, but instead occurrent
desire arose spontaneously and dissolved when satisfied, then it would be hard to explain
why we consistently make similar choices in similar situations—​for instance, why some
people almost always choose chocolate ice-​cream, even when lots of other flavors are
available. On the other hand, if we did not have occurrent desires, but instead all of our
desires were equally poised to influence behavior at all times, it would be hard to explain
why we are sometimes highly motivated to eat ice-​cream, but at other times indiffer-
ent to doing so, even when we know it is readily available and we have no other very
pressing tasks.
A deeper reason for recognizing the distinction is that, where standing desires enable us
to store permanent knowledge about the values of outcomes, occurrent desires enable us
to register the ways in which the values of outcomes sometimes change dramatically with
circumstances and in particular with our physiological needs. This need for two levels
of responsiveness to the values of outcomes is particularly clear when we consider that
outcomes that are sorely needed in some circumstances may be dangerous in others. For
example, when an animal is very salt deprived, it may be crucial to survival that it experi-
ences strong desires for salty foods; but if this boost to the desires persisted in the long
term, the risk to the animal’s health could be considerable.
The distinction between standing and occurrent desires is supported by our under-
standing of how the brain works. There is some evidence for identifying standing desires
with relatively stable, gradually evolving neural structures in the orbitofrontal cortex
(Balleine and O’Doherty 2010; Plassmann et al. 2007). Occurrent desires, on the other
hand, are naturally equated with current patterns of neural activity, in line with imaging
results. Given this picture, we can view behavior selection by the desire system in terms
of competing coalitions of activity that represent associations between currently valued
outcomes and salient available actions (cf. Cisek 2007).
It is no doubt the proximal role of occurrent desires in behavior control that renders
the task faced by the desire system computationally tractable. If all standing desires
needed to be taken into account when selecting behaviors, the desire system would
continually need to carry out unmanageably complex calculations. However, if it
works solely with occurrent desires that are activated only when perceptual experi-
ences or ongoing thoughts bring their objects to mind, then it will only need to deal
with a limited number of outcomes. This picture fits with the introspectively plausible
idea that the strength of an occurrent desire is affected by the degree of attention that
its object attracts (Hare et al. 2011). We want things less when they are more distant
from our thoughts, and more when they are closer. Given that our attention is also
guided by our desires, it seems that a positive feedback loop may operate here, which
may explain why addicts sometimes find themselves constantly thinking about the
things they are addicted to.
  105

Systems for behavioral control 105

2.5  The planning system

The two systems we have discussed so far are both for deciding what to do now. However,
we humans also have the ability to think about what to do in the future, to form plans,
intentions, and resolutions, and indeed the ability to reflect in a more general way on
our own motives, purposes, and concerns. These abilities, discussed by Michael Bratman
(1987) and Richard Holton (2009), make up the planning system.
The planning system does not always fix how we will behave in the future—​we do not
always stick to our plans—​but it certainly has a significant influence. Conflicts frequently
arise between what we have planned in advance, and what the desire system or the habit
system recommend when the time comes. Sometimes the older habit and desire systems
dominate in such cases of conflict, but this is by no means the general rule.
The other two systems we have discussed so far both promote rewards, and so are rela-
tively easy to understand as adaptations. Both habits and desires are learnt through the
association of their objects—​behaviors or outcomes—​with reward, and reward signals are
prompted in the first instance by outcomes that are relatively easy to detect and conducive
to survival and reproduction, such as food, water, sex, and (in many animals, including
humans) apparent social success. The habit system allows animals to learn new rewarding
behaviors, and the desire system adds to the sophistication with which reward is pursued.
While there is room for debate about the nature of some of the advantages brought by the
desire system over the habit system, one clear advantage is that it allows relevant informa-
tion learnt in a wider variety of ways to be brought to bear in deciding how to behave—​as
in the outcome devaluation experiments. However, in the case of the planning system, it
may be less obvious why the system is adaptive.
There is much to be said on this topic. Here we shall focus on three putative advantages
that come with the ability to form long-​term plans. First, the planning system helps us
to coordinate our actions more effectively, not only with other people, but also with our
successive selves over time, so to speak. To appreciate the latter point, consider what may
be viewed as an intrapersonal prisoner’s dilemma: the question of whether or not to take
exercise today (Ainslie 2001; Gold 2014). For many people, this question is like the pris-
oner’s dilemma. From their present perspective, things will go best for them if they do
not exercise today, but do on many other days; things will go fairly well if they exercise
today and on many other days; they will go fairly badly if they neither exercise today nor
on other days; and they will go very badly if they exercise today but do little otherwise.
So it looks as though, on any given day, it will be rational to choose not to exercise, since
this will mean things go better whether or not they exercise on other days. However, the
planning system allows us to adopt a longer-​term perspective, and recognize that in the
face of this situation the only way to get the best overall outcome is to come up with a
training plan, and stick to it.
A second advantage is that the planning system allows us to draw on a far wider range
of information about means to valued ends than either the habit or desire systems. We
have seen that the desire system outstrips the habit system in forming desires on the basis
106  

106 Normal and addictive desires

of new information about which outcomes are valuable. But this is consistent with the
desire system being highly limited in its access to information about which behaviors
will be effective means to those valued ends. For all that has been said so far, the desire
system may be limited in this respect to means–​end information that comes directly from
personal experience of given behaviors leading to valued outcomes. If this is right, then a
system that allows careful and prolonged deliberation before fixing on a plan will have the
benefit of being able to draw on far wider informational resources, including, crucially,
the testimony of others and bodies of cultural tradition.
Finally, there is the point that the planning system allows us to reflect on our desires, and
privilege the pursuit of ones that we judge more worthy. This is in contrast to the desire
system, in which the desires that influence us most strongly are just those that happen to be
most strongly aroused by currently encountered cues. We can mark this contrast by saying
that the planning system is orientated to what we judge to be genuinely valuable, where
the desire system simply pursues those outcomes that are currently desired. Reflection
may lead us to judge that some things we desire are of little value—​cigarettes, say—​or to
judge that things that we do not desire much are of significant value—​like spending time
exercising outdoors. This does not necessarily commit us to the view that our judgments
of value are radically independent of our desires. Some philosophical theories of value do
have this consequence, and they may well be right. But for present purposes, it will make
no difference if judgments of value are simply the upshot of reflection about the ways in
which different desires facilitate, enhance, obstruct, or diminish one another. For example,
you might currently desire both good health and cigarettes. But reflection will tell you
that your good health is something that is also conducive to satisfying many of your other
desires, whereas cigarette smoking is not, and will thus indicate that the former desire is
more worthy that the latter. Note that, even on this deflationary understanding of the dif-
ference between value and desire, a planning system that focuses on values will have a clear
adaptive advantage over one that is restricted to pursuing currently desired outcomes.

2.6  Systems in conflict

So far we have said little about the way the three systems of behavioral control interact,
and in particular about the way the human planning system relates to the habit and desire
systems we share with other animals. There are two possible models here.
On a first parallel systems model, the three systems each operate independently, fixing
on their preferred behavioral response at any one time, and then fighting it out, so to
speak, as to which one prevails. On this model, the planning system will presumably man-
age to dominate the other two in normal cases, but there will be special circumstances
where it is outfought.
On a second nested systems model, the planning system will rather seek to control
behavior, when it does so seek, not directly on its own behalf, but by controlling the
processes of the older behavioral control systems, perhaps by somehow increasing the
strength of desires for certain outcomes, and reducing desires for other outcomes. On
this model, the planning system can again be expected to determine behavior in most
  107

A theory of drug addiction 107

cases, but will fail to do so if it loses control of the desire system. This might happen, for
example, if more immediate influences on the formation of desires are strong enough to
eclipse the modulation of desires stemming from the planning system.
Holton (2009, ch. 6) argues for a view of the relationship between the planning and
desire systems that amounts to the parallel systems model, on the grounds that resisting
temptation can be effortful, and that this implies a struggle between the two systems. But
we see no reason to commit to this model. Even on the nested systems model, resisting
temptation will involve a process in which different influences—​the planning system and
current cues—​compete to generate stronger occurrent desires. These influences may be
mediated by effects on attention, since this is a major factor in determining the strength
of occurrent desires, and the conscious control of attention is typically effortful.
On either model, then, the planning system will on occasion fail to determine behavior,
either because its directive is countermanded by an independently operating desire sys-
tem, or because it loses control of the desire system that it is normally able to manipulate.
So, on either model, there will be a threat to the planning system at any time when the
independent influences that direct the desire system push strongly for a course of action
other than chosen by the planning system.
Our discussion so far has indicated a number of ways in which divergence between
the planning system and the desire system can arise. For a start, we have seen how the
planning system will use a wide range of testimonial and cultural information to choose
means to its ends, where the desire system is arguably limited to information derived
from the organism’s experience. Moreover, the planning system normally chooses strate-
gies for possible future circumstances, and so will often fail to take into account changes
in circumstances that occur in the interim and influence the later desire system. Finally,
the planning system will seek to pursue considered values, where the desire system is
directed by unmodulated occurrent desires, which are largely determined by what is sali-
ent to us at any given time, with the result that, at least in modern environments, we often
experience strong temptations that fluctuate rapidly. Given that the habit system is also
capable of affecting how we feel inclined to behave, and will also sometimes conflict with
the other two systems, it is not surprising that we often fail to stick to our resolutions, or
behave inconsistently in other ways.

3  A theory of drug addiction

3.1  Dopamine and addiction
We now turn to the standing desire theory of drug addiction, which has been most clearly
advocated by Richard Holton and Kent Berridge (2014; see also Chapter 9, this volume).
The theory claims that many addictive drugs boost dopamine levels in the brain abnor-
mally, which causes the formation of excessively strong desires for those drugs. This the-
ory is concerned with what it is about the drugs that makes them addictive, and why
addiction to these drugs generates the patterns of motivation and behavior that it does,
rather than with other issues, such as why some drug users become addicted and others
108  

108 Normal and addictive desires

do not. It does not incorporate any claims about the extent to which addicts are in control
of their actions, unlike proposals by Frankfurt (1971) and Foddy and Savulescu (2010),
but does provide empirical foundations from which such claims could be developed. It
also illustrates the value of the framework for understanding behavior that we have just
described. As we will explain, this theory is elegant and promising, although it is in ten-
sion with some prominent accounts of the function of dopamine. We start by describing
the role of dopamine in the theory.
Dopamine is a neurotransmitter that is thought to have a small number of specific
cognitive functions, although there is considerable debate about exactly what those
functions are (we will discuss aspects of the debate in section 3.4). However, one point
that is well established is that many addictive, psychoactive drugs boost levels of dopa-
mine in the brain, by a variety of mechanisms. These include amphetamines, alco-
hol, nicotine, opiates, cocaine, cannabis, and benzodiazepines (Nestler 2005; Tan et al.
2010). Given that in other respects these drugs have variable effects, a natural hypoth-
esis is that it is this effect on dopamine which explains why they are addictive, and this
claim now forms part of many different theories of the mechanism of drug addiction.
Even if this is correct, it is likely that the details of the ways in which different drugs
act are important for understanding the various specific symptoms that addiction can
incorporate, but we will concentrate on the wider picture. In addition, it is entirely pos-
sible that some psychoactive drugs are addictive and abused despite not having this effect,
or primarily in virtue of other effects, but for the sake of convenience we will ignore this
possibility when using terms like “drugs” and “drug addiction.”
A further reason to view dopamine as central to drug addiction is that, despite the
uncertainty about the exact functions of dopamine, it is known to be centrally involved
in processes relating to reward and behavior selection. In particular, there is considerable
(although perhaps not conclusive) evidence that phasic dopamine is some form of reward
signal. Dopamine is released continuously by the midbrain dopamine neurons, at what
is called a “tonic” level, but its release is also frequently punctuated by brief bursts and
pauses that are believed to be signals of some kind, and are known as “phasic” dopamine
signals. Some of the most influential studies suggesting that phasic dopamine signals
represent reward levels were conducted by Wolfram Schultz (e.g. 1998), who found that
bursts of dopamine release occurred in monkeys’ brains when they were given unex-
pected rewards. Further to this, Schultz also found that when the monkeys were trained
to expect the rewards subsequent to a cue, the bursts stopped occurring when the reward
was delivered, and instead came to coincide with the cue. When the cue was delivered
but the reward omitted, a phasic pause in dopamine release followed the omission, inter-
preted as a negative signal.
As a result of these studies, phasic dopamine has come to be thought of as a reward
prediction error signal, representing the difference between the level of reward received
and the level expected at that time. So, on this hypothesis, drugs of abuse that cause arti-
ficially high dopamine levels would—​provided they boosted phasic rather than tonic
dopamine—​lead the brain to treat them as always much better than expected.
  109

A theory of drug addiction 109

3.2  The standing desire theory

The theory we are interested in further proposes, in keeping with this idea, that the pha-
sic dopamine signal is used for updating the strength of standing desires (see esp. Holton
and Berridge 2014, sect. IV). This point, rather than the claim that addiction is caused
by abnormal dopamine signals, is what distinguishes the standing desire theory. On this
view, positive dopamine signals at a given time strengthen standing desires for those out-
comes that were salient at that time, while negative dopamine signals cause such desires
to be weakened. So the standing desire theory proposes that addiction is the result of
dysfunctional, exceptionally strong standing desires to take drugs, which are produced
as a result of artificially boosted dopamine signals. We will first discuss some features of
this theory, then describe how it relates to existing accounts of drug addiction.
A first point is that the standing desire theory denies that drug addiction can be
explained purely in terms of strong habits. It attributes addiction to acquired desires, and
accordingly sees drug-​seeking behavior as an adoption of behavioral means to desired
ends, as directed by the desire system and perhaps also the planning system. In support
of this perspective, Berridge and his co-​authors have emphasized how addicts often per-
form complex sequences of carefully planned behaviors in order to get the opportunity
to take drugs (Berridge and Robinson 2011; Holton and Berridge 2014). They argue that
this counts against the thought that drug addictions are literally habits, because the habit
system involves no capacity for means–​ends reasoning.
Further, the standing desire theory yields an immediate explanation of why addicts
are strongly attracted to drugs: strong standing desires generate strong occurrent desires,
when appropriately cued. On this point, it is an advantage of the standing desire theory
that it makes motivation to take drugs depend on addicts’ circumstances, to about the
right degree. For addicts who are trying to quit or cut down, avoiding drug cues is a
valuable—​perhaps crucial—​strategy, and the theory can explain this, because standing
desires motivate us only when they become occurrent, and occurrent desires are strong-
est when they are most strongly cued. This also explains the possibility of relapse even
in addicts who describe themselves as having lost their desire for drugs, since standing
desires alone may be difficult to recognize in oneself. Other addicts, however, may spend
considerable amounts of time seeking and taking drugs, and the theory can also explain
this, because it predicts that people who spend a lot of time engaging with drugs and in
circumstances they associate with drugs will be almost continuously motivated to remain
in those circumstances, and take more drugs. Notably, the standing desire theory predicts
not only that addicts will form strong standing desires for drugs, but also for other out-
comes that are salient when they take drugs, because these desires will also be artificially
boosted by the drugs’ effect on dopamine. Further, because standing desires persist over
time, the theory can explain why even addicts who have stopped taking drugs remain at
risk of relapse much later in life.
Finally, the standing desire theory can also explain why addicts are so prone to suc-
cumb to temptation even when they have resolved to abstain. Along with everybody else,
110  

110 Normal and addictive desires

addicts will have many other standing desires, and many of these will conflict with their
desires to take drugs, especially if their addictions become harmful. The planning system
means that addicts are likely to recognize this and to form plans, intentions, or resolutions
which involve reducing or ending their drug use. However, the role of dopamine in the
formation of their desires to consume drugs means that their occurrent desires for drugs
will on occasion be exceptionally strong. This can influence the desire system to operate
in opposition to the planning system, along the lines discussed in section 2.6, with the
result that the addicts find themselves unable to stick to their resolution to quit.
Overall, then, the standing desire theory provides a particularly straightforward
account of addiction, which explains all of its most striking features. By focusing on the
desire system, the theory avoids the difficulties that face habit-​based accounts of addic-
tion, and also leaves room for the planning system—​and the desire system, in circum-
stances that do not evoke drug-​related occurrent desires—​to motivate addicts to quit.
So crucially, the standing desire theory in the context of the three-​system view accounts
for the conflict in addiction. The theory does not imply that this conflict is a necessary
feature of addiction, but only that it will arise if addicts come to believe that their addic-
tions are damaging.
The standing desire theory also has the advantage of making very limited claims about
the neurobiology of addiction. It claims only that all addictive drugs boost dopamine
levels, and the dopamine has the function of updating standing desires. Addictive drugs
certainly have many effects on the brain, which may contribute in a variety of ways to
the behavior and experiences of those who consume them, but it is an advantage for a
theory to make minimal claims about these matters, since we are looking for the common
mechanism behind drug addiction in general.

3.3  Comparison with existing theories

In this section, we shall briefly discuss how the standing desire theory relates to some
other existing accounts of drug addiction.
One well-​known account which has strong affinities to the standing desire theory is
by Hyman (2005). Hyman argues that drug addiction is caused by distorted dopamine
signals, that phasic dopamine is a reward prediction error, and that it explains drug addic-
tion because it causes long-​term changes to the motivational effects of drug cues and
to dispositions to perform drug-​seeking behaviors. However, Hyman’s account is less
specific about the effects of dopamine than is the standing desire theory. On his view
addiction seems to involve the distortion of habits, desires, and possibly some other states
involved in behavioral control as well.
Arpaly and Schroeder (2014) also give an account of addiction that focuses on the
role of dopamine. However, they do not see this as generating artificially strong standing
desires, but rather as affecting habits—​including habits of thought—​by accentuating the
strength of positive and negative reward signals. They also claim that these signals would
produce very pleasant and very unpleasant conscious experiences, respectively, which
would also influence behavior. From the perspective we have adopted, their theory seems
  111

A theory of drug addiction 111

unnecessarily complex; it is unclear why they resist the straightforward suggestion that
dopamine updates standing desires.
Other recent accounts have tended to emphasize the degree to which addictive choice
is like ordinary choice. Foddy and Savulescu (2010) argue that this follows from the
claim that addicts are motivated by unusually strong desires to take drugs, and write
that “addictive desires are just strong, regular appetitive desires” (p. 14). Pickard (2012)
acknowledges that addictive desires are strong and “habitual,” but argues that addiction
is explained in large part by the genuine benefits drugs bring to some addicts with co-​
morbid psychiatric disorders, and also by the poor quality of life that addicts reasonably
expect to experience on quitting. Both of these accounts therefore emphasize the way that
taking drugs is rational for addicts, given their strong desires. Up to a point, we agree with
this perspective. From our point of view, addictive behavior should indeed be understood
in terms of the systems of behavior control that addicts share with everybody else, and in
particular in terms of the operation of the addicts’ desires within these systems. But on its
own this observation leaves out the crucial point that addicts’ desires are abnormally and
artificially strong, due to the special effects addictive drugs have on dopamine, and that
this explains why addicts find it so hard to stick to their resolutions.
In addition to these, a further prominent account has been put forward by Berridge
and his collaborators, which is based on his view of dopamine function. This view does
not sit entirely comfortably with the standing desire theory, so we attempt to disentan-
gle these issues in section 3.4. In our view, it is best to think of Berridge’s incentive sen-
sitization theory as related to, but distinct from, Holton and Berridge’s standing desire
theory.

3.4  The incentive salience theory

While we find the standing desire theory plausible and explanatorily attractive, we do not
take it to be mandated by the evidence. As we have explained, what makes the standing
desire theory distinctive is not the claim that drug addiction is caused by drugs’ effects
on dopamine, but the further claim that the function of dopamine signals is to update
standing desires. However, this is not the only hypothesis about dopamine function that
is defended in the current literature. Two prominent alternatives are the incentive salience
view, and reward learning accounts.
The incentive salience theory is advocated by Berridge (2007, 2012). According to this
theory, the function of phasic dopamine signals is to motivate animals to act at times
when they encounter either rewards or cues that predict reward, and to pursue the per-
ceived or cued reward. As Berridge puts it, dopamine is a “wanting” signal—​it makes ani-
mals occurrently “want” perceived or cued rewards. Dopamine does not, according to this
view, have the function of producing lasting effects, either by contributing to learning new
actions, or by updating stored information about reward. Berridge argues that dopamine
is neither necessary nor sufficient for habit learning, since mice with almost no dopamine
can learn to perform new actions (Hnasko et al. 2005), and giving mice extra dopamine
does not seem to enhance learning (Cagniard et al. 2005).
112  

112 Normal and addictive desires

If the incentive salience theory described dopamine’s only effect then drug addiction
would be hard to explain. The incentive salience theory predicts that, when we take drugs,
we will become unusually strongly motivated to take more of the same drugs, because
of the artificially high levels of dopamine release that this causes. So the incentive sali-
ence theory is particularly well suited to explaining drug binges. However, the incentive
salience theory does not itself predict that drugs will cause any longer-​term changes, or
that addicts will feel strongly motivated to take drugs at any time other than once they
have already begun a session of use. So the incentive salience theory of dopamine func-
tion certainly needs to be supplemented with further claims about dopamine’s effects, in
order to contribute to a theory of drug addiction. (Without such claims, one might even
argue that the phenomenon of drug addiction undermines the incentive salience theory
of dopamine function. After all, the phenomenon of being abnormally attracted to drug
use even before starting a session seems to be common to a range of dopamine-​involving
drugs. So we would expect the role of dopamine to explain this phenomenon, and not just
the continuation of binges once they have started.)
Berridge and his colleagues are not unaware of this difficulty, and deal with it by add-
ing a further element to their theory. They argue that repeated exposure to drugs causes
sensitization of the incentive salience system. As Berridge and Robinson (2011) present
matters, addiction occurs because the incentive salience system becomes hyperreactive to
drug cues, after those cues have been repeatedly followed by unusually strong dopamine
signals. Since this is a long-​lasting effect, it can potentially explain addiction. The idea
would be that exposure to cues will activate the addicts’ incentive salience system, even
before they have ingested the drug. This would then explain the addicts’ abnormal ten-
dency to succumb to temptation in the first place, and not just their disposition to carry
on bingeing once they have succumbed.
It is worth emphasizing the differences between the standing desire theory and the
incentive salience (plus-​sensitization) theory. For a start, the standing desire theory
thinks of what it is to desire something as, roughly, to value it for the purposes of goal-​
directed control, as identified by outcome devaluation experiments. The incentive sali-
ence (plus-​sensitization) theory, in contrast, thinks of “wanting” something as, roughly,
the state of being attracted to it when one encounters it. “Wanting” in this form is possible
in creatures that lack the desire system, because they can respond to “wanted” objects by
unlearnt behaviors such as approaching and consuming them. Perhaps the same kind of
state does track the values of outcomes and objects for these two purposes in creatures
that do have desire systems, but this is not to be taken for granted. A further point is that
Berridge and Robinson (2011) distinguish incentive salience “wants” from what they call
“cognitive incentives” or “cognitive desires,” which they say are involved in “goal-​directed
planning” and responsible for subjective ratings of desire. It is not clear that they are here
thinking of behavioral control in quite the same tripartite way as we do, but in any case
their “wantings” seem distinct from our desires.
Finally, since the incentive salience theory claims that dopamine does not have the
function of causing long-​term changes to behavioral control systems, it presumably
also claims that such long-​term changes are not elicited by natural rewards, at least
  113

Conclusion 113

in normal circumstances. So a major difference between the incentive salience (plus-​

sensitization) theory and the standing desire theory is that on the former the persisting
psychological states responsible for addiction seem to be of a kind distinctive to drug
addictions; on the latter these are normal psychological states—​standing desires—​in an
abnormal form.

3.5  Reward learning accounts

Reward learning accounts of dopamine function are probably the most popular current
class of theories (e.g. Wise 2004; Balleine et al. 2008; Glimcher 2011). These theories support
the standing desire theory of addiction to the extent that they claim that phasic dopamine is
a reward prediction error signal, with the function of producing long-​term changes to moti-
vational systems. But, like the incentive salience theory, they do not back up the crucial claim
for the standing desire theory, which is that phasic dopamine has the function of updating
standing desires. Instead, they argue that dopamine signals update habits (and perhaps also
Pavlovian values, which are associations between stimuli and innate behaviors).
This does not mean that they agree with Berridge’s incentive salience account; Berridge
claims that dopamine bursts signal that a stimulus with high incentive salience is pre-
sent, and generate enhanced effects only in the moment, whereas reward learning accounts
claim that dopamine bursts teach the incentive salience system to come to treat certain
stimuli as action guiding even when they are encountered in the future. So reward learning
accounts of dopamine function, unlike the incentive salience account, predict that drugs
would cause the formation of strong habits controlling drug consumption. These accounts
therefore fit well with Hyman’s and Arpaly and Schroeder’s accounts of drug addiction.
In principle, it could be that dopamine signals update habits and Pavlovian values as
well as desires—​or, for that matter, that they both update desires and have the incentive
salience function ascribed by Berridge. If this were so, it would be a partial vindication of
the standing desire theory. But at the same time it would mean that this theory would not
be the whole story about dopamine and addiction.

4 Conclusion
Our strategy in this chapter has been to understand addiction within the framework of
three interlocking systems of behavioral control—​the habit, desire, and planning systems.
We take the analysis we have offered to illustrate the virtues of this approach.
In the first instance, our framework allows a clear specification of what is distinctive
about addicts. They experience abnormally strong occurrent desires to consume the rel-
evant drugs. As a result, they will characteristically—​though not necessarily—​fail to stick
to any long-​term abstention plans they may form. As we have explained, while the plan-
ning system will normally function to override any contrary behavior prompted by con-
flicting occurrent desires, this is by no means inevitable, and is crucially sensitive to the
strength of the conflicting desires.
Moreover, our framework allows a plausible and attractive explanation for why addicts
experience abnormally strong cravings. We attribute this to the subversion of the desire
114  

114 Normal and addictive desires

system by the dopamine that is produced by the ingestion of drugs. In our view, this leads
to permanent changes in addicts, in the form of strong standing desires for drugs.
Other theories also attribute addiction to abnormal effects resulting from the produc-
tion of dopamine, but not via the formation of standing desires. We have explained above
how these theories face certain empirical difficulties. In addition, it is not always clear
exactly how these theories view the overall psychological effects produced by dopamine,
for lack of any explicit discussion of mechanisms of behavior control, along the lines of
our three-​system model.
Given this, we recommend the standing desire theory as the most elegant explanation
currently available of the phenomenon of drug addiction. It explains why addicts experi-
ence lasting attraction to drugs, which nonetheless is sensitive to a significant degree to
their circumstances; it leaves space for addicts to have a wide range of attitudes to their
own addictions, depending on the judgments of value made by the planning system; and
it can explain readily why it is hard (but possible) to quit and why relapse remains possi-
ble. It also requires only a single, straightforward claim about the function of dopamine—
that it constitutes the reward signal that is required for desire updating.

References
Adams, C.D. (1981). Variations in the sensitivity of instrumental responding to reinforcer devaluation.
Quarterly Journal of Experimental Psychology, 34B, 77–​98.
Adams, C.D. and Dickinson, A. (1981). Instrumental responding following reinforcer devaluation.
Quarterly Journal of Experimental Psychology, 33B, 109–​22.
Ainslie, G. (2001). Breakdown of Will. Cambridge: Cambridge University Press.
Arpaly, N. and Schroeder, T. (2014). In Praise of Desire. New York: Oxford University Press.
Balleine, B., Daw, N. and O’Doherty, J.P. (2008). Multiple forms of value learning and the
function of dopamine. In: P. Glimcher (ed.), Neuroeconomics: Decision Making and the Brain.
London: Academic Press.
Balleine, B.W. and Dickinson, A. (1998). Goal-​directed instrumental action: contingency and incentive
learning and their cortical substrates. Neuropharmacology, 37, 407–​19.
Balleine, B.W. and O’Doherty, J.P. (2010). Human and rodent homologies in action
control: corticostriatal determinants of goal-​directed and habitual action. Neuropsychopharmacology
Reviews, 35, 48–​69.
Berridge, K.C. (2007). The debate over dopamine’s role in reward: the case for incentive salience.
Psychopharmacology, 191, 391–​431.
Berridge, K.C. (2012). From prediction error to incentive salience: mesolimbic computation of reward
motivation. European Journal of Neuroscience, 35, 1124–​43.
Berridge, K.C. and Robinson, T.E. (2011). Drug addiction as incentive sensitization. In: J. Poland and
G. Graham (eds), Addiction and Responsibility. Cambridge, MA: MIT Press, pp. 21–​54.
Bratman, M. (1987). Intention, Plans and Practical Reason. Cambridge, MA: Harvard University Press.
Cagniard, B., Balsam, P.D., Brunner, D., and Zhuang, X. (2005). Mice with chronically
elevated dopamine exhibit enhanced motivation, but not learning, for a food reward.
Neuropsychopharmacology, 31, 1362–​70.
Cisek, P. (2007). Cortical mechanisms of action selection: the affordance competition hypothesis.
Philosophical Transactions of the Royal Society Series B, 362, 1585–​99.
  115

Conclusion 115

Frankfurt, H. (1971). Freedom of the will and the concept of a person. Journal of Philosophy, 68, 5–​20.
Foddy, B. and Savulescu, J. (2010). A liberal account of addiction. Philosophy, Psychiatry and
Psychology, 17, 1–​22.
Glimcher, P. W. (2011). Understanding dopamine and reinforcement learning: the dopamine reward
prediction error hypothesis. Proceedings of the National Academy of Sciences, 108, 15647–​54.
Gold, N. (2014). Team reasoning, framing, and self-​control: an Aristotelian account. In: N. Levy
(ed.), Addiction and Self-​Control: Perspectives from Philosophy, Psychology and Neuroscience.
New York: Oxford University Press, pp. 48–​66.
Hare, T.A., Malmaud, J., and Rangel, A. (2011). Focusing attention on the health aspects of food
changes value signals in the vmPFC and improves dietary choice. Journal of Neuroscience, 31,
11077–​87.
Hnasko, T.S., Sotak, B.N., and Palmiter, R.D. (2005). Morphine reward in dopamine-​deficient mice.
Nature, 438, 854–​57.
Holton, R. (2009). Willing, Wanting, Waiting. Oxford: Oxford University Press.
Holton, R. and Berridge, K. (2014). Addiction between compulsion and choice. In: N. Levy
(ed.), Addiction and Self-​Control: Perspectives from Philosophy, Psychology and Neuroscience.
New York: Oxford University Press.
Hull, C.L. (1943). Principles of Behavior: An Introduction to Behavior Theory. Oxford: Appleton-​Century.
Hyman, S.E. (2005). Addiction: a disease of learning and memory. American Journal of Psychiatry, 162,
1414–​22.
Neal, D., Wood, W., Wu, M., and Kurlander, D. (2011). The pull of the past: when do habits persist
despite conflicts with motives? Personality and Social Psychology Bulletin, 37, 1428–​37.
Nestler, E.J. (2005). Is there a common molecular pathway for addiction? Nature Neuroscience, 8,
1445–​49.
Niv, Y., Dayan, P., and Joel, D. (2006). The effects of motivation on extensively trained behavior. Leibniz
Technical Report, Hebrew University 2006–​6.
Pickard, H. (2012). The purpose in chronic addiction. AJOB Neuroscience, 3(2), 40–​49.
Plassmann, H., O’Doherty, J.P., and Rangel, A. (2007). Orbitofrontal cortex encodes willingness to pay
in everyday economic transactions. Journal of Neuroscience, 27, 9984–​88.
Schroeder, T. (2004). Three Faces of Desire. New York: Oxford University Press.
Schultz, W. (1998). Predictive reward signal of dopamine neurons. Journal of Neurophysiology, 80, 1–​27.
Skinner, B.F. (1938). The Behavior of Organisms. New York: Appleton-​Century.
Tan, K.R., Brown, M., and Labouèbe, G., et al. (2010). Neural bases for addictive properties of
benzodiazepines. Nature, 463, 769–​74.
Tricomi, E., Balleine, B.W. and O’Doherty, J.P. (2009). A specific role for posterior dorsolateral
striatum in human habit learning. European Journal of Neuroscience, 29, 2225–​32.
Wise, R. (2004). Dopamine, learning and motivation. Nature Reviews Neuroscience, 5, 483–​94.
Yin, H.H., Knowlton, B.J., and Balleine, B.W. (2004). Lesions of dorsolateral striatum preserve outcome
expectancy but disrupt habit formation in instrumental learning. European Journal of Neuroscience,
19, 181–​89.
Yin, H.H., Knowlton, B.J., and Balleine, B.W. (2006). Reversible inactivation of dorsolateral striatum
enhances sensitivity to changes in action-​outcome contingency in instrumental conditioning.
Behavioral Brain Research, 66, 189–​96.
Yin, H.H., Ostlund, S.B., Knowlton, B.J., and Balleine, B.W. (2005). The role of dorsomedial striatum
in instrumental conditioning. European Journal of Neuroscience, 22, 513–​23.
116  

Chapter 7

Addiction, compulsion, and weakness

of the will: A dual-​process perspective
Edmund Henden

Abstract
How should addictive behavior be explained? In terms of neurobiological
illness and compulsion, or as a choice made freely, even rationally, in the face
of harmful social or psychological circumstances? Some of the disagreement
between proponents of the prevailing medical models and choice models
in the science of addiction centers on “loss of control” as a normative
characterization of addiction. I examine two of the standard interpretations
of loss of control in addiction, one according to which addicts have lost
free will, the other according to which their will is weak. I argue that both
interpretations are mistaken and propose therefore an alternative based on a
dual-process approach. This alternative neither rules out a capacity in addicts
rationally to choose to engage in drug-oriented behavior, nor the possibility
that addictive behavior can be compulsive and depend upon harmful
changes in their brains caused by the regular use of drugs.

1 Introduction
Addiction is a paradoxical phenomenon. On the one hand, addicts seem knowingly to
expose themselves to harm, including emotional distress, legal and financial problems,
health problems, and so on. Many report feeling miserable and wanting to quit. They
are commonly thought of as having lost control over their drug-​taking behavior. On
the other hand, many addicts seem to be susceptible to a wide range of ordinary incen-
tives, including money, and counter-​incentives such as the risk of harm, suggesting that
they do in fact exercise a substantial degree of control over this same behavior. These
apparently conflicting observations are frequently used by adherents of what are held
to be the two main opposing models of addiction. The “medical model” sees addictive
behavior as compulsive and symptomatic of neurobiological illness. The “choice model”
sees it as the manifestation of harmful social or psychological circumstances that have
led addicts to freely and, in a certain sense, rationally choose to engage in this behavior.
What exactly is it that the proponents of these two models disagree about? Some of
their disagreement concerns the correct normative characterization of addiction where
  117

Addiction and freedom of the will 117

the bone of contention is the notion of “loss of control.” Loss of control refers to a
highly heterogeneous phenomenon for which a precise definition is difficult to give, the
trouble being that the empirical evidence does not seem to favor one normative charac-
terization more than any other. In this chapter I want to suggest a way of thinking about
addiction that resolves the apparent paradox. If correct, there will be no need for any
normative conflict between medical models and choice models of addiction.
The structure of the chapter is as follows:  in the following two sections I  examine
two alternative interpretations of addicts’ control problem. According to the first, they
have lost free will; according to the other, their will is weak. Contrary to what many
have claimed, I argue that addicts typically neither have lost free will nor display ordi-
nary weakness of will. In the subsequent section I propose an alternative interpretation
of addicts’ control problem that is based on a dual-​process approach. According to this
proposal, addicts’ lack of control stems from a certain sort of malfunctioning of their
will.1 This sort of malfunctioning does not rule out the capacity in addicts to act freely
and, in a certain sense, rationally choose to engage in drug-​oriented behavior, nor does
it rule out the possibility that this behavior can be compulsive and depend upon harmful
changes in their brains caused by the regular use of drugs.

2  Addiction and freedom of the will

What does it mean to “lose control of one’s actions?” It certainly means not doing what one,
in some way, “wants to do,” i.e. where what one wants to do expresses the content of one’s
will. Saying of someone that she has “lost control of her actions” is, therefore, to say that she
is acting against her own will. Very generally, the term “will” refers to a cognitive state that
has some action as part of its content, as well as an executive capacity to bring the world into
conformity with that content.2 It is common to distinguish, broadly, between two different
senses in which persons may be said to “act against their own will.” They can do so in the
sense that their will is too weak to sustain the performance of what they want to do. Such
persons are said to suffer from “weakness of the will.” But they can also act against their will
in the sense of being compelled by external forces to do other than they want. Such persons
are said to lack “freedom of the will.” In both cases, although the person fails to express their
will in action, the normative implications are very different. In the former, the person is held
to be blameworthy (or at least rationally criticizable), in the latter she is not. That is because
it is only in the latter case that persons are held to have been prevented from translating
the content of their will into action. Much of the philosophical discussion about addic-
tion has consequently revolved around determining to which of these normative categories
addicts’ control problems belong. That is, does their loss of control typify weakness of will

1 The sort of view I am proposing is, of course, not new. For different versions see, e.g., Redish et al.
(2008), Holton and Berridge (2013), Schroder and Arpaly (2013).
2 There are different views of exactly what sort of capacity this involves. According to one common view,
it can be characterized, roughly, as the capacity to choose and to act in the light of reasons (see, e.g., Raz
1997). This will be the view I shall assume in what follows.
118  

118 Addiction, compulsion, and weakness of the will

or unfreedom of the will? My own view, which I shall argue for in a moment, is that it typi-
fies neither—​at least not as they are ordinarily understood. Let me start with the claim that
addicts’ control problem consists in a loss of freedom of the will.

2.1  Addictionas loss of free will: irresistible desires

and rational incapacity
According to a plausible view, one (at least) necessary condition of a person having
free will with respect to a particular action at a given time is that she has the capacity to
refrain from that action at that time. If addicts’ control problem is lack of free will, it fol-
lows that addicts typically lack the capacity to refrain from drug-​oriented behavior. Those
who accept this claim tend to assume that what prevents addicts from exercising their
will is that their addiction creates such an overwhelming desire for drugs as to be liter-
ally irresistible.3 To say that a person’s desire is “irresistible” is to say that she is unable to
resist acting on it. This is commonly interpreted as meaning that she has lost a rational
capacity—​broadly construed, the capacity to respond to reasons of a certain type (Fisher
and Ravizza 1998). Given the plausible view that capacities are general or multi-​track in
nature—​they don’t simply manifest themselves as single possibilities, but rather as whole
rafts of possibilities—​a person can be said to have lost such a capacity if a whole host of
relevantly similar counterfactuals are true of her (Smith 2003). Consider, for example, a
cocaine addict. What makes it the case that she has lost the capacity to resist cocaine?
According to the reasons-​responsiveness view, it would be true to say of her that she would
not respond to reasons to refrain from cocaine—​cocaine of ever so slightly different kinds,
in ever so slightly different circumstances—​in a suitable context of relevantly similar pos-
sible worlds in which she was given what she took to be good and sufficient reasons to
refrain from cocaine. In short, addicts lack free will with respect to their drug-​oriented
choices and actions because they have lost the capacity to respond to reasons to abstain.
Is there any evidence that addicts suffer from this sort of rational incapacity? No. On
the contrary, there appears to be plenty of evidence that they do not. The evidence dem-
onstrates that a wide range of counterfactuals of the sort just mentioned is actually false
of many addicts. Several studies suggest, for example, that addictive behavior varies as a
function of costs, benefits, and cultural values, and many addicts appear to quit sucessfully
unaided when they reach their early thirties (Heyman 2009; Pickard 2015). One particu-
larly interesting illustration of addicts’ capacity to control their drug-​oriented motivation
is provided by a form of behavioral therapy known as contingency management (CM)
treatment. In CM, every time a desired behavior occurs, tangible and immediate rein-
forcers are applied. When it does not occur, they are withdrawn (see Petry et al. 2011).
When the method is applied to substance disorders, they commonly use an exchange
system. Patients receive vouchers for metabolic evidence of drug abstinence. Every time
an addict submits a scheduled urine sample (typically three times a week) that tests nega-
tive for drugs, she earns a voucher that can then be traded in for desirable but inexpensive

3 For a classic statement of this view, see Frankfurt (1971).

  119

Addiction and freedom of the will 119

goods, such as restaurant gift certificates, clothing, or electronics. The value of the vouch-
ers increases with each consecutive instance of proven drug abstinence. Conversely, if
the patient uses drugs or fails to submit a scheduled sample, the value of the voucher is
reset to the starting point. In combination with counseling, CM has proved a surprisingly
effective method (Heyman 2009). Several studies have shown that, compared with control
groups receiving traditional psychological counseling, a significantly larger number of
subjects in the voucher groups remain abstinent during and after CM treatment.
Now, it is reasonable to infer that CM treatment must impact on addicts’ capacity for
reasons-​responsiveness, either by strengthening their motivation to exercise this capacity
or by strengthening the capacity itself—​or perhaps by some combination of the two. For
present purposes, what is interesting about it is that it suggests that some capacity must
be present also before treatment with CM. In fact, voluntarily submitting to CM treatment
and complying with the therapist’s instructions seems evidence of such capacity. The
effectiveness of CM, in other words, appears difficult to explain unless one assumes that
many addicts do in fact possess a rational capacity, at least to some extent. If this capacity
is diminished as a result of drug abuse, it would presumably explain why they find it dif-
ficult to abstain. But if many addicts typically possess rational capacity, it is not plausible
that addiction typically causes a loss of free will with respect to drug-​related choices and
actions. The view that addicts necessarily lack free will is therefore likely to be false.
Should the control problem be conceptualized then in terms of ordinary weakness of
will? Some have argued that it should (e.g. Benn 2007). In section 2.2 I present two dif-
ferent versions of this view (corresponding to two different accounts of weakness of will)
and argue that both are mistaken.

2.2  Addiction and weakness of the will

The claim that addiction is a species of weakness of the will can be interpreted in two
ways, corresponding to two different views of the nature of weakness of will. According
to what, perhaps, has become the dominant view, weakness of will refers to the violation
of a “resolution,” where the resolution is a kind of decision, intention, plan, or policy that
is formed precisely in order to remain firm in the face of contrary desires one expects to
arise when the time to act comes (Holton 1999). According to this view, weakness of will
is manifested in an over-​readiness to abandon one’s resolutions for exactly the type of
reasons they were meant to overcome. If addicts typically are weak-​willed in this sense,
their control problem consists in a persistent lack of resolve with respect to abstinent
behavior. So, for illustration, imagine an addict who now forms a resolution to abstain
from taking drugs in the future. As the opportunity for consumption draws near, she suc-
cumbs to the temptation to rationalize her reluctance to abstain by changing her mind
about what would be best—​for example, by giving too much weight to certain considera-
tions that appear to provide reasons for consumption, such as the immediate pleasures
of the drug—​and then revises her resolution to abstain accordingly, usually regretting it
afterwards. This sort of weakness involves diachronic conflict, the irrationality of which
is displayed in a form of incoherence between her motivational state at the time of action
and her long-​term attitudes toward these states.
120  

120 Addiction, compulsion, and weakness of the will

However, not all agree that weakness of will must involve violation of a resolu-
tion. The traditional view, dating back to Aristotle, is that weakness of will is the
same as “akrasia,” i.e. intentional action freely performed against the person’s con-
sciously held judgment that, on the whole, it would be best not to perform the action
(Mele 2009). According to this view, weakness of will is expressed in a failure to com-
ply with one’s own best judgments while still holding them. It involves, in other
words, a synchronic conflict, the irrationality of which is displayed in a form of inco-
herence in the person’s attitudes at the time of action. If addicts are weak-​willed in
this sense, they typically might not be forming resolutions to give up their addictions.
Rather, their control problem might consist, precisely, in their failure to form such resolu-
tions. Typically, what they find hard might be to decide to do or commit themselves to
doing what their own judgment is telling them would be the best option all things consid-
ered. There need be no sense, therefore, of them constantly changing their minds about
what is best. Rather, their problem is that they cannot make up their minds to actually do
it. Their defect consists in a failure to comply with their best judgment.
Before proceeding, let me add a brief remark about these two views, as there has been
some controversy about which of them truly captures the nature of weakness of will
(Holton 1999; Mele 2009; May and Holton 2012). I  shall assume, for present purposes,
that there can be both akratic and non-​akratic forms of weakness of will. I do so because it
seems to me that, insofar as we care about this phenomenon at all, we care about it because
of its normative significance, because weak-​willed people are held to be rationally criti-
cizable (or blameworthy) for succumbing to temptation.4 Thus, it is widely agreed to be
reasonable to expect or demand that persons displaying weakness of will resist acting on
their contrary desires—​something that is commonly taken to imply the possession (unlike
persons lacking in free will) of capacities of resistance.5 In other words, weakness of will
is displayed by persons insofar as they are criticizable (or blameworthy) for making insuf-
ficient effort to exercise capacities they are believed to possess in order to do what they in
some sense want to do. If the importance of weakness of will—​at least for most practical
purposes—​is associated with its normative characterization, I think it would be a reason-
able hypothesis to assume that the ordinary notion of weak-​willed action corresponds to a
more general notion of action lacking in self-​control due to insufficient effort rather than
to any of the more technical definitions discussed in the philosophical literature.6 This
means, of course, that weakness of will is displayed both by persons who exert insufficient
effort to make up their mind about what to do, i.e. to make decisions, form resolutions, or
commitments, as well as by persons who exert insufficient effort to stick to such decisions,
resolutions, or commitments, in the face of temptation. One might wonder, though, why

4 By “we” here I mean us ordinary folk. I take weakness of will to be, at its core, a folk-​psychological
notion. There may still, of course, be other reasons why someone might care about it (e.g. philosophical
or scientific reasons).
5 For a different view, however, see Watson (2004).
6 For empirical evidence that the folk notion of weakness of will may not be identical with any one of the
two standard definitions in the philosophical literature, see May and Holton (2012).
  121

Addiction and freedom of the will 121

it matters whether addicts typically display weakness of will in the akratic or non-​akratic
sense. There is, however, a practical reason why it might matter: helping addicts overcome
a tendency unreasonably to revise their resolutions to give up drugs may seem to require
a different kind of approach than helping them make such resolutions in the first place.
While the former requires helping them learn ways of avoiding drug-​associated thoughts,
cues, or situations that might lead to a reconsideration of their resolutions, the latter seems
more a matter of strengthening their motivation to actually do what they themselves judge
is best, that is, to support and encourage them to translate their best judgments into practi-
cal plans or commitments to maintain abstinence.
Now, deciding whether addiction typifies akratic or non-​akratic weakness of will is
obviously problematic because it is very difficult to assess whether addicts in general
retain the judgment that it would have been better to abstain or, by the time they have
embarked on their addictive behavior, whether they have revised it. Nevertheless, both
alternatives seem to have some evidence in their favor. In sections 2.2.1 and 2.2.2 I shall
present some of this evidence. However, I shall argue that even if addiction is frequently
associated with both diachronic and synchronic forms of irrationality, it is still misleading
to treat it as a case of ordinary weakness of will.

2.2.1  Addiction as non-​akratic weakness of

the will: hyperbolic discounting
The best evidence that addicts tend to be weak-​willed in the (non-​akratic) resolution-​
violating sense is that their discount rates tend not to remain constant over time. “Discount
rate” is the rate at which they discount the utility of future rewards, such as, for example,
the benefits of a drug-​free life. There is plenty of research suggesting that addicts have a
tendency to discount the utility of future rewards, not by a fixed proportion per period
of time (exponentially)—​which would have led to temporally stable preferences—​but by a
proportion that declines as the length of the delay increases (hyperbolically), leading to reg-
ular and systematic preference reversals (Ainslie 2001; Bickel and Marsch 2001). Consider
an alcoholic who prefers before breakfast not to drink at tonight’s dinner party but still
has a strong craving for alcohol. As the opportunity to drink draws closer in time and the
prospect of drink begins to weigh more heavily with her, the rate at which she discounts
abstinence rises rapidly relative to the rate at which she discounts consumption. When the
opportunity to imbibe finally presents itself, her estimate of its utility outweighs her estimate
of the utility of abstinence with the result that her preference reverses. It may seem plausi-
ble that such “hyperbolic discounting” leads to weakness of the will, that is, an unreason-
able over-​readiness to abandon one’s resolutions. This view depends on the assumption that
hyperbolic discounting causes addicts to change their judgment of what is best at the time
of action, i.e. when the alcohol or drugs become available for consumption.7

7 Levy (2006) defends a version of this view. It is worth noting that if motivating preferences are revealed
in actual behavior and a gap may obtain between such preferences and a person’s evaluative ranking
of her alternatives, preferences can reverse in the absence of judgment shifts. The view that hyperbolic
discounting causes weakness of the will depends, therefore, on the assumption that motivating prefer-
ences always track considered judgments. For a different view, see, e.g., Lowenstein (1999).
122  

122 Addiction, compulsion, and weakness of the will

Now, although addicts find it extremely hard to remain abstinent over the longer term
(many of them probably due to hyperbolic discounting), the view that loss of control in
addiction should be conceptualized in terms of non-​akratic weakness of will faces a dif-
ficulty: in many cases addictive behavior does not seem to involve judgment shifts at the
time of action. Three key observations support this: first, addictive behavior sometimes dis-
plays a strongly habitual element—​a fact that seems important for understanding relapses
among users who have quit and are attempting to remain abstinent (Schroeder and Arpaly
2013). Such relapses seem frequently triggered by environmental cues and often appear to
occur without much conscious awareness of the behavior undertaken. Explaining them
in terms of a shift in considered judgments about what it would be best to do therefore
seems implausible. Second, addicts sometimes seem to retain a strong sense of the disvalue
of their drug-​oriented behavior even as they are carrying it out. For instance, there is evi-
dence suggesting that they sometimes make conscious and strenuous efforts to resist, even
indeed while seeking or taking drugs. As Robert West (2006) puts it, “when the restraint
fails, there is often (but not always) no sense of the addict having changed his mind and
deciding to engage in the behavior as a positive step; rather the sense is of a failure to exert
control followed by regret and a feeling of having let oneself down” (p. 133). But this seems
difficult to explain on the assumption that addicts typically abandon their judgment that
abstaining is best while satisfying their addiction. Third, addicts may continue to seek and
take drugs even when they derive no pleasure from their consumption, even in the absence
of withdrawal—​even, in fact, when they are convinced that taking drugs is a disastrous
course of action for them (Robinson and Berridge 1993). Once again, addicts would seem
on this observation to retain at times their judgment of abstinence as the most valuable
alternative while carrying out their drug-​oriented behavior.

2.2.2 Addiction as akratic weakness of the will: “wanting” and “liking”

The above observations may be taken as supporting the view that addicts’ control prob-
lem must be one of akratic rather than of non-​akratic weakness of the will. Akratic weak-
ness of will, remember, is action carried out contrary to the person’s consciously held
judgment of what it would be best to do all things considered. An initial objection against
this view could be that there simply is no evidence showing that it is even possible inten-
tionally to perform actions while at the same time consciously judging that it would be
best to refrain. All we can rely on to support this possibility are introspective reports of
the persons themselves and such reports are notoriously unreliable (Levy 2011a). But
there is other evidence. Of particular interest for our present purposes is the series of
neuroscientific experiments conducted by Terry E. Robinson and Kent C. Berridge (see,
e.g., Robinson and Berridge 1998). Based on these experiments, they distinguish two
components of motivation that are mediated by different psychological processes and
neural substrates, what they call “liking,” which is associated with an affective or cogni-
tive value (e.g. subjective pleasure, goodness, or predicted utility); and “wanting,” which
is associated with incentive salience—​the degree to which a stimulus is action driving.
While “liking” and “wanting” normally go together so that we “want” the things we “like”
  123

Addiction and freedom of the will 123

(e.g. the value associated with some environmental cue serves as a trigger to activate and
direct “wanting”), Robinson and Berridge provide evidence that they actually come apart
in addiction, often making addicts “want” things they do not “like.” The possibility of
a decoupling between “wanting” and “liking” supports the possibility of a dissociation
between intentional action and best judgment; if “wanting” can make addicts perform
actions contrary to what they “like,” and if judging some action as the best option entails
“liking” it, then this is evidence that it is possible to perform actions contrary to one’s best
judgment (see also Holton and Berridge 2013).
So, granting that intentional action contrary to one’s best judgment is possible, should
we, on the basis of the observations mentioned in the last section, conclude that addic-
tion typically involves akratic rather than non-​akratic weakness of will? I think there are
reasons to resist that conclusion. But let me begin by addressing a worry some readers may
have at this point. Why do we need to choose between the view of addiction as either a
species of akratic or non-​akratic weakness of will? Why can addicts not exhibit both forms
of weakness of will? I think both questions are reasonable. Addicts are not all alike. They
differ in circumstances, the drugs they use, their social and personal resources (e.g. abilities
and motivation), and their beliefs about the value of their options. Moreover, as we have
seen, there is credible evidence suggesting that they often violate their resolutions as well
as that they sometimes act contrary to their best judgments. Addicts’ control problem often
appears, then, to involve diachronic as well as synchronic forms of irrationality. Still, I think
there is an important sense in which addiction and ordinary weakness of the will differ. To
see this we need to go back to what is arguably the common-​sense notion of weakness of
will according to which persons who try too little to exercise capacities they are assumed to
possess in order to do what they in some sense want to do are criticizable (or blameworthy).
Now, what determines an effort’s “insufficiency,” and thereby criticizability, is a normative
matter. Presumably, there are certain shared expectations and norms guiding what counts
as sufficient effort in various contexts. Thus, when we observe people failing to do what
they want to do, we seem to have ways of correctly answering questions such as: did she try
hard enough? I have elsewhere argued that the notion of sufficient effort is normative in the
sense of being relative to the level of effort which, other things being equal, would have been
sufficient for a normal person successfully to perform an action of the same type if she was
as strongly motivated to perform it (Henden 2013). By “normal person” I mean someone
whose capacity for self-​control equals that of the majority of adults, and whose motivational
system is congruent. In other words, if a person, according to this view, fails to do what she
wants to do because her effort to do so is insufficient relative to this ordinary standard, she
can be said to display weakness of will. In such cases, she is not, as we may say, sincerely
trying. Hence she is criticizable (or blameworthy) for her behavior.
This gives us, I think, a way of distinguishing addiction from ordinary weakness of the
will. The reason is that many addicts who try to abstain appear to try very hard to give up
drugs, and many of them are wholly sincere in their effort to quit. Nevertheless, they fail
again and again. If this failure cannot be explained by lack of capacity, it must be explained
by insufficient effort. However, even though the effort they put into it is insufficient relative
124  

124 Addiction, compulsion, and weakness of the will

to what is actually required of them to abstain, it might still count as sufficient relative to
ordinary standards. That is, if a normal person had made the same effort in similar cir-
cumstances, it would seem reasonable to expect her to succeed. Assuming, then, that many
addicts display a strong will to give up drugs, it would seem unfair to criticize them for
weakness of will. Rather, they seem instead to face some extraordinary obstacle peculiar to
them that evades most normal persons. One observation supporting this is that we seem
inclined to consider addicts much less criticizable (or blameworthy) for their drug-​oriented
behavior than weak-​willed persons for their weak-​willed behavior. Consider, for example, a
cocaine addict and a weak-​willed non-​addict both of whom take cocaine.8 Suppose both act
contrary to their best judgments (or fail to keep to their resolutions). Their drug-​oriented
motivations, we will assume, were resistible at the time of action. Now, we are clearly more
inclined to criticize the non-​addict than the cocaine addict for their wrongful behavior. The
reason seems obvious:  while we assume it would be relatively easy for the non-​addicted
user to refrain from cocaine in this situation, it is, we assume, comparatively harder for the
cocaine addict to do the same. This normative difference affects our attitudes towards the
two cases, suggesting that our attitudes to addiction and ordinary weakness of will differ
significantly. Admittedly, this observation alone is not conclusive evidence that addiction is
not a species of weakness of will. Still, it gives us, I think, good reason to explore a different
approach to the loss of control in addiction. It is to this approach I now turn.

3  Addiction and compulsion

If we cannot conceive of addicts’ loss of control as loss of free will nor as ordinary weak-
ness of will, how should we conceive of it? To answer this question, we might begin by not-
ing that an important feature of addictive behavior is the regularity with which it occurs.
The property of “being addicted” refers to a certain kind of relation a person has, not to
some isolated act of consumption, but to a pattern of behavior, enacted on a regular basis
in characteristic circumstances, which the person finds extremely difficult to override by
intentional effort. Explaining an action in terms of addiction usually involves seeing it as
part of such a behavioral pattern (in fact, it is not clear it even makes sense to speak of
one-​off addictive actions).
Lack of control over behavioral patterns is, of course, also what tends to be emphasized
in clinical descriptions of compulsive behavior. Such behavior is characterized as strongly

8 On what grounds, it could be asked, can one distinguish a non-​addicted user from an addicted one
when their actual mental states and behavioral patterns are the same? The difference, I think it is plau-
sible to say, resides in certain counterfactuals being true when the person is an addicted and false when
he is a non-​addicted user. For example, it would be true of the addict but false of the non-​addict that,
as his supply of drugs wanes, he would begin obsessing over them, perhaps to the point of making an
extraordinary effort to obtain them (often at great cost to himself). And were his drug use to become
associated with displeasure, emotional distress, or health problems, it would be true of the addict but
false of the non-​addict that he would continue to consume the drug, often experiencing a physical
compulsion to do so. See Skog (2003) for defense of a counterfactual definition of addiction.
  125

Addiction and compulsion 125

cue-​dependent in the sense that it is regularly triggered by certain situations, places, or

people associated with the type of behavior in question; there is a feeling of being driven
again and again to behave in precisely that particular way (often in spite of oneself), and
it is a common experience that resistance, however sincere, becomes increasingly difficult
over time. Since all these features are typically also present in addictive behavior, it seems
reasonable to infer that addictive behavior is compulsive in the clinical sense. Indeed, this
is part of the standard medical definition of addiction (American Psychiatric Association
1994).9 If this is correct, it serves to further distinguish addiction from ordinary weakness
of will since the latter is not definitionally tied to patterns of behavior exhibiting features
of compulsivity. The question then becomes in what sense individuals who are exhibiting
compulsive patterns of behavior are “acting against their own will.”

3.1  Attention, bias, and the will

One important aspect of the will concerns its relation to attention. That is because “execu-
tive control,” an umbrella term used in psychology for top-​down cognitive processes that
regulate, coordinate, and control other cognitive processes that are necessary for the initi-
ation and monitoring of goal-​directed actions (such as reasoning, planning or inhibition),
involves directed attention, the capacity to voluntarily focus or shift attention. In fact, in
some theories of executive control, they are not even clearly separated—​executive control
is treated as more or less identical to the mechanism controlling the deployment of atten-
tion (Miyake et al. 2000). Why this is relevant to our understanding of the loss of control
in addiction is because there appears to be plenty of evidence that addicts’ attention is
biased toward drug-​associated stimuli (for a review, see, e.g., Field and Cox 2008; see also
Chapter 15, this volume). An attentional bias is a certain sort of disruption of a person’s
attentional selection process. This process is “biased” toward a particular kind of stimulus
if the stimulus intrudes on her experience by capturing her attention and this capturing is
part of a pattern and occurs in a systematic rather than random fashion. Attentional bias
is, in other words, a statistical tendency or inclination to direct attention at a particular
kind of stimulus (McKay and Efferson 2010).
Now evidence for drug-​related attentional bias in addiction is provided by a variety
of experiments measuring implicit cognition. In the addiction Stroop task, for example,
addicts are asked to name the color of drug-​related words, and the time they take to name
the color of these words is compared to those for drug-​unrelated words (e.g. words related
to musical instruments). The challenge is to focus attention on color while blocking out the
word’s meaning. Attentional bias is indexed as the difference between participants’ mean
color-​naming reaction time in trials with drug-​related and those with drug-​unrelated words

9 It is worth noting that while, in the philosophical literature, the term “compulsive behavior” tends to
be used to characterize behavior caused by “irresistible desires,” there is no assumption in the clinical
literature that behavior must be caused by “irresistible desires” in order to count as compulsive. When
I speak of “compulsion” in what follows, it is the clinical notion I shall have in mind. For a discussion,
see Henden et al. (2013).
126  

126 Addiction, compulsion, and weakness of the will

(Cox et al. 2006). Using this approach, researchers have found that, while addicts exhibit
significantly slower reaction times and are more prone to error when naming the color of
drug-​related words, control participants do not exhibit this pattern. The result has been
demonstrated in users of a variety of different drugs, including alcohol, cannabis, cocaine,
heroin, and tobacco. The standard interpretation of this “Stroop interference” is that, com-
pared to non-​addicts, drug-​related words capture addicts’ attention, causing excessive pro-
cessing of the semantic content of these words, thereby disrupting their color naming.
Based on the addiction Stroop and other paradigms developed to measure attentional
bias, many researchers believe attentional bias toward drug-​associated stimuli plays an
important part in explaining the maintenance or escalation of drug-​oriented behavior,
including relapses among users who have quit and are attempting to remain abstinent.10
Robinson and Berridge’s influential theory, briefly discussed in section 2.2.2, explains why
repeated drug use may lead to drug-​related attentional bias. According to this theory,
it is caused by a process they call “incentive-​sensitization” in which repeated drug use
produces a dopaminergic response that becomes sensitized (i.e. progressively larger) by
making certain regions in the brain involved in the motivation of behavior more easily
activated by drugs or drug-​associated stimuli. As a consequence, these drug-​associated
stimuli acquire powerful “incentive properties” by drastically enhancing their capac-
ity to grab the person’s attention and be perceived as particularly salient. According to
Robinson and Berridge, it is this process of incentive-​sensitization that causes a decou-
pling of “wanting” (incentive salience) from “liking.”
In human addicts, drug-​related attentional bias caused by incentive-​sensitization pre-
sumably works through the human decision-​making system, as evidenced in the ability
of addicts, in general, to delay, alter and in some cases, substitute, their drug-​oriented
behavior based on deliberation. Typically, addicts plan when and how to obtain the
drugs, taking all sorts of considerations into account. In order fully to explain how
drug-​related attentional bias produces drug-​oriented behavior in humans, therefore,
we need to know how the attentional bias interacts with the addicts’ decision-​making
system.11 In section 3.1.1, I present in broad outline what I think might be a plausible
view of this interaction.

10 One caveat is in order: there has been some debate in the psychological literature concerning the inter-
pretation of the evidence from the addiction Stroop, and not all agree that it demonstrates a bias in
addicts’ attentional selection processes. Some have suggested, for example, that the delayed color nam-
ing could result from attempts to suppress the processing of drug-​related words, or from a generic slow-
down in cognitive processing as a consequence of experienced craving induced by the drug-​related
words rather than from a selective attention to those words (see, e.g., Algom et al. 2004). The standard
interpretation does, however, appear to be corroborated even when more direct measures of attentional
bias are used, such as the monitoring of eye movements while the subject completes a visual probe task
in which drug-​related and control pictures are presented. For discussion, see Field and Cox (2008).
11 Incentive sensitization theory does not deny a role for top-​down cognitive control. Robinson and
Berridge are clear that whether or not a sensitized response is actually expressed may depend on con-
textual factors. See, e.g., Robinson et al. (2013).
  127

Addiction and compulsion 127

3.1.1 Addiction as a malfunctioning of

the will: a dual-​process perspective
It has become common in cognitive psychology to distinguish between two modes of
decision-​making: one fast, intuitive and effortless that is shaped by biology and implicit
learning, the other slow, analytical and effortful, and shaped by culture and formal tui-
tion.12 While the former mode—​often referred to as type-​1 processes—​depends on envi-
ronmental cues, is associative, automatic, and can control behavior directly without the
need for controlled attention, the latter—​often referred to as type-​2 processes—​depends
on decontextualization, is rule-​based, and requires controlled attention and effort.13 To
achieve rational decision-​making and reliably contribute to the person’s goal achieve-
ment, the two modes have to work well together. This requires, first, that the person’s
type-​2 processes can exert an executive function and override the impulsive output of her
type-​1 processes. For this to happen, her type-​2 process must be able to generate a more
considered response that is in line with her normative reasons, as well as involve inhibi-
tory mechanisms to suppress the response tendencies of her type-​1 processes. Second, the
person’s type-​1 processes must be able to select adequate and relevant information about
the practical situation as an input to her type-​2 processes (Saunders and Over 2009).
The sort of top-​down processes that are associated with executive control exemplify
type-​2 processes. They are inferentially integrated with the person’s propositional attitude
system and draw on all her background knowledge and beliefs. In contrast, the processes
underpinning attentional bias are examples of type-​1 processes. They consist of associative
relations in memory between environmental cues and behavioral propensities—​relations
that can be activated during critical decision points without the person’s conscious inten-
tion, deliberation, or even awareness. The latter processes are modular in the sense of
involving highly specified mechanisms that are inferentially isolated from the person’s
propositional attitude system.
Now, assuming this dual-​process model of decision-​making, drug-​related attentional
biases in type-​1 processes might be hypothesized to affect addicts’ behavior in a variety
of ways depending, most likely, on individual differences between addicts (e.g. differ-
ences in personal and social resources, abilities, and motivation, type of drug used and so
on). For example, one way might be by entering their type-​2 processes and shaping their
beliefs, desires, and reasoning about what to do. Thus, by persistently directing the per-
son’s attention to drug-​associated features of their immediate physical and social environ-
ment, the processes might cause an over-​appreciation of these features as well as blindness
or indifference to longer-​range goals (Dill and Holton 2014). The result could be that the
rate at which these addicts discount the utility of future rewards, such as the benefits of

12 Partsof what follows are based on Henden et al. (2013). A wide variety of evidence has converged on
the conclusion that some sort of dual-​process notion is needed to explain how the overall process of
decision-​making works. For a brief review, see Frankish and Evans (2009).
13 Thereis some disagreement on precisely how these processes should be characterized and distin-
guished. I cannot enter into this debate here. For discussion, see Stanovich (2009).
128  

128 Addiction, compulsion, and weakness of the will

abstinence, is increased drastically relative to the rate at which they discount consump-
tion, leading to regular and systematic preference reversals (and perhaps judgment shifts)
of the sort associated with hyperbolic discounting.
However, drug-​associated attentional bias might also shape addicts’ beliefs, desires, and
reasoning without causing hyperbolic discounting. Not all addicts have unstable prefer-
ences. Some never make any effort to abstain from drugs because they have no desire
to quit. Does that mean they are in full control of their drug-​oriented behavior? I think
most people would be inclined to say no. There is something odd about the idea of a per-
son addicted to a drug being in “full control” of her behavior with respect to that drug.
Simple reflection on the meaning of the notion of “being addicted” seems to speak against
this possibility. Plausibly, addicts with stable preferences (who do not fight their addic-
tion) will still show symptoms of compulsivity and obsession with respect to drug seeking
and drug taking (e.g. drug-​oriented considerations will always have precedence in their
practical reasoning), or would show such symptoms if their drugs were to become una-
vailable or associated with an increase in negative costs (see footnote 8). Compulsivity
and obsession—​despite their superficial appearance of “too much control”—​seem on a
deeper level to indicate the opposite of control. The dual-​process approach can explain in
what sense control is lacking: even if some addicts have stable preferences, all their beliefs
and desires will still be infused by drug-​associated attentional bias; hence, by taking these
beliefs and desires as inputs, their practical reasoning itself will in a sense be “out of control.”
But drug-​related attentional bias in a type-​1 process might also affect an addict’s behav-
ior in a much more direct way. As we have seen, there is evidence that addicts some-
times retain their judgment that abstinence is the most valuable alternative even as they
are carrying out their drug-​oriented behavior. From a dual-​process perspective, the
problem in these kinds of cases is not that the type-​2 processes are internally biased,
but rather that they regularly fail to suppress responses generated by their biased type-​
1 processes despite their conflicting with considered normative responses generated by
their type-​2 processes. The Stroop effect serves as a good example of such a dual-​process
conflict: an automatic type-​1 process directs the subject’s attention to one feature of the
stimulus (word meaning), which disrupts a type-​2 process (deciding the color of the
word). Berridge and Robinson’s decoupling of “wanting” and “liking” is another example.
As they describe “wantings,” they are clearly modular. Not only are the processes that
underpin them associative—​and presumably operating at high speed and requiring low
effort—​they are implicit in the sense of being able to drive addicts’ behavior indepen-
dently of their propositional attitude system (although in some cases they might operate
by increasing their experienced cravings). It is reasonable to infer that “wantings” must
be a kind of type-​1 process. “Likings,” by contrast, seem underpinned by processes that
work through addicts’ propositional attitude system. That is because they typically mani-
fest themselves in cognitive-​affective states such as desires, beliefs, and value judgments.
Addicts who “want” to engage in drug seeking and drug taking without “liking” it can
therefore plausibly be viewed as experiencing a dual-​process conflict between a type-​1
and a type-​2 process.
  129

Addiction between illness and choice 129

Now the question was how drug-​related attentional bias in a type-​1 process might
affect addicts’ behavior more directly. The answer, presumably, is that it upsets their
will in some way. According to a theory that has gained wide acceptance in psychol-
ogy, executive control functions all draw on the same limited resource. The more this
resource is consumed, the more depleted it becomes and the poorer the person per-
forms on subsequent tasks requiring executive control. This so-​called “limited-​resource
model” of executive control is supported by numerous experiments showing that when
people engage in executive control, later attempts at executive control will be less suc-
cessful (for a review, see Muraven and Baumeister 2000; see also Chapter 16, this vol-
ume). Now the Stroop task is commonly used as a measure of executive control and
cognitive depletion. The faster the participants name the colors of the words (which
requires suppressing the initial tendency to name the meaning of these words), the
higher their level of current executive control; the slower they are, the more depleted
they are assumed to be. Since addicts perform poorly on the addiction Stroop com-
pared to non-​addicts, it is reasonable to infer that their executive functions are impaired
by cognitive depletion. Plausibly, a drug-​oriented type-​1 process directing attention
toward drug-​associated features of the environment has become fixed—​most likely due
to incentive-​sensitization—​in entrenched dispositions and patterns of perception and
response, thereby increasing the cognitive load on executive functions associated with
type-​2 processes requiring directed attention. A  dramatic increase of cognitive load
on directed attention means a corresponding increase in the consumption of execu-
tive control resources, which, given the limited supply of these resources, suggests they
must detract from other executive functions. This impairs the performance of tasks
involving inhibition, reasoning, or planning. Cognitive depletion, then, might explain
why addicts regularly fail to override their type-​1 processes in cases in which they expe-
rience a conflict between type-​1 and type-​2 processes (see also Levy 2011b). In com-
bination, attentional bias and cognitive depletion increase cognitive inflexibility, that
is, they reduce addicts’ capacity to switch their thinking and attention among different
tasks, operations, or practical perspectives in response to changing goals or circum-
stances. Their drug-​related decision-​making becomes, as a result, more stimulus-​bound
and less responsive to reasons to abstain from drugs. Since cognitive flexibility is plau-
sibly part of a well-​functioning will, drug-​associated attentional bias can thus be said
to disrupt addicts’ will by drastically reducing cognitive flexibility and, in this sense,
causing their will to malfunction.

4  Addiction between illness and choice

This account no doubt needs more development and defence than I  can provide here,
but assuming that it is on the right tracks, what are the implications for the supposed
opposition between medical models and choice models in addiction science, and for the
disagreement between those who claim addicts have lost free will and those who claim
they suffer from weakness of will?
130  

130 Addiction, compulsion, and weakness of the will

First, a malfunctioning of the will of the sort proposed neither implies that addicts’
will must be unfree nor that it must be weak. Although cognitive inflexibility makes it
much harder for addicts to revise or abandon their drug-​oriented decision-​making pat-
tern, there is no reason to assume it makes them unable, in general, to resist their desires
for drugs. Unlike a loss of capacities of resistance (and hence free will), cognitive inflex-
ibility might be offset, for example, by a sharp increase of intentional effort, i.e. by force-
fully and actively redirecting attention towards drug-​unrelated aspects of the situation,
or to the normative reasons. Further, it seems hardly plausible that such malfunction-
ing constitutes ordinary weakness of the will. By drastically reducing cognitive flexibility,
drug-​related attentional bias creates persistent obstacles to addicts’ decision-​making—​
obstacles requiring a sustained and extraordinary effort to overcome. Even addicts who
exhibit a strong will to give up drugs frequently fail due to the difficulties of maintaining
abstinence. Insofar as they do not seem criticizable for weakness of will, it seems unfair to
blame them for their lack of success.
Second, regarding the supposed opposition between medical models and choice mod-
els, a malfunctioning of the will of the sort proposed here neither rules out intention and
choice in addictive behavior nor compulsion in the clinical sense. That addictive behavior
patterns are sustained by cognitive inflexibility does not mean that addicts cannot choose
to perform drug-​oriented actions on the basis of their beliefs and desires. What causes
their lack of control is the shaping of these beliefs and desires by processes that are com-
pletely dissociated from their propositional attitude system. Even in cases in which they
perform drug-​oriented actions contrary to what they desire, believe, or even judge is best,
their actions can still be guided by intentions and they can retain the capacity to refrain
from performing them. There is no reason, therefore, to assume that addicts do not inten-
tionally engage in addictive behavior.
Finally, drug-​associated attentional bias is underpinned by type-​1 processes it is easy to
imagine could create behavior patterns that exhibit features of compulsivity. The “compul-
sion” often reported by addicts might simply be the subjective experience of cue-​triggered
decision-​making, that is, decision-​making underpinned by associative relations in mem-
ory activated without their conscious intention, deliberation, or even awareness. If this is
correct, what distinguishes compulsive from non-​compulsive actions would not simply be
the causal strength of their motivational antecedents (as is standardly assumed)—​strong
desires, for example, are felt by most normal persons from time to time—​but rather their
frequency, computational speed, cue-​dependence, and dissociated nature, features that
together disrupt the normal functioning of their will and therefore make it extremely
hard to maintain resistance over time.

5 Conclusion
In this chapter I  have argued that a dual-​process approach can explain the seemingly
paradoxical features of addictive behavior, that is, its appearance of control and of non-​
control. It shows how addiction can involve intentional actions that are freely performed,
  131

Conclusion 131

why addicts can make choices, as well as why addictive behavior still counts as, in an
important sense, “compulsive” and “out of control.” It therefore resolves the apparent
conflict between medical and choice models of addiction. Addiction, on this view, is a
varied and multi-​determined behavioral phenomenon. Many different factors—​ranging
from social and psychological factors shaping reasons and choice to biological changes
disrupting the capacity for attentional control—​are likely to affect its manifestations in
individual cases, the difficulties facing the addict wanting to quit, and what are the best
ways of helping her achieve that goal. But if the argument of this chapter is correct, then
every addict can be said to suffer from a malfunctioning of the will brought about by the
regular use of drugs.

References
Ainslie, G. (2001). Breakdown of Will. Cambridge: Cambridge University Press.
Algom. D., Chatjut. E., Lev, S. (2004). A rational look at the emotional Stroop phenomenon: a generic
slowdown, not a Stroop effect. Journal of Experimental Psychology: General, 133, 323–​38.
American Psychiatric Association (APA) (1994). Diagnostic and Statistical Manual of Mental Disorders
(4th ed.). Washington, DC: APA.
Benn, P. (2007). Disease, addiction and the freedom to resist. Philosophical Papers 3, 465–​81.
Bickel, W.K. and Marsch, L.M. (2001). Toward a behavioral economic understanding of drug
dependence: delay discounting processes. Addiction, 96, 73–​86.
Cox, W.M., Fadardi, J.S., and Pothos, E.M. (2006). The addiction-​Stroop test: theoretical considerations
and procedural recommendations. Psychological Bulletin 132(3), 443–​76.
Dill, B. and Holton, R. (2014). The addict in us all. Front. Psychiatry, 09 October.
Field, M. and Cox, W.M. (2008). Attentional bias in addictive behaviors: a review of its development,
causes, and consequences. Drug and Alcohol Dependence, 97, 1–​20.
Fisher, J.M. and Ravizza, M. (1998). Responsibility and Control: An Essay on Moral Responsibility.
Cambridge: Cambridge University Press.
Frankfurt, H. (1971). Freedom of the will and the concept of a person. Journal of Philosophy, 68, 5–​20.
Frankish, K. and Evans, J. St. B.T. (2009). The duality of mind: an historical perspective. In: J. St.
B.T. Evans and K. Frankish (eds), In Two Minds: Dual Processes and Beyond. New York: Oxford
University Press, pp. 1–​29.
Henden, E. (2013). Addictive actions. Philosophical Psychology, 26(3), 362–​82.
Henden, E., Melberg, H.O., and Røgeberg, O.J. (2013). Addiction: Choice or compulsion? Frontiers in
Psychiatry, 4(77), 1–​11.
Heyman, G.M. (2009). Addiction, a Disorder of Choice. Cambridge, MA: Harvard University Press.
Holton, R. (1999). Intention and weakness of will. Journal of Philosophy, 96, 241–​62.
Holton, R., Berridge, K. (2013). Addiction between compulsion and choice. In: N. Levy (ed.), Addiction
and Self-​Control: Perspectives from Philosophy, Psychology, and Neuroscience. New York: Oxford
University Press, pp. 239–​68.
Levy, N. (2006). Autonomy and addiction. Canadian Journal of Philosophy, 36, 427–​48.
Levy, N. (2011a). Resisting “weakness of the will.” Philosophy and Phenomenological Research, 82(1),
134–​55.
Levy, N. (2011b). Addiction, Responsibility, and Ego Depletion. In: J. Poland and G. Graham (eds.),
Addiction and responsibility. Cambridge, MA: The MIT Press, pp. 89–​112.
132  

132 Addiction, compulsion, and weakness of the will

Lowenstein, G. (1999). A visceral account of addiction. In: J. Elster and O.-​J. Skog (eds), Getting
Hooked: Rationality and Addiction. Cambridge, MA: MIT Press, pp. 235–​64.
May, J. and Holton, R. (2012). What in the world is weakness of will? Philosophical Studies, 157(3),
341–​60.
McKay, R and Efferson C. (2010). The subtleties of error management. Evolution and Human Behavior,
5(31), 309–​19.
Mele, A. (2009). Weakness of will and akrasia. Philosophical Studies, 150(3), 391–​404.
Miyake, A., Friedman N.P., Emerson, M.J., et al. (2000). The unity and diversity of executive functions
and their contributions to complex “frontal lobe” tasks: a latent variable analysis. Cognitive
Psychology, 41, 49–​100.
Muraven, M. and Baumeister, R.F. (2000). Self-​regulation and depletion of limited resources: does self-​
control resemble a muscle? Psychological Bulletin, 126(2), 247–​59.
Petry, N.M., Sheila, M., Alessi, S.M., and Rash C.J. (2011). Contingency management treatment of
drug and alcohol use disorders. In: J. Poland and G. Graham (eds), Addiction and Responsibility.
Cambridge, MA: MIT Press, pp. 225–​45.
Pickard, H. (2015). Psychopathology and the Ability to Do Otherwise. Philosophy and
Phenomenological Research, 90(1), 135–​63.
Raz, J. (1997). Incommensurability and agency. In R. Chang (ed.), Incommensurability, Incomparability,
and Practical Reason. Cambridge, MA: Harvard University Press, pp. 110–​28.
Redish, D.A., Jensen, S., and Johnson, A. (2008). A unified framework for addiction: vulnerabilities in
the decision process. Behavioral and Brain Sciences, 31, 415–​87.
Robinson, T.E. and Berridge, K.C. (1993). The neural basis of drug craving: an incentive-​sensitization
theory of addiction. Brain Research Reviews, 18, 252–​55.
Robinson, T.E. and Berridge, K.C. (1998). What is the role of dopamine in reward: hedonic impact,
reward learning, or incentive salience? Brain Research Reviews, 28, 309–​69.
Robinson, M.J.F., Robinson, T.E., and Berridge, K.C. (2013). Incentive salience and the transition to
addiction. Biological Research on Addiction, 2, 391–​99.
Saunders, C. and Over, D.E. (2009). In two minds about rationality. In: J. St. B. T. Evans and K. Frankish
(eds), In Two Minds: Dual Processes and Beyond. New York: Oxford University Press, pp. 317–​34.
Schroeder, T. and Arpaly, N. (2013). Addiction and blameworthiness. In: N. Levy (ed.), Addiction and
Self-​Control: Perspectives from Philosophy, Psychology, and Neuroscience. New York, NY: Oxford
University Press, pp. 214–​38.
Skog, O.-​J. (2003). Addiction: definitions and mechanisms. In: R.E. Vuchinich and N. Heather (eds),
Choice, Behavioral Economics and Addiction. Oxford: Elsevier, pp. 157–​82.
Smith, M. (2003). Rational capacities, or: How to distinguish recklessness, weakness, and compulsion.
In: S. Stroud and C. Tappolet (eds), Weakness of Will and Practical Irrationality. New York: Oxford
University Press, pp. 17–​38.
Stanovich, K.E. (2009). Distinguishing the reflective, algorithmic, and autonomous minds: is it time for
a tri-​process theory? In: J. St. B. T. Evans and K. Frankish, (eds), In Two Minds: Dual Processes and
Beyond. Oxford: New York, pp. 55–​88.
Watson, G. (2004). Skepticism about weakness of the will. Agency and Answerability. New York: Oxford
University Press, pp. 33–​58.
West, R. (2006). Theory of Addiction. Oxford: Blackwell.
  133

Chapter 8

Addiction as a form of akrasia

Nick Heather

Abstract
This chapter argues that addiction can usefully be seen as a form of akrasia,
an ancient philosophical concept referring to acting against one’s better
judgment. Ordinary akrasia and addiction can be located at extremes
on a continuum of frequency of akratic actions. Four criteria essential to
“core akratic action” also apply to addiction—that it must be: (1) free; (2)
intentional; (3) contrary to the agent’s better judgment based on practical
reasoning; (4) consciously recognized as contrary to better judgment at
the time of action. The chapter concludes by briefly describing four lines
of current theory and research on addiction with parallels to philosophical
accounts of akrasia: (1) behavioral economic theories; (2) dual-process
cognitive theory; (3) “willpower” and ego depletion; and (4) neuroscientific
research on deficits to self-regulation. Finding ways these lines of enquiry
can converge is an exciting task for theory and research on addiction.

1 Introduction
“Akrasia” was Aristotle’s word for acting against one’s better judgment (Peiknenburg
1996) or contrary to what one considers it is in one’s best interests to do. It is sometimes
translated as “incontinence,” in the sense of “lacking control over one’s actions.” Another,
more familiar synonym is “weakness of will.” The converse of akrasia is “enkrateia” or
self-​control (continence, strength of will), meaning that behavior is consistent with bet-
ter judgment. This and the preceding terms can sometimes be used to mean a character-
ological trait, referring to a more or less enduring property of someone’s behavior across
a wide range of situations. This may or may not be a valid dimension of character or per-
sonality but it is not what is under discussion here. Rather, this chapter is concerned with
discrete acts or relatively short sequences of behavior that can be described as “akratic,”
the person committing such acts being known as “the akrates.” The aim of this chapter is
to argue that addiction can usefully be seen as a form of akrasia.

2  Ordinary akrasia
Last evening I noticed there was an interesting movie on late-​night TV that I had not
seen but, because I knew that I had to begin work on this chapter the next morning and
134  

134 Addiction as a form of akrasia

wanted to make an early start, I thought it better to miss the film, with the hope I could
catch it on some later occasion. Because I knew that, when the time came, I might be
tempted to stay up and see the film, I made a deliberate decision not to start watching it.
However, around the time I usually go to bed, I did start to watch the movie and quickly
became engrossed in the plot, with the result that I watched it till the end and got to bed
much later than I had intended. I will discuss the requirements for a formal definition of
akrasia later but, for now, I merely point out that my continuing to watch the film with
the conscious awareness that it contradicted my longer-​term and best interests to do so is
an instance of akrasia. There are other aspects of this story that, whether or not they are
thought essential to a definition of akrasia, are at least frequent accompaniments of it: we
can say that I gave in to the temptation to watch the movie when I believed it would have
been better for me not to; or that the following morning I experienced regret at having
given in to this temptation because I felt more tired than I wanted to be. These features
of the episode in question make clear, I suggest, that we are talking here about a common
human experience, one with which all adults and children past infancy (Diaz and Fruhauf
1991) are familiar. Let’s call this ordinary akrasia.
There is another aspect of this story that should be emphasized. Not only did I watch the
film knowingly against my best interests, I also failed to adhere to my prior decision not to
start watching it. As we shall see, whether or not this failure to maintain a prior commit-
ment is essential to a definition of weakness of will, or whether, if it does appear essential,
the phenomenon in question should not be called akrasia, has recently been a matter of
dispute between philosophers. To roughly anticipate a conclusion of this chapter, I believe
that, in order to show that addiction can usefully be viewed as a form of akrasia or weak-
ness of will, a failure to maintain a prior decision and a failure to conform to a presently
held judgment should both be an accompaniment of the addictive behavior. In this way,
addiction might be seen as akrasia but as a special kind of akrasia. For this purpose, there-
fore, the characterization of akrasia with which this chapter began as “acting against one’s
better judgment” should be modified to “knowingly acting against one’s present judgment
and one’s prior resolution.” The grounds for this modification will be explained below.

3  Philosophical accounts of akrasia

While they would presumably have agreed that my behavior in the story above repre-
sented a common human experience, ancient Greek philosophers were not all agreed that
shortcomings of this kind should be understood as acting against one’s better judgment.
Socrates, for example, believed that akratic action was conceptually and psychologically
impossible. According to Plato, he asserted that “No-​one goes willingly toward the bad”
(Plato 1976), i.e. no-​one intentionally acts against her best interests. However oddly
someone behaved, we could be sure that she considered this to be the best course of action
open to her in the prevailing circumstances; if the action appeared to us clearly contrary
to her best interests, it could only be because she was ignorant of what was best or good
to do. This dispute—​whether or not the notion of acting against one’s better judgment
  135

Mele’s definition of akrasia 135

involves an inherent contradiction—​has occupied the minds of philosophers down the

ages ever since (Peiknenburg 1996). An heroic attempt to resolve the dispute was made by
the late American philosopher, Donald Davidson, in an essay originally written in 1969
and entitled, “How is weakness of the will possible?” (Davidson 1980). Davidson set out
to prove that akrasia was not an illogical concept and that, in acting against his better
judgment, the akrates was irrational but not internally illogical. However, Davidson’s par-
ticular defense of akrasia has not been accepted by all philosophers (e.g. Charlton 1988;
Kennett 2001), and it is disputed by some whether a meaningful concept of akrasia can
be formulated (e.g. Watson 1977; Pugmire 1982). To contribute to this dispute is beyond
the scope of the present chapter. It is assumed here that akratic action does exist, that the
concept of akrasia is not self-​contradictory, and that it can be used to refer to an impor-
tant class of irrational human behavior. What is at issue is whether or not it makes sense
to claim that addiction belongs to this class of behavior and is therefore a form of akrasia.

4  Mele’s definition of akrasia

Over the last 30 years, a series of intensive and minute analyses of the concept of akrasia
has been undertaken by the American philosopher, Alfred R. Mele (e.g. Mele 1987, 1992,
2001, 2002, 2012). In the latest of these, published in 2012, he arrives, following some jus-
tificatory arguments, at the following definition of what he calls core akratic action: “free,
sane, intentional action that, as the non-​depressed agent consciously recognizes at the
time of action, is contrary to his better judgment, a judgment that is based on practical
reasoning” (Mele 2012, p.8).
Two aspects of this definition—​that akratic action must be sane and must not be a
result of depression—​are not essential for present purposes and can be quickly dispensed
with. These properties of weak-​willed action, says Mele, are typically taken for granted
but rarely made explicit. Mele (2012, ch. 1)  provides examples of behavior that might
qualify for a definition of akrasia were it not for the fact that they are a consequence of
either insanity or clinical depression.1 Although an argument might be made that these
examples should be seen as forms of akrasia, they do not conform to what Mele wants to
call core akratic action, the focus of the present chapter.
We are left with four definitional requirements of core akratic action. It must be: (1) free;
(2) intentional; (3) contrary to the agent’s better judgment based on practical reasoning;
and (4) consciously recognized as contrary to better judgment at the time of action. These

1 This should not be taken to mean that psychotic or depressed persons are not capable of core akratic
action and therefore, according to my argument here, cannot act addictively. It means only that when
an agent acts against what he believes to be in his best interests as a consequence of some delusional or
clinically depressed motivation, this is not the kind of core akratic action of which addiction is held to
be an extreme example. See the examples of akratic action linked to insanity or depression provided by
Mele (2012, pp. 8–​10). At other times, presumably, psychotic or depressed persons are capable of core
akratic actions and, therefore in principle, addictive behavior. Thanks to Gabriel Segal for pointing out
the need for this clarification.
136  

136 Addiction as a form of akrasia

four requirements will be taken in turn and I  will ask whether addiction, or addictive
behavior, satisfies them and can therefore be seen as within the category of core akratic
action.

5  Three theoretically uninteresting differences

between ordinary akrasia and addiction
Before turning to these requirements, however, there are three obvious but theoretically
uninteresting differences between ordinary akrasia and addiction, differences that do not
undermine the aim of this chapter and need not detain us long. The first is that addiction
represents repeated instances of akratic action. In the example of akrasia with which this
chapter began, my weakness of will in failing to resist the temptation to watch the late-​
night movie caused some inconvenience the following morning but was presumably a
one-​off event or, at least, an infrequent occurrence. If it began to happen regularly and to
impair my ability to get work done the following day, and if I made repeated but unsuc-
cessful attempts to avoid it, then it might conceivably be called an addiction—​in this case
a kind of behavioral addiction to watching late-​night movies on TV. The simple point
is that what we commonly think of as addictive behavior involves persistent episodes
of similar manifestations of akratic behavior. Thus, ordinary akrasia and addiction are
located towards opposite ends of a continuum of frequency of akratic episodes.2
The second difference concerns the distinction, to use biblical terms, between “sins
of commission” and “sins of omission” (Heather 1994). A certain kind of akrasia can
be thought of as a failure to do what one judges it better to do; for example, academ-
ics exhibit akrasia when they delay starting on a large pile of marking they have to do,
putting it off to the last minute, and then having to work late into the night to get it
finished. “There may be cases in which, even though it is intentional on a person’s part
that he does not A, his not A-​ing is not an action and therefore is not an intentional
action” (Mele, 2012, pp. 10–​11), thus infringing one of the requirements for addiction
to be regarded as a form of akrasia to be discussed below. The example Mele gives is
when someone intentionally does not vote in an election despite judging it best that
she should vote. Colloquially, it might be possible to speak coherently of an addic-
tion to laziness or procrastination. For the purposes of the present argument, however,
it would be better to restrict addiction to the commission of behavior that the agent
judges it better not to do.
The third obvious difference between ordinary akrasia and addiction concerns the
agent’s reactions to the contravention of better judgments. In my example of akratic late-​
night TV watching, I experienced regret the following morning, mild discontent, and

2 Edmund Henden would disagree that the repetitiveness of addictive behaviour is theoretically unin-
teresting. This is because he sees a repetitive pattern of behavior as one of the features constituting the
compulsive nature of addictive behavior—​see Henden (2012) and also Chapter 7, this volume. Even if
the kind of compulsivity described by Henden is a valid characterization of addictive behavior, however,
whether it qualifies as a clear qualitative difference from ordinary weakness of will is another question.
  137

Requirements of a concept of addiction as akrasia 137

a degree of self-​criticism, but my reaction could not be described as involving strong

negative emotions. In cases of addictive behavior, however, the typical reaction can be
one of severe depression and self-​loathing, not to say despair. Flanagan (2013) has devel-
oped a twin normative failure model of addiction. One of these failings is similar to
what is being proposed here—​a failure in executing normal powers of rational effective
agency in that the addict “decides not to use and uses” (p. 1). The other is a failure “to
live up to the hopes, expectations, standards, and ideals she has for a good life for herself
because of her addiction” (p. 1). This results in shame, which, for Flanagan, is part of
the normal phenomenology of addiction and often a source of motivation leading to
recovery. The point here is that, while Flanagan eloquently describes the negative emo-
tional consequences for the addict from failing to adhere to her better judgment, these
consequences are different in degree but not in kind from those that follow ordinary
episodes of akrasia.

6  Requirements of a concept of addiction as akrasia

I will begin with the two requirements least problematic for my thesis and end with the
two that cause the most difficulty for it.

6.1  Akratic action is intentional action

The standard model of intentional action in philosophy is that, when a person acts with an
intention, she sets a positive value on some state of affairs, she believes that an action of a
kind open to her to perform will realize the valued state of affairs, and so she acts because
of her value or desire and her belief (Davidson 1980). Her desire and belief together con-
stitute her reason for acting. A simple example is that if I wish to know the time, I believe
that by looking at my watch I will succeed in knowing the time, and so I intentionally look
at my watch. It seems clear that what is normally regarded as addictive behavior—​at least,
consummatory addictive behavior (e.g. injecting a needle into a vein; ordering, paying for
and consuming an alcoholic drink; putting money into a fruit machine and pulling the
handle)—​fits this model of intentional action.
One point to note is that asserting that incontinent action is intentional action does not
entail that the incontinent agent intends to act akratically. Someone eating a large portion
of dessert after deciding to go on a diet may know that he is acting akratically without
intending to act akratically (Mele 1987, p. 5).

6.2  Akratic
action is against one’s better judgment based
on practical reasoning
In many philosophical discussions of akrasia in the literature, the judgment that the
akratic action contravenes is an “all-​things-​considered” judgment (Davidson 1980), i.e.
one that takes into account all facts, beliefs, and values the agent thinks relevant to the
decision to act or not to act in a certain way. Mele (1987, p. 7) speaks of “decisive better
judgment”, meaning judgment that commits the agent to action, and in his later book he
138  

138 Addiction as a form of akrasia

says that “an agent who judges it best to A is thereby rationally committed to A-​ing, in the
sense that (as long as the judgment is retained) the uncompelled, intentional performance
of any action he believes to be incompatible with his A-​ing would open him to the charge
of irrationality” (2012, p. 16, italics original).
Note that the rationality here is subjective rationality, i.e. rational from the agent’s own
point of view. Better judgments may be but are not necessarily moral judgments nor judg-
ments based on prevailing social norms. Further, they are judgments that philosophers
call “declarative propositions”, i.e. propositions that may be either true or false. The practi-
cal reasoning upon which the better judgment is based is simply reasoning about what it
is best to do here and now. It is not a mere academic exercise of reasoning but one that is
tightly connected to motivation and intending to do what it is best to do.
The judgments against which addiction is usually recognized are not different in kind
from those that define akrasia. They are all of the form: given all I know about the harms
(to me, my family, or other people) from my engaging in this behavior (injecting drugs,
drinking heavily, gambling on fruit machines), and given that these harms exceed any
actual or potential benefits to me, it would be better not to engage in this behavior but
rather to desist from it. This is an all-​things-​considered and a decisive better judgment
based on the agent’s own values that could be either true or false but which the agent at
the time believes to be true.

6.3  Akratic
action is consciously recognized by the agent
as contrary to better judgment at the time of action
This has been one of the most controversial issues in philosophical discussions of akra-
sia and is one that is crucial to an understanding of addiction as akrasia. The idea that,
in order to qualify as akratic action, it is necessary that the agent continues knowingly
to judge that it would be better to refrain from the action at the time it is carried out is
referred to by philosophers as “clear-​eyed akrasia” (Setiya 2007). Such a necessary condi-
tion for akrasia has been disputed but forms part of the view of akrasia held by Davidson
(1980) and Mele (2012), two of the leading authorities in this branch of philosophy, and
can be considered part of the standard philosophical model of akrasia or weakness of will.
A recent but already influential criticism of this standard model has come from Richard
Holton (1999, 2009). He first asks what it is that non-​philosophers mean by “weakness of
will” and finds that they make no mention of judgments about a better or worse course
of action but rather think that weak-​willed people are irresolute, do not persist in their
intentions, and are too easily deflected from resolutions they have previously formed.
He therefore concludes that “central cases of weakness of will are best characterized not
as cases in which people act against their better judgment but as cases in which they fail
to act on their intentions” (2009, p. 241). Although in most previous accounts they were
regarded as synonymous, Holton wishes to make a distinction between weakness of will
as described above and akrasia; people can show weakness of will without showing akra-
sia and vice versa, and Holton provides examples of both kinds of case. Put crudely, in
Holton’s version of weakness of will, the agent “changes his mind” about what it is best to
  139

Requirements of a concept of addiction as akrasia 139

do, although this change of mind occurs too easily, is unreasonable, and is not justified
by any change in the prevailing circumstances. He allows that cases conforming to the
traditional model can exist but argues that many cases of weakness of will are captured
by his account; elsewhere, he states his view that the traditional account involving action
contrary to better judgment is plain wrong (Holton 1999, p. 258).
Holton (2009) describes the intention that is abandoned in cases of weakness of will as a
resolution (p. 9). Further, such a resolution is made “in the attempt to overcome contrary
desires that one believes one will have when the time comes to act” (p. 77); a resolution
in this sense is thus a “contrary-​inclination-​defeating intention.” I  believe it is resolu-
tions like this that are implicit in addicts’ attempts to break free from their addiction.
Perhaps the most famous resolutions of this kind known to students of alcohol problems
are the signed pledges against ever again taking intoxicating liquors by adherents of the
nineteenth-​century Temperance Movement (Harrison 1970). The suggestion here is that,
while they are not so formally stated, resolutions made by the addicted individual typi-
cally approach such pledges in their force, their intensity, and their attempt to forestall
future transgressions.
Mele (2012) takes issue with Holton’s account of weakness of will and maintains that,
even in cases that do clearly involve a breakdown of a previous resolution, there is always
some present judgment that is contravened (see also Levy 2011a). Both philosophers use
empirical data to support their case, reporting the results of surveys in which students and
others are asked about their understanding of what weakness of will means. This is not the
place to attempt to resolve this disagreement concerning the empirical evidence bearing
on the nature of akrasia or weakness of will and whether they refer to the same or differ-
ent phenomena. It can be noted in passing, however, that Holton’s account is much more
in tune than Mele’s with behavioral-​economic theories of impulsive behavior (including
addiction). As pointed out by Elster (1999), in Ainslie’s (1992) theory there is a prefer-
ence reversal in impulsive choice such that, while she may previously have judged that
the larger, later reward would be better, at the time she acts the impulsive agent believes
that the smaller, sooner reward is better. The same seems true of Herrnstein and Prelec’s
(1992) and Rachlin’s (2000) theories.
The key issue for present purposes is which of these two accounts is most applicable
to addiction and which can be used to relate addiction to ordinary akrasia/​weakness of
will. Clinical experience of working with addictive disorders (together with my own past
experience of struggling to quit cigarette smoking) leads me to the view that addiction,
in the sense of repeated unsuccessful attempts to radically change behavior, implies both
a breakdown down of a prior resolution and an infringement of an all-​things-​considered
judgment. In other words, the irrationality in addiction is both diachronic and syn-
chronic. In this view, a paradigm case demonstrating the link between addiction and
ordinary akrasia is the dieter who, when offered a slice of chocolate cake says, “I know
I shouldn’t but I will” and then proceeds to take and eat it. As this implies, the dieter has
formed an all-​things-​considered judgment that eating chocolate cake is bad for him and
has resolved at some time in the past not to do so, but, in giving in to the temptation to
140  

140 Addiction as a form of akrasia

accept the offer of cake and eat it, he is aware that he is thereby going against his better
judgment and at the same time breaking his prior resolution. My suggestion is that this
is precisely the situation in which the addict finds himself. Does it seem likely that the
addict, having clearly recognized that the behavior in question is harmful to him and
is one from which he wishes to desist in future, suddenly abandons this insight and the
accompanying resolution and decides that it’s “a jolly good thing” after all?3 Judgment
shifts of this kind no doubt occur for various reasons in cases of ordinary weakness of
will and are interestingly illustrated by Holton (2009), but my suspicion is that they do
not occur in cases of addiction, at least under a characterization of addiction as involving
repeated breakdowns of prior resolutions.4 This is admittedly an empirical issue and it is
an interesting and important question by what means it could be put to test.5
As implied, I do not suggest that all cases of ordinary akrasia or weakness of will are of
the kind I  have described for addiction. There may well be examples characterized only
by a contravention of a presently held judgment without a prior resolution or decision of
any kind. On the other hand, there may be examples characterized only by the breaking of
a prior resolution, as described by Holton in his (2009) book. All that is necessary for my
argument is that there are some cases with both features and I further suggest that it is obvi-
ous there are—​as shown by the familiar example of the failed dieter of the previous para-
graph. It is this kind of akrasia/​weakness of will that, I submit, is continuous with addiction.
As regards terminology, accepting for the moment Holton’s argument that the two
terms can be taken to refer to different phenomena, it is a moot point whether the joint
occurrence of judgment infringement and resolution breakdown should be called akrasia
or weakness of will. If we accept, however, that the addict is aware of going against his bet-
ter judgment at the time of acting, we also accept a key element of Mele’s view of akrasia
and we may as well stick with that term.
A further complication is introduced by the nature of psychoactive substances them-
selves, one of the effects of which may be to persuade the addict that her previous misgiv-
ings about consuming the substance were misplaced and that it is, after all, the best thing

3 Yaffe (2013) argues that addicts “value use over abstention at the moment they choose to use, but
value abstention over use moments before and even moments after” (p. 194). If Yaffe is right about this
momentary reversal at the time of action, it applies to “values” as he defines them, not to judgments.
Gabriel Segal believes that this kind of reversal is indeed possible among “addicts” and includes it in the
“insanity” implied in the Alcoholics Anonymous Step 2. But does this apply to addiction manifested as
repeated failures to change behavior? This is, I suppose, an empirical issue.
4 Robert West (2006) takes a similar position but only in relation to some addictive behavior: “When the
restraint fails, there is often (but not always) no sense of the addict having changed his mind and decid-
ing to engage in the behavior as a positive step; rather the sense is of a failure to exert control followed
by regret and a feeling of having let oneself down” (p. 133). The difference between his position and
mine probably arises from different characterizations of addiction.
5 Neil Levy (2011b) discusses findings from experiments on ego-​depletion that he interprets as evi-
dence supporting a judgment-​shift explanation of breakdowns of resolutions in addiction. See also
Fitzpatrick (2008).
  141

Requirements of a concept of addiction as akrasia 141

to do at the present time. There is some evidence, for example, that alcohol priming has
the effect of changing one’s view of the risks of behavior (Rose et al. 2013). The same might
be an effect of all popular psychoactive substances and even of the experiences arising
from engagement in typical behavioral addictions, like problem gambling. The experi-
ence in question may lead to a “change of mind,” though it is unclear whether this would
satisfy Holton’s conditions for weakness of will. Be that as it may, the contention here is
that an awareness that the addictive activity is not the best thing to do applies to all initial
choices to engage in addictive behavior and to thereby break a resolution to desist from it.
In these times when an increasing number of non-​substance behaviors—​gambling,
work, shopping, exercise, Internet use, various forms of sexual activity, etc. (Rosenberg
and Feder 2014)—​are becoming popularly and scientifically recognized as forms of
addiction, the question arises where to draw the line, how to decide whether a candidate
behavior is usefully seen as an addiction or is merely an unusual and possibly harmful
preference. One simple way is to insist that to qualify as an addiction there must have
been repeated and unsuccessful attempts to change the behavior because of the harm the
individual recognizes it as causing. An attempt to use these ideas as a basis for a defini-
tion of addiction was made elsewhere in this volume (see Chapter 1), but I hope it will be
seen here that a view of addiction as “repeated and continuing failures to refrain from or
radically reduce a specified behavior despite prior resolutions to do so” fits easily with the
notion of addiction as a form of akrasia characterized by the joint occurrence of a break-
down of a prior resolution and an infringement of a current judgment.

6.4  Akratic action is freely chosen

This is perhaps the most difficult and crucial criterion of akrasia that addiction must fulfill
to be considered a form of akrasia. This is because in traditional accounts of addiction,
and in psychiatric accounts in particular, addictive behavior is not seen as freely chosen
but, rather, as compelled. Indeed, that it is an example of compulsive behavior is the defin-
ing characteristic of psychiatric accounts of addiction (Leshner 1997). In like fashion, the
temptation to engage in addictive behavior presented to addicts is said to be irresistible,
and the craving they experience in the face of temptation is said to be overpowering.
It can be argued that, in the past, philosophers have too readily accepted the psychiatric
perspective on addiction and have simply assumed that addiction is compulsive because
psychiatrists and other authorities say it is (Levy 2006). More recently, younger philoso-
phers have begun to dispute this version of addiction and to argue instead that, while
addictive behavior may be characterized by impairments to autonomy and constraints
upon choice, it is freely chosen at the point of action (Levy 2006; Foddy and Savulescu
2010; Pickard 2012). The most obvious way to show that addiction is not best seen as
compulsive behavior is to point to the mountain of evidence showing that addictive
behavior is an operant (Heather and Robertson 1981; Heyman 2009; Hart 2013; Pickard
forthcoming), i.e. in the classical sense of being responsive to changes in environmental
contingencies (Skinner 1953), and is therefore voluntary behavior rather than compelled
and involuntary. Other evidence and arguments could be adduced to demonstrate the
142  

142 Addiction as a form of akrasia

fallacy of believing that addictive behavior is compelled (e.g., Pickard and Pearce 2012;
Pickard forthcoming; Heather, Chapter  25, this volume). Here, however, I  will adapt a
philosophical argument taken from Mele (1987, pp. 22–​29) to make this point.
Let us first consider in full the example that Mele focuses on:
On New Year’s Eve Fred resolved not to eat an after-​dinner snack for the entire month of January.
It is now January 15 and he still has not succumbed to temptation. Tonight, however, Fred finally
gives in. While watching the NBC Sunday night movie, he experiences a desire to eat the piece of
chocolate pie that he has set aside for tomorrow’s dessert. He is not at all surprised by the desire’s
nagging presence since he has had similar desires almost every night for the last two weeks. He has
been able successfully to resist them by rehearsing the reasons for not acting upon them and by
making it evident to himself that these reasons far outweigh any competing reasons that he has.
The reasons are now quite familiar to Fred, and the calculation is easy. Fred weighs the competing
reasons on the basis of their respective merits, judges that it would be best, all things considered,
not to eat the piece of pie, and decides to save the tempting tidbit for tomorrow’s dessert. However,
a short time later, during an advertisement for Michelob Light, Fred walks to his refrigerator for a
beer (which neither Fred nor I count as a snack), spies the pie on the middle shelf, and, still think-
ing it would be best not to eat it, removes it and a container of Dream Whip from the refrigerator.
Fred carefully spreads the whipped cream over the pie, carries the dessert to his seat in front of the
television, and, admitting to himself that, all things considered, he ought not to do what he is about
to do, proceeds to eat the pie.

(Reproduced from Irrationality: An Essay on Akrasia, Self-​deception and Self-​control by

Alfred R. Mele (1987): 275w (p. 22) by permission of Oxford University Press, USA)

Mele uses this example as a test case to ascertain “whether intentional action against a
consciously held better judgment about something to be done here and now can be free”
(p. 22), or, in other words, whether ordinary akratic action can be considered free. I con-
tend, however, that it can also be used to discuss whether, mutatis mutandis, addictive
behavior can be seen to be free. Fred’s behavior, I suggest, is not different in kind from
what is normally considered addictive behavior and, if it were repeated, would certainly
fit a part of my provisional definition above of a failure to refrain from (drug use) despite
prior resolutions to do so. It may differ from typical addictive behavior in the strength
of the “contrary-​inclination-​defeating” resolution that is broken, the frequency of such
breakdowns in the past, the duration of the period of time when it was resolved to resist
temptation, and in other inessential ways, but in terms of nature of the akratic action
itself it is not qualitatively different from someone who has resolved to quit smoking on
New Year’s Day and has failed to keep that resolve, i.e. has relapsed to addictive behavior.
Thus, if we can follow Mele in accepting that Fred’s behavior was freely chosen, we have
the basis for concluding that addictive behavior is freely chosen also.6
The crucial question is whether it was within Fred’s power successfully to resist acting on
the desire to eat the pie. If it was within his power, then the desire did not compel him to
eat the pie and, other things being equal, his eating the pie was a free action. An argument

6 All that is meant by freely chosen here is that the action is not compelled and vice versa. It is possible
to make this distinction without becoming embroiled in the free will–​determinism problem.
  143

Requirements of a concept of addiction as akrasia 143

to support the view that the desire was irresistible, due to Neely (1974), arises from the
proposal that “a desire is irresistible if and only if it is the case that, if the agent had been
presented with what he took to be good and sufficient reason for not acting on it, he would
still have acted on it” (p. 47, italics original) and that someone “is free with respect to some
action which he performed only if it is true that, if he had been presented with what he
took to be a good and sufficient reason for not doing what he did, he would not have done
it” (p. 48). The first of these conditions would presumably be held to apply to addiction by
those who believe that addictive behavior is compulsive, and the second would be held to
apply to behavior that was not compulsive and therefore not an example of addiction.
A more modest version of this argument was advanced by Pugmire (1982), who asserts
that when a person had “put himself through a deliberation that opened him, as much as
anything in his power could, to what he was doing, and … reached a dissuasive all-​things-​
considered value-​judgment, resolved and set himself against what he then did anyway,” he
did all he could have done to resist (p. 189). But “if available resources for resistance failed,
it would be arbitrary to insist that the desire was resistible on the occasion and his action
clearly voluntary” (p. 189). In this way, the desire the agent confronted was “as good as
irresistible” because “it defeated his best efforts” (p.  188). Again, this seems consistent
with the view of addictive behavior as compulsive.
However, the argument assumes that, if an agent’s belief that there is a good and suffi-
cient reason for not acting on a desire fails to generate successful resistance to that desire,
it was not within the agent’s power successfully to resist acting on it. This assumption,
says Mele, is false, because it depends on the premise that there is nothing more the agent
could have done to resist the temptation to act akratically.
Successful resistance is resistance that prevents the agent’s final motivational balance from fall-
ing on the side of incontinent action. And there is more that Pugmire’s agent could have done to
prevent this. An agent can, for example, refuse, at the time of action, to focus his attention on the
attractive aspects of the envisioned akratic action and concentrate instead on what is to be accom-
plished by acting as he judges best. He can attempt to augment his motivation for performing the
action judged best by promising himself a reward … for doing so. He can refuse to entertain second
thoughts about the judgment that he has just very carefully reached. He can practice more sophis-
ticated self-​control techniques prescribed by his behavioral therapist.

(Mele 1987, pp. 23–​24)

In this paragraph, Mele refers to techniques that humans naturally use to resist temptation,
and that are potentially available to all of us and within our power to use. They go beyond
the “brute resistance” (Mele 1987, p. 26) that Neely and Pugmire appear to believe is the
only kind of resistance available to the agent, and include the deliberate efforts to change
the focus of attention and the conditions for self-​reinforcement that Mele mentions. They
clearly reflect the methods reported to be used by pre-​school children to resist tempta-
tion in the famous “delay of gratification” experiments by Walter Mischel and colleagues
(Mischel 2014); the fact that these methods of resisting temptation are apparently available
to young children suggests that they were available to the adult Fred should he have chosen
to use them (although it does not follow from this, of course, that he would necessarily
144  

144 Addiction as a form of akrasia

have used them successfully). Such methods are indeed described by Ainslie (1992) and
Elster (1984) as ways that normal adults often use to avoid “impulsive” behavior in their
longer-​term interests. As Mele implies, explicit and articulated forms of these techniques
formed the basis of the cognitive-​behavioral self-​control therapies that were developed and
introduced into clinical practice in the 1970s (Thoresen and Mahoney 1974; Meichenbaum
1977) and continue to be successfully employed in the treatment of addiction today (e.g.
Carroll et al. 2008). It could be argued that improvements in psychosocial treatments for
addiction depend on the development in clinical research of more sophisticated and effica-
cious methods of self-​control training (Ryan 2013; see also Chapter 25, this volume).
The key contention here is that such methods of resisting temptation, and possibly oth-
ers, were available to Fred had he chosen to use them in an attempt to avoid eating the
chocolate pie. The fact that he did not choose to use them (or chose but did not use them
successfully) is precisely what needs to be explained, but what is inadequate as an expla-
nation is to say that he had no choice in the matter. With regard to addiction per se, and
in the most general terms, the very fact that addicts do often succeed in resisting tempta-
tion by using the kinds of strategies Mele refers to, with or without professional help (and
without pharmacotherapy), is powerful evidence in itself that it is not useful to regard
addictive behavior as compulsive.7
Mele’s argument on the freedom condition of akratic action is far more detailed and
complex than it has been possible to convey here, and the interested reader is referred to
the original (Mele 1987, pp. 22–​29). It may be necessary to add that, in suggesting that
addictive behavior is freely chosen at the time of action and that addicts are therefore in
some way responsible for their actions, there is no accompanying implication that addicts
thereby deserve punishment (Pickard 2012); it is a scientific, not a moral, case that is
being proposed here. Nor should it be inferred that, on this account, the temptations and
cravings that addicts are subject to are any less severe than typically portrayed in the lit-
erature on addiction; though freely chosen and resistible in principle, addictive behavior
can be extremely and painfully difficult to resist. For example, addictive behavior has been
seen as being chosen, in the legal concept, while under duress (Watson 1999), by analogy
with the situation of someone who is threatened with the use of deadly force or grievous
bodily harm to himself or loved ones unless he commits some unwanted act.
Although, as I have tried to show, the idea that addictive behavior is compulsive is mis-
taken, there is certainly a sense in which addicts can experience themselves as being out
of control of their behavior. For example, using Davidson’s (1980) account of akrasia, a
colleague and I have argued that addicts cannot give reasons, to themselves or to others,
for why they have persistently failed to keep to their resolutions to desist from addictive

7 Henden (2012) has proposed a novel way in which addictive behavior may be said to be compulsive.
On this account, compulsive actions do not occur because of irresistible desires and it is not the case
that addicts have lost control over addictive behavior. (See also Chapter 7, this volume.) There will
not be space here to do justice to this highly original argument for compulsivity in addiction and the
interested reader is referred to the relevant sources.
  145

Links with current theory and research on addiction 145

behavior (Heather and Segal 2013, 2015), and that this gives rise to a “subjective sense of not
being able to understand one’s own past behavior and therefore feeling that one must have
been driven by some extrapersonal force to carry it out” (Heather and Segal 2013, p. 451).
The main aim of all reflection and enquiry on addiction is, of course, to explain why
it is that addicts continue to behave in ways they know are harmful to them and which
they sincerely say they do not want. The idea that the answer to this question is that it is
because addicts are compelled to behave the way they do explains nothing and is merely
question begging. Once the idea is accepted that addictive behavior is in some way chosen,
the task for theory and research becomes to discover and describe the nature and deter-
minants of this choice and why it goes against a previous resolution to choose otherwise.
My general thesis here is that the foundation for this task is laid by including addiction
within the wider class of irrational human behavior known as akrasia.

7  Links with current theory and research on addiction

There are several current lines of theory and research that, whether or not this is recog-
nized by the researchers in question, can be viewed as instances of seeing addiction as a
form of akrasia, and these will very briefly be listed here. It will be possible to provide only
a rough sketch of possible links between akrasia as a focus of philosophical enquiry and
the theories in question.

7.1  Behavioral economic theories of addiction

One link arises from a broad theoretical viewpoint that has already been mentioned in
passing above and is called behavioral economic theories or behavioral choice theories
of addiction (Vuchinich and Heather 2003). The most prominent of these are by George
Ainslie (1992, 2001; see also Chapter 13, this volume), Howard Rachlin (2000; see also
Chapter 14, this volume), and Herrnstein and Prelec (1992). The last-​named theory has
been taken forward and developed as an account of addiction by Gene Heyman (2009; see
also Chapter 21, this volume). The parallels between these theories of addiction, which
are all concerned in their own ways with the conflict between immediate or shorter-​
term, smaller rewards and longer-​term, larger rewards, and akrasia, where the conflict is
between incontinent and continent action, are obvious.

7.2  Dual-​process theories

A related line of research by cognitive psychologists is located in the area of dual process
theory or dual systems theory (Evans and Frankish 2009) and the tension between auto-
matic, unconscious (System 1) responses and conscious, controlled (System 2) responses,
recently popularized by Nobel prize-​winner, Daniel Kahneman (2011). The first applica-
tion of dual process theory to addiction was Tiffany’s (1990) cognitive craving model.
This has been followed by a great deal of research on addiction within the framework of
dual-​process theory, especially with regard to attentional bias (e.g. Field and Cox 2008).
A unified framework for understanding addiction based on dual process theory has been
146  

146 Addiction as a form of akrasia

offered by Redish et  al. (2008), and the literature on cognitive processes in substance
use disorders from a dual process perspective has recently been reviewed by Wiers et al.
(2014; see also Chapter 15, this volume).
Regarding the link to akrasia, the idea is that incontinent action is related in some way
to automatic, System 1 processes while continent action is within the sphere of controlled,
System 2 processes, although the precise way in which this distinction is best made is the
focus of much theoretical debate and empirical enquiry. It is interesting, however, that an
attempt by Donald Davidson to provide the basis for an explanation of akrasia (Davidson
1982) postulates the existence of semi-​autonomous partitions of the mind and of non-​
logical causal relationships between elements of these partitions. Again, the parallels with
a dual-​process account of addiction are obvious.

7.3  Ego depletion

A unique line of research with strong resonances to akrasia is the work on self-​regulation,
or “willpower”, by Roy Baumeister and his colleagues (e.g. Baumeister 2003; Gaillot and
Baumeister 2007; see also Chapter  16, this volume). This begins with the assumption,
supported by plenty of experimental evidence, that the capacity for self-​regulation can be
thought of as a limited resource akin to strength or mental energy. When people expend
this energy in efforts at self-​control, the energy resource becomes depleted (known as
“ego depletion”), making further tasks needing self-​control more difficult to accomplish.
It is this phenomenon, in particular, that is thought to be relevant to addiction because the
effort to resist temptation and craving reduces the capacity for further resistance, leading
eventually to relapse. On the other hand, the repeated exercise of successful self-​control
leads over time to the strengthening of the energy resource and an increased ability to
resist temptation, reminiscent of an idea as old as William James’ (1890, p. 213) discussion
of willpower and habit.

7.4  Neuroscience theories of addiction

Lastly, modern neuroscientific research on addiction can be linked to philosophical
accounts of akrasia or weakness of will. Much of neuroscientific research on addiction has
been concerned with how addiction is thought to “hijack” the dopamine reward system
(e.g. Wilson and Kuhn 2005) but there is another strand of research on impairments to
self-​regulation in addiction involving dysfunctions in the prefrontal cortex (e.g. Kalivas
and Volkow 2005; Bechara 2005; Hyman 2007; Goldstein and Volkow 2011). Both these
areas of neuroscientific research are relevant to the general idea that addiction repre-
sents a motivational disorder of the kind that is suggested by the notion of akrasia or
weakness of will (see c­ hapters  9-​12, this volume). The possibilities for the intersection
of neuroscience theories of addiction and philosophical discussions of weakness of will
are illustrated by a recent collaboration between the philosopher Richard Holton and the
neuroscientist Kent Berridge (Holton and Berridge 2013; see also Chapter 9, this volume)
and between philosophers Hanna Pickard and Bennett Foddy and neuroscientist Serge
  147

Summary and conclusions 147

Ahmed (Pickard et al. 2015; see also Chapter 2, this volume). Other philosophers (e.g.
Schroeder 2010; Yaffe 2013) have based their discussions of addiction on neuroscientific
evidence on the role of dopamine in the reward system.

8  Summary and conclusions

In this chapter I have argued that addiction can usefully be seen as a form of akrasia, an
ancient philosophical concept referring to acting against one’s better judgment. I gave an
example of ordinary akrasia, which I suggest is an experience familiar to us all, and then
proposed that addiction, defined as a “repeated and continuing failures to refrain from a
specified behavior despite prior resolutions to do so,” can be located at one extreme of a
continuum of frequency of akratic actions.
An explanation of addiction based on this premise eschews “pathological” processes
that apply only to those with a putative disease of addiction and not to those who do not
suffer from this disease. Thus, one advantage of regarding addiction as an extreme but
continuous form of akrasia is that it becomes possible to relate addiction to explanatory
processes governing normal human behavior. The consequences for methods of treat-
ment and self-​change of addictive behavior are that such methods would be based on the
study of normal processes of self-​control.
I then took four criteria nominated by a leading philosopher of akrasia (Mele) as essen-
tial to “core akratic action” and tried to show that they also applied to addiction. These
criteria were that akrasia (and therefore addictive behavior) must be: (1) free; (2) inten-
tional; (3)  contrary to the agent’s better judgment based on practical reasoning; and
(4) consciously recognized as contrary to better judgment at the time of action. Two of
these criteria presented the greatest difficulties. The insistence that the addict must be
aware of transgressing a better judgment at the time of carrying out the addictive behav-
ior stands in contrast to a recent and influential philosophical view of weakness of will
by Holton (2009) and, possibly, to some behavioral economic theories of addiction (e.g.
Ainslie, 2001). In avoiding this difficulty I argued that addictive behavior shows both a
breakdown of a prior resolution and an infringement of a current judgment.
The criterion that addictive behavior must be freely chosen runs directly counter to
the standard psychiatric view of addiction as compulsive behavior reflecting irresistible
temptation and overpowering desires. This is perhaps the most radical implication of see-
ing addiction as a form of akrasia—​the conclusion that the alleged compulsive nature of
addiction is question begging and unhelpful to scientific enquiry. From the addiction-as-
akrasia viewpoint, the task for theory and research is to discover and describe the nature
and determinants of the processes of choice involved in addiction and exactly in what
ways choice may be compromised, subverted, or “disordered.”
I concluded by very briefly describing four lines of theory and empirical research on
addiction that had parallels with the philosophy of akrasia. These were:  (1)  behavioral
economic theories of addiction; (2)  research on addiction from within a framework
of dual-​process cognitive theory; (3)  research on “willpower” and ego depletion; and
148  

148 Addiction as a form of akrasia

(4)  neuroscientific research on deficits to self-​regulation in addiction. Finding ways in

which these lines of enquiry can converge, together with their philosophical underpin-
nings, is an exciting task for the future of theory and research on addiction.

Acknowledgments
This chapter is based on a paper presented at a meeting of the Kettil Bruun Society in
Turin, Italy on 9 July, 2014. I am grateful to Robin Room for his remarks as Discussant of
my paper and to Edmund Henden and Gabriel Segal for useful comments on an earlier
draft of this chapter.

References
Ainslie, G. (1992). Picoeconomics: The Strategic Interaction of Successive Motivational States within the
Person. Cambridge: Cambridge University Press.
Ainslie, G. (2001). Breakdown of Will. Cambridge: Cambridge University Press.
Baumeister, R.F. (2003). Ego depletion and self-​regulation failure: a resource model of self-​control.
Alcoholism: Clinical & Experimental Research, 27, 1–​4.
Bechara, A. (2005). Decision making, impulse control and loss of willpower to resist drugs: a
neurocognitive perspective. Nature Neuroscience, 8, 1458–​63.
Carroll, K.M., Ball, S.A., Martino, S., et al. (2008). Computer-​assisted delivery of cognitive-​behavioural
therapy for addiction: a randomized trial of CBT4CBT. American Journal of Psychiatry, 165, 881–​88.
Charlton, W. (1988). Davidson and his critics. In: W. Charlton (ed.), Weakness of Will Edinburgh: Basil
Blackwell, pp. 114–​34.
Davidson, D. (1980). How is weakness of the will possible? In: D. Davidson (ed.), Essays on Actions and
Events. Oxford: Clarendon Press, pp. 21–​42.
Davidson, D. (1982). Paradoxes of irrationality. In: R. Wollheim and J. Hopkins (eds), Philosophical
Essays on Freud. Cambridge: Cambridge University Press, pp. 289–​305.
Diaz, R.M. and Fruhauf, A.G. (1991). The origins and development of self-​regulation: a developmental
model on the risk for addictive behaviours. In: N. Heather, W.R. Miller, and J. Greeley (eds), Self-​
control and the Addictive Behaviours. Sydney: Maxwell Macmillan Publishing Australia, pp. 83–​106.
Elster, J. (1984). Ulysses and the Sirens: Studies in Rationality and Irrationality (rev. ed.).
Cambridge: Cambridge University Press.
Elster, J. (1999). Davidson on weakness of will and self-​deception. In: L.E. Hahn (ed.), The Philosophy of
Donald Davidson. Chicago: Open Court Publishing Company, pp. 425–​41.
Evans, J. St. B.T. and Frankish, K. (eds). (2009). In Two Minds: Dual Processes and Beyond.
New York: Oxford University Press.
Field, M. and Cox, W.M. (2008). Attentional bias in addictive behaviours: a review of its development,
causes and consequences. Drug and Alcohol Dependence, 97, 1–​20.
Fitzpatrick, W.J. (2008). Moral reponsibility and normative ignorance. Ethics, 118, 518–​613.
Flanagan, O. (2013). The shame of addiction. Frontiers in Psychiatry, 4, 120.
Foddy, B. and Savulescu, J. (2010). A liberal account of addiction. Philosophy, Psychiatry and
Psychology, 17, 1–​22.
Gaillot, M.T. and Baumeister, R.F. (2007). The physiology of willpower: linking blood glucose to self-​
control. Personality & Social Psychology Review, 11, 303–​27.
Goldstein, R.Z. and Volkow, N.D. (2011). Dysfunction of the prefrontal cortex in addiction: findings
and clinical implications. Nature Reviews Neuroscience, 12, 652–​69.
  149

Summary and conclusions 149

Harrison, B. (1970). Drink and the Victorians. London: Faber.

Hart, C. (2013). High Price: Drugs, Neuroscience and Discovering Myself. London: Penguin.
Heather, N. (1994). Weakness of will: a suitable topic for scientific study? (Editorial). Addiction
Research, 2, 135–​39.
Heather, N. and Robertson, I. (1981). Controlled Drinking. London: Methuen
Heather, N. and Segal, G. (2013). Understanding addiction: Donald Davidson and the problem of
akrasia. Addiction Research and Theory, 21, 445–​52.
Heather, N. and Segal, G. (2015). Is addiction a myth? Donald Davidson’s solution to the problem of
akrasia says not. International Journal of Alcohol and Drug Research, 4, 77–​83.
Henden, E. (2012). Addictive actions. Philosophical Psychology, iFirst, 1–​21, DOI: 10.10-​80/​
09515089.09512012.09660668.
Herrnstein, R. and Prelec, D. (1992). A theory of addiction. In: G. Loewenstein and J. Elster (eds),
Choice Over Time. New York: Russell Sage Foundation, pp. 331–​60.
Heyman, G.M. (2009). Addiction: A Disorder of Choice. Cambridge, MA: Harvard University Press.
Holton, R. (1999). Intention and weakness of will. Journal of Philosophy, 96, 241–​62.
Holton, R. (2009). Willing, Wanting, Waiting. Oxford: Oxford University Press.
Holton, R. and Berridge, K. (2013). Addiction between compulsion and choice. In: N. Levy
(ed.), Addiction and Self-​control: Perspectives from Philosophy, Psychology and Neuroscience.
Oxford: Oxford University Press, pp. 239–​68.
Hyman, S.E. (2007). The neurobiology of addiction: implications for voluntary control of behaviour.
American Journal of Bioethics, 7, 8–​11.
James, W. (1890). Principles of Psychology (vol. 1). New York: Holt.
Kahneman, D. (2011). Thinking, Fast and Slow. London: Penguin Books.
Kalivas, P.W. and Volkow, N.D. (2005). The neural basis of addiction: a pathology of motivation and
choice. American Journal of Psychiatry, 162, 1403–​13.
Kennett, J. (2001). Agency and Responsibility: A Common-​Sense Moral Psychology. Oxford: Oxford
University Press.
Leshner, A.I. (1997). Addiction is a brain disease, and it matters. Science, 278, 45–​47.
Levy, N. (2006). Addiction and autonomy. Canadian Journal of Philosophy, 36, 427–​47.
Levy, N. (2011a). Resisting “weakness of will.” Philosophy and Phenomenological Research, 82, 134–​55.
Levy, N. (2011b). Addiction, responsibility, and ego depletion. In: J. Poland and G. Graham (eds),
Addiction and Responsibility. Cambridge, MA: MIT Press, pp. 89–​112.
Meichenbaum, D.H. (1977). Cognitive-​Behavioural Modification. New York: Plenum.
Mele, A.R. (1987). Irrationality: An Essay on Akrasia, Self-​Deception, and Self-​Control. Oxford: Oxford
University Press.
Mele, A.R. (1992). Akrasia, self-​control and second-​order desires. Nous, 26, 281–​302.
Mele, A.R. (2001). Automomous Agents: From Self-​Control to Autonomy. Oxford: Oxford
University Press.
Mele, A.R. (2002). Akratics and addicts. American Philosophical Quarterly, 39, 153–​67.
Mele, A.R. (2012). Backsliding: Understanding Weakness of Will. Oxford: Oxford University Press.
Mischel, W. (2014). The Marshmallow Test. Place: Publisher.
Neely, W. (1974). Freedom and desire. Philosophical Review, 83, 32–​54.
Peiknenburg, J. (1996). Acting Against One’s Best Judgement: An Enquiry into Practical Reasoning,
Dispositions and Weakness of Will. Unpublished Doctoral Thesis. Groningen: Rijksuniversiteit.
Pickard, H. (2012). Responsibility without blame: philosophical reflections on clinical practice.
In: B. Fulford, R. Gipps, and J. Sadler (eds), Oxford Handbook of Philosophy of Psychiatry. Oxford:
Oxford University Press, pp. 1134–​54.
150  

150 Addiction as a form of akrasia

Pickard, H. (forthcoming). Addiction. In: N. Levy, M. Griffith, and K. Timpe (eds), Routledge Handbook
of Free Will. London: Routledge.
Pickard, H. and Pearce, S. (2012). Addiction in context: philosophical lessons from a personality
disorder clinic. In: N. Levy (ed.), Addiction and Self-​Control: Perspectives from Philosophy, Psychology
and Neuroscience. New York: Oxford University Press, pp. 165–​89.
Pickard, H., Ahmed, S.H., and Foddy, B. (2015). Alternative models of addiction. Frontiers in
Psychiatry, 6, doi: 10.3389/​fpsyt.2015.00020.
Plato (1976). Protagoras (358d). Trans. C.C.W. Taylor. Oxford: Clarendon Press.
Pugmire, D. (1982). Motivated irrationality. Proceedings of the Aristotelian Society, 56, 179–​96.
Rachlin, H. (2000). The Science of Self-​Control. Cambridge MA: Harvard University Press.
Redish, A.D., Jensen, S., and Johnson, A. (2008). A unified framework for addiction: vulnerabilities in
the decision process. Behavioral and Brain Sciences, 31, 415–​87.
Rose, A.K., Jones, A., Clarke, N., and Christiansen, P. (2013). Alcohol-​induced risk taking
on the BART mediates alcohol priming. Psychopharmacology, Open access, doi: 10.1007/​
s00213-​00013-​03377-​00211.
Rosenberg, K.P. and Feder, L.C. (2014). Behavioral Addictions: Criteria, Evidence and Treatment.
London: Academic Press.
Ryan, F. (2013). Cognitive Therapy for Addiction: Motivation and Change. Chichester: Wiley-​Blackwell.
Schroeder, T. (2010). Irrational action and addiction. In: D. Ross, H. Kincaid, D. Spurrett, and P. Collins
(eds), What is Addiction? Cambridge, MA: MIT Press, 391–​407.
Setiya, K. (2007). Reasons Without Rationalism. Princeton, NJ: Princeton University Press.
Skinner, B.F. (1953). Science and Human Behavior. New York: Macmillan.
Thoresen, C.E. and Mahoney, M.J. (1974). Behavioral Self-​Control. New York: Holt, Rinehart and
Winston.
Tiffany, S.T. (1990). A cognitive model of drug urges and drug use behaviour: role of automatic and
nonautomatic processes. Psychological Review, 97, 147–​68.
Vuchinich, R.E. and Heather, N. (eds) (2003). Choice, Behavioral Economics and Addiction.
Oxford: Elsevier Science.
Watson, G. (1977). Skepticism about weakness of will. Philosophical Review, 86, 316–​39.
Watson, G. (1999). Excusing addiction. Law and Philosophy, 18, 589–​619.
West, R. (2006). Theory of Addiction. Oxford: Blackwell.
Wiers, R.W., Field, M., and Stacy, A.W. (2014). Passion’s slave? Conscious and unconscious processes
in alcohol and drug abuse. In: K.J. Sher (ed.), Oxford Handbook of Substance Use Disorders.
Oxford: Oxford University Press; doi: 10.1093/​oxfordhb/​9780199381678.013.009.
Wilson, W.A. and Kuhn, C.M. (2005). How addiction hijacks our reward system. Cerebrum: The Dana
Forum on Brain Science, 7, 53–​66.
Yaffe, G. (2013). Are addicts akratic? Interpreting the neuroscience of reward. In: N. Levy
(ed.), Addiction and Self-​Control: Perspectives from Philosophy, Psychology and Neuroscience.
New York: Oxford University Press, pp. 190–​213.
  151

Section III

Perspectives
from neuroscience
152  
  153

Chapter 9

Compulsion and choice in addiction

Richard Holton
Kent Berridge

Abstract
Most accounts of addiction either see it as a disease, and hence outside the
addict’s control, or else as involving chosen behavior, and hence within the
addict’s control. Contrary conclusions about responsibility are held to follow.
We argue that this involves a false dichotomy. There is good evidence
that addiction does involve a pathological decoupling of wanting and
liking—addicts want—indeed, crave—their drugs without necessarily liking
them. However, whilst cravings are outside addicts’ control, this does not
mean that they have no control over whether they act on those cravings.
We draw on findings about self-control to argue that self-control in these
circumstances is possible but hard; and we consequently argue for a graded
notion of responsibility.

1 Introduction
Despite a wealth of recent empirical findings, the debate on addiction remains polar-
ized along traditional lines.1 In one camp stand those who see the characteristic actions
of the addict as driven by something very much like a disease: a pathologically intense
compulsion that they can do nothing to resist. In a book published in 1812, Benjamin
Rush quoted an alcoholic giving powerful expression to this approach: “Were a keg of
rum in one corner of a room, and were a cannon constantly discharging balls between
me and it, I could not refrain from passing before that cannon, in order to get at the
rum” (Rush 1812, p. 266). The example impressed Williams James, who quoted it and
continued with the story of an alcoholic who supposedly chopped his hand off in order
to be given brandy (James 1890, vol. II, p.  543). The understanding found its way
into literature. Saying that he was “following what psychologists tell us,” Oscar Wilde
described the lure of opium on Dorian Gray:  “Men and women at such moments
lose the freedom of their will. They move to their terrible end as automatons move.

1 This is a shortened and simplified version of (Holton and Berridge 2013). In addition to the people we
thank there, we would like to thank the editors of the current volume for many helpful suggestions.
154  

154 Compulsion and choice in addiction

Choice is taken from them, and conscience is either killed, or, if it lives at all, lives
but to give rebellion its fascination and disobedience its charm” (Wilde 1891, ch. 16).
Modern expressions tend to be less dramatic, but the basic conception remains much
the same. Many contemporary theorists insist that the addict is in the grip of a brain
disease that removes control over their actions and so requires treatment rather than
condemnation.
In the opposing camp stand those who see addictive behavior as involving ordi-
nary choices, and so as something that takes place within the domain of ordinary
intentional action. This approach sees an addict’s decision to take a drug as moti-
vated by a standard structure of beliefs and desires. It harks back to an earlier under-
standing that saw alcoholics as simply overfond of their drink, but in recent years
it has received new impetus in the hands of certain economists and behavioral
psychologists.2
Members of the disease camp point to the extraordinarily self-​destructive behavior that
addicts exhibit, and to the burgeoning literature that suggests that their brains are func-
tioning in abnormal ways. Members of the ordinary choice camp point to findings that
show that addicts often respond to incentives in normal ways. For example, most suc-
ceed in getting over their addictions by their mid-​30s, often with minimal help.3 Further,
many addicts beyond that age stop taking drugs if the incentives are great enough and
clear enough. Anesthesiologists and airline pilots who, having been once detected in their
addiction, are required to pass random and frequent drug tests on pain of dismissal, are
remarkably good at giving up.
The two approaches are typically seen as quite incompatible. If addiction is a brain dis-
ease, then there is no role for willpower or self-​control. To take a representative example,
the book from a recent television series lists as one of the “seven myths of addiction” the
idea that “addiction is a willpower problem,” and goes on to say:
This is an old belief, probably based upon wanting to blame addicts for using drugs to excess. This
myth is reinforced by the observation that most treatments for alcoholism and addiction are behav-
ioral (talk) therapies, which are perceived to build self-​control. But addiction occurs in an area of
the brain called the mesolimbic dopamine system that is not under conscious control.

(Hoffman and Froeke 2007, p. 37)

We agree that the mesolimbic dopamine system is centrally involved in addiction, and
that the workings of that system do not appear to be under direct conscious control (in
the sense that there doesn’t seem to be much that one can deliberately do to directly
affect it). But it is one thing to say that people cannot control their mesolimbic dopamine
system; quite another to say that they cannot control how it influences their actions. In a

2 The idea that addiction involved some kind of compulsion doesn’t really take hold until the eighteenth
century. For discussion see (Sournia 1990).
3 This point is made very forcefully in (Heyman 2009, ch. 4). He draws his conclusion from examination
of national population surveys—​not just surveys of addicts. He argues that most of those who remain
addicted do so because they suffer from other psychiatric illnesses.
  155

Introduction 155

parallel way, there isn’t much that people can deliberately do to influence their perceptual
systems, but that doesn’t mean that there is nothing they can do to control the effects of
what they perceive on their actions.
Our contention here is that, once they are shorn of their claims to the whole truth, the
disease approach and the choice approach are not so incompatible as has been made out.
Our aim is to present the outlines of a middle path. The findings from brain science are
solid enough. There is good evidence that the brain of an addict is importantly differ-
ent from that of a normal non-​addicted individual—​indeed, there is even some reason
to think that the addict’s brain might have started out with a vulnerability to addiction.
Certainly, once addiction is under way, the desire for the addictive drug (amongst which
we include alcohol and nicotine) takes on a life of its own, with an intensity that is particu-
larly, perhaps uniquely, high. The desire becomes insulated from factors that, in normal
intentional behavior, would undermine it, and so persists even when the addict knows
that acting on it would be highly damaging. The addict may recognize that taking the
drug again will incur the loss of family, friends, job, and most that makes life worth living,
and yet still continue to take it. More surprisingly, addicts need not even like the thing
that they are addicted to: they need gain no pleasure from it, nor anticipate that they will.
Nor need they be motivated by a desire to avoid the horrors of withdrawal. Alcohol or
heroin addicts often relapse long after withdrawal is over, and cocaine addiction is no
less potent for having a relatively mild withdrawal syndrome. Addicts may relapse when
they see nothing good in their drug whatsoever. They may see it as nasty, damaging, and
worthless in every respect. Yet they may still want it, and want it, moreover, in a particu-
larly immediate and intense way—​perhaps more immediately and more intensely than
most other people ever experience.
Nevertheless, the intensity and power of an addictive desire does not mean that addicts
are automata, powerless spectators moved by their desires. For whilst addictive desires
are very strong, the human capacity for self-​control is also highly developed—​much more
developed, it seems, than in rats. Addicts do not actually cross into the paths of cannon-
balls or their equivalents, despite William James’ colorful assertion. They go around or
wait for a lull. Smokers on airplanes postpone their urge to smoke until the flight is over.
As these cases show, addictive urges can be controlled, at least for a short while, and
sometimes for longer if the stakes are high enough and clear enough. The experience of
self-​control that everyone has at certain moments is a veridical one: self-​control is a real
phenomenon, something that can be used to control acting on addictive desires, even if
at a considerable cost and, for most addicts, subject to occasional failure. We should thus
not be thinking of addictive desires as things that are impossible to resist, but as things
that are very difficult to resist.4 Our moral evaluations should reflect this fact, and our
scientific account of addiction should explain why resistance is difficult and why failure
happens on the occasions it does.

4 Could they sometimes be truly irresistible? It seems rash to rule that out, although it is hard to be sure
quite what the claim means: That no incentive would overcome it? That no incentive could?
156  

156 Compulsion and choice in addiction

Our aim here is to articulate such a model, one that explains why addictive desires have
the distinctive features they have, but that also explains how they can be controlled. The
account has two parts. The first explains what goes wrong in addiction, making the case
for the incentive salience approach, which sees addiction as primarily a disorder of the
wanting system. The second explains why, nevertheless, addicts remain agents. We start
by outlining what we think is wrong with the pure choice model.

2  Pure choice models
We cannot hope to survey all of the different pure choice models here, but some brief com-
ments will serve to show why we think that they cannot provide a complete explanation of
addiction. A choice model can, of course, easily explain the behavior of those who willingly
and knowingly take addictive drugs. But addicts frequently say that they have been some-
how captured by the addiction—​that they wish that they could escape it but that something
is making it very hard for them to do so. Some listeners might dismiss these comments as
disingenuous or self-​deceived, but we think there is something in what they say.
How can choice models make sense of this capture? They have two approaches. One is
to ascribe to addicts abnormal desires; the other is to ascribe to them mistaken beliefs.
Advocates of the first approach typically see addicts as having steep temporal discount
curves—​they see them as having much stronger desires for the present and immediate
future than for the more distant future. Since addictive drugs normally involve a nasty
period of withdrawal, an already addicted agent whose focus is on the immediate future
will want to avoid embarking on the suffering that such a process will involve, even if
they know that the long-​term effects will be beneficial. Of course they might well prefer
not to have started consuming the drug in the first place—​in this way advocates of this
approach can make sense of the idea that they are really addicted and are not simply will-
ing consumers—​but given the state that they are in now, continuing to take the drugs is
preferable to withdrawal.5
Advocates of the second approach typically see addicts as mistaken, at least initially,
about the effects of their drugs (they believe that they will not become addicted, or that
addiction will not be so bad); or they see them as failing to take into account the conse-
quences of current consumption for their future state: by focusing only on their current
options, addicts fail to see that consuming addictive drugs now will lower their overall
wellbeing in the future.6
The two approaches may be combined: mistaken beliefs might explain why addicts fall
into addiction, and then the steep discount curves might explain why they stay there; and

5 The most influential presentation of this line is from Becker and Murphy; for a simplified presentation
see (Skog 1999). Becker and Murphy give no explanation of how addicts get into the state of addiction;
that is left to be explained by exogenous factors.
6 See Loewenstein (1999) and Herrnstein and Prelec (1992) for versions of the first approach; and
Heyman (2009, ch. 6) for a detailed development of the second. An addict, understood on Heyman’s
lines as one who fails to think about their future, will be behaviorally equivalent to the steep discounter
who doesn’t care about it; but this will derive from features of their beliefs rather than of their desires.
  157

Disease models 157

elements from these approaches might be used to supplement other accounts. Indeed, we
ourselves are inclined to think that there are important insights to be had here. In particu-
lar, there is good evidence that ignorance has an important role in the process of acquiring
an addiction. But we do not think that an ordinary choice account can provide the fun-
damental explanation of what is distinctive about addiction. For if they were right, then a
former addict who had been through the pains of withdrawal should be the least likely to
consume again. They would no longer have the cost of withdrawal to endure; and they, of
all people, would be well informed of their own vulnerability to addiction, of how nasty it
is, and of the cost of not looking to the future. We are not talking here of the person who
really prefers to be addicted; they will just start consuming again, although such a person
would be unlikely to have put themselves through the process of withdrawal in the first
place. But the person who genuinely wanted to be free of the drugs should be uniquely
well qualified to ensure that they remain so.
Yet that is not what we find. People who have come through withdrawal, and gained
much self-​knowledge in the process, are much more likely to take up drugs again than
those who never started, a process that is typically triggered by cues that are associated
with the previous addiction. Indeed, withdrawal seems largely irrelevant in the process of
maintaining addiction. Not only do people consume again after having gone through it,
cravings are experienced long before it comes in, and some highly addictive drugs—​most
notably, cocaine—​have minimal withdrawal symptoms. A pure choice model struggles to
explain these features. So let us turn to the disease models that do better with them.

3 Disease models
There are many disease models of addiction. To get some traction on the debate, we divide
these into four, at the cost of some simplification. The first sees addiction as habit: drug-​
taking actions are triggered automatically in particular situations, independently of the
subject’s beliefs and desires. The second sees it as involving distorted pleasure:  addic-
tive drugs “hijack” the subject’s pleasure circuits, or conversely, as a desperate attempt to
escape the intense suffering of withdrawal states, distressing life circumstances, and/​or
continuing existential anguish. The third, using reinforcement learning theory, sees the
distortion as affecting not the pleasure itself, but the subject’s beliefs about what will give
them pleasure. The final account, which we endorse (whilst denying that this provides the
whole story about addiction), involves desire: consumption of addictive drugs gives rise
to pathologically intense desires or cravings, states that are largely insulated from the sub-
ject’s beliefs and other desires. We start by briefly outlining those with which we disagree.

3.1  Habit accounts

In its simplest form the habit model follows the classic stimulus-​response account laid
down in the early twentieth century by Thorndike. In its early behaviorist guise, this
approach was linked with skepticism about mental states altogether; but such an approach
has few supporters now, and we shall say nothing about it. More interesting is the idea
that habits stand alongside, but independent of, the agent’s beliefs and desires (Wise 2004;
158  

158 Compulsion and choice in addiction

Everitt et al. 2008). Contemporary versions of the theory hold that drugs induce brain
systems of action to perform particular behaviors when cued, behaviors that have been
established by previous drug-​taking episodes—​much like a shoe-​tying habit but even
more strongly automatic.
If addictive states were understood this way, this would provide some explanation of
why they are insensitive to addicts’ desires to stop. But the account implies that drug taking
is unmotivated, and most likely to surface when the addict’s attention is distracted. That
belies the intensely motivated nature of addictive urges, and the observation that attentively
thinking about drugs, rather than being distracted, is the most dangerous situation for an
addict. Whilst some aspects of habitual behavior might be important—​reaching unthink-
ingly for a cigarette—​the account cannot easily explain why an agent will take drugs in full
awareness of what they are doing, but quite contrary to their views of what is best.7

3.2  Pleasure accounts

So let us move the second class of accounts, those premised on excessive pleasure.8
Clearly many addicts do get great pleasure from the drugs they take. If drugs can “hijack”
the pleasure circuit, giving a disproportionate amount of pleasure to those who consume
them, then this would give rise to a very strong learned desire for them.9 And if the pleas-
ure per unit decreased over time, as tolerance developed, the agent would want more and
more of the drugs to compensate (Wise 1980, 1985).
This account was once thought to be bolstered by the finding that the addictive drugs
boost the mesolimbic dopamine system: either by stimulating the production of dopa-
mine (in the case of amphetamine, nicotine, caffeine); by reducing the production of sub-
stances like GABA that themselves reduce the amount of dopamine released (opiates,
perhaps THC); by reducing the level of substances that break down dopamine (alcohol);
or by reducing the activity of the system that reabsorbs dopamine (cocaine, and perhaps
amphetamine). Add the premise that the mesolimbic dopamine system is the pleasure
system, and we have what looks like a compelling picture.10

7 Forwork on the areas in which pure habit accounts do provide good explanations see Wood and Neal
(2007).
8 Thorndike’s original account of learning was in terms of pleasure, though he later came to talk purely
in terms of stimulus and response. Historically then, pleasure-​based accounts represent something of
a reversion to an earlier idea.
9 We speak of “hijacking” and “disproportionate pleasure” here, but accounts that think that there is no
rational constraint on what gives one pleasure will find it hard to make sense of this. To that extent, this
approach will lapse back into a choice account, in which the agent acts on desires for their strongest
pleasure.
10 For a recent popular presentation of such an approach by a neuroscientist, see Linden (2011, ch. 2).
Linden writes “Addictive drugs, by co-​opting the pleasure circuitry and activating it more strongly than
any natural reward, create deeply ingrained memories that are bound up in a network of associations”
(p. 53).
  159

Disease models 159

A converse “withdrawal” version of the pleasure account focuses on addictive escape

from the displeasure of withdrawal, or of other distressing conditions of life, which, it is
posited, lead the addict to pursue drugs to regain an acceptable level of pleasure. This view
is supported by findings that the brains of addicts often have lower levels of dopamine
receptors available (specifically the D2 type of dopamine receptors that can be measured
by current neuroimaging techniques). Repeated taking of drugs induces “downregula-
tion,” a neuronal compensation in which neurons reduce their production of the D2
dopamine receptors in response to receptor overstimulation by those drugs (Volkow et al.
2004). Further, it has been suggested that some addiction-​prone individuals have lower
levels of D2 dopamine receptors to begin with—​possibly associated with an innate ten-
dency to have higher levels of dopamine responses to reward-​related events, which might
play the same role as drugs in providing overstimulation, and consequent downregula-
tion, of their D2 receptors.
But the essential premise that brain dopamine is a mechanism of pleasure is false, in our
view. A host of findings have now shown conclusively that the primary role of the dopa-
mine system is not to do with pleasure. In rats, suppressing the dopamine system does
not result in a lack of pleasure responses to sweet substances. Likewise, human subjects
whose dopamine systems are suppressed artificially, or as a result of Parkinson’s disease,
give normal pleasure ratings to sugar. Conversely, elevated dopamine levels in either rats
or human beings do not result in greater pleasure (Berridge 2012, p. 1132).
Dopamine thus does not seem to be directly concerned with liking. We will suggest that
it is concerned with the creation of wanting. This might not matter if there were none-
theless a very tight correlation between liking and wanting: if liking invariably resulted
in wanting, and if wanting were invariably the result of prior liking. But the very results
that show that they are distinct states also show that, whilst they might typically be linked
by causal connections, sometimes those connections will fail. We will argue that this is
crucial for understanding addiction.

3.3  Learning accounts

So let us move to those models that see addiction as resulting from learning. Admittedly,
in a simple behaviorist model learning is not a very substantial notion: there isn’t much
more to it than the idea that a subject’s behavior changes as a result of what happens to
them, and hardly anyone could disagree that that is true of addiction. But in more cogni-
tivist models, the idea of learning is much more specific: it is the idea of forming predic-
tive associations, that is, beliefs.
These accounts see addiction as stemming, not from heightened pleasure itself, but
from mistaken beliefs about pleasure. Addictive drugs hijack the circuits that learn about
pleasure, and so they distort the memories that are used to guide future desires. One pop-
ular theory of reward learning holds that dopamine spikes indicate “reward prediction
errors”: dopamine is released whenever an outcome is better than expected (Sutton and
Barto 1998). Applied to addiction, the idea is that dopamine-​stimulating drugs cause an
exaggerated prediction error (Schultz et al. 1997; Redish 2004). Consumption of the drug
160  

160 Compulsion and choice in addiction

itself doesn’t have to be especially pleasurable, since the effect on the dopamine system is to
trigger a large prediction error as if it were pleasurable, with the result that the “memory”
of the pleasure greatly exceeds any actual pleasure. On this approach, addicts’ fundamental
desires are desires for pleasure. Since they mistakenly believe that consumption of the drug
will give them pleasure, this results in a strong instrumental desire to consume.
We think that this is mistaken. We will present instead a model—​the incentive sali-
ence model—​that sees addiction as driven by desires that have no essential connection
to beliefs about what will be liked, or about what will be beneficial in other ways. The key
idea here is that the dopamine signals are not learning signals, in the sense that they do
not give rise to beliefs, predictions, or memories (real or apparent) at all. Instead, they give
rise to desires directly—​or, more accurately, to a sensitivity to experience desires when
cued with appropriate stimuli. The desire felt is not an instrumental desire, driven by an
intrinsic desire for pleasure; instead, it is an intrinsic desire for the drug, a desire that may
lead to action even in the face of contrary desires, and in the face of beliefs that consump-
tion will have terrible consequences.
Before we explain the evidence for such an account in any detail, let us get clearer on
the distinctions we have just outlined: that between wanting and liking; and that between
the formation of beliefs and the acquisition of desires.

4  Distinguishing wanting and liking

In one sense it is obvious that wanting and liking are distinct, at least if we think of liking
in terms of pleasure: wanting typically comes before one gets the thing wanted, whereas
the pleasure typically (though not invariably) comes once one has got it. Liking and
wanting can also come apart as a result of false beliefs. We can want something that we
believe we will like, even though we won’t in fact like it: perhaps we haven’t tried it before,
or have forgotten that we didn’t like it, or we believe for some reason that our reaction
will be different to last time. For parallel reasons, we can like something and not want
it: we might not realize that we like it, or we might have other reasons for forgoing it.
So the real issues do not concern the identity of wanting and liking. Instead, we think
that they are two-​fold. One concerns the causal relations between wanting and liking, and
their embodiment in particular brain mechanisms. The second concerns the relation of
wanting to expected liking. We take these in turn.

4.1  Causal relations between wanting and liking

Does liking invariably cause wanting? Conversely, is wanting always caused by liking? We
answer “no” to both questions. While we accept that brain activations that cause increases
in liking typically cause increases in wanting too, we think that these mechanisms are
separable, so that under the right conditions liking can be generated without wanting,
and wanting without liking.
The evidence here came originally from studies of rats. Since rats can’t talk we need
to have some non-​verbal behavioral indicators of wanting and of liking. Wanting is
  161

Distinguishing wanting and liking 161

straightforward: rats want something if they try to get it. (This is where we assume that
issues of self-​control will not intrude; things are more complicated with human beings as
we shall see later.) Liking is harder to identify. But a set of results indicates that a range of
evolved facial expressions—​including tongue protrusions and lip sucking—​are correlated
with liking for the sensory pleasure of tastes across a wide range of species including rats,
monkeys, and human infants (Berridge 2000; Berridge and Kringelbach 2008).
Once we have distinct criteria for wanting and liking, we find that one can be induced
without the other. If rats’ dopamine levels are suppressed, they are no longer prepared to
work to gain food rewards that they would previously have worked for. At the extreme, they
will not eat pleasant foods that are freely available, even though they still display strong
liking for them once the foods are placed in their mouths. Indeed, rats who have 98% of
the dopamine neurons in their nucleus accumbens and neostriatum chemically destroyed,
and who would starve to death were they not intragastrically fed, nevertheless maintain
their normal liking reactions, indicating that pleasure in food is unchanged. So liking is
not sufficient for wanting. Conversely, by boosting rats’ dopamine levels we find that their
wanting can be increased without their liking being increased—​we will discuss an example
of this shortly (Berridge 2007; Smith et al. 2011). Indeed, wanting can be artificially engen-
dered in rats without any signs of liking at all (Peciña et al. 2003; Wyvell and Berridge 2000;
Faure et al. 2010; Smith et al. 2011; Tindell et al. 2005; Berridge and Valenstein 1991). In
the past decade, the distinction between liking and wanting has also been confirmed in a
number of human studies based on ratings of their own experience of sensory pleasures,
such as cocaine and other addictive drugs (Leyton 2010; Lawrence et al. 2003).

4.2  Relation of wanting to expected liking

The second issue concerns the relation of wanting to expected liking. Can subjects want
something whilst believing they will not like it? And conversely, can they believe that they
will like something and not want it? This is where the talk of learning fits in: can subjects
come to learn that they like something, and yet not go on to form a desire for it?
Here again the empirical evidence suggests that wanting without expected liking is
indeed possible. Consider first a set of experiments done by Cindy Wyvell, which showed
that levels of wanting could be manipulated even though the associated levels of liking
or expected liking remained constant. The experiment worked in two stages. The first
got rats to associate a random stimulus (a noise) and an activity (lever pressing) by pair-
ing each with a sugar reward. As a result of this pairing, the rats acquired a conditioned
response: they came to press the lever when they heard the noise, even if no sugar was
present.11 Their facial responses showed that they liked the sugar when they did get it.
The second stage involved probing the effect of changes in dopamine level on this behav-
ior. To achieve this, cannulas were inserted into the rats’ brains, enabling their mesolimbic
dopamine systems to be affected directly by microinjection. A  control group received

11 See Robinson and Berridge (2003, pp. 41–​43) for further discussion of why this feature is important.
162  

162 Compulsion and choice in addiction

an inert substance through this cannula, whilst the other group received amphetamines,
which greatly increase dopamine release. The effects of the dopamine could then be deter-
mined by observing the differences between the two groups.
Both groups continued to like the sugar. Indeed, they liked it to the same degree,
evidence that dopamine does not produce pleasure. Importantly though, the ampheta-
mines did not seem to increase anticipated pleasure from the sugar either: when given
the lever to press, the amphetamine group did not press it any more frequently than the
control group.
The difference came when the rats heard the noise that they had been conditioned
to associate with sugar. Now both groups increased their lever pressing. But rats in the
amphetamine group pressed it dramatically more: more than four times as frequently as
before, and more than 50% more than the rats in the no-​amphetamine group. And this
effect was switched on and off as the noise went on and off.12
It appears that the increased dopamine levels resulted in a massive amplification of
the conditioned response that was already present. It is very hard to explain this result
in terms of changes in expectation, for we have no reason to think that hearing the noise
caused change in the rats’ beliefs about how pleasurable the sugar would be. The rats were
not learning anything new; and the effects fell off as soon as the noise ceased.
There are many other experiments that confirm that learning, which is what would be
involved in expected liking, is different from wanting. For example, the rats discussed
above who have lost nearly all of their mesolimbic dopamine due to neurochemical 6-​
OHDA lesions are still quite capable of learning new values about food rewards. When a
previously liked food is made unpalatable by inducing nausea, the dopamine depleted rats
will learn to react to it with signs of disgust, in just the same way as normal rats. Similarly,
mice who have been genetically engineered to lack dopamine are still able to learn basic
Pavlovian reward associations (Berridge 2012, pp. 1139–​40). Learning, in the sense of the
formation of new beliefs or of the formation of new behavior, does not seem to be essen-
tially dependent on the dopamine system.
So what exactly is dopamine doing? As we saw from the Wyvell experiments, it is
involved in the generation of desires in response to certain stimuli. To see why such a
system might have evolved, let us start by employing some relatively a priori considera-
tions about creatures like us and about the kind of wanting system that would benefit us.

5  Modeling the wanting system
Some creatures are tightly locked into a specific pattern of consumption: an insect that
eats the leaves from a single plant species, or a koala that eats the leaves from four. Such
creatures can have their tastes hardwired. Other creatures are more opportunistic, adapt-
ing their consumption patterns to what is available. Human beings, like rats, are at the far

12 It’s
an interesting question why the sight of the lever didn’t itself work as a cue. Clearly not all cues are
created equal.
  163

Addiction as malfunction of the wanting system 163

end of this continuum. Although some of our desires are perhaps hardwired, most are
highly plastic.
Let us think in the abstract about how a creature with plastic desires will structure its
consumption. We assume that it has some way of telling, when it samples a given food,
how good that food is in providing it with what it needs. The goal of the creature is to
maximize its consumption of things that are good for it. How could it go about that?
One way would be for the creature simply to try each thing that it comes across to see
how good it is and then consume it if it is; but obviously that would be highly inefficient,
since it would involve constantly retrying things which had already been shown to be bad.
A second would be to learn what is good for it, in the full sense of that term: the creature
would develop beliefs about which foods are good, and then, given its desire to consume
what is good, it would form instrumental desires for just those foods (Dickinson and
Balleine 2010). A third possibility would be to avoid forming the beliefs at all. Instead, the
creature could directly form its desires on the basis of what it had discovered to be good.
That is, it could form intrinsic desires for the good foods, without recourse to any beliefs
or predictions about them.
This third possibility would have some advantages. It could be simpler and easier to
implement than a belief-​based system, and in some ways more robust. So let us consider
how it might work. It would need to do two things. First, it would need to form long-​term
intrinsic desires for the kinds of food it has found good; second, it would need to act on
those desires in the presence of the foods. For this it would need two systems: a desire for-
mation system that would create intrinsic desires, and a consumption system that would
regulate the creature’s consumption in accord with those desires.
We suggest that both rats and human beings exhibit both of these systems, working
together in just this way. To see this for rats, return to the findings of Cindy Wyvell. As
mentioned, Wyvell found that boosting dopamine caused huge increases in short-​term
wanting in the presence of the relevant cue. But this was not all. She also ran a paral-
lel set of experiments on rats who had received earlier amphetamine injections during
the initial conditioning, rather than at the time of the later stimulus. She found that this
sensitized their brains in a long-​term way. Despite being free of the drug for ten days, the
later stimulus of the sound elicited twice the rate of lever pressing from these rats as it
did from a control group who had not received the sensitizing injections. The dopamine
was creating long-​term intrinsic desires (Wyvell and Berridge 2001; see also Tindell et al.
2005; Smith et al. 2011).
So here we have evidence of exactly the kind of model mentioned. Dopamine is involved
in two processes: the laying down of long-​run intrinsic desires; and then the triggering of
those desires in the presence of the relevant cues.

6  Addiction as malfunction of the wanting system

Now that we have the model in place, our account of addiction can be quick. What would
happen if a subject consumed a substance that caused an artificial boost in dopamine
levels? The effect would be two-​fold. First, it would experience a boost in its immediate
164  

164 Compulsion and choice in addiction

desire for the substance. Second, it would experience a boost in its long-​run desire for
it. That desire would be cued by the substance itself, or by other cues that were around
at the time that the substance was initially consumed. If the dopamine signal was strong
enough, the ongoing sensitization could be very great, potentially persisting indefinitely.
Our claim is that this is just what happens in cases of addiction. Since the addictive
drugs artificially stimulate the dopamine system so powerfully, they give rise to long-​
lasting dispositional desires. The dispositional desires are triggered by cues surrounding
the consumption of the drugs: the drugs themselves, but also, given the associative nature
of the process, the places in which they are consumed, the paraphernalia surrounding
their consumption, and so on. Since these are intrinsic and not instrumental desires, they
are not undermined by the belief that consumption of the drugs will not be pleasurable,
or that it will be harmful in some other way. These dispositional desires may persist long
after the subject has stopped taking the drugs, and has gone through any associated with-
drawal. A  cue provided by seeing the drug, or the environment in which it was once
taken, or even by imagining it, may provoke a powerful occurrent desire for it; and if this
results in further consumption, the whole pattern will be repeated.
This seems to fit the facts very well. Or at least, it fits some of the facts very well: the
pathological facts, those concerning the way that addiction differs from ordinary behav-
ior. But it might seem that this has taken us too far. For what are we to make of those
aspects of addiction that make it seem very much like ordinary behavior? Can we pre-
serve the idea that addicts are nonetheless sensitive to standard incentives?
The crucial point here is that, in human beings, the incentive salience process that we
have sketched does not necessarily lead directly to behavior.13 It typically leads instead
to cravings:  to powerful desires that tend to crowd out other considerations. (See
Loewenstein (1999) for a good discussion of how cravings tend to narrow one’s focus.)
Many philosophers make a sharp contrast between desires and intentions. Desires are
the inputs to deliberation; it is quite rational to have many that conflict. Intentions are
the outputs of deliberation; they are insulated from reconsideration and lead directly to
action, and so they generally need to be consistent (Holton 2009). Cravings seem to come
somewhere between the two. Whilst they have many of the features of standard desires,
they are not easily thought of as inputs to deliberation. Rather, they lead directly to action
unless something stops them. Stopping them requires self-​control; to this we now turn.

7  Self-​control
Both philosophers and psychologists tend to view desires as a fundamentally uniform
class. Roughly, they are the states that move an agent to action. In contrast, we think that
they are heterogeneous. So far, we have focused just on one kind, the desires, or cravings,

13 This is not to deny that incentive salience effects can work unconsciously in a way that takes them fairly

directly to behavior. See Winkielman and Berridge (2004). But such behavior is still susceptible to self-​
control; it is just that the subject doesn’t see the need to exert it.
  165

Self-control 165

that result from the incentive salience process. As we mentioned at the beginning, we also
have other, more rationally tractable desires: a desire to take a holiday in St Petersburg,
say, or to be healthy, or to treat a particular person well. And many of these are intimately
connected with our beliefs. If we come to think that St Petersburg is too Western to reveal
the true Russia, and it is the true Russia that we are after, then our desire to visit will be
undermined. In contrast, the cravings that result from the incentive salience process are
not typically undermined by the belief that they are harmful. (For an excellent discussion
of such desires see Railton (2012).) Many actual desires may combine an element of both
types; indeed the very case that Railton uses as illustrative of the more cognitive desire—​a
desire for an espresso—​is very plausibly a case in point.)
But if we have at least two different sorts of desires—​together perhaps with other factors
that also influence our behavior, like our habits—​then the question arises of what it is that
determines what we will do. This is a difficult and complex question that we cannot hope
fully to answer here. But one thing that we think has become clear in recent years is that
it is not fully determined by the relative strength of the different sorts of desires. We also
need to factor in a more active control on the part of the agent.
Although the details remain controversial, a wealth of psychological research supports
the idea that self-​control should be taken seriously. Self-​control develops in children after
the development of desires; it is effortful; it is depleted by various factors including stress,
fatigue, and its prior exercise; and it can be developed and deployed more or less success-
fully (for general discussion see Holton (2009)). A failure to behave a certain way might
indicate a lack of desire to behave that way. Alternatively, it might indicate that a desire,
even the kind of craving that results from addiction, is being held in check by self-​control.
To say that self-​control is real is not to deny that its exercise is sensitive to the agent’s
beliefs and desires. Agents can be well motivated to employ it, if they think that there is
something to be gained from it, and that its employment will be successful. Alternatively,
if they think that it will bring little benefit, or that the benefits can be gained more easily
another way, or that it is unlikely to succeed, they will be far less likely to employ it, and
even if they do initially employ it, given that it is effortful, they will be far more likely to
give up.
As we have seen, the pathology of addiction means that addicts will experience strong
cue-​driven cravings long after withdrawal is over, especially at particular moments such
as when a drug cue is encountered in a moment of stress or emotional excitement. But
this is not the end of the story. Whilst there is some evidence that addictive drugs can
diminish self-​control by damaging the prefrontal cortex (Volkow et al. 2004), there is no
reason to think that addicts lose it altogether. Indeed, the fact that addicts can get them-
selves off their addictions is strong evidence that they do not. Controlling cravings may
be tremendously hard work, but that it is not to say that it is impossible. Understanding
when it is that addicts will continue to consume and when they will not thus requires
an understanding of how their cravings interact with their self-​control. Whilst we do
not have even the beginnings of a real account here, we identify the following factors as
very likely to be relevant to the pattern of activity that we remarked on at the outset, in
166  

166 Compulsion and choice in addiction

particular the responsiveness of addicts to incentives, and their tendency to escape their
addictions in their late 20s or early 30s.

7.1  The strength of the self-​control system

There is evidence that self-​control, regulated primarily by the prefrontal cortex, continues
to develop in strength into the mid-​20s, typically maturing rather earlier in women than
in men (see, e.g., Luna and Sweezy 2004; Goldstein et al. 2009; and, for a popular review,
Sabbagh 2006).

7.2  The
efficiency with which the self-​control system
is employed
A great deal of research indicates that there are techniques that enable agents to better
deploy their self-​control. Forming prior intentions and then acting on them without reo-
pening the question of what to do seems important, as does avoiding tempting situations.
Similarly, mindfulness techniques can enable agents to stand back from their desires in
ways that make their self-​control more effective. It is still an open question how effective
such techniques can be against the kinds of cravings engendered by addiction, but ini-
tial research indicates that they can make a difference (Prestwich et al. 2006; Kober et al.
2010). Again, skill in using the self-​control system is something that we might expect to
increase with age.

7.3  The role of desires

Addicts who have strong motivations for giving up rather than continuing are more likely
to employ their self-​control to overcome cravings. And it does seem likely that concerns
about partners, families, and careers will become more pressing as people reach their late
20s and early 30s. Conversely, since dopamine levels start to fall from the teenage years
onwards, the power of the cravings may themselves diminish.

7.4  The role of belief

If addicts think that there is little reason to give up today, since giving up tomorrow will be
just as good, there will be little motivation to employ self-​control. Vague concerns about
health and wellbeing are often of that form; there can be a sense that, whilst giving up is
something that needs to be done at some point, one more dose won’t hurt. In contrast, the
incentives that have been shown to work well—​for instance, the knowledge that certain
dismissal from a much-​valued job will follow a single positive drug test—​guarantee an
immediate cost or benefit. We suspect that much the same is true of a price rise; whilst it is
true that paying the higher price just one more time is probably within the addict’s reach,
there is no escaping the fact that a higher price is being paid. The other set of relevant
beliefs concern the efficacy of exerting self-​control. If the addict is convinced that they
will succumb despite their best efforts—​if not today, then surely soon—​the motivation
to try will be much reduced. And here, presumably, the addict’s own theory of addiction
will have a part to play. If they think of the addiction as resulting in behavior that is quite
  167

The extent of addiction, and its rationality 167

outside their control, they will be far less motivated to try to control it (a point that has
been noted many times by Albert Bandura; see, e.g., Bandura (1999)).

8  The extent of addiction, and its rationality

We have talked about addictions that are caused by drugs—​by substances that interfere
directly with the dopamine system, and gain their incentive salience effect from that
interference. But what of the many other kinds of behavioral addictions—​addictions to
gambling, shopping, sex, or the Internet—​that feature so prominently in current discus-
sion. Can we give an account of them? Or is the theory we have given bound to say that
they these are not really addictions?14
Clearly, our account is bound to say that there is an important difference between sub-
stance and behavioral addictions. The latter do not, so far as we know, involve mecha-
nisms that short-​circuit the dopamine system in the way the former do. Nevertheless,
there is good reason to think that they too work through the incentive salience system,
and so that they too can result in cue driven cravings that are relatively insulated from
other desires and from beliefs about what is good. Of course, if the dopamine system
has not been short-​circuited, then these behavioral addictions must have originated from
behavior that was pleasurable, or was in some other way recognized by the agent’s dopa-
mine system as being beneficial. But the assessment of the dopamine system might be at
odds with the agent’s more cognitive beliefs about the value of the activity; and even if it
is not, once the intrinsic desires have been established, they will tend to persist through
changes in the agent’s assessment at any level. Even if the agent stops liking the thing
concerned, a well-​established incentive salience desire will degrade very slowly. The result
can be behavior that looks very like the addiction engendered by drugs (Dill and Holton
2014). (Further evidence that drug and behavioral addictions have much in common
comes from the cases of Parkinson’s patients who respond to their dopamine supplement
by developing addictive behavior (O’Sullivan et al. 2009). We leave open the question of
whether other behaviors that look rather like chemical addictions—​those resulting from
obsessive compulsive disorder, for instance—​should also be understood in the same way.)
This brings us finally to an issue that we have largely skirted up till now, that of the
rationality of addicts. Most ordinary choice models see addicts as quite rational, though
working with unusual desires or false beliefs (perhaps there is some irrationality in how
they arrived at those beliefs, but that doesn’t affect the rationality of how they act upon
them). Most disease models see the addict as largely arational: addictive actions hardly
count as intentional actions at all, and so fall outside the scope of rationality. In contrast,

14 We have made the traditional division between substance addiction and behavioral addiction, but it
could be that some substances give rise to addiction-​like behavior without hijacking the dopamine
system in the way we have discussed, and so should be grouped with the behavioral addictions. Sugar
might be like that, and perhaps, though here the findings are controversial, cannabis. So a more careful
distinction would be between the dopamine-​hijacking addictions, and those that are not. But we will
stick with the more traditional terminology.
168  

168 Compulsion and choice in addiction

the account that we have developed here sees the addict as potentially irrational in two
ways. One is familiar: if considered views about what would be best diverge from action,
then both substance addicts and behavioral addicts will frequently be akratic, in ways
that have at least a prima facie claim to irrationality. The second is rather less familiar. If
what we have said is right, then something goes badly wrong with the process by which
substance addicts (but not behavioral addicts) form their desires: substances come to be
desired independently of any pleasure or other benefits that they bring. There has been
much discussion in philosophy of whether intrinsic desires can be irrational. What we
are suggesting is that substance addiction results from the malfunctioning of a normally
rational system for creating intrinsic desires. This seems to us as clear a case of an irra-
tional intrinsic desire as one is ever likely to find.

9 Conclusion
We started by stressing the need to find a middle path. Our attempt to find one has
involved exploring the interaction between two different systems: one that regulates our
desires, and one that controls which desires we act on. Addiction results from the mal-
function of the first; insofar as it does not result in a complete loss of agency, that is,
thanks to the second. In a sense then, both the disease model and the choice model are
describing something real; but each gives a picture that is partial. We hope that we have
gone some way to putting them together.

References
Bandura, A. (1999). A sociocognitive analysis of substance abuse. Psychological Science, 10, 214–​7.
Berridge, K. (2000). Taste reactivity: measuring hedonic impact in infants and animals. Neuroscience
and Biobehavioral Reviews, 24, 173–​98.
Berridge, K. (2007). The debate over dopamine’s role in reward: the case for incentive salience.
Psychopharmacology, 191, 391–​431.
Berridge, K. (2012). From prediction error to incentive salience: mesolimbic computation of reward
motivation. European Journal of Neuroscience, 35, 1124–​43.
Berridge, K. and Kringelbach, M. (2008). Affective neuroscience of pleasure: reward in humans and
animals. Psychopharmacology, 199, 457–​80.
Berridge, K. and Valenstein, E. (1991). What psychological process mediates feeding evoked by
electrical stimulation of the lateral hypothalamus? Behavior Neuroscience, 105, 3–​14.
Dickinson A. and Balleine, B. (2010). Hedonics: the cognitive-​motivational interface. In Kringelbach,
M. and Berridge, K. (eds), Pleasures of the Brain. Oxford: Oxford University Press, pp. 74–​84.
Dill, B. and Holton, R. (2014). The addict in us all. Frontiers in Psychiatry, 5, 139. doi: 10.3389/​
fpsyt.2014.00139.
Everitt, B., Belin, D., Economidou, D., Pelloux, Y., Dalley, J., and Robbins, T. (2008) Neural
mechanisms underlying the vulnerability to develop compulsive drug-​seeking habits and addiction.
Philosophical Transactions of the Royal Society of London B Biological Science, 363, 3125–​35.
Faure, A., Richard, J., and Berridge, K. (2010). Desire and dread from the nucleus accumbens: cortical
glutamate and subcortical GABA differentially generate motivation and hedonic impact in the rat.
PLoS One, 5, e11223.
Goldstein, R., Craig, A., Bechara, A., et al. (2009). The neurocircuitry of impaired insight in drug
addiction. Trends in Cognitive Science, 13, 372–​80.
  169

Conclusion 169

Herrnstein, R. and Prelec, D. (1992). A theory of addiction. In: G. Loewenstein and J. Elster J. (eds),
Choice over Time. New York: Russell Sage Press, pp. 331–​60.
Heyman, G. (2009). Addiction: A Disorder of Choice. Cambridge, MA: Harvard University Press.
Hoffman, J. and Froeke S. (eds) (2007). Addiction: Why Can’t They Just Stop? Emmaus, PA: Rodale Press.
Holton, R. (2009). Willing, Wanting, Waiting. Oxford: Clarendon Press.
Holton, R. and Berridge, K. (2013). Addiction between compulsion and choice. In: N. Levy (ed.),
Addiction and Self-​Control. Oxford: Oxford University Press, pp. 239–​68.
James, W. (1890). Principles of Psychology. New York: Henry Holt.
Kober, H., Kross E., Mischel M., Hart C., and Ochsner K. (2010). Regulation of craving by cognitive
strategies in cigarette smokers. Drug and Alcohol Dependence, 106, 52–​55.
Lawrence, A., Evans A., and Lees A. (2003). Compulsive use of dopamine replacement therapy in
Parkinson’s disease: reward systems gone awry? Lancet Neurology, 2, 595–​604.
Leyton, M. (2010). The neurobiology of desire: dopamine and the regulation of mood and motivational
states in humans. In: M. Kringelbach and K. Berridge, Pleasures of the Brain. New York: Oxford
University Press, pp. 222–​43.
Linden, D. (2011). The Compass of Pleasure. New York: Viking.
Loewenstein, G. (1999). A visceral account of addiction. In: J. Elster and O. Skog (eds), Getting Hooked.
Cambridge: Cambridge University Press, pp. 235–​64.
Luna B. and Sweezy J. (2004). The emergence of collaborative brain function: fMRI studies of the
development of response inhibition. Annals of the New York Academy of Science, 1021, 296–​309.
O’Sullivan, S, Evans, A., and Lees, A. (2009). Dopamine dysregulation syndrome: an overview of its
epidemiology, mechanisms and management. CNS Drugs, 23, 157–​70.
Peciña, S., Cagniard, B., Berridge, K., Aldridge, J., and Zhuang, X. (2003). Hyperdopaminergic
mutant mice have higher “wanting” but not “liking” for sweet rewards. Journal of Neuroscience, 23,
9395–​402.
Prestwich, A., Conner, M., and Lawton, R. (2006). Implementation intentions: can they be used to
prevent and treat addiction? In: R. Wiers and A. Stacy (eds), Handbook of Implicit Cognition and
Addiction. Thousand Oaks, CA: Sage, pp. 455–​71.
Railton, P. (2012). That obscure object, desire. Proceedings and Addresses of the American Philosophical
Association, 86, 22–​46.
Redish, A.D. (2004). Addiction as a computational process gone awry. Science, 306, 1944–​47.
Robinson, T. and Berridge, K. (2003). Addiction. Annual Review of Psychology, 54, 25–​53.
Rush, B. (1812) Medical Inquiries and Observations, Upon the Diseases of the Mind.
Philadelphia: Kimber and Richardson.
Sabbagh, L. (2006). The teen brain, hard at work. Scientific American Mind, August/​September, 20–​25.
Schultz, W., Dayan, P., and Montague, R. (1997). A neural substrate of prediction and reward. Science,
275, 1593–​99.
Skog, O. (1999). Rationality, irrationality and addiction: notes on Becker’s and Murphy’s theory of
addiction. In: J. Elster and O. Skog (eds,) Getting Hooked. Cambridge: Cambridge University Press,
pp. 173–​207.
Smith, K., Berridge, K., and Aldridge, J. (2011). Disentangling pleasure from incentive salience and
learning signals in brain reward circuitry. Proceedings of the National Academy of Science, 108,
E255–​64.
Sournia, J. (1990). A History of Alcoholism. Oxford: Blackwell.
Sutton, R. and Barto, A. (1998). Reinforcement Learning. Cambridge, MA: MIT Press.
Tindell, A., Berridge, K., Zhang, J., Peciña S., and Aldridge J. (2005). Ventral pallidal neurons code
incentive motivation: amplification by mesolimbic sensitization and amphetamine. European
Journal of Neuroscience, 22, 2617–​34.
170  

170 Compulsion and choice in addiction

Volkow, N., Fowler, J., and Wang G. (2004). The addicted human brain viewed in the light of imaging
studies. Neuropharmacology, 47, 3–​13.
Wilde, O. (1891). The Picture of Dorian Gray. London: Ward Locke and Co.
Winkielman P. and Berridge, K. (2004). Unconscious emotion. Current Directions in Psychology,
13, 120–​3.
Wise, R. (1980). The dopamine synapse and the notion of “pleasure centers” in the brain. Trends in
Neuroscience, 3, 91–​5.
Wise, R. (1985). The anhedonia hypothesis: Mark III. Behaviour and Brain Science, 8, 178–​86.
Wise, R. (2004). Dopamine, learning and motivation. National Review of Neuroscience, 5, 483–​94.
Wood, W. and Neal, D. (2007). A new look at habits and the habit–​goal interface. Psychological Review,
114, 843–​63.
Wyvell, C. (2001). Incentive sensitization by previous amphetamine exposure: increased cue-​triggered
“wanting” for sucrose reward. Journal of Neuroscience, 21, 7831–​40.
Wyvell, C. and Berridge K. (2000). Intra-​accumbens amphetamine increases the conditioned incentive
salience of sucrose reward. Journal of Neuroscience, 20, 8122–​30.
  171

Chapter 10

Choice in addiction: A neural tug

of war between impulse and insight
Marc D. Lewis

Abstract
The neuroscience of addiction does not frame addiction as a disease. In
fact, an examination of brain processes helps clarify the role of choice in
addiction, by explaining how habitual perceptions and time-based biases
influence decision-making for addicts and non-addicts alike. The impact
of addictive cues is enhanced by dopamine signals to the striatum and
orbitofrontal cortex, due to repeated experiences of pleasure or relief.
Moreover, dopamine-based craving peaks when addictive rewards become
available, in the moment. Finally, the consolidation of synaptic networks
modifies brain structures to maximize the appeal of addictive activities over
competing rewards. Given these neural realities, the choice to engage in
an addictive activity is biased in favor of momentary fulfillment, but it’s
still a choice. Choice is always pulled between momentary and long-term
rewards. Thus choice is never “rational”—it is a spontaneous resolution
to unbalanced goals. Addiction merely epitomizes this inescapable human
conundrum.

1 Introduction
In our efforts to make sense of addiction, we often settle on one of two definitions: either
it’s a pathology or a disease, or, by contrast, it’s some kind of choice. In this chapter, I peel
back both of these stick-​on labels, these too-​easy answers to the problem of addiction,
and argue that addiction is neither a disease nor a rational choice. Rather, I see addic-
tive choices as decisions, deliberately made, but far from rational, constrained as they
are by deeply learned habits of thought and shifting from one extreme to the other with
one’s perception of time. The temporal landscape of addictive choice is riven with ridges
and valleys: time does not flow evenly across it. So the decision to abstain can be main-
tained for days or months, until time buckles and the resolution suddenly vanishes, to be
replaced by the determination to use, right now, tonight. When that happens, the hori-
zons of the present moment have expanded into an overarching present tense, discon-
nected from the past and the future.
172  

172 Choice in addiction

Psychologists have a name for the problem of temporal fluctuations in decision-​

making: intertemporal choice. However, a biological account of addiction requires us to
trace this psychological phenomenon to its roots in the structure and function of the
brain, and to see how brain changes in addictive learning make intertemporal choice that
much more problematic. Brains were designed by evolution to focus our attention on
presently available goals, especially when immediate goals with high hedonic impact are
weighed against long-​term goals, even though they may be advantageous in the long run.
As a result, addicts find it particularly difficult to continue to abstain from using when
addictive rewards (e.g. drugs, alcohol, gambling, sex, etc.) suddenly appear on the radar
screen of the mind.
According to the disease model in its current form, addiction is a brain disease that
affects motivation, behavior, and self-​control. It is characterized by changes in spe-
cific brain systems, especially those that process rewards (i.e. valued outcomes). Brain
systems responsible for anticipating rewards, motivating us to go after them, and
evaluating and re-​evaluating the worth of those rewards are reshaped by the repeated
use of drugs, including alcohol. Researchers have found additional brain changes in
systems underlying cognitive control, delayed gratification, and abstract skills like
comparing and predicting outcomes and selecting best choices. Perhaps most impor-
tantly, the disease model rests on the assumption that addiction is compulsive. The
urge to use is more than an urge; it is a command that cannot be resisted, not only
because of diminished cognitive control but because of a fundamental shift in the
structure of the motivational core of the brain. According to the disease model, these
changes are caused by exposure to drugs of abuse. Thus, taking drugs fundamentally
damages the brain.
Of course the disease model builds on a biological framework, and it does a good job
of explaining why some individuals are more vulnerable to addiction than others, based
on genetic differences and other dispositional factors. In addition, addiction is currently
viewed as a chronic disease. But that’s not a problem for the model, because it’s also true of
many well-​known illnesses, including heart disease, diabetes, and some forms of cancer.
For those too there are treatments but not cures.
The idea that addiction is a choice comes from a cognitive (rather than a biological)
perspective, emphasizing changes in thought processes. Researchers in behavioral eco-
nomics, which blends social psychology with economic thinking, try to understand why
people make the choices they make, including the choice to take addictive substances.
While few people imagine that addiction is a good choice, it is often considered a rational
one, at least in the short run—​as when the pleasure or relief derived from one’s addiction
seems to outweigh the value of other possible choices. Unfortunately, the choice model
provides a convenient platform for those who consider addicts indulgent and selfish. If
addiction is a choice, they reason, then addicts are deliberately inflicting harm on them-
selves and, more seriously, on others. Yet other proponents of the choice model point to
environmental or economic conditions beyond the addict’s control, including poverty
and social isolation (see Alexander 2008).
  173

Introduction 173

The choice model does a better job than the disease model of explaining how addicts
quit. When conditions change with time and circumstances, so do choices. As empha-
sized by Gene Heyman (2009) and others, it should not be surprising if people choose
to quit when life circumstances improve, or when the financial or social costs of remain-
ing addicted exceed the benefits of being high. Either explanation could account for
the undisputed finding that a majority of heroin-​addicted US veterans stopped using
heroin when they returned from Vietnam (e.g. Heyman 2009; see also Chapter 21, this
volume).
The ongoing tug of war between the disease model and the choice model has its advan-
tages. It compels us to closely examine the mechanisms that maintain addiction and those
by which people quit. However, as with many conflicts that drag on too long, some terri-
tory has been heavily, perhaps irrevocably, damaged—​in this case an important region of
intellectual inquiry. I’m talking about the insidious absorption of neuroscience, or per-
haps biological science more generally, into the dominion of the disease model, and its
abandonment by those who emphasize choice.
Is the alliance between neuroscience and the disease model authentic? Does neuro-
science inevitably support the disease model? A  lot of experts seem to think so. Nora
Volkow, the head of the National Institute on Drug Abuse (NIDA) in the USA, is a power-
ful spokesperson for this view, but it’s probably safe to say that a bond between the disease
model and brain science is assumed by most workers in the field: neuroscientists them-
selves, geneticists and other biological researchers, physicians and other medical person-
nel, psychiatrists to be sure, and addiction workers far and wide, from the directors of
rehab centres, waiting for addicts to come in to get cured (or at least to pay handsomely
for the chance), to the raised voices in some—​not all—​AA meetings, which maintain that
the disease of addiction will inevitably overtake those who don’t abstain and comply with
the program.
Yet neuroscience doesn’t have to cast addiction as a disease. Rather, it can offer a way
out of the disease-​versus-​choice controversy, by identifying the biological processes that
connect reward to decision-​making in the moment and goal-​seeking habits over time.
Choice model advocates often paint volition as a deliberate, rational function we can
apply at will—​which sounds like a descendant of Cartesian dualism, whereby mind stuff
and body stuff are distinctly different in their composition and essence. In fact, the possi-
bility of free will is generally disputed by philosophers, but let’s leave that one for now. The
point is that the choice to return to one’s addiction time after time, even if it’s voluntary,
perhaps especially if it’s voluntary, is in no away independent of the brain. After all, choice
is executed by the same brain that gives rise to hope, need, fear, and uncertainty, a brain
that’s highly sensitive to learned associations and contextual cues, a brain that forges new
connections based on the activation of existing connections and the strong emotions they
render. Because choice is complex, nuanced, and frequently ill-​conceived, choice advo-
cates need to dismiss the idea of rationality and focus instead on the fleshy (neural) con-
ditions responsible for the erratic character of our planned behavior. I would urge choice
model advocates not to surrender the territory of neuroscience to the enemy—​the disease
174  

174 Choice in addiction

camp—​but rather to send in their own scouts and try to win it back. Once choice is rein-
tegrated with the biological and emotional mushiness that truly characterizes human cog-
nition, then the role of individual choice can be described realistically and persuasively.

2  Cues, conditioning, and dopamine

According to some advocates of choice, addicts voluntarily stop using drugs when the
consequences get bad enough. For example, both Ainslie (2001) and Heyman (2009)
extol the benefits of contingencies and incentives—​that is, using the consequences of
behavior—​in programs for helping addicts abstain. It is notable that such a view ech-
oes Skinner’s own views on self-​destructive behavior. In Walden Two, Skinner (1948)
argued that strategic modification of contingencies is the best cure for self-​indulgence.
Yet a good deal of research on the neural responses underlying addiction points to cues,
not consequences, as determinants of drug-​seeking behavior and relapse. Can we view
addiction in terms of volition, yet recognize that this volition is deeply affected by drug-​
(and alcohol-​) related cues? That cues greatly outweigh consequences as the causal basis
of goal seeking?
Berridge and his colleagues (e.g. Berridge and Aldridge 2008; Robinson and Berridge
2003), recognized leaders in the neuroscience of addiction, propose incentive salience or
“cue-​triggered wanting” as the sine qua non of addiction. Cues associated with drugs,
such as paraphernalia, vodka ads, drug-​taking buddies, or auditory tones for animals in
cages, become more salient with recurring consumption and pairing of substance and
stimulus. This is because drug taking triggers the release of dopamine, sucked up from
the ventral tegmental area (VTA) in the midbrain (I’ll just say midbrain from here on).
In turn, dopamine increases the processing of these particular stimuli in several regions
of the striatum, involved in reward seeking, the amygdala, involved in emotional condi-
tioning, and the orbitofrontal cortex (OFC), where expectancies are encoded and value is
assigned (Robinson and Berridge 1993, 2003; Rosenkranz and Grace 2002; Schultz 2007;
Volkow and Fowler 2000; Wyvell and Berridge 2000). The pairing of cues with anticipated
drug experiences leads to their increasing power to induce psychophysiological responses
(e.g. “wanting” or craving). The cues come to signify reward through Pavlovian condi-
tioning, not instrumental conditioning. They evoke responses that are at least partially
independent of expected outcomes. In which case, “choice” is nothing like sitting back
and planning a strategy based on best expected results. Rather, choice is the thing we do
to resolve conflicts between impulse, compulsion, planning, and strategy formation.
Choice advocates might assume that there can’t be anything like choice in a brain that’s
caught up in this kind of juggling act. But this assumption is based on the incorrect premise
that addiction is either a choice or a biological problem (e.g. a disease). This is a false dichot-
omy. Biological explanations are not incompatible with choice. Human motivation cannot
be reduced to fixed action patterns, and cues do not trigger human behavior the way the
image of a fly triggers the flick of a frog’s tongue. Rather, motivation and emotion influence
behavior by narrowing appraisals and urging actions that address them (this is the party
line in emotion theory; e.g. Frijda 1986; Izard 1991). Acknowledging the motivational force
  175

Cues, conditioning, and dopamine 175

of drug-​related cues doesn’t dehumanize addiction; it helps capture the biological vulner-
abilities that are part of what it means to be human—​a human trying hard to steer his or her
behavior in a direction that feels right, whether or not it makes sense objectively.
The central place of dopamine in addiction is strongly supported by the dopamine-​
enhancing effects of an entire class of addictive drugs (methamphetamine and cocaine),
the impact of dopamine antagonists, dopamine-​knockout, and other manipulations on
drug taking and orbitostriatal activation (e.g. Nestler 2005; Volkow and Fowler 2000),
and by numerous other findings, including the potentiation of gambling problems
in Parkinson’s patients treated with dopamine agonists (Dodd et  al. 2005). Moreover,
Pavlovian conditioning in the amygdala is directly enhanced by dopamine (Rosenkranz
and Grace 2002), where it has been found to correlate with cocaine self-​administration
(Hurd et  al. 1997). Yet many neuroscientists no longer believe that dopamine equals
reward in the brain, and there is copious evidence that reward is largely independent
of dopamine (see the recent review by Nutt et al. 2015). While “rewardingness” might
depend in part on dopamine, the increased activation of dopamine-​powered networks in
the striatum is thought to correspond more closely with the feeling of desire, anticipation,
and craving, not the feeling of being high (Risinger et al. 2005), or “wanting” rather than
“liking” in Robinson and Berridge’s model.
We have to be a little careful not to overgeneralize or blur distinctions when consider-
ing the role of dopamine in addiction. First, we have to keep in mind that dopamine has
many different effects in different parts of the brain, and even in the same parts, due to
its reception in at least five distinct receptor types. Second, the neurochemistry of dopa-
mine is incredibly complex (so is neurochemistry in general), and dopamine release and
absorption is just one part of a more elaborate chemical bouquet. Third, dopamine is a
key neuromodulator for motivating and directing goal-​related behavior of all sorts. Its
functions are not restricted to addictive behavior. Yet much of the “natural” reward seek-
ing subserved by dopamine can have addiction-​like characteristics. Eating is an obvious
example. In addition, dopamine-​powered augmentation of the perceived value of imme-
diate rewards may be a general characteristic of goal-​seeking behavior, as argued later.
Finally, the dopamine metabolism implicated in addiction includes cell changes both in
the dopamine source (in the midbrain) and in the areas that receive it, so the effect of
dopamine activity changes over time (in weeks and months).
Addiction is characterized by a particular quality of motivation usually termed “crav-
ing,” which amounts to intense, sometimes unrelenting desire for something that is not
immediately available. It is also characterized by loss of control, but that may be a sec-
ondary effect of intense and recurrent craving. If it weren’t for craving, addiction just
wouldn’t be a problem, and the choice to indulge or not would look and feel entirely dif-
ferent. Craving for a particular substance or activity implies a severely narrowed range of
attractive goals. There may be nothing else in the repertoire that seems worth pursuing.
The very narrowness of the goal hierarchy means that a particular goal has been thought
about, sought, and/​or experienced enough times to form a robust synaptic pattern repre-
senting that goal and the expected value of attaining it. Because craving relies on synaptic
176  

176 Choice in addiction

entrenchment, it is most likely to result from the excessive and repeated emotional high-
lighting of an attractive experience—​i.e. one that brings either pleasure or else relief from
anxiety or other negative states. As noted earlier, dopamine release from the midbrain to
the striatum is the key neural event thought to underpin addictive craving, and dopamine
release by addictive cues becomes increasingly likely over the course of the addiction.
Dopamine release to the prefrontal cortex also changes with repeated addictive behavior,
so that actions are increasingly triggered by the impulsive or compulsive urges mediated
by the striatum and less by prefrontally mediated insight and reflection. What’s behind
this self-​perpetuating, self-​amplifying sequence? Why does it happen?
The greater the hedonic impact of a stimulus and the more it is repeated, the more dopa-
mine will amplify its salience and the more rapidly the relevant synaptic networks will
become consolidated. That amplification/​consolidation process is a result of increasingly
targeted dopamine signals from the midbrain to the striatum whenever addictive cues are
encountered—​for example, when you drive by the liquor store or meet someone who can
obtain drugs. Repetition of addictive behavior results from two related factors: (1) sali-
ent, emotionally compelling, pleasureful, or relief-​providing experiences are more likely
to be repeated than any other kind of experience; and (2) addictive rewards are by their
nature short-​lived. Drugs wear off, alcohol leads to excessive sedation and sleep, sexual
pyrotechnics become boring, and there’s only so much you can eat. These short-​lived
impactful experiences leave us feeling empty and unsatisfied at the very least. They also
frequently leave us feeling shame, guilt, and/​or remorse. With physically addictive sub-
stances, these outcomes are joined by bodily discomfort and anxieties about how to stem
them. All of these negative outcomes (the “dark side” of addiction; Koob 2009) cause us
to want to have more of what makes us feel better. Hence repetition. In sum, repetition
of hedonically intense experiences consolidates synaptic connections (e.g. in the stria-
tum and OFC) that represent the addictive goal and strengthens synaptic connections
between these regions and the brainstem, thus sucking up more dopamine.
I started this section by suggesting that a biological model of addiction can still accom-
modate choice, as long as we recognize that choice isn’t necessarily rational or thoughtful.
I suggested that we look to cues rather than consequences as the chief data source when it
comes to goal seeking in addiction. In section 3, I add to the already stacked deck in favor
of addictive choices—​hedonic appeal, augmented focus, and synaptic structuring—​by
looking at one of the most insidious aspects of dopamine-​powered attention: its amplifi-
cation of the perceived value of what’s in front of our noses.

3  Craving and “now appeal”

Addiction is highly sensitive to time. It is a problem that pits the present tense against the
future tense. Most people know this anecdotally, and those of us who have lived through
addiction know it through that most convincing source, personal experience. Addicts
are far more likely to violate their commitment to abstinence and go after drugs or booze
when available in the moment rather than planned for next week. What is it about the
  177

Craving and “now appeal” 177

moment that distorts our decision-​making in favor of rewards that end up doing us more
harm than good?
Delay discounting is a commonly observed behavioral tendency, in humans, other
mammals, and even birds, to value immediate rewards over long-​term benefits. Delayed
rewards are discounted. Their value is reduced. Delayed negative consequences are also
discounted—​in other words, delayed punishments seem less severe than immediate ones.
Delay discounting has been researched extensively by psychologists and researchers in
behavioral economics. Psychologists study it by conditioning lab animals or humans to
expect two or more rewards, each at a different delay interval, modifying the actual value
of each reward, and recording the frequency of choosing one reward over the other. For
example, the reward expected sooner may be designed to be objectively less valuable (less
food or sucrose for rats, less money for humans) than the one expected after a delay.
The delay discounting phenomenon is seen when participants go for the short-​term gain,
despite an overall reduction in net value. I’ll refer to delay discounting as “now appeal”
for the rest of this chapter.
It’s well known that addictive choices are far more likely when the payoff is immediate.
Let’s get high tonight sounds a lot more appealing than let’s get high next Wednesday. In
fact, addiction is a paradigm case of now appeal run rampant. Short-​term rewards are
obviously the pleasure, relief, or excitement afforded by the choice to get high, gamble,
eat, or whatever it is that you do too much of. These rewards are amplified in their per-
ceived value—​even for long-​time addicts who know intellectually that it’s just not going
to be that great (“wanting” rather than “liking” in Robinson and Berridge’s model). Long-​
term benefits include happier, healthier relationships, physical health, money in the bank,
self-​respect, and the likelihood of staying out of jail. They include the avoidance of long-​
term suffering, the inevitable descent into misery forewarned by messages from the War
on Drugs, or into cancer and death, as threatened by words and images on cigarette boxes
all over the world. Yet these future rewards (and probable disasters) are minimized, dis-
sipated, stripped of value, by the intrusive, glittering promise of the immediate goal.
Not surprisingly, dopamine activation of the striatum and its neighbors is to blame for
now appeal—​an unfortunate bias in the firmware that’s come to us over evolution. Like
the slouching that results in backaches—​a painful spin-​off of the achievement of upright
locomotion—​now appeal is an evolutionary side effect. Yet how could it be otherwise?
The striatum evolved to get the animal to go for low-​hanging fruit, available sexual part-
ners, whatever is most accessible—​a habit it has kept to this day. Dopamine rises with
anticipation, rushing in to rev up the striatum, when rewards are just around the corner.
The release of dopamine by immediately available goodies distorts the perspective, the
considered insight, we could have (otherwise) achieved using our more advanced cogni-
tive abilities (generally housed in the prefrontal cortex). We lunge for the immediate. With
now appeal, the attention supported by currently activated synapses contracts to a narrow
beam. A lot of that narrowing is underpinned by the network of roads between the stria-
tum and the OFC. So it’s no surprise that research subjects lying in an fMRI scanner show
increased activation in both the striatum and the OFC when they choose an immediate
178  

178 Choice in addiction

payoff over a later reward (Tanaka et al. 2004; see review by Wittmann et al. 2007). In real
life, these two partners in crime focus your attention on the attractive traveller smiling at
you from the next table, not the comforting hug of your spouse waiting at home.
Addicts are excessively now-​oriented, more prone to delay discounting than the popula-
tion average (Bickel and Marsch 2001). But nobody knows quite why. According to some
of the research, this is a brand of impulsivity, a stable personality trait, evident since child-
hood. Trait-​like now appeal helps explain why impulsive individuals are more suscepti-
ble to addiction throughout their lives. Other researchers have found that a diagnosis of
attention deficit disorder frequently overlaps both with measures of now appeal and with
vulnerability to addiction (e.g. Benningfield et al. 2014). Thus, now appeal can be consid-
ered an aspect of an attentional style that makes it difficult to learn from experience. It’s
also likely that now appeal gets more severe when life is lived on the edge and all possible
rewards and punishments are partitioned into two categories—​score or go without, now
or never. But neuroscience offers us other important clues:  one of the most revealing
studies of drug taking in time found that addicts’ self-​reports of craving and their striatal
and orbitofrontal activity (measured with fMRI) both peaked just before the moment of
drug administration and declined immediately after (Risinger et al. 2005). Thus, at both
behavioral and neural levels of description, immediacy is fundamental, and craving is its
motivational correlate. At a very different time scale, researchers have observed a discon-
nection between regions of the PFC and the striatum (or related regions) when addiction
progresses over a period of years (Goldstein and Volkow, 2002, 2011). The dorsal and
lateral regions of the PFC are critical for judgment, perspective taking, comparing alter-
native outcomes—​in a word, insight. So the diminished capacity for perspective taking
that we see in now appeal has a concrete neural parallel.
Conventional utility theory (a branch of behavioral economics) models delay discount-
ing (now appeal) as an exponential curve in the perceived value of a reward as it gets
closer in time. When the reward is in the future, perceived value is low. But the increas-
ing perceived value as the rewards get closer in time produces an accelerating upward
slope, which reaches its peak just before the expected payoff. According to George Ainslie
(2001), this is why conventional utility theories do such a poor job of explaining addic-
tion. In his elegant and powerful book, Breakdown of Will, Ainslie shows that discount-
ing curves do not change smoothly (see also Chapter 13, this volume). Rather, the curve
representing perceived value remains low until just before the reward becomes available,
then rises suddenly, tracing a hyperbola. This universal tendency has one chief implica-
tion. Not only are nearby rewards overvalued, but their overvaluation accelerates rapidly
just as they appear on the subjective horizon, that is, when they become imminent. For
the addict, drinker, or gambler, the value of the immediate indulgence suddenly looms
when drugs, booze, or big stakes become available, dwarfing the value of clean living that
one aspires to in the hazy future. This is why addicts continually “choose” what Heyman
(2009) calls the local choice. However, Heyman and others construe this choice as rational
or logical (at least in the short term). This seems at odds with the urgency addicts feel
when their rewards become available—​not next week, not tomorrow, but now!
  179

Brain changes over time 179

Although Ainslie has no interest in the brain or its chemicals, he describes a fact about
motivated choice that fits perfectly with Risinger et  al.’s (2005) findings:  craving peaks
just before the expected experience and it does so in sync with orbitostriatal activation.
There is nothing gradual about it. Berridge’s model makes the same prediction: just before
the reward is likely to appear, when the brain is flooded with dopamine, attraction sud-
denly skyrockets (Tindell et al. 2005). As noted earlier, the anatomy of this mechanism
is thought to be centered in the ventral striatum, where attention and action are focused
by rewards, and/​or the orbitofrontal cortex, where events are evaluated as rewarding (or
not). Both systems are directly fueled by dopamine from the midbrain. In sum, there is
a rapid rise in craving and attraction, just before a reward becomes available, subserved
by brain systems devoted to appraising the rewardingness of the immediate environment
and directing one’s actions accordingly. It seems likely that there have been distinct evo-
lutionary advantages to the motivational highlighting of immediate gains. Thus, craving,
and its amplification in now appeal, may be the byproduct of a brain designed to be maxi-
mally responsive to immediate rewards.
Addiction epitomizes bad judgment and impulsivity, if not stark irrationality. But we
can forgive the addict his addiction, and the brain its impulsivity, if we recognize that the
dopaminergic machinery of the forebrain wasn’t designed for rational decision-​making.
Dopamine’s job is to activate the power of cues. It evolved in vertebrates (hundreds of
millions of years ago) because it highlights the value of immediate rewards and directs
learning (Schultz 2007). Thus, addiction may be a common ailment in brains such as
ours. Other (more dorsal and more lateral) regions of the prefrontal cortex have to correct
dopamine’s impatience through a more reflective, insightful consideration of options and
their consequences, and these regions are currently of great interest in the study of addic-
tion. However, a sudden shift in the psyche triggered by the most rewarding of substances
(i.e. stuff designed to make you feel good) cannot be reduced to plain stupidity. Rather,
a dopaminergic time machine ensnares the addict in a brain process far older than the
capacity to reflect.

4  Brain changes over time

There is ample evidence that the brain is modified by addiction, structurally as well as func-
tionally, and some of these modifications hinder self-​regulatory capabilities and insight
itself. Nestler (2005) reviews data from many laboratories indicating that drugs of abuse
induce lasting changes in regions associated with goal seeking. These changes include
alterations in the responsiveness of the midbrain, such that dopamine release is enhanced
in relation to the drug and its cues, yet decreased in relation to other stimuli—​consistent
with Robinson and Berridge’s (1993) incentive-​sensitization hypothesis. There are also
shifts in other neuromodulators that interact with dopamine: for example, a major reduc-
tion in acetylcholine, leaving dopamine to dominate effortful behavior (Bechara and van
der Kooy 1992; Koob 2009; Nader and van der Kooy 1997). Because acetylcholine is the
neuromodulator responsible for normal alertness and attention, the character of behavior
180  

180 Choice in addiction

now shifts from exploration, alertness, and volition to single-​minded, desperate pursuit.
Brain changes also include a primary target of dopamine within the ventral striatum,
the nucleus accumbens, which begins to respond differently to inputs due to dendritic
restructuring as well as altered receptivity to dopamine. Other regions associated with the
midbrain-​striatal loop may also be altered, and these include the amygdala (emotional
memory), hippocampus (declarative or explicit memory), hypothalamus (involuntary
behavior and the autonomic nervous system), as well as cortical regions. Volkow and col-
leagues have engaged in two decades of research into cortical changes underlying addic-
tion. They conclude that prefrontal regions responsible for judging options and selecting
among them lose grey matter volume (reduced synaptic density) and become partially
dysfunctional over the course of addiction (e.g. Goldstein and Volkow 2002; Volkow and
Fowler 2000). They dub the resulting cognitive dysfunction “impaired response inhibi-
tion”—​not a hopeful starting place for building new strategies.
Given the brain changes that correspond with addiction, why not just call it a disease
and have done with it? There are two main reasons. First, every experience that is repeated
enough times because of its motivational thrust will change the wiring of the striatum
(and related regions) while adjusting the uptake and reception of dopamine. Much of the
cortex and limbic system changes with learning as well, but striatal changes are particu-
larly central when it comes to the motivated learning of novel goal-​directed sequences.
My favourite example of brain change underlying normal, motivated learning concerns
sexual attraction, romantic love, and mating. According to Fisher et al. (2002), “increased
levels of central dopamine contribute to the lover’s focused attention on the beloved and
the lover’s tendency to regard the beloved as unique.” Burkett and Young (2012) go on
to state the following: “Like with drugs of abuse, mesolimbic dopamine [which ends up
in the striatum] is a major contributor to the formation of pair bonds in prairie voles
and particularly in the nucleus accumbens region.” Moreover, just as in addiction, “when
these early interactions with the object of addiction produce rewarding outcomes, dopa-
mine is released in the nucleus accumbens, which acts to increase the salience of incentive
cues that predict the reward.”
The neural and behavioral parallels between love and addiction are far more complex
than indicated by this brief summary, but it is clear that dopamine-​mediated striatal acti-
vation is a common denominator. In fact, these authors use the phrase “partner addiction,”
and that certainly seems to capture the difficulties endured by many couples. Conversely,
people who have struggled with substance addiction often describe their attraction to a
particular substance as feeling like being in love. But are the brain changes that mediate
falling in love, and even mating for some period of time, permanent or impermanent?
Given current divorce rates, we might assume (and hope) that brains continue to alter
themselves when people fall out of love, allowing them to fall deeply and genuinely in love
with new partners. This issue leads directly to the second reason to stop calling addiction
a disease.
The second reason is that brain changes in addiction are not permanent; and brain
changes induced by (repeated) experience make more sense interpreted as developmental
  181

Brain changes over time 181

changes than as pathology. There has been a great deal of recent work on the ongo-
ing plasticity of synaptic patterns underlying trauma, dysfunctional learning habits
(Arrowsmith-​Young 2012), damage due to strokes and other physical insults, and neuro-
motor conditions including obsessive-​compulsive disorder and even Parkinson’s Disease
(see reviews by Schwartz and Begley 2002; Doidge 2007, 2015). With respect to addic-
tion, Connolly et al. (2013) showed that the reduction of grey matter volume in specific
regions of the prefrontal cortex, thought to progress with the length of addiction, reversed
over several months of abstinence. Grey matter volume returned to a normal baseline
level within six months to a year of abstinence (from heroin, cocaine, and alcohol), and
similar results have been found by other studies as well. Moreover, grey matter volume in
these regions continued to increase, beyond the normal baseline level (recorded for peo-
ple who’ve never been addicted)—​a fascinating result that seems related to the increase in
top-​down cognitive control or “insight” required for sustained abstinence.
But even if the brain changes underlying addiction are not permanent, might we not
still see addiction as a curable or temporary disease state, like the flu or common cold?
(NIDA consistently defines addiction as a chronic brain disease, but that’s hardly the last
word on the matter.) To resolve this issue convincingly, we should characterize the man-
ner in which the brain changes out of addiction. Does it return to a previous healthy state,
like an organ returning to homeostasis, or does it continue to change, such that it acquires
a structure that allows it to function in some novel way, overriding the longstanding func-
tional habits that got its user into so much trouble?
The answer is beautifully captured in a metaphor provided by Alvaro Pascual-​Leone,
a renowned neuroscientist who helped cultivate our current understanding of neuro-
plasticity, recounted at length by Doidge (2007). To paraphrase, brain change is like
shaping and reshaping a hunk of Play-​Doh. Plasticity is crucially different from elastic-
ity. The brain does not spring back to its former shape, like a rubber band, after it has
been structurally altered. Rather, like Play-​Doh, its shape is always novel. Each new
configuration is a product of what has gone before, and its shape always bears traces of
its previous structure. Thus, even if we take a hunk of Play-​Doh in the shape of a cube,
mess it up, and then reshape it as a cube, it will not be the same cube it was previously.
For Pascual-​Leone, “even when a patient with a neurological or psychological problem
is ‘cured,’ that cure never returns the patient’s brain to its pre-​existing state” (Doidge
2007, p. 208).
Much more can be said about brain change, but the critical point is that neuroplasticity
has particular properties: it is a growth process, like the spread of ivy on a stone wall. New
configurations are always and only modifications of previous configurations, because
brain change builds on itself. It is self-​perpetuating. Unlike other organs, the brain does
not return to some normative state after its shape has been altered. That’s why addiction
can’t be seen as a disease—​even a curable one. Each brain finds its own (relatively) stable
structure at each phase of its development—​a product of experience in the context of its
pre-​existing structure—​and then it goes on to find a new stable structure with further
development. And so on. There is no norm to return to. In this sense, the word “recovery”
182  

182 Choice in addiction

does not accurately describe the end of addiction. Recovery is what a rubber band does
after it has been stretched. Rather than recover, the brain develops out of addiction, modi-
fying its pre-​existing synaptic patterns even while it discovers and consolidates new syn-
aptic patterns, allowing for the pursuit of entirely different goals.
However, it should be noted that addiction, viewed as a developmental outcome, can
never be entirely outgrown. Because developmental trajectories never completely out-
run their origins, there is no such thing as a clean slate. This general principle may help
explain why former addicts remain sensitive to the stimuli that once provoked them to
dive for their wallet, pipe, bottle, or syringe. When times get tough or negative emo-
tions knock healthy mental habits out of balance, addiction can return. Like riding a bike,
deeply learned psychological habits never disappear completely.
The fact that addiction changes the brain does not make it a disease. It makes it a bio-
logical as well as psychological condition, based on an unusually strong (but not “unnat-
ural”) connection between hedonic experience and neural restructuring, facilitated by
dopamine’s pull for the immediate. Instead of ignoring addiction-​related brain changes,
I suggest that we look at them more closely. And since choice itself remains a mystery to
philosophers as well as psychologists, let’s hope that neuroscience can teach us something
important about choice through the insights it has gained about addiction.

5 Conclusion
Based on the temporal properties of dopamine-​induced craving and the neural changes
that result from repeated dopamine enhancement, addiction and abstinence are not
choices made freely and flexibly. Yet, rather than lead to the notion of disease, these neu-
ral lessons suggest an approach to addiction compatible with Ainslie’s model of inter-
temporal bargaining. Ainslie (2001) shows how temporal unevenness in assessing reward
gives rise to a bargaining process between different “selves” or “positions.” Each position
takes the perspective of a different point in time—​the immediate present, the anticipated
future, or something in between—​and choice arises from bargaining among these posi-
tions. Addiction is a consequence of poor bargaining, says Ainslie, and I think Heyman
would agree. Heyman also highlights competing temporal perspectives, leading to “con-
flicting motives that are experienced as ambivalence” (Heyman 2009, p.  172), and he
recommends “conscious self-​reflection” to achieve the best outcome. The best outcome
is, of course, quitting (or abstaining), or (at least for some individuals) finding a level
of controlled use that no longer threatens the health or happiness of oneself or others.
Such an outcome is captured by the phrase, “intertemporal cooperation” (Ainslie 2001,
p. 104): lasting goodwill between the capricious urges of the current self and the realistic
concerns of the future self.
The concepts of intertemporal bargaining and intertemporal cooperation point to a
logical and deeply humanistic perspective for understanding the challenges faced by
addicts every day. Thus, they provide a valuable platform for making sense of “choice” in
addiction. However, all of us—​addicts and non-​addicts alike—​engage in some form of
bargaining between immediate rewards and long-​term consequences whenever we make
  183

Conclusion 183

choices. We make choices at the juncture between urges for immediate gratification and
more thoughtful considerations for long-​term outcomes.
What I want to emphasize is that we are all—​addicts, lovers, and those whose lives are
more boring and possibly more pleasant than either—​subject to some form of intertem-
poral bargaining when we make choices. Because dopamine-​enhanced synaptic changes
and now appeal impact on our decisions, we are none of us rational. Moreover, every
individual has a tendency to make choices—​for immediate versus delayed rewards—​that
fit a certain mathematical pattern, a “discounting curve” as it’s called. That means that we
all tend to pick immediate rewards of questionable value against delayed rewards of more
objective value … at a certain ratio. We are all impulsive at times. Except perhaps for those
rare individuals who always think before acting. (And I would argue that most or all of
us are even to some extent compulsive, when performing over-​rehearsed actions such as
checking that the car is locked or the lights have been turned off.)
You might stop at the orange light most of the time. But there will be the odd occasion
when you race through the intersection, for the thrill, to make your appointment on time,
or simply because you’re tired of doing what you’re supposed to be doing. You might do
that one time in five, or one time in ten, or one time in two when you’ve had a drink or
a fight with your partner. Or won a lottery that very morning. Or recently visited Los
Angeles. If all your choices fall somewhere between a clear-​minded insight and a thought-
less plunge, what should we call them? Should we even call them choices? Or should we
see you as being in the throes of a disease process with greater or lesser severity?
No, I think we should call them choices. But we should recognize that those choices—​
all choices—​are hatched between two opponents in an interminable tug of war:  now
appeal versus conscientiousness. If either of those forces dictated our behavior with com-
plete authority, then there would be no choice, no freedom, whatsoever. We would simply
respond to the requirements of one or another rule. Either pursue the most salient goal or
compute the best net outcome. Then choice would be nonexistent.
But we humans aren’t slaves, either to salience or to insight. We do make choices, and
those choices have consequences for future outcomes, no matter how much they were
triggered by momentary cues. And because all human choices are compromises between
immediate gratification and longer-​term gains, they fall out along a continuum of ration-
ality. None are perfect. What that means is that addictive choices are simply more extreme
than other, more normal, choices. They gravitate toward the “immediate” or “impulsive”
end of the scale. But they are still choices. They are cognitive acts that sometimes lead to
relief or pleasure, sometimes to suffering or even disaster, most often to some combina-
tion. Yet, as with other human acts, their consequences don’t vanish into thin air. Rather,
their consequences can help give rise to better choices next time around.

References
Ainslie, G. (2001). Breakdown of Will. Cambridge: Cambridge University Press.
Alexander, B.K. (2008). The Globalisation of Addiction: A Study in Poverty of the Spirit. New York:
Oxford University Press.
184  

184 Choice in addiction

Arrowsmith-​Young, B. (2012). The Woman who Changed her Brain. And Other Inspiring Stories of
Pioneering Brain Transformation. New York: Free Press.
Bechara, A. and van der Kooy, D. (1992). A single brain substrate mediates the motivational effects of
both opiates and food in non-​deprived, but not in deprived animals. Behavioral Neuroscience, 106,
351–​63.
Benningfield, M.M., Blackford, J.U., Ellsworth, M.E., Samanez-​Larkin, G.R., et al. (2014). Caudate
responses to reward anticipation associated with delay discounting behaviour in healthy youth.
Developmental Cognitive Neuroscience, 7, 43–​52.
Berridge, K.C. and Aldridge, J.W. (2008). Decision utility, the brain, and pursuit of hedonic goals.
Social Cognition, 26, 621–​46.
Bickel, W.K. and Marsch, L.A. (2001). Toward a behavioral economic understanding of drug
dependence: Delay discounting processes. Addiction, 96, 73–​86.
Burkett, J.P. and Young, L.J. (2012). The behavioral, anatomical and pharmacological parallels between
social attachment, love, and addiction. Psychopharmacology, 224(1), 1–​26.
Connolly, C.G., Bell, R.P., Foxe, J.J., and Garavan, H. (2013). Dissociated grey matter changes with
prolonged addiction and extended abstinence in cocaine users. PLOS ONE, 8(3), e59645.
Dodd, M.L., Klos, K.J., Bower, J.H., Geda, Y.E., Josephs, K.A., and Ahlskog, J.E. (2005). Pathological
gambling caused by drugs used to treat Parkinson disease. Archives of Neurology, 62, 1377–​81.
Doidge, N. (2007). The Brain the Changes Itself: Stories of Personal Triumph from the Frontiers of Brain
Science. New York: Penguin.
Doidge, N. (2015). The Brain’s Way of Healing: Remarkable Discoveries and Recoveries from the Frontiers
of Neuroplasticity. New York: Penguin.
Fisher, H.E., Aron, A., Mashek, D., Hafang, L, and Brown, L.L. (2002). Defining the brain systems of
lust, romantic attraction, and attachment. Archives of Sexual Behavior, 31(5) , 413–​19.
Frijda, N.H. (1986). The Emotions. Cambridge: Cambridge University Press.
Goldstein, R.Z. and Volkow, N.D. (2002). Drug addiction and its underlying neurobiological
basis: neuroimaging evidence for the involvement of the frontal cortex. American Journal of
Psychiatry, 159, 1642–​52.
Goldstein, R.Z. and Volkow, N.D. (2011). Dysfunction of the prefrontal cortex in addiction:
neuroimaging findings and clinical implications. Nature Reviews Neuroscience, 12, 652–​69.
Heyman, G.M. (2009). Addiction: A Disorder of Choice. Cambridge, MA: Harvard University Press.
Hurd, Y.L., McGregor, A., and Pontén, M. (1997). In vivo amygdala dopamine levels modulate cocaine
self-​administration behavior in the rat: D1 dopamine receptor involvement. European Journal of
Neuroscience, 9, 2541–​48.
Izard, C.E. (1991). The Psychology of Emotion. New York: Plenum.
Koob, G. F. (2009). Dynamics of neuronal circuits in addiction: reward, antireward, and emotional
memory. Pharmacopsychiatry, 42(Suppl. 1), S32–​41.
Nader, K. and van der Kooy, D. (1997). Deprivation state switches the neurobiological substrates
mediating opiate reward in the ventral tegmental area. Journal of Neuroscience, 17, 383–​90.
Nestler, E.J. (2005). Is there a common molecular pathway for addiction? Nature Neuroscience, 8,
1445–​49.
Nutt, D.J., Lingford-​Hughes, A., Erritzoe, D., and Stokes, P.R.A. (2015). The dopamine theory of
addiction: 40 years of highs and lows. Nature Reviews Neuroscience, 16, 305–​12.
Peele, S. (2000). What addiction is and is not: the impact of mistaken notions of addiction. Addiction
Research, 8, 599–​607.
Risinger, R.C., Salmeron, B.J., Ross, T.J., Amen, S.L., Sanfilipo, M., Hoffmann, R.G., et al. (2005).
Neural correlates of high and craving during cocaine self-​administration using BOLD fMRI.
Neuroimage, 26, 1097–​108.
  185

Conclusion 185

Robinson, T.E. and Berridge, K.C. (1993). The neural basis of drug craving: An incentive-​sensitization
theory of addiction. Brain Research Reviews, 18, 247–​91.
Robinson, T.E. and Berridge, K.C. (2003). Addiction. Annual Review of Psychology, 54, 25–​53.
Rosenkranz, A.J. and Grace, A.A. (2002). Dopamine-​mediated modulation of odour-​evoked amygdala
potentials during pavlovian conditioning. Nature, 417, 282–​87.
Schaler, J.A. (2000). Addiction is a Choice. Chicago: Open Court.
Schultz, W. (2007). Behavioral dopamine signals. Trends in Neuroscience, 30, 203–​10.
Schwartz, J.M. and Begley, S. (2002). The Mind and the Brain: Neuroplasticity and the Power of Mental
Force. New York: HarperCollins.
Skinner, B.F. (1948). Walden Two. Indianapolis: Hackett.
Tanaka, S.C., Doya, K., Okada, G., Ueda, K., Okamoto, Y., and Yamawaki, S. (2004). Prediction
of immediate and future rewards differentially recruits cortico-​basal ganglia loops. Nature
Neuroscience, 7, 887–​93.
Tindell, A.J., Berridge, K.C., Zhang, J., Peciña, S., and Aldridge, J.W. (2005). Ventral pallidal neurons
code incentive motivation: amplification by mesolimbic sensitization and amphetamine. European
Journal of Neuroscience, 22, 2617–​34.
Volkow, N.D. and Fowler, J.S. (2000). Addiction, a disease of compulsion and drive: Involvement of the
orbitofrontal cortex. Cerebral Cortex, 10, 318–​25.
Wittmann, M., Leland, D.S., and Paulus, M.P. (2007). Time and decision making: differential
contribution of the posterior insular cortex and striatum during a delay discounting task.
Experimental Brain Research, 179, 643–​53.
Wyvell, C.L. and Berridge, K.C. (2000). Intra-​accumbens amphetamine increases the conditioned
incentive salience of sucrose reward: enhancement of reward “wanting” without enhanced “liking”
or response reinforcement. Journal of Neuroscience, 20, 8122–​30.
186

Chapter 11

Assessing drug choice in human

addiction: Costs, benefits, and findings
from current research paradigms
Scott J. Moeller
Rita Z. Goldstein

Abstract
Individuals with drug addiction pursue and choose drug-related reinforcers
over other pleasant, non-drug reinforcers, and the extent of such drug-
related choice often correlates with addiction severity. However, assessing
the propensity to choose drugs in human addiction poses special challenges,
including issues of ethics and generalizability. This chapter first provides an
overview of classical paradigms of drug choice and cue reactivity, which
are often employed in active users and abstinent users respectively, and
describes some of their strengths and limitations. In an attempt to address
these limitations, we developed and implemented recent behavioral and
neuroimaging paradigms that model actual choice behavior while using
drug-related stimuli (pictures) instead of actual drugs. We suggest that such
simulated drug choice can reveal novel neurocognitive markers inaccessible
with other paradigms, helping to clarify the addiction process and
potentially predict key clinical outcomes.

1 Introduction
A core tenet of drug addiction is that addicted individuals pursue and choose drugs and
drug-​related stimuli over other positive reinforcers; uncontrolled drug-​taking behavior
persists despite reduced pleasure derived from the drug and even in the face of cata-
strophic personal, social, and legal consequences (Goldstein and Volkow 2011; Volkow
et al. 2011). An understanding of this reward imbalance, which forms part of the basic
phenomenology of the addiction process, is vital for developing appropriate research
models and paradigms. Such paradigms can be readily accomplished in animal stud-
ies. In these studies, initially drug-​naïve animals are given a choice between drugs or
another potent alternative reinforcer (Ahmed et al. 2013; Banks and Negus 2012; Lenoir
et al. 2013; Nader and Banks 2014). Yet an important question persists regarding how to
  187

Drug choice self-administration paradigms 187

effectively model drug-​seeking behavior in human drug addiction, as clinical research in

drug-​addicted individuals poses many additional challenges. In particular, it is often ethi-
cally proscribed to provide treatment-​seeking or abstaining addicted individuals, who
express a desire to curb or eliminate their problematic drug use, the option of receiving
a drug infusion. Accordingly, research in these treatment-​seeking or abstaining popula-
tions often relies on tasks of drug-​cue reactivity. Although these tasks effectively measure
biased processing of drug-​related stimuli, biased attention toward drugs is only one fac-
tor in the eventual choice to seek or use them. Thus, important variance in drug-​related
choice may be missed.
The goal of the current chapter is to review several paradigms available for use in human
research, which broadly aim to predict whether addicted individuals will ultimately use
drugs outside the laboratory. We do not aim to review any particular paradigm compre-
hensively; rather, we provide a broad overview of a number of different paradigms and
review select (mostly recent) studies within each paradigm as exemplars. We begin by dis-
cussing the most straightforward approach for modeling choice in human addiction: the
human self-​administration laboratory. These studies use sophisticated, often economically
inspired analyses to inform the conditions under which individuals will choose drugs.
Next, we discuss drug-​cue reactivity paradigms. Cue-​reactivity paradigms assess the
degree to which drug-​related cues, including actual drug paraphernalia or more abstract
cues such as words or images, capture attention. Some of these paradigms entail passive
cue exposure; others incorporate an additional executive function such as inhibitory con-
trol, including the assessment of whether addicted individuals can halt a (presumably)
pre-​potent tendency to respond to drug stimuli. We then describe a simulated drug choice
research protocol that, while not without its own limitations, tries to capture many of the
strengths of some of these prior paradigms while also trying to address some of their weak-
nesses. Throughout this review, we concentrate on human studies to highlight special con-
siderations inherent in working with a clinical population; for detailed coverage of animal
choice paradigms, we refer the reader to other highly informative reviews (Ahmed 2010;
Ahmed et al. 2013; Banks and Negus 2012; Nader and Banks 2014). An important thesis of
the current chapter is that many of the available and well-​studied paradigms, while highly
important and influential for elucidating various aspects of the addiction process, typically
involve trade-​offs that limit their impact and/​or generalizability. Simulated choice of drug-​
related stimuli, rather than actual drugs, may address some of these limitations.

2  Drug choice self-​administration paradigms

In these paradigms, addicted individuals are asked to choose between receiving a drug
infusion or an alternative reward, such as money or chocolate (Bennett et al. 2013; Donny
et al. 2004; Hart et al. 2000; Hogarth and Chase 2011; Martinez et al. 2007; Martinez et al.
2009; Stoops et al. 2010, 2011, 2012; Vosburg et al. 2010) (see Figure 11.1a). Often, a key
goal of these paradigms is to expose participants to varying drug and alternative rein-
forcer schedules to determine the precise inflection point(s) where a shift to an alternative
188  

188 Assessing drug choice in human addiction

(A) (B)

r = .47, p <.00001, n = 156

Choose a key 100

G

Percent choice of tobacco

80

vs. chocolate key

You win 1/4 of 60
a cigarette/
chocolate bar/
40
nothing

20

ITI 0
4 6 8 10 12 14 16 18 20 22
Cigarette dependence scale

Figure 11.1  Human self-administration paradigms. (A) Schematic of a human drug choice para-
digm, juxtaposing choice for cigarettes versus choice for chocolate (or a control condition, where
no reward is available). Participants press a key to indicate their choice, and feedback of their
choice is provided. (B) The higher this choice for cigarettes over chocolate, the higher the nicotine
dependence
Adapted from Hogarth, L. and Chase, H. W. (2011), Parallel goal-directed and habitual control of human drug-
seeking: implications for dependence vulnerability, Journal of Experimental Psychology General: Animal Behavior
Processes, 37 (3), 261–76, with permission from the American Psychological Association.

reinforcer becomes desirable; these metrics can then be related to addiction severity (see
Figure 11.1b). Accordingly, this approach leans heavily on the perspectives of behavio-
ral economics, which views drugs through the prism of being powerful reinforcers that,
not unlike other reinforcers, have both benefits and costs. That is, in a natural environ-
ment that offers many concurrent behavioral opportunities, an individual’s motivation
to consume a drug depends on the value of that drug compared with the values of other
potential opportunities in that environment (i.e. its relative value); a substance use dis-
order, then, entails an abnormal overvaluation of the drug’s relative value (Plebani et al.
2012) and/​or undervaluation of the relative value of other non-​drug rewards (Goldstein,
Tomasi, Alia-​Klein, et al. 2007).
Consistent with this framework, the availability and escalation of alternative reinforcers
has been shown to reduce cocaine self-​administration (Donny et al. 2004; Hart et al. 2000;
Stoops et al. 2010). This shift to alternative reinforcers is accentuated when such reinforc-
ers are easier to obtain (Stoops et al. 2012) or more likely to be available (Vosburg et al.
2010). An opposite effect is achieved by increasing the attractiveness or salience of the drug
reward. For example, an immersive virtual reality environment, during which cigarette
smokers were exposed to the sights and smells of cigarettes, increased craving for cigarettes
as indexed by elevated Omax (i.e. the maximum amount the participants were willing to
pay for cigarettes) and Breakpoint (i.e. the price at which choosing cigarettes is prohibitive
  189

Drug-cue reactivity tasks 189

and completely eliminated) (Acker and MacKillop 2013). Similarly, relative to neutral cues,
exposure to alcohol cues prior to the critical choice phase increased both craving and the
monetary amount heavy drinkers were willing to spend on alcohol (MacKillop, O’Hagen,
et al. 2010). One can also vary the attractiveness of the drug through levels of deprivation
(e.g. withdrawal). For example, a higher degree of smoking deprivation (12 hours versus
one hour without a cigarette) was associated with increased Omax and various other indi-
cators of cigarette demand in habitual cigarette smokers (MacKillop et  al. 2012). These
informative tasks remain to be undertaken in other drug use disorders (e.g. heroin).
Interestingly, some choice studies have also incorporated a neuroimaging component,
enabling researchers to measure drug choice in correlation with neurobiological markers
of drug-​associated impairments. In a line of work using positron emission tomography
(PET), the choice for cocaine over money in cocaine abusers correlated with mark-
ers of compromised dopamine neurotransmission (e.g. blunted striatal dopaminergic
release to a stimulant challenge and lower striatal dopamine D1 receptor availability)
(Martinez et  al. 2007; Martinez et  al. 2009). Similar dopaminergic abnormalities were
seen in heroin abusers, although here these abnormalities were uncorrelated with her-
oin choice (Martinez et al. 2012). Glutamate signaling has also been probed (specifically,
the metabotropic glutamate receptor type 5 (mGluR5) expression in striatum); however,
despite a 20–​22 percent reduction in mGluR5 availability in the cocaine abusers as com-
pared with matched healthy controls, here again this neural alteration was uncorrelated
with the choice to self-​administer cocaine (Martinez et al. 2014). To summarize, these
studies point to the importance of dopamine neurotransmission (but not glutamate) in
underlying drug-​related choice in cocaine (but not heroin) addiction. Although addi-
tional work is warranted, these paradigms promise to offer important clarity into the
underlying neurobiology of human drug choice.

2.1  Limitations of choice self-​administration studies

Despite the clarity and elegance of these paradigms, ethical constraints typically preclude
their implementation in treatment seekers or abstainers. This issue presents generalizabil-
ity concerns because fundamental differences in drug-​related behavior and/​or underlying
neural circuitry may exist between active users and treatment seekers. These differences
could include, among others, treatment motivation (Levesque et al. 2007), recent drug use
that may prime further use (Donny et al. 2004), the expectation of receiving an imminent
drug reward (Wilson et al. 2004, 2012, 2013), and/​or functional changes attributable to
abstinence (Moeller, Tomasi, et  al. 2012; Nestor et  al. 2011; Volkow et  al. 2001; Wang
et al. 2004).

3  Drug-​cue reactivity tasks
Due in part to these limitations, a major thrust in human addiction research has been to study
the phenomenon of cue reactivity. These paradigms expose human drug abusers to stimuli
associated with their respective drug of abuse (e.g. pictures, words, paraphernalia, imagery,
190  

190 Assessing drug choice in human addiction

etc.). In the more passive instantiation of these paradigms, participants are instructed to view
passively or experience the particular cue. Matched non-​drug stimuli in the same sensory
modality, which are often neutral in valence and arousal, typically serve as a within-​subjects
control condition to be subtracted from the drug condition. This computation yields a drug
minus neutral contrast, which can then be compared between study groups (e.g. addicted
individuals versus controls, heavily dependent individuals versus non-​dependent social
users, etc.). These paradigms can be accomplished behaviorally or while also incorporat-
ing functional neuroimaging (Jasinska et al. 2014). Although variability exists, meta-​analytic
studies, which quantitatively aggregate the results of multiple studies, have demonstrated that
the neural correlates of cue reactivity often include the dorsal and ventral striatum, orbito-
frontal cortex (OFC), anterior cingulate cortex (ACC), superior and inferior frontal gyri,
amygdala, precuneus, and posterior cingulate (Chase et  al. 2011; Engelmann et  al. 2012;
Kuhn and Gallinat 2011; Schacht et al. 2013). Other regions that are engaged during cue
reactivity include those that reflect the sensory and motor representations needed to actu-
ally consume drugs, such as the supplementary motor area and cerebellum (Yalachkov et al.
2010; Yalachkov et al. 2013). Activations in these same regions have further correlated with
markers of addiction severity and treatment success (Jasinska et al. 2014).
A step beyond passive viewing reactivity tasks are paradigms that instruct addicted
individuals to engage actively with the drug cues, often with the goal of implement-
ing top-​down inhibition, self-​regulation, or a targeted cognitive function (e.g. working
memory). The central logic of such paradigms is that the elevated salience of drug cues,
as demonstrated in passive viewing tasks, will capture attention in addicted individuals
and will interfere with their ability to exert inhibitory control or perform a demanding
cognitive task. For example, participants may be asked to ignore the drug-​related con-
tent of a stimulus and instead respond to one of its peripheral characteristics (e.g. border
color, number of horizontal lines) (Fadardi and Cox 2009; Luijten et al. 2011). They may
also be asked to regulate reactivity to the cue itself by exerting top-​down cognitive con-
trol (Brody et al. 2007; Hartwell et al. 2011; Kober et al. 2010).
An especially well-​studied paradigm in this “active” cue reactivity category is the drug
Stroop task, an emotional Stroop task variant targeted to substance users. Drug Stroop
tasks typically ask study participants to press with speed and accuracy for the ink color of
drug or neutral words while ignoring their semantic content (see Figure 11.2a). Studies
using such tasks have found that addicted individuals are slower in responding to drug-​
associated stimuli relative to control stimuli, especially under conditions of deprivation or
in individuals with more severe drug use (meta-​analysis: Cox et al. 2006). As pertains to
the former, drug Stroop paradigms support the hypothesis that a state of drug deprivation
potentiates failed inhibition. In one study, smokers exhibited more smoking-​Stroop inter-
ference when deprived of a cigarette for 13–​16 hours than when deprived for one hour
(Canamar and London 2012). As pertains to the latter, a cocaine-​word Stroop was admin-
istered to 50 stimulant-​dependent individuals, 27 non-​dependent recreational cocaine
users, and 52 healthy control participants during functional magnetic resonance imaging
(fMRI) (Smith et al. 2014). Behaviorally, only the dependent users showed an attentional
  191

Drug-cue reactivity tasks 191

(a)

crack razor dealer desk

(b)
100

50
GLM Contrast Activation

–50

–100

–150

–200

Controls Recreational SDIs

Group

(c)

X = 16 Y = 48 Z = –4
Figure 11.2 Attention bias paradigms. (A) Schematic of a drug-Stroop task, during which partici-
pants are asked to respond to the ink color of a presented word while ignoring its semantic con-
tent. In this example, “crack,” “razor,” and “dealer” are drug-related words; “desk” is a neutral
word. (B–C) Recreational users, but not dependent users or healthy controls, displayed decreased
activation to cocaine minus neutral words in the right anterior cingulate and orbitofrontal corti-
ces, indicating this task’s ability to discriminate pathological from recreational drug use
Figure 11.2b and c reproduced with permission from Smith, D. G. et al. (2014), Enhanced orbitofrontal cortex
function and lack of attentional bias to cocaine cues in recreational stimulant users, Biological Psychiatry, 75 (2),
124–31. Copyright © 2014 Society of Biological Psychiatry. Published by Elsevier Inc.

bias during the task. In contrast, recreational users did not differ from controls, indicating
that this bias occurs especially in individuals with more pathological drug use patterns.
Interestingly, however, brain responses during the task showed a different pattern of
results. Specifically, relative to the other two study groups, the recreational users displayed
192  

192 Assessing drug choice in human addiction

deactivation of the OFC and ACC in response to the drug cues (see Figure 11.2b–​c), an
effect that was interpreted as potentially protective in this recreationally using group (i.e.
perhaps helping to prevent a transition to addiction) (Smith et al. 2014). Nevertheless,
other studies have indicated that, when task performance is matched between addicted
individuals and controls, ACC differences indeed emerge between the groups (Goldstein,
Alia-​Klein, et al. 2009). Future research is thus needed to explore this ACC–​Stroop rela-
tionship more closely (e.g. is ACC deactivation in this context serving as a compensa-
tory mechanism to maintain comparable performance between groups?). Importantly,
the drug Stroop also predicts treatment outcomes. In one study, dorsal ACC response
during a cocaine Stroop task positively correlated with prospective cocaine use over three
months following detoxification (Marhe et al. 2013).

3.1  Limitations of drug-​cue reactivity tasks

Cue reactivity paradigms are useful and important for assessing the relative value or sali-
ence of drug stimuli that might indicate an inclination to use. Nevertheless, such paradigms
remain an important step removed from choosing to seek out or use a particular drug. For
example, while active users are more likely to seek out drugs than abstinent users, it is the
abstinent users who often display the highest activation during cognitive and emotional
tasks in the rostral ACC (Connolly et al. 2012; Moeller, Tomasi, et al. 2012; Nestor et al.
2011). Because this rostral ACC region is implicated in personal relevance (van der Meer
et al. 2010) and the resolution of emotional conflict (Etkin et al. 2011; Kanske and Kotz
2011), this activation to drug cues could reflect a state of elevated vigilance that helps the
abstinent addicted individual avoid drug-​related contexts or effectively manage craving. In
contrast to cue reactivity, the choice to use drugs, beyond attending to drug stimuli, requires
an active decision-​making process. For this reason, choice depends on additional variables
such value, availability, and cost (Padoa-​Schioppa 2011). Choice can also be governed by
factors beyond emotional reactivity or drug craving, such as habits (Everitt et al. 2008).

4  Simulated drug choice
In recognition of these multiple challenges, several years ago we developed behavioral
tasks to assess simulated drug choice in individuals with drug addiction. We reasoned
that such tasks have the potential to capture strengths of both self-​administration para-
digms and cue-​reactivity tasks. First, these simulated choice tasks invoke active decision-​
making: beyond attention processes, participants execute a drug-​related choice. Second,
because no drug is actually administered, these tasks are safe to administer even to abstain-
ers, treatment seekers, or younger individuals without a history of problematic drug use.
Our lab developed two different versions of the simulated choice task—​one with explicit
task contingencies and one with more probabilistic task contingencies—​that reflect
complementary approaches (see Figure  11.3a, c). On both versions, study participants
indicate their choice for viewing images of cocaine/​people using cocaine versus standard-
ized pleasant, unpleasant, and neutral images (e.g. babies, disfigurement, and household
   193
(A) (B) 400
350

Total Button Presses

300
*
250
200
150
100
50
0
Onset 5 sec Pleasant Unpleasant Neutral Cocaine
Picture Category

Cocaine Active

(C) (D) Cocaine Abstinent

30 Control

(c) 25
*

Total Selections
20

15

10

5
Deck1 Deck 2 Deck 3 Deck 4
0
Pleasant Unpleasant Neutral Cocaine
Picture Category
Figure 11.3  Simulated drug choice paradigms. (A) Schematic of a simulated drug choice task with “explicit” contingencies (side-by-side picture display); on this task,
(B) actively using cocaine-dependent individuals chose the most cocaine images for viewing, followed respectively by abstinent dependent individuals and healthy controls.
(C) Schematic of a simulated drug choice task with “probabilistic” contingencies (upside-down cards, where deck identities are not fully known); on this task and similarly to
the “explicit” task, (D) actively using cocaine-dependent individuals chose the most cocaine images for viewing, followed respectively by abstinent dependent individuals and
healthy controls. Also on this task, there was an inverse pattern of responding among the groups for pleasant images.
Figure 11.3a and c adapted with permission from Moeller, S. J. et al. (2009), Enhanced choice for viewing cocaine pictures in cocaine addiction, Biological Psychiatry, 66 (2), 169–76. Copyright © 2009 Society
of Biological Psychiatry. Published by Elsevier Inc.
Figire 11.3b and d adapted with permission from Moeller, S. J. et al. (2010), Impaired insight in cocaine addiction: laboratory evidence and effects on cocaine-seeking behaviour, Brain, 133 (Pt 5), 1484–93.
Copyright © 2010, Oxford University Press.
194  

194 Assessing drug choice in human addiction

objects, respectively) (Lang et al. 2008). During the “explicit task,” participants choose via
continued button pressing between two fully visible side-​by-​side images from different
picture categories. Choice for a desired image enlarges this chosen image to cover the
screen fully, which participants can view for the trial duration of 5000 milliseconds (ms)
by continued button pressing; 500 ms of non-​response, however, returns the side-​by-​side
image display. After each trial, a new trial with new images ensues. Button pressing (i.e.
“working”) for images is an important design feature of this task, meant to echo classical
animal drug choice studies. We typically analyze the total number of button presses exe-
cuted for each picture type (pleasant, unpleasant, neutral, and cocaine), although other
analytic approaches have also been explored (Moeller, Beebe-​Wang, et al. 2013).
On the “probabilistic task,” participants choose via a single button press to view pictures
hidden under flipped-​over cards, arranged in four decks. Immediately after choosing
from a particular “deck,” an image is revealed that covers the entire screen for 2000 ms of
passive viewing. The images are arranged such that each deck contains a preponderance,
but not an entirety, of one type of image (e.g. pleasant). Participants complete four runs
of this task, during which the deck location of the four picture categories shifts. Due to
this shifting throughout the task, participants must learn and then relearn the location of
their preferred deck (e.g. cocaine images). We sum the total number of deck selections
across the four total task runs, separately for each of the four picture categories (pleasant,
unpleasant, neutral, cocaine). Unlike the “working” component of the explicit task, the
probabilistic task was instead meant to tap into more standard notions of choice. Thus,
from a behavioral economic perspective, the explicit choice can be thought to reflect Omax
(e.g. MacKillop, Miranda, et al. 2010). In contrast, the probabilistic choice can be thought
to reflect intensity (preference for cocaine images over the alternative images when there
is no cost associated with the choice).
Our initial report of these simulated choice tasks showed that cocaine abusers chose
to view more cocaine images than healthy controls on both tasks, consistent with our
hypotheses. On the probabilistic task specifically, choice for cocaine pictures also came
at the expense of choice for pleasant pictures:  drug choice was increased, while pleas-
ant choice was decreased. Interestingly, the specific cocaine minus pleasant contrast also
correlated with more frequent cocaine use over the past 30 days (Moeller et al. 2009). In
a subsequent study, this same cocaine minus pleasant choice at baseline longitudinally
predicted drug use over the next six months in treatment seekers (Moeller, Beebe-​Wang,
et  al. 2013). Also of interest, we found that cocaine picture choice was modulated by
recent abstinence:  those testing positive for cocaine in urine, indicating use within 72
hours of study time, chose to view more cocaine images on both tasks (and fewer pleas-
ant images on the probabilistic task) than those testing negative for cocaine in urine and
healthy controls (Moeller et al. 2010) (see Figure 11.3b, d).
We have also found that simulated drug choice is sensitive to abstinence-​related neu-
ral and genetic modulation. For the former, we studied a sample of treatment-​seeking
cocaine abusers at a baseline study session and again at a six-​month follow-​up session. At
both sessions, participants performed our simulated drug choice tasks and a salient fMRI
  195

Simulated drug choice 195

drug-​word task (Goldstein, Tomasi, Rajaram, et al. 2007; Goldstein, Alia-​Klein, et al. 2009).
Results revealed increased blood-​oxygenation-​level-​dependent (BOLD) fMRI activity in
the midbrain from baseline to follow-​up during the fMRI task, which in turn correlated
with decreased drug-​related choice on the explicit task (Moeller, Tomasi, et al. 2012). For
the latter, a larger sample of cocaine abusers was genotyped for a common single nucleo-
tide polymorphism (SNP) in the dopamine transporter gene (SLC6A3). Consistent with
prior research (Brody et al. 2006; Erblich et al. 2004, 2005; Franklin et al. 2011; Wetherill
et al. 2014), we compared individuals with the 10R/​10R genotype versus individuals with
either the 9R/​10R genotype or the 9R/​9R genotype. The 9R-​allele of this SNP has been
functionally associated with lowered tonic but increased phasic dopamine firing (van Dyck
et al. 2005), which in turn has been associated with increased reward-​related responsive-
ness (Dreher et al. 2009; Forbes et al. 2009). Results revealed the greatest drug minus pleas-
ant choice in the abusers who were carriers of a 9R-​allele and who were in acute cocaine
withdrawal (short-​term abstinence) on study day (Moeller, Parvaz, et al. 2013).

4.1  Self-​awareness
Another interesting facet of simulated drug choice, which cannot be readily accomplished
with drug-​cue reactivity tasks, is the ability to tap into deficits of insight and self-​awareness.
These neurocognitive deficits are marked by a compromised ability to self-​monitor one’s
behavior and symptoms. We have suggested that insight impairment might comprise a
core, yet previously underappreciated deficit in drug addiction that is potentially medi-
ated by abnormal functioning of select brain regions involved in self-​referential process-
ing, including the ACC (Goldstein, Craig, et al. 2009; Moeller and Goldstein 2014). This
type of deficit is not easily captured with attention bias paradigms, yet it could impor-
tantly contribute to compulsive, excessive, and disadvantageous drug use.
To assess self-​awareness in our studies, participants are instructed immediately at the
conclusion of the probabilistic choice task to self-​report their most selected picture type
(pleasant, unpleasant, neutral, or cocaine). We then compare participants’ self-​perceptions
of choice behavior with their actual choice behavior. Study participants who show cor-
respondence between these subjective and objective measures (e.g. executing the highest
number of presses for cocaine images and responding “cocaine” when asked about their
most selected picture category) are subgrouped as having intact self-​awareness; addicted
individuals lacking such correspondence (e.g. executing the highest number of presses for
cocaine images and responding “pleasant” when asked about their most selected picture cat-
egory) are subgrouped as having impaired self-​awareness. Whereas approximately 90 percent
of healthy controls showed correspondence between their choices and their self-​perceptions
of such choices, only approximately 60 percent of cocaine abusers showed such correspond-
ence (Moeller et al. 2010). Importantly, this deficit was driven by impaired self-​awareness of
cocaine choice (Moeller, Hajcak, et al. 2012). This finding suggests that drug-​choice during
this task occurred without individuals’ complete knowledge of their ongoing behavior.
A recent study in our lab explored the functional, structural, and emotional cor-
relates of this self-​awareness deficit. Cocaine abusers with and without compromised
196  

196 Assessing drug choice in human addiction

self-​awareness completed the following study procedures: (1) an fMRI color-​word Stroop

task to probe into error-​related processing (which has implications for self-​awareness
because of the need to self-​monitor behavior upon or to avoid error commission);
(2) structural MRI; and (3) the Levels of Emotional Awareness (LEAS) task to gauge indi-
viduals’ self-​awareness of their own emotional experience and awareness of the emotional
experiences of others (Lane et al. 1990). The LEAS asks participants to write open-​ended
responses to complex socioemotional scenarios (e.g. a friend receives a work-​related
award for which you were also competing). Results of this study revealed that cocaine
abusers with impaired self-​awareness, relative to healthy controls and even other abusers
whose self-​awareness was intact, had reduced rostral ACC activity during error commis-
sion, and this reduced rostral ACC activity further correlated with more frequent cocaine
use (Moeller et al. 2014). Also in this study, cocaine abusers with impaired self-​awareness,
again compared with the other two study groups, had reduced rostral ACC gray matter
integrity and LEAS performance (Moeller et al. 2014). Thus, addicted individuals whose
self-​awareness of their drug-​choice behavior was limited showed discernable functional
and structural abnormalities in a brain region that has been previously implicated in self-​
referential processing (Murray et al. 2012; van der Meer et al. 2010).
These findings of a self-​awareness deficit also raise the important question of whether
using self-​reports to predict drug-​seeking behavior and other important clinical out-
comes may be inadvisable in some addicted individuals (Goldstein, Alia-​Klein, Tomasi,
et  al. 2007; Goldstein et  al. 2008; Hester et  al. 2007; Moeller et  al. 2009; Moeller et  al.
2010). We hypothesized that, for individuals in whom self-​awareness is impaired, pre-
diction of behavior could be improved via quantitative, objective measures that do not
rely on self-​report. For this purpose, we used scalp-​recorded event-​related potentials
(ERPs), concentrating on the late positive potential (LPP) component. The LPP is a non-​
lateralized midline ERP component that appears approximately 300 ms after presentation
of motivationally salient stimuli (Hajcak et al. 2010), and it has been previously used to
index the salience of drug stimuli in addicted individuals (Dunning et al. 2011; Franken
et al. 2008). In addicted individuals with impaired choice self-​awareness on the probabil-
istic task, we hypothesized that cocaine-​elicited LPPs would positively predict simulated
cocaine choice on the explicit task. In contrast, in individuals with intact self-​awareness,
we hypothesized that self-​reports would predict simulated cocaine choice. Results sup-
ported these hypotheses: LPPs predicted choice in abusers with impaired self-​awareness,
whereas self-​reports predicted choice in abusers with intact self-​awareness (Moeller,
Hajcak, et al. 2012). Moreover, a similar pattern of effects was observed for indices of drug
use severity:  LPPs correlated with addiction severity in cocaine abusers with impaired
self-​awareness, whereas self-​reports were associated with addiction severity in abusers
with intact self-​awareness (Moeller, Hajcak, et al. 2012).

5  Future directions
An important step in the validation of simulated drug choice procedures is to move
beyond simply associating choice with self-​reported drug use. Initial progress on this
front has been achieved based upon the association between choice and objective markers
  197

Future directions 197

of recent use, such as cocaine urine status (Moeller et al. 2010). Nevertheless, urine sta-
tus is potentially confounded with other factors such as craving and withdrawal, and
it therefore does not conclusively establish simulated drug choice as a model of actual
drug choice. It will be important to administer the simulated choice tasks in conjunc-
tion with an actual self-​administration paradigm to ensure that the two measures cohere.
Such coherence is indeed anticipated based upon the very high correlations that have
been reported between hypothetical purchases of alcohol and actual purchases of alcohol
(Amlung et al. 2012). It would also be interesting to test whether drug-​related choice is
modulated by laboratory challenges, such as drug infusions, drug cues, or stress.
Another important future direction is to design an fMRI task that can directly inter-
rogate the neural correlates of simulated drug choice. We anticipate involvement of the
regions that assign a decision value to options under consideration, and that then compare
these values to make a choice (Padoa-​Schioppa 2011; Rangel and Hare 2010; Rushworth
et al. 2011). A large number of studies, using different species and techniques, consistently
link computation of decision values during choice with activity in the medial OFC/​ven-
tromedial prefrontal cortex (VMPFC), which typically includes Brodmann Areas (BAs)
32, 10, and 11 (Padoa-​Schioppa 2011; Rangel and Hare 2010; Sokol-​Hessner et al. 2012;
Wallis and Kennerley 2010). We also expect an important contribution of the dorsolateral
prefrontal cortex (DLPFC), typically defined as BA 9 and 46. Similarly to the VMPFC,
the DLPFC frequently tracks decision values (Kable and Glimcher 2007; Litt et al. 2011;
Plassmann et al. 2007; Plassmann et al. 2010; Sokol-​Hessner et al. 2012), but it also par-
ticipates in top-​down control during choice (Hare et al. 2009: Hare, Malmaud, et al. 2011).
Finally, simulated drug choice may engage regions that compare stimulus values to deter-
mine the optimal course of action, such as the dorsal ACC (Blair et al. 2006; Hare, Schultz,
et al. 2011). Importantly, these same regions show functional abnormalities in addicted
individuals during non-​drug-​related decision-​making tasks (e.g. ACC, OFC, and DLPFC;
Goldstein and Volkow 2011; Gowin et al. 2013).
Relatedly, although we anticipate that simulated drug choice will account for unique vari-
ance in real-​life drug choice or clinical outcomes not already accounted for by drug cue
reactivity, this hypothesis requires empirical verification. For example, one could directly
compare responding on an fMRI drug-​choice task with responding on an fMRI passive cue-​
reactivity task. Both fMRI drug choice and passive drug-​cue viewing may engage regions
relevant to cue reactivity, including the striatum (Litt et al. 2011; Martinez et al. 2011), mid-
brain (Moeller, Tomasi, et al. 2012), and insula (Naqvi et al. 2007; Naqvi and Bechara 2009).
However, fMRI drug choice might uniquely identify activations in the dorsal ACC, DLPFC,
OFC, and VMPFC. Given the functions of these regions, as described in the preceding para-
graph, activation of these regions could reflect the engagement of additional cognitive pro-
cesses, including evaluating, planning, and deciding among alternative courses of action.
Another direction, which taps into aversive processing and withdrawal effects in addic-
tion, is to examine simulated drug choice in relation to unpleasant choice. This framework
might identify other interesting brain regions or genetic factors that may be modulated
by different, yet related neurocircuitry, such as the kappa/​dynorphin system (Trifilieff
and Martinez 2013). This system appears to regulate, via tuning of striatal dopamine
198  

198 Assessing drug choice in human addiction

neurotransmission, the neurobiological processes that occur during drug withdrawal

(Carlezon et al. 2006; Gehrke et al. 2008; Muschamp and Carlezon 2013; Wee and Koob
2010). For example, withdrawal from amphetamine is associated with upregulation of
prodynorphin (Horner et al. 2009). Also relevant to this aversive reactivity framework,
one could test whether addicted individuals continue choosing drugs and drug stimuli
even when doing so incurs punishment (e.g. negative evaluative feedback, loss of money,
or mild electric shock). This framework taps into continued drug use despite adverse
consequences, a core symptom of addiction. This symptom is also exemplified in animal
studies, which have shown that a small but important minority of rats continues to self-​
administer drugs despite receiving an aversive foot-​shock (Deroche-​Gamonet et al. 2004;
Everitt et al. 2008). These research directions could help clarify which individuals might
be at special risk for continuing to take drugs despite catastrophic life circumstances that
follow from the addiction process (e.g. exacerbation of a medical condition).
Finally, future studies can aim to recruit adolescents, who can complete the simulated
drug-​choice tasks and then be longitudinally followed. This type of project could enable
testing, for example, of whether enhanced simulated choice for addictive substances over
other pleasant rewards predicts a future transition to addiction. This idea is consistent
with a prior study that used a drug Stroop task: in non-​addicted undergraduate social
drinkers, an individualized alcohol Stroop correlated with a measure of alcohol use sever-
ity (Christiansen and Bloor 2014).

6 Conclusion
In this chapter, we reviewed paradigms of drug choice in human drug addiction, while
also highlighting some of the unique challenges of these clinical studies. Although the
most valid choice paradigms will always involve giving addicted individuals the oppor-
tunity to self-​administer actual drugs versus receive other valuable reinforcers, these
options raise ethical dilemmas for individuals seeking treatment or attempting to abstain.
In contrast, drug-​cue reactivity paradigms, while simplistically elegant and enormously
successful in measuring drug salience, are nonetheless suboptimal for measuring drug-​
relevant decision-​making. For these reasons, we proposed simulated drug choice to
bridge this gap. Although many questions remain, simulated drug choice could reflect an
exciting model of real-​world drug seeking in clinical populations, used as a proxy and/​
or predictor of important clinical outcomes. It will be highly interesting and important
to evaluate whether simulated drug-​choice can objectively, yet non-​invasively, identify
the individuals who may require additional therapeutic resources for achieving sustained
abstinence and better longer-​term outcomes.

Acknowledgments
This work was supported by the National Institute on Drug Abuse (grants R21DA034954
and K01DADA037452). We thank Nelly Alia-​Klein, Anna B. Konova, and Muhammad
A. Parvaz for helpful discussions on ideas presented in this chapter.
  199

Conclusion 199

Correspondence should be addressed to: Scott J. Moeller, One Gustave L. Levy Place,

Box 1230, New  York, NY 10029-​6574; Tel:  212-​241-​6231; Fax:  212-​803-​6743; scott.
moeller@mssm.edu. Correspondence may also be addressed to: Rita Z. Goldstein, One
Gustave L.  Levy Place, Box 1230, New  York, NY 10029-​6574; tel. (212) 824-​9312; fax
(212) 996-​8931; rita.goldstein@mssm.edu.

References
Acker, J. and MacKillop, J. (2013). Behavioral economic analysis of cue-​elicited craving for tobacco: a
virtual reality study. Nicotine and Tobacco Research, 15, 1409–​16.
Ahmed, S.H. (2010), Validation crisis in animal models of drug addiction: beyond non-​disordered drug
use toward drug addiction. Neuroscience and Biobehavioral Reviews, 35, 172–​84.
Ahmed, S.H., Lenoir, M., and Guillem, K. (2013), Neurobiology of addiction versus drug use driven by
lack of choice. Current Opinion in Neurobiology, 23, 581–​87.
Amlung, M.T., Acker, J. Stojek, M.K., Murphy, J.G., and MacKillop, J. (2012). Is talk “cheap?” An
initial investigation of the equivalence of alcohol purchase task performance for hypothetical and
actual rewards. Alcoholism: Clinical and Experimental Research, 36, 716–​24.
Banks, M.L. and Negus, S.S. (2012). Preclinical determinants of drug choice under concurrent
schedules of drug self-​administration. Advances in Pharmacological Sciences, article ID 281768.
Bennett, J.A., Stoops, W.W., and Rush, C.R. (2013). Alternative reinforcer response cost impacts
methamphetamine choice in humans. Pharmacology Biochemistry and Behavior, 103(3), 481–​86.
Blair, K., Marsh, AA., Morton, J., et al. (2006). Choosing the lesser of two evils, the better of two
goods: specifying the roles of ventromedial prefrontal cortex and dorsal anterior cingulate in object
choice. Journal of Neuroscience, 26 (44), 11379–​86.
Brody, A.L., Mandelkern, M.A., Olmstead, R.E., et al. (2006). Gene variants of brain dopamine
pathways and smoking-​induced dopamine release in the ventral caudate/​nucleus accumbens.
Archives of General Psychiatry, 63 (7), 808–​16.
Brody, A.L., Mandelkern, M.A., Olmstead, R.E., et al. (2007). Neural substrates of resisting craving
during cigarette cue exposure. Biolgical Psychiatry, 62 (6), 642–​51.
Canamar, C.P. and London, E. (2012). Acute cigarette smoking reduces latencies on a Smoking Stroop
test. Addictive Behaviors, 37 (5), 627–​31.
Carlezon, W.A., Jr, Beguin, C., DiNieri, J.A., et al. (2006). Depressive-​like effects of the kappa-​opioid
receptor agonist salvinorin A on behavior and neurochemistry in rats. Journal of Pharmacology and
Experimental Therapeutics, 316(1), 440–​47.
Chase, H.W., Eickhoff, S.B., Laird, A.R., and Hogarth, L. (2011). The neural basis of drug stimulus
processing and craving: an activation likelihood estimation meta-​analysis. Biological Psychiatry,
70(8), 785–​93.
Christiansen, P. and Bloor, J.F. (2014). Individualised but not general alcohol Stroop predicts alcohol
use. Drug and Alcohol Dependence, 134, 410–​13.
Connolly, C.G., Foxe, J.J., Nirenbert, J., Shpaner, M., and Garavan, H. (2012). The neurobiology of
cognitive control in successful cocaine abstinence. Drug and Alcohol Dependence, 121(1–​2), 45–​53.
Cox, W.M., Fadardi, J.S., and Pothos, E.M. (2006). The addiction-​Stroop test: theoretical considerations
and procedural recommendations. Psychological Bulletin, 132(3), 443–​76.
Deroche-​Gamonet, V., Belin, D., and Piazza, P.V. (2004). Evidence for addiction-​like behavior in the
rat. Science, 305 (5686), 1014–​17.
Donny, E.C., Bigelow, G.E., and Walsh, S.L. (2004). Assessing the initiation of cocaine self-​
administration in humans during abstinence: effects of dose, alternative reinforcement, and
priming. Psychopharmacology, 172(3), 316–​23.
200  

200 Assessing drug choice in human addiction

Dreher, J.C., Kohn, P., Kolachana, B., Weinberger, D.R., and Berman, K.F. (2009). Variation in
dopamine genes influences responsivity of the human reward system. Proceedings of the National
Academy of Sciences of the United States of America, 106(2), 617–​22.
Dunning, J.P., Parvaz, M.A., Hajcak, G., et al. (2011). Motivated attention to cocaine and emotional
cues in abstinent and current cocaine users—​an ERP study. European Journal of Neuroscience, 33(9),
1716–​23.
Engelmann, J.M., Versace, F., Robinson, J.D., et al. (2012). Neural substrates of smoking cue
reactivity: a meta-​analysis of fMRI studies. Neuroimage, 60(1), 252–​62.
Erblich, J., Lerman, C., Self, D.W., Diaz, G.A., and Bovbjerg, D.H. (2004). Stress-​induced cigarette
craving: effects of the DRD2 TaqI RFLP and SLC6A3 VNTR polymorphisms. Pharmacogenomics
Journal, 4(2), 102–​109.
Erblich, J., Lerman, C., Self, D.W., Diaz, G.A., and Bovbjerg, D.H. (2005). Effects of dopamine D2
receptor (DRD2) and transporter (SLC6A3) polymorphisms on smoking cue-​induced cigarette
craving among African-​American smokers. Molecular Psychiatry, 10(4), 407–​14.
Etkin, A., Egner, T., and Kalisch, R. (2011). Emotional processing in anterior cingulate and medial
prefrontal cortex. Trends in Cognitive Sciences, 15(2), 85–​93.
Everitt, B.J., Belin, D., Economidou, D., Pelloux, Y., Dalley, J.W., et al. (2008). Review. Neural
mechanisms underlying the vulnerability to develop compulsive drug-​seeking habits and addiction.
Philosophical Transactions of the Royal Society of London Series B: Biological Sciences, 363(1507),
3125–​35.
Fadardi, J.S. and Cox, W.M. (2009). Reversing the sequence: reducing alcohol consumption by
overcoming alcohol attentional bias. Drug and Alcohol Dependence, 101(3), 137–​45.
Forbes, E.E., Brown, S.M., Kimak, M., Ferrell, R.E., Manuck, S.B., et al. (2009). Genetic variation in
components of dopamine neurotransmission impacts ventral striatal reactivity associated with
impulsivity. Molecular Psychiatry, 14(1), 60–​70.
Franken, I.H. , Dietvorst, R.C., Hesselmans, M., et al. (2008). Cocaine craving is associated with
electrophysiological brain responses to cocaine-​related stimuli. Addiction Biology, 13(3–​4), 386–​92.
Franklin, T.R., Wang, Z., Li, Y., Suh, J.J. et al. (2011). Dopamine transporter genotype modulation of
neural responses to smoking cues: confirmation in a new cohort. Addiction Biology, 16(2), 308–​22.
Gehrke, B.J., Chefer, V.I., and Shippenberg, T.S. (2008). Effects of acute and repeated administration of
salvinorin A on dopamine function in the rat dorsal striatum. Psychopharmacology, 197(3), 509–​17.
Goldstein, R.Z. and Volkow, N.D. (2011). Dysfunction of the prefrontal cortex in addiction: neuroimaging
findings and clinical implications. Nature Reviews Neuroscience, 12(11), 652–​69.
Goldstein, R.Z., Alia-​Klein, N., Tomasi, D., et al. (2007). Is decreased prefrontal cortical sensitivity
to monetary reward associated with impaired motivation and self-​control in cocaine addiction?
American Journal of Psychiatry, 164(1), 43–​51.
Goldstein, R.Z., Tomasi, D., Alia-​Klein, N., et al. (2007). Subjective sensitivity to monetary gradients is
associated with frontolimbic activation to reward in cocaine abusers. Drug and Alcohol Dependence,
87(2–​3), 233–​40.
Goldstein, R.Z., Tomasi, D., Rajaram, S., et al. (2007). Role of the anterior cingulate and medial
orbitofrontal cortex in processing drug cues in cocaine addiction. Neuroscience, 144(4), 1153–​59.
Goldstein, R.Z., Parvaz, M.A., Maloney, T., et al. (2008). Compromised sensitivity to monetary reward
in current cocaine users: an ERP study. Psychophysiology, 45(5), 705–​13.
Goldstein, R.Z., Alia-​Klein, N., Tomasi, D., et al. (2009). Anterior cingulate cortex hypoactivations to
an emotionally salient task in cocaine addiction. Proceedings of the National Academy of Sciences of
the United States of America, 106 (23), 9453–​58.
Goldstein, R.Z., Craig, A. D., Bechara, A., et al. (2009). The neurocircuitry of impaired insight in drug
addiction. Trends in Cognitive Sciences, 13(9), 372–​80.
  201

Conclusion 201

Gowin, J.L., Mackey, S., and Paulus, M.P. (2013). Altered risk-​related processing in substance
users: imbalance of pain and gain. Drug and Alcohol Dependence, 132(1–​2), 13–​21.
Hajcak, G., MacNamara, A., and Olvet, D.M. (2010). Event-​related potentials, emotion, and emotion
regulation: an integrative review. Developmental Neuropsychology, 35(2), 129–​55.
Hart, C.L., Haney, M., Foltin, R.W., and Fischman, M.W. (2000). Alternative reinforcers differentially
modify cocaine self-​administration by humans. Behavioral Pharmacology, 11(1), 87–​91.
Hare, T.A., Camerer, C.F., and Rangel, A. (2009). Self-​control in decision-​making involves modulation
of the vmPFC valuation system. Science, 324(5927), 646–​48.
Hare, T.A., Malmaud, J., and Rangel, A. (2011). Focusing attention on the health aspects of foods
changes value signals in vmPFC and improves dietary choice. Journal of Neuroscience, 31(30),
11077–​87.
Hare, T.A., Schultz, W., Camerer, C.F., O'Doherty, J.P., and Rangel, A.(2011). Transformation of
stimulus value signals into motor commands during simple choice. Proceedings of the National
Academy of Sciences of the United States of America, 108(44), 18120–​5.
Hartwell, K.J., Johnson, K.A., Li, X., et al. (2011). Neural correlates of craving and resisting craving for
tobacco in nicotine dependent smokers. Addiction Biology, 16(4), 654–​66.
Hester, R., Simões-​Franklin, C., and Garavan, H. (2007). Post-​error behavior in active
cocaine users: Poor awareness of errors in the presence of intact performance adjustments.
Neuropsychopharmacology, 32(9), 1974–​84.
Hogarth, L. and Chase, H.W. (2011). Parallel goal-​directed and habitual control of human
drug-​seeking: implications for dependence vulnerability. Journal of Experimental Psychology
General: Animal Behavior Processes, 37(3), 261–​76.
Horner, K.A., Noble, E.S., and Lauterbach, E.C. (2009). Differential regulation of prodynophin, c-​fos,
and serotonin transporter mRNA following withdrawal from a chronic, escalating dose regimen of
D-​amphetamine. Synapse, 63(4), 257–​68.
Jasinska, A.J., Stein, E.A., Kaiser, J., Naumer, M.J., and Yalachkov, Y. (2014). Factors modulating neural
reactivity to drug cues in addiction: a survey of human neuroimaging studies. Neuroscience and
Biobehavioral Reviews, 38, 1–​16.
Kable, J.W. and Glimcher, P.W. (2007). The neural correlates of subjective value during intertemporal
choice. Nature Neuroscience, 10(12), 1625–​33.
Kanske, P. and Kotz, S.A. (2011). Emotion speeds up conflict resolution: a new role for the ventral
anterior cingulate cortex? Cerebral Cortex, 21(4), 911–​19.
Kober, H., Mende-​Siedlecki, P., Kross, E.F., et al. (2010). Prefrontal-​striatal pathway underlies cognitive
regulation of craving. Proceedings of the National Academy of Sciences of the United States of America,
107(33), 14811–​16.
Kuhn, S. and Gallinat, J. (2011). Common biology of craving across legal and illegal drugs—​a quantitative
meta-​analysis of cue-​reactivity brain response. European Journal of Neuroscience, 33(7), 1318–​26.
Lane, R.D., Quinlan, D.M., Schwartz, G.E., Walker, P.A., and Zeitlin, S.B. (1990). The Levels of
Emotional Awareness Scale: a cognitive-​developmental measure of emotion. Journal of Personality
Assessment, 55(1–​2), 124–​34.
Lang, P.J., Bradley, M.M., and Cuthbert, B.N. (2008). International Affective Picture System
(IAPS): Affective ratings of pictures and instruction manual. Technical Report A-​8. (Gainsville,
FL: University of Florida).
Lenoir, M., Augier, E., Vouillac, C., and Ahmed, S.H. (2013). A choice-​based screening method for
compulsive drug users in rats. Current Protocols in Neuroscience, Chapter 9, Unit 9 44.
Levesque, C.S., Williams, G.C., Elliot, D., et al. (2007). Validating the theoretical structure of the
treatment self-​regulation questionnaire (TSRQ) across three different health behaviors. Health
Education Research, 22(5), 691–​702.
202  

202 Assessing drug choice in human addiction

Litt, A., Plassmann, H., Shiv, B., and Rangel, A. (2011). Dissociating valuation and saliency signals
during decision-​making. Cerebral Cortex, 21(1), 95–​102.
Luijten, M., Veltman, D.J., van den Brink, W., et al. (2011). Neurobiological substrate of smoking-​
related attentional bias. Neuroimage, 54(3), 2374–​81.
MacKillop, J., O'Hagen, S., Lisman, S.A., et al. (2010). Behavioral economic analysis of cue-​elicited
craving for alcohol. Addiction, 105(9), 1599–​607.
MacKillop, J., Miranda, R., Jr., Monti, P.M., et al. (2010). Alcohol demand, delayed reward discounting,
and craving in relation to drinking and alcohol use disorders. Journal of Abnormal Psychology,
119(1), 106–​14.
MacKillop, J., Brown, C.L., Stojek, M.K., et al. (2012). Behavioral economic analysis of withdrawal-​
and cue-​elicited craving for tobacco: an initial investigation. Nicotine and Tobacco Research, 14(12),
1426–​34.
Marhe, R., Luijten, M., van de Wetering, B.J., Smits, M., and Franken, I.H. (2013). Individual
differences in anterior cingulate activation associated with attentional bias predict cocaine use after
treatment. Neuropsychopharmacology, 38(6), 1085–​93.
Martinez, D., Narendran, R., Foltin, R.W., et al. (2007). Amphetamine-​induced dopamine
release: Markedly blunted in cocaine dependence and predictive of the choice to self-​administer
cocaine. American Journal of Psychiatry, 164(4), 622–​29.
Martinez, D., Slifstein, M., Narendran, R., et al. (2009). Dopamine D1 receptors in cocaine dependence
measured with PET and the choice to self-​administer cocaine. Neuropsychopharmacology, 34(7),
1774–​82.
Martinez, D., Carpenter, K.M., Liu, F., et al. (2011). Imaging dopamine transmission in cocaine
dependence: link between neurochemistry and response to treatment. American Journal of
Psychiatry, 168(6), 634–​41.
Martinez, D., Saccone, P.A., Liu, F., et al. (2012). Deficits in dopamine D(2) receptors and presynaptic
dopamine in heroin dependence: commonalities and differences with other types of addiction.
Biological Psychiatry, 71(3), 192–​98.
Martinez, D., Slifstein, M., Nabulsi, N., et al. (2014). Imaging glutamate homeostasis in cocaine
addiction with the metabotropic glutamate receptor 5 positron emission tomography radiotracer
[(11)C]ABP688 and magnetic resonance spectroscopy. Biological Psychiatry, 75(2), 165–​71.
Moeller, S.J. and Goldstein, R.Z. (2014). Impaired self-​awareness in human addiction: deficient
attribution of personal relevance. Trends in Cognitve Sciences, 18(12), 635–​641.
Moeller, S.J., Maloney, T., Parvaz, M.A., et al. (2009). Enhanced choice for viewing cocaine pictures in
cocaine addiction. Biological Psychiatry, 66(2), 169–​76.
Moeller, S.J., Maloney, T., Parvaz, M.A., et al. (2010). Impaired insight in cocaine addiction: laboratory
evidence and effects on cocaine-​seeking behavior. Brain, 133(Pt 5), 1484–​93.
Moeller, S.J., Hajcak, G., Parvaz, M. A., et al. (2012). Psychophysiological prediction of
choice: relevance to insight and drug addiction. Brain, 135(Pt 11), 3481–​94.
Moeller, S.J., Tomasi, D., Woicik, P.A., et al. (2012). Enhanced midbrain response at 6-​month follow-​
up in cocaine addiction, association with reduced drug-​related choice. Addiction Biology, 17(6),
1013–​25.
Moeller, S.J., Beebe-​Wang, N., Woicik, P. A., et al. (2013). Choice to view cocaine images predicts
concurrent and prospective drug use in cocaine addiction. Drug and Alcohol Dependence, 130 (1–​3),
178–​85.
Moeller, S.J., Parvaz, M. A., Shumay, E., et al. (2013). Gene x abstinence effects on drug cue reactivity
in addiction: multimodal evidence. Journal of Neuroscience, 33 (24), 10027–​36.
Moeller, S.J., Konova, A. B., Parvaz, M.A., et al. (2014). Functional, structural, and emotional correlates
of impaired insight in cocaine addiction. JAMA Psychiatry, 71(1), 61–​70.
  203

Conclusion 203

Murray, R. J., Schaer, M., and Debbane, M. (2012). Degrees of separation: a quantitative neuroimaging
meta-​analysis investigating self-​specificity and shared neural activation between self-​and other-​
reflection. Neuroscience and Biobehavioral Reviews, 36(3), 1043–​59.
Muschamp, J.W. and Carlezon, W.A., Jr (2013). Roles of nucleus accumbens CREB and dynorphin in
dysregulation of motivation. Cold Spring Harbor Perspective in Medicine, 3(2), a012005.
Nader, M.A. and Banks, M.L. (2014). Environmental modulation of drug taking: nonhuman primate
models of cocaine abuse and PET neuroimaging. Neuropharmacology, 76(Pt B), 510–​17.
Naqvi, N.H. and Bechara, A. (2009). The hidden island of addiction: the insula. Trends in Neuroscience,
32(1), 56–​67.
Naqvi, N.H., Rudrauf, D., Damasio, H., and Bechara, A. (2007). Damage to the insula disrupts addiction
to cigarette smoking. Science (New York, N.Y.), 315, 531–​4. DOI: 10.1126/science.1135926.
Nestor, L., McCabe, E., Jones, J., Clancy, L., and Garavan, H. (2011). Differences in ‘bottom-​up’ and ‘top-​
down’ neural activity in current and former cigarette smokers: evidence for neural substrates which
may promote nicotine abstinence through increased cognitive control. Neuroimage, 56(4), 2258–​75.
Padoa-​Schioppa, C. (2011). Neurobiology of economic choice: a good-​based model. Annual Review of
Neuroscience, 34, 333–​59.
Plassmann, H., O’Doherty, J., and Rangel, A. (2007). Orbitofrontal cortex encodes willingness to pay in
everyday economic transactions. Journal of Neuroscience, 27(37), 9984–​88.
Plassmann, H., O’Doherty, J.P., and Rangel, A. (2010). Appetitive and aversive goal values are encoded
in the medial orbitofrontal cortex at the time of decision making. Journal of Neuroscience, 30(32),
10799–​808.
Plebani, J.G., Ray, L.A., Morean, M.E., et al. (2012). Human laboratory paradigms in alcohol research.
Alcoholism: Clinical and Experimental Research, 36(6), 972–​83.
Rangel, A. and Hare, T. (2010). Neural computations associated with goal-​directed choice. Current
Opinion in Neurobiology, 20(2), 262–​70.
Rushworth, M.F., Noonan, M.P., Boorman, E.D., Walton, M.E., and Behrens, T.E. (2011). Frontal
cortex and reward-​guided learning and decision-​making. Neuron, 70(6), 1054–​69.
Schacht, J.P., Anton, R.F., and Myrick, H. (2013). Functional neuroimaging studies of alcohol cue
reactivity: a quantitative meta-​analysis and systematic review. Addiction Biology, 18(1), 121–​33.
Smith, D.G., Simon Jones, P., Bullmore, E.T., Robbins, T.W., and Ersche, K.D. (2014). Enhanced
orbitofrontal cortex function and lack of attentional bias to cocaine cues in recreational stimulant
users. Biological Psychiatry, 75(2), 124–​31.
Sokol-​Hessner, P., Hutcherson, C., Hare, T., and Rangel, A. (2012). Decision value computation in
DLPFC and VMPFC adjusts to the available decision time. European Journal of Neuroscience, 35(7),
1065–​74.
Stoops, W.W., Lile, J.A., and Rush, C.R. (2010). Monetary alternative reinforcers more effectively
decrease intranasal cocaine choice than food alternative reinforcers. Pharmacology Biochemistry and
Behavior, 95(2), 187–​91.
Stoops, W.W., Poole, M.M., Vansickel, A.R., and Rush, C.R. (2011). Influence of escalating alternative
reinforcer values on cigarette choice. Behavioural Processes, 87(3), 302–​305.
Stoops, W.W., Lile, J.A., Glaser, P.E., Hays, L.R., and Rush, C.R. (2012). Alternative reinforcer response
cost impacts cocaine choice in humans. Progress in Neuropsychopharmacology and Biological
Psychiatry, 36(1), 189–​93.
Trifilieff, P. and Martinez, D. (2013). Kappa-​opioid receptor signaling in the striatum as a potential
modulator of dopamine transmission in cocaine dependence. Frontiers in Psychiatry, 4, 44.
van der Meer, L., Costafreda, S., Aleman, A., and David, A.S. (2010). Self-​reflection and the brain: a
theoretical review and meta-​analysis of neuroimaging studies with implications for schizophrenia.
Neuroscience and Biobehavioral Reviews, 34(6), 935–​46.
204  

204 Assessing drug choice in human addiction

van Dyck, C.H., Malison, R.T., Jacobsen, L.K., et al. (2005). Increased dopamine transporter availability
associated with the 9-​repeat allele of the SLC6A3 gene. Journal of Nuclear Medicine, 46(5), 745–​51.
Volkow, N.D., Chang, L., Wang, G.J., et al. (2001). Loss of dopamine transporters in methamphetamine
abusers recovers with protracted abstinence. Journal of Neuroscience, 21(23), 9414–​18.
Volkow, N.D., Baler, R.D., and Goldstein, R.Z. (2011). Addiction: pulling at the neural threads of social
behaviors. Neuron, 69(4), 599–​602.
Vosburg, S.K., Haney, M., Rubin, E., and Foltin, R.W. (2010). Using a novel alternative to drug choice
in a human laboratory model of a cocaine binge: a game of chance. Drug and Alcohol Dependence,
110(1–​2), 144–​50.
Wallis, J.D. and Kennerley, S.W. (2010). Heterogeneous reward signals in prefrontal cortex. Current
Opinion in Neurobiology, 20(2), 191–​98.
Wang, G.J., Volkow, N.D., Chang, L., et al. (2004). Partial recovery of brain metabolism in
methamphetamine abusers after protracted abstinence. American Journal of Psychiatry, 161(2),
242–​48.
Wee, S. and Koob, G.F. (2010). The role of the dynorphin-​kappa opioid system in the reinforcing effects
of drugs of abuse. Psychopharmacology, 210(2), 121–​35.
Wetherill, R.R., Jagannathan, K., Lohoff, F.W., et al. (2014). Neural correlates of attentional bias for
smoking cues: modulation by variance in the dopamine transporter gene. Addiction Biology, 19(2),
294–​304.
Wilson, S.J., Sayette, M.A., and Fiez, J.A. (2004). Prefrontal responses to drug cues: a neurocognitive
analysis. Nature Neuroscience, 7(3), 211–​14.
Wilson, S.J., Sayette, M.A., and Fiez, J.A. (2012). Quitting-​unmotivated and quitting-​motivated
cigarette smokers exhibit different patterns of cue-​elicited brain activation when anticipating an
opportunity to smoke. Journal of Abnormal Psychology, 121(1), 198–​211.
Wilson, S.J., Creswell, K.G., Sayette, M.A., and Fiez, J.A. (2013). Ambivalence about smoking and cue-​
elicited neural activity in quitting-​motivated smokers faced with an opportunity to smoke. Addictive
Behaviors, 38(2), 1541–​49.
Yalachkov, Y., Kaiser, J., and Naumer, M.J. (2010). Sensory and motor aspects of addiction. Behavioural
Brain Research, 207(2), 215–​22.
Yalachkov, Y., Kaiser, J., Gorres, A., Seehaus, A., and Naumer, M. J. (2013). Sensory modality of
smoking cues modulates neural cue reactivity. Psychopharmacology, 225(2), 461–​71.
  205

Chapter 12

The role of the insula in goal-​directed

drug seeking and choice in addiction
Nasir H. Naqvi
Antoine Bechara

Abstract
This chapter describes our current understanding of the role of the insula
in addiction. It begins by summarizing the initial work by the authors
showing the effects of insula lesions on smoking behavior. The chapter
then summarizes more recent work demonstrating a role for the insula in
affective, motivational, interoceptive, and decision-making functions that
are relevant for addiction. A theoretical model is presented in which the
insula plays a role in motivating addictive choices when there are high levels
of conflict or risk. The clinical implications of this model are discussed.

1 Background
Addiction, defined as compulsive drug use despite significant negative consequences, is a
major cause of morbidity and mortality globally (here we use the term “drug” to encom-
pass all addictive substances, including alcohol). While there are some moderately effective
treatments for addiction, there is no cure. Identifying specific neural targets for addiction
treatments should help improve their efficacy and may someday lead to a cure. There has
been an explosion of research on the neuroscience of addiction over the last three decades.
Much of this work has implicated the striatum (both ventral and dorsal), along with its
inputs from the ventromedial prefrontal cortex, amygdala and the mesolimbic dopamine
system, in a variety of appetitive motivational processes that drive drug seeking and drug
taking (Everitt and Robbins 2005; Koob and Le Moal 2001; Robinson and Berridge 2008).
A parallel line of research has highlighted the role of impairments in prefrontal cortical
systems in impulse control and decision-​making processes that normally rein in drug
seeking and drug taking (Bechara 2005; Goldstein et  al. 2004; Jentsch and Pennington
2013; Rogers et al. 1999; Liu et al. 1998). The broad clinical implication of this body of
research is that interventions that downmodulate the function of motivational systems or
that enhance the functioning of control systems should reduce addictive behavior (Potenza
et al. 2011). Despite this remarkable progress, no effective treatments have been derived
from our increasingly sophisticated understanding of the brain basis of addiction.
206  

206 The role of the insula in goal-directed drug seeking and choice in addiction

In 2007, we discovered that damage to the human insula, a brain region that had until
that point been largely overlooked by addiction researchers, led to an abrupt and profound
disruption of addiction to cigarette smoking (Naqvi et  al. 2007). To some, this finding
confirmed the notion that addiction was not so much a choice, but rather a disease that
depended critically upon the function of this brain region; that no matter how strong the
motivation either to continue or to stop using drugs, insula damage would halt it in an
obligatory fashion. This led to a number of sensationalistic pronouncements in the popular
media of a “cure” for addiction, one that would require little on the part of the addicted indi-
vidual beyond a willingness to undergo brain surgery to resect his insula (Vorel et al. 2007).
In this chapter, we present a model of the role of the insula in addiction that is con-
siderably more nuanced (and more interesting) than this rather simplistic account. We
begin by summarizing our initial findings on the effects of insula lesions on smoking
behavior, as well as more recent human lesion studies that replicate and extend our work.
We then review functional imaging and animal studies examining the role of the insula
in (1) appetitive motivational processes that drive addictive behavior; (2) interoceptive
functions that are relevant to addiction; and (3) decision-​making processes that modify
addictive behavior according to long-​term goals and risks. We then present a model that
integrates these seemingly disparate aspects of insula function, focusing on the role of
the insula in a specific goal-​directed mode of drug seeking that predominates when sub-
jective risk is high or there is a high degree of conflict between drug use and other goals.
We conclude by describing the implications of this model for understanding addiction
as a choice, and for understanding behavior change mechanisms in addiction treatment.

2  The effects of insula damage on addiction

to cigarette smoking
In our 2007 study (Naqvi et al. 2007), we retrospectively examined the effects of insula
damage on the likelihood of two outcomes: (1) quitting smoking after the onset of brain
damage; and (2) undergoing a “disruption of smoking addiction” after the onset of brain
damage, which we operationalized as not only quitting but being able to do so imme-
diately, easily, without relapsing and without a persistent urge to smoke. We identified
19 patients with insula lesions, along with 50 comparison patients with lesions that did
not include the insula, to control for non-​specific effects of having brain damage. All
were smoking heavily at the time of lesion onset. We found that patients with insula
lesions, either on the right or left side, were somewhat more likely to quit smoking after
lesion onset than comparison patients, though this was not a statistically significant
difference. However, when we focused on the patients who actually quit after lesion
onset, we found that the likelihood of undergoing a disruption of addiction was signifi-
cantly higher in the insula lesion group than in the lesion comparison group (Figure
12.1). Thus, while insula lesions did not necessarily increase the likelihood of quit-
ting, they did increase the likelihood that quitting was easy, compared to non-​insula
lesions. This makes sense when one considers that any form of brain damage that leads
  207

The effects of insula damage on addiction to cigarette smoking 207

(a) (b) Non- Total

insula insula
69 4
Total 6
15

31 12
1

37 32 Left Right
Non-quitters Quitters insula insula
1
0
5

7
16
No 16
disruption Disruption 1 5
of smoking of smoking
addiction addiction
Non-quitters
Quitters with no disruption of smoking addiction
Quitters with disruption of smoking addiction

Figure 12.1  Results from Naqvi et al. (2007)

to hospitalization, etc. (especially stroke, which is caused by smoking), would prompt

someone to quit smoking, but only damage in a region that is critical for addiction
would make quitting so easy.
This finding has subsequently been corroborated by two prospective human lesion stud-
ies. The first, by Suñer-​Soler et al. (2011), showed that insula lesions increased the likeli-
hood of successful quitting, while also increasing measures of motivation for remaining
abstinent, compared to non-​insula lesions. They also found that having an intention to
quit smoking prior to the stroke predicted the likelihood of quitting after stroke, though
they did not examine the interaction between this effect and lesion location. A second
study, performed by our group (Gaznick et  al. 2013)  examined the combined effects
of insula and basal ganglia lesions in cigarette smokers. These are regions that, due to
shared vascular supply, are frequently damaged together. It was found that lesions to the
basal ganglia alone were associated with an increased likelihood of quitting smoking and
greater reduction in self-​report measures of addiction severity, compared to comparison
lesions that did not affect either the basal ganglia or the insula. Furthermore, lesions that
included both the basal ganglia and the insula led to an even greater likelihood of quitting
smoking and even greater reductions of addiction score, compared to lesions of the basal
ganglia alone, suggesting a particular role for the insula in smoking addiction.
Yousefzadeh-​Fard et al. (2013) completed a study in Iran in which they examined the
effects of insula strokes and basal ganglia strokes on the rates of quitting heroin use.
Similarly to the Gaznick et al. (2013) study, they found that basal ganglia strokes alone, as
208  

208 The role of the insula in goal-directed drug seeking and choice in addiction

well as insula strokes alone, increased the rate of reducing heroin use over one month, three
months, and six months, compared to comparison lesions that did not affect either the
insula or the basal ganglia. They also found that strokes that affected the insula alone led
to greater effects than strokes that affected the basal ganglia alone. Furthermore, patients
in the insula lesion group actually quit heroin use entirely at a higher rate than patients in
the comparison group, whereas patients in the basal ganglia group did not. This finding
corroborates our earlier work, and also extends the finding to indicate that the insula is a
critical neural substrate for opioid addiction, in addition to tobacco addiction.
Bienkowski et al. (2010) performed the only published study that has contradicted our
initial results. They found that patients with insula strokes were not more likely to quit
cigarette smoking than patients with non-​insula strokes. They also found that none of
the patients who quit smoking after a stroke—​in any lesion group—​underwent a dis-
ruption of smoking addiction, as we had originally defined it. One explanation for this
discrepancy is that nearly half of smokers in Poland who are prompted to quit because of
health problems related to smoking (e.g. stroke) do not perceive the link between the two
(Sieminska et al. 2008). As we discuss below, the perception that smoking has negative
consequences may be an important psychological factor mediating the effects of insula
lesions on smoking addiction.

3  The insula as an appetitive motivational system

in addiction
Our initial 2007 study was motivated in part by a number of functional neuroimaging
studies showing that exposure to drug cues activated the insula, and that insula activ-
ity was frequently correlated with cue-​induced craving, which is an appetitive emotion
that promotes drug use. Since our study, a number of meta-​analyses of functional imag-
ing studies have confirmed the insula’s involvement in cue-​induced craving (Kühn and
Gallinat 2011; Schacht et al. 2012; Engelmann et al. 2012; Chase et al. 2011). Functional
imaging studies have also shown relationships between cue-​elicited insula activity and
clinical variables related to addiction severity and treatment outcome (Janes et al. 2010;
Claus et al. 2013; Claus et al. 2011a), along with moderating effects of genes that predis-
pose to addiction on cue-​elicited insula activity (Blaine et al. 2013; Janes et al. 2012). These
are studies across a wide variety of substances, indicating that the insula is involved in a
number of different forms of addiction. The patterns of activation vary slightly between
the different drugs of abuse, and the reader is directed to a recent comprehensive review
for these and other details of these studies (Naqvi et al. 2014).
Animal models of drug motivation, coupled with anatomically targeted manipulation of
insula functions, provide further evidence for the role of the insula in specific motivational
functions that drive addiction (Contreras et al. 2007, 2012; Forget et al. 2010; Hollander et al.
2008; Pushparaj et al. 2012; Scott and Hiroi 2011; Seif et al. 2013). These studies are heteroge-
neous with respect to the specific insular subregions targeted, the method of manipulation
of insula function, and the behavioral assays used to assess the effects of the manipulations.
  209

The insula as a locus for interoceptive representation in addiction 209

Despite this heterogeneity, they all show that manipulations that disrupt insula functioning
reduce drug-​seeking behavior, which is broadly consistent with the effects of insula lesions
in humans. There also appears to be a dissociation of function between the posterior (gran-
ular) insula and the anterior (agranular) insula: whereas the posterior insula appears neces-
sary for registering the reinforcement value of drugs and for the learning of drug-​context
associations, the anterior insula appears necessary for the retrieval and reconsolidation of
drug-​context associations. Furthermore, most of these studies showed that whereas insula
lesions disrupt drug-​related motivation, they appear to spare food-​related motivation, simi-
lar to what was found in our initial study. Again, the reader is referred to a recent compre-
hensive review for details of these studies (Naqvi et al. 2014).
One study by Seif et al. (2013) deserves particular attention because it provides a highly
detailed anatomical, physiological, and behavioral characterization of the role of the
insula in specific forms of addictive behavior. After our initial 2007 study, we hypoth-
esized that the insula, in particular its inputs into the nucleus accumbens, plays a role in
addictive behavior that persists in the face of negative consequences (Naqvi and Bechara
2009, 2010). Seif et al. tested this hypothesis directly using a rodent model of “aversion
resistant” alcohol intake. They trained rats to self-​administer alcohol and then switched
some of the rats to alcohol that was adulterated with quinine (an aversive taste). Normally,
once rats learn to self-​administer alcohol, i.e. once they are addicted, adulterating it with
quinine does not reduce their intake. This then serves as a model of addictive alcohol
use that persists despite negative consequences. Using optogenetic techniques, Seif et al.
(2013) selectively disrupted excitatory inputs from the insula into the nucleus accumbens.
This reduced drinking from the quinine-​adulterated alcohol, while sparing drinking from
the quinine-​free alcohol. This confirmed the hypothesis that the insula is critical for moti-
vating alcohol use that persists in the face of an aversive consequence. Similar results
were obtained when excitatory inputs to the nucleus accumbens from the ventromedial
prefrontal cortex were optogenetically disrupted, indicating that this region may also play
a role in similar motivational processes as the insula.

4  The insula as a locus for interoceptive representation

in addiction
Interoception is the transduction, transmission, and central representation of signals that
are related to the physiological state of peripheral tissues, such as temperature, organ dis-
tention, acid/​base balance, inflammation, and glucose concentration. Interoception also
includes sensory signals related to motivated behavior, such as nociception (pain), taste,
genital sensation, and pleasant touch. Nearly all drugs of abuse exert highly distinctive
interoceptive effects. For example smoking tobacco, drinking alcohol, and intranasal
cocaine or heroin use all stimulate chemosensory afferents within the oropharyngeal and
nasal mucosa, including taste receptors. Cocaine, amphetamines, nicotine, opioids, and
alcohol all exert powerful effects on the autonomic nervous system. Interoceptive effects
are notably different from rewarding CNS drug effects, i.e. dopamine release, which do not
210  

210 The role of the insula in goal-directed drug seeking and choice in addiction

discriminate between different drugs, or between drugs and other rewards. Interoceptive
drug effects are an important source of pleasure and reinforcement from drug use. This has
been shown extensively for tobacco (Rose 2006; Carpenter et al. 2007; Naqvi and Bechara
2005, 2006), and more recently for cocaine (Wise et al. 2008). The first experiences with
the interoceptive effects of a drug are usually unpleasant (ask anyone who has smoked a
cigarette), and become pleasurable over time. This indicates a hedonic learning process in
which interoceptive drug effects “switch” from being aversive to being pleasurable, desir-
able, and, in addicted individuals, needed.
According to Craig (2002, 2010), interoceptive signals reach the central nervous system
through a dedicated set of peripheral pathways that converge on the insula. The right
anterior insula plays a special role in integrating interoceptive awareness into conscious
emotional feelings, as well as a variety of motivational, executive, social, and self-​aware
processes. The anterior insula sends projections to the basolateral amygdala (Stefanacci
and Amaral 2002), nucleus accumbens (Reynolds and Zahm 2005), entorhinal cortex/​
hippocampal formation (Augustine 1996), anterior cingulate cortex (Augustine 1996),
and orbitofrontal cortex (Mesulam and Mufson 1982), which together are likely to medi-
ate these affective, motivational, social and executive functions (Figure 12.2).
Interoception plays a critical role in positive hedonic emotions, specifically the sub-
jective pleasure that is derived from obtaining homeostatic goals. For example, without
taste or genital sensations there would be few hedonic feelings of pleasure from eating or

Anterior Posterior

Interoceptive thalamus

Granular
Dysgranular

Nucleus accumbens Agranular

Basolateral amygdala
Hippocampus
Orbitofrontal cortex
Anterior cingulate cortex
Figure 12.2  The insula, its anatomical subdivisions, and major inputs/outputs
Adapted from Gray (Standring 2008), Craig (Craig 2008), Bonthius et al. (Bonthius et al. 2005)
  211

The insula as a decision system in addiction 211

sexual intercourse, respectively. The evidence for the role of the insula and related systems
in sensory and hedonic aspects of taste is extensive, and has been reviewed elsewhere
(Small 2010). Functional MRI studies have also shown that the insula is activated by geni-
tal sensations (Georgiadis and Holstege 2005) and by sensual touch (Morrison et al. 2011;
Olausson et al. 2002), which signal copulation and social affiliation, respectively. In addi-
tion to its role in representing interoceptive stimuli that impinge directly upon the body,
the insula, in particular the right anterior insula, plays a role in a variety of emotional,
motivational, and social processes that involve the anticipation or mentalization of inter-
oceptive states. For example, fMRI studies have shown that the right anterior insula is
activated during empathy for pain in a loved one, whereas the posterior insula is activated
by the direct experience of pain (Singer et al. 2004). Similarly, the right anterior insula is
activated by the anticipation of sensual touch, whereas the posterior insula is activated
by the touch itself (Lovero et al. 2009). Single unit recordings have shown that the ante-
rior insula of monkeys contains neurons that respond to sweet tastes, as well as neurons
that respond during the anticipation of sweet tastes, with the dynamics of neural activity
predicting the timing of taste delivery (Mizuhiki et al. 2012). Together, these findings sug-
gest that, whereas the posterior insula represents the sensory properties of interoceptive
stimuli during hedonic experience, the anterior insula, in particular the right anterior
insula, is involved in recalling these representations from memory and holding them in
mind while simulating the hedonic experience of the self and of others.
A number of functional MRI studies have shown that the insula is activated by the
interoceptive effects of drugs, including the chemosensory effects of nicotine (Albrecht
et al. 2009) and the taste of alcohol (Filbey et al. 2008a), and, for alcohol, that this activa-
tion is moderated by genes that are involved in dopamine-​signaling of reward (Filbey
et al. 2008b). We have shown in a preliminary study that insula damage disrupts the abil-
ity of smokers to discriminate the airway sensory effects of nicotine (Naqvi and Bechara
2010). In rodents, inactivation of the insula disrupts the ability to discriminate the che-
mosensory effects of alcohol (Brasser et al. 2015). Together, these data suggest that the
insula plays an important role in the sensory processing of the interoceptive effects of
drugs of abuse. From what we understand about the role of the insula in interoceptive
functions more generally, the insula may function to encode into memory, recall, hold in
mind, and translate into conscious experience the various interoceptive effects of drugs of
abuse. Later in this chapter, we speculate on the role that interoceptive processes play in
the pathophysiology and mechanisms of behavior change in addiction.

5  The insula as a decision system in addiction

The insula has been implicated in a number of executive function processes that involve
weighing the pursuit of certain rewards against the avoidance of uncertain negative conse-
quences, i.e. decision-​making under risk. Such decision-​making processes are likely to play
an important role in addiction, where addicted individuals are continuously faced with the
choice of either using in order to obtain the short-​term rewarding effects of the drug or
212  

212 The role of the insula in goal-directed drug seeking and choice in addiction

abstaining in order to avoid long-​term social, financial, medical, and legal consequences.
Damasio (1994) initially proposed that the anterior insula is part of a network of brain areas
that guides decision-​making under risk and uncertainty by “marking” various options for
behavior in terms of their potential positive and negative consequences. Subsequent work
using both functional MRI (Kuhnen and Knutson 2005; Preuschoff et al. 2008; Xue et al.
2010; Kohno et al. 2013; Mohr et al. 2010) and lesion methods (Clark et al. 2008; Ishii et al.;
2012) have corroborated the role of the insula in risk-​based decision-​making, primarily in
the promotion of decisions where the advantageous choice involves taking risks.
Abnormalities in insula function during executive processes may play a role in promot-
ing addiction. Our group postulated that abnormalities in somatic marker representation
by the insula during risky decisions are a core deficit that underlies continued drug use in
the face of negative consequences (Verdejo-​Garcia and Bechara 2009). Functional MRI
studies in a number of substance use disorders have demonstrated abnormal insula func-
tion during risky decision-​making tasks, compared to healthy controls (Paulus et al. 2008;
Claus et al. 2011b; Devito et al. 2013), as well as relationships between these abnormalities
and genes that predispose to addiction (Villafuerte et al. 2012). These task-​related func-
tional abnormalities may be related to structural and resting state functional abnormali-
ties in the insula that have been found in a variety of substance use disorders (Ersche et al.
2012; Franklin et al. 2002; Chanraud et al. 2007; Cisler. et al. 2013; Gardini and Venneri
2012; Makris et al. 2008; Morales et al. 2012; Moreno-​López et al. 2012; Sullivan et al.
2013; Sutherland et al. 2012; Zhang et al. 2011). What is not known is whether these struc-
tural and functional abnormalities in the insula represent a “loss” of risk-​related functions
that would otherwise reduce drug use (e.g. not taking a drug because it is risky) or a “gain”
of functions that tend to promote drug use (e.g. overcoming risk to take a drug). In either
case, the extent of the abnormality would correlate with the severity of addiction.

6  Automatic versus goal-​directed drug seeking

Tiffany (1999) was the first to propose a distinction between what he termed “automatic”
drug seeking and “controlled” drug seeking, based upon the observation that craving, a sub-
jective feeling of desire to take a drug, is often difficult to correlate with actual drug use
behavior. According to Tiffany, most drug use in an addicted individual occurs automati-
cally, i.e. “absent-​mindedly,” without attention to antecedents or consequences, and without
conscious cravings. Cravings are a reflection of a “controlled” mode of drug seeking that is
engaged when there are impediments to drug use, such as the unavailability of the drug or the
imminent threat of severe negative consequences, which require that the individual become
aware of antecedents and consequences in order to obtain the drug. In this conception, crav-
ing does not drive drug seeking in individuals who are not facing impediments to using.
At a neural level, the distinction between automatic and controlled drug seeking can
be tied to the distinction that a number of authors have made between “goal-​directed”
or “voluntary” actions and “habitual” action, which they have also extended into the
understanding of the neural substrates for drug motivation (Berridge and Robinson 2003;
Everitt and Robbins 2005; Balleine and O’Doherty 2010). Specifically, habitual action
  213

An integrative model of the role of the insula in addiction 213

is highly stimulus-​bound and divorced from outcome values, such that it persists even
when the outcome is no longer pleasurable or even when it is painful. Habitual action
depends upon neural systems that register the presence of rewarding stimuli, such as
the central nucleus of the amygdala, and initiate and sustain highly stereotyped appe-
titive behaviors, such as the dorsal striatum and the substantia nigra pars compacta. In
contrast, goal-​directed action, also known as instrumental action-​outcome contingency
learning, involves the explicit representation of contingencies between specific actions
and the hedonic value of their specific outcomes, including negative outcomes (i.e. risk).
Goal-​directed or voluntary action depends upon several integrated functions, including
the ability to form associations between cues and the primary hedonic value of the spe-
cific rewards that they predict, mediated by the basolateral amygdala; the ability to the
predict the hedonic value of future rewards, mediated by the ventromedial prefrontal cor-
tex; the ability to integrate predictive reward representations with reward-​seeking motor
programs in order to initiate and sustain goal-​directed actions, mediated by the nucleus
accumbens and its dopaminergic innervation from the ventral tegmental area; working
memory and attentional functions that hold representations of current goals online while
suppressing representations of competing goals, mediated by the dorsolateral prefron-
tal cortex; and monitoring of conflict between current goals and competing goals that
may require a shift in behavioral set, mediated by the dorsal anterior cingulate cortex.
According to Everitt and Robbins (2005), the transition from casual drug use to addic-
tion is due largely to irreversible dopamine-​induced neural plasticity that causes a shift
in control over drug seeking from goal-​directed/​voluntary systems to habitual systems.
Work in rodents has shown that the insula plays an important role in goal-​directed
action. Specifically, the insula is necessary for using information about the sensory impact
of rewards in order to retrieve their specific hedonic value from memory (Balleine and
Dickinson 2000; Parkes and Balleine 2013; Balleine and O’Doherty 2010). This is consist-
ent with evidence reviewed above that the anterior insula holds interoceptive representa-
tions of predicted homeostatic rewards in mind over delays, and suggests that the insula
may function to retrieve these interoceptive representations from memory and hold them
in mind during the planning of goal-​directed behavior. This function is likely to depend
upon the strong interconnections between the insula and other neural systems involved
in goal-​directed behavior, including the ventromedial prefrontal cortex, the amygdala,
and the nucleus accumbens. This is an important function for “natural” rewards, such
as food and sex, in which the attainment of the goal-​outcomes is signaled through their
highly specific sensory (interoceptive) effects. Below, we discuss how these functions may
play a role in addiction, where the rewarding effects of the drug taking (the goal of drug
seeking) are derived in large measure from interoceptive drug effects.

7  An integrative model of the role of the insula

in addiction
How is it that the insula can play a role in both motivational processes that drive ongo-
ing drug use and relapse, such as craving, as well as executive functioning processes that
214  

214 The role of the insula in goal-directed drug seeking and choice in addiction

rein in drug use in the face of negative consequences? Furthermore, what role, if any, do
the interoceptive functions of the insula play in craving and relapse? Here, we attempt
to answer these questions within an integrated theoretical framework for the role of the
insula in addiction (see Figure 12.3). In this framework, the insula is critically involved in
a mode of goal-​directed drug seeking that engages representations of interoceptive drug
effects in order to overcome subjective risks and conflicts. We have discussed elements of
this framework elsewhere (Naqvi et al. 2014; Naqvi and Bechara 2009, 2010).
We propose that the specific function of the insula in addiction is to link drug seeking
(an action) to the value of drug taking (a goal, or outcome) through specific, discrimina-
tive interoceptive representations of drug taking rituals. This occurs through convergence
of inputs from the anterior insula, amygdala, and the ventromedial prefrontal cortex
onto the nucleus accumbens. This leads to a coherent goal representation of drug seek-
ing, which integrates exteroceptive features (e.g. the people, places, and things that are
associated with drug taking), interoceptive features (e.g. the taste of alcohol, the feeling
of tobacco smoke in the back of the throat) and a specific value (i.e. a higher value of the
drug-​taking goal under conditions of withdrawal or stress). The dorsolateral prefrontal
cortex functions to hold this coherent goal representation of drug taking in mind dur-
ing planning and decision-​making processes, while the anterior cingulate cortex con-
tinuously detects conflicts between this goal and the threat of imminent consequences
or competing goals. Together, this results in a highly targeted form of goal striving that
remains “on track” because of the specificity of the interoceptive representations of drug
taking. The conscious feeling of cue-​induced craving is the subjective manifestation of
the goal-​striving process, and is phenomenally derived from embodied memories for
interoceptive drug effects that are represented by the insula. This explains why one of
the patients with an insula lesion whom we interviewed in our original study (patient N)
described how his “body forgot the urge to smoke.”
The goal-​directed system is brought online when the anterior cingulate detects either
(a) a high subjective risk, which involves an emotional appreciation that a salient negative
consequence is imminent; or (2) high subjective conflict, in which drug use interferes with
the attainment of an alternative goal with high homeostatic value for the individual at that
time. This signals a need to decide that the short-​term reward from drug taking outweighs
the long-​term risks, or that the reward from drug taking outweighs the competing rewards.
Such functions require the ability to recall and hold in mind explicit representations of
the reward value of drug taking, which we propose occurs through interoceptive repre-
sentations. Without high subjective risk or conflict, an automatic mode of drug seeking
predominates, one in which the individual is largely unaware of the antecedents or conse-
quences of drug use and in which there is little conscious craving felt, or where conscious
craving has no relationship to drug-​seeking behavior. The automatic mode of drug seeking
is mediated primarily by the central amygdala/​dorsal striatum/​substantia nigra system, i.e.
the habitual action system proposed by Everitt and colleagues (Everitt and Robbins 2005).
The goal-​directed mode of drug seeking predominates when drug users are experi-
menting with drugs, prior to becoming addicted. At this stage, drug seeking is driven
  215

(A)
Cortical

ACC dlPFC

Insula vmPFC

dSTR CNA

DA
SNpc
Subcortical

(B)

Conflict
Cortical

ACC dlPFC

Insula vmPFC

NAcc BLA

DA
VTA
Subcortical

Figure 12.3 A model for the insula’s role in addiction. (A) Without the presence of subjective risk or
conflict, drug seeking proceeds automatically, i.e. it is divorced from the value of drug taking and it is
not subjectively experienced as craving. Automatic drug seeking is largely stimulus (cue) driven. The
stimulus-response process is mediated by subcortical systems, including the central nucleus of the
amygdala (CNA) and its outputs to the dopaminergic systems in the substantia nigra pars compacta,
which innervate the dorsal striatum (dSTR) and thereby initiate habits and automatic cognitive
processes. Automatic drug seeking does not engage cortical systems for goal-directed behavior. As
a result, it is not affected by insula lesions. (B) The cortical network for goal-directed drug seeking is
brought online by the anterior cingulate cortex (ACC), which detects conflict and risks. The insula
holds representations of interoceptive drug effects in mind and allows access to the current value
of drug taking, which is represented in the ventromedial prefrontal cortex (vmPFC). The dorsolateral
prefrontal cortex (dlPFC) coordinates working memory and attentional functions that are necessary for
this integrated goal representation. The output of the cortical goal-directed system is directed through
the nucleus accumbens (NAcc), which links integrated goal representations with motivational signals
provided by the basolateral amygdala (BLA) and dopamine (DA) release from the ventral tegmental
area (VTA). The resulting behavior is goal-directed, i.e. it is tied to the value of drug taking, as well as
subjectively experienced as craving. Insula lesions decouple drug taking from its predictive (imagined)
value, and thereby abolish goal-directed drug seeking and its associated experience of craving.
216  

216 The role of the insula in goal-directed drug seeking and choice in addiction

primarily by a desire to obtain positive hedonic effects, and it is relatively easily to rein
in because it is under control of a system that weighs the seeking of pleasure against
its negative consequences and against other, more adaptive pleasures. Over time, with
repeated use of the drug, drug seeking shifts from the goal-​directed mode to the auto-
matic mode, largely as a result of dopamine-​induced neural plasticity. As a result, drug
seeking becomes increasingly harder to stop, even as negative consequences accu-
mulate. This does not mean, however, that the goal-​directed system ceases to play a
role. Rather, the goal-​directed mode is brought online in severely addicted individu-
als when the automatic mode is interrupted by high subjective risk and conflict. This
often occurs when an addicted individual has just experienced or is about to experi-
ence a significant negative consequence (e.g. arrest, illness) or is presented with the
opportunity to obtain a highly valued reward that conflicts with drug use (e.g. money,
social approval). These situations often bring about attempts to reduce drug-​seeking
behavior. Thus, the goal-​directed mode is a platform for behavior change, both within
and outside of treatment.
Note that in this model the insula does not play a role in the automatic mode of drug
seeking, which may explain why in our initial study (Naqvi et  al. 2007)  and others
(Bienkowski et  al. 2010), insula damage did not necessarily increase the likelihood of
quitting smoking. It may be that, for some smokers who have just sustained a medical
illness that is attributable to smoking (e.g. stroke), there is a heightening of subjective risk
that brings about the goal-​directed mode. By itself, this would increase the likelihood of
quitting but would not make quitting easy, since the goal-​directed mode is manifested
by cravings and the tendency to relapse. However, for patients whose stroke damaged
the insula, quitting would have been easier because craving and the tendency to relapse
would be greatly reduced. Other smokers may not have experienced a heightened subjec-
tive risk after sustaining a stroke, due to a lack of awareness of the role of smoking in their
stroke. These patients would probably have continued to smoke after their stroke and, for
them, insula lesions would appear to have no effect.
Why are some patients less aware of risks and conflicts, and therefore less likely to
engage the goal-​directed mode after an event such as a stroke? Impaired insight, which
is a key clinical feature of addiction and a major impediment to behavior change, can
be understood as an inability to subjectively experience conflicts and risks, which
leads to a failure to switch from automatic to goal-​directed modes. Goldstein and col-
leagues (2009) have suggested that impaired insight in addiction is tied to a loss of
saliency detection functions of the insula. We propose that the lack of insight in addic-
tion, rather than being “caused” by a loss of insula function, is due to impairments in
upstream processes, such as conflict detection functions of the anterior cingulate cor-
tex, which in turn prevent the goal-​directed mode from becoming engaged. As a result
of this loss of insight, addicted individuals tend to remain in the automatic mode, even
when objective risks and conflicts are high. For this reason, the insula would be less
active in individuals with low insight, but this would be an effect of low insight, rather
than a cause.
  217

Implications for understanding addiction as choice 217

8  Clinical implications
If the insula’s role in addiction is limited to the goal-​directed mode of drug seeking that
is promoted by high subjective risk and conflict, then treatments aimed at increasing the
subjective riskiness of drug use, or that increase conflict between drug use and other
goals, should, by promoting the goal-​directed mode, render addiction more labile to dis-
ruption through manipulations of the functioning of the insula and other components of
the goal-​directed system. A number of existing evidence-​based treatments may promote
the goal-​directed mode. For example, motivational interviewing (Miller and Rollnick
2013) is thought to work in part by increasing “discrepancy” (i.e. conflict) between drug
use and adaptive goals. Brief interventions (Heather 2003)  typically focus on increas-
ing the subjective riskiness of drug use by making patients aware of potential negative
consequences. Contingency management (Petry 2010)  provides incentives for absti-
nence, which serve as alternative rewarding goals that conflict with drug use. Cognitive-​
behavioral therapy (Parks et al. 2003) is focused on increasing awareness of antecedents
and consequences of drug use, which tends to decrease automaticity. Disulfiram (Diehl
et al. 2010), a medication for alcohol use disorder, reduces drinking by causing a severe
visceral reaction if alcohol is consumed while it is being taken, markedly increasing the
subjective riskiness of drinking.
Once the goal-​directed mode becomes engaged, the next step would be to manipu-
late insula function using non-​invasive methods. These include neurostimulation tech-
niques such as repetitive transcranial magnetic stimulation (rTMS), which has already
been shown to alter addictive behavior when targeted at regions such as the dorsolateral
prefrontal cortex (Li et al. 2013; Rose et al. 2011) and the insula (Dinur-​Klein et al. 2014).
Insula function may also be manipulated pharmacologically, either directly through
drugs that bind receptors in the insula, or indirectly through drugs that bind elsewhere
but have downstream effects on insula function. Thus, the most effective treatments may
be those that combine psychosocial interventions that promote the goal-​directed mode
with biologically based treatments that disrupt insula function.

9  Implications for understanding addiction as choice

In our view, addiction is a choice when the goal-​directed mode of drug seeking predomi-
nates. Conversely, addiction is not a choice when the automatic mode predominates. It
is the goal-​directed mode that mediates the decision to abstain versus relapse, and that
is experienced subjectively as craving and also as willful control over drug seeking. The
shift from automatic to goal-​directed modes is not a binary switch but rather a spec-
trum, where a number of factors can influence the balance between automatic and goal-​
directed modes. These include individual differences in addiction severity and the level of
insight, as well as, within an individual, the level of motivation for change, the extent of
social support for abstinence and the availability of treatment. All of these factors inter-
act with the level of subjective risk and conflict experienced by an addicted individual
218  

218 The role of the insula in goal-directed drug seeking and choice in addiction

at any given time. In the framework we have elaborated, the insula plays a specific role
in the goal-​directed mode of drug seeking, especially in the subjective experiences that
derive from this mode. Thus, the degree of choice in addiction is inextricably linked to
the level of feeling, the experience of free will and the sense of self, all of which can vary
within an individual depending upon the demands placed on them by their environment.
We believe that this is a more humane view of addiction than one that sees addiction as
either entirely a choice and thereby subject to ethical opprobrium (“all addicts are evil”)
or entirely involuntary and thereby unchangeable through personal agency (“all addicts
are weak”).

Acknowledgments
Supported by NIAAA K23 AA02277 (N.H.N.) and NIDA R01 DA16708 (A.B.)

References
Albrecht, J., Kopietz, R., Linn, J., et al. (2009). Activation of olfactory and trigeminal cortical areas
following stimulation of the nasal mucosa with low concentrations of S(−)-​nicotine vapor-​An
fMRI study on chemosensory perception. Human Brain Mapping, 30, 699–​710. DOI: 10.1002/​
hbm.20535.
Augustine, J.R. (1996). Circuitry and functional aspects of the insular lobe in primates including
humans. Brain Research. Brain Research Reviews, 22, 229–​44.
Balleine, B.W. and Dickinson, A. (2000). The effect of lesions of the insular cortex on instrumental
conditioning: evidence for a role in incentive memory. Journal of Neuroscience, 20, 8954–​64.
Balleine, B.W. and O’Doherty, J.P. (2010). Human and rodent homologies in action
control: corticostriatal determinants of goal-​directed and habitual action. Neuropsychopharmacology,
35, 48–​69. DOI: 10.1038/​npp.2009.131.
Bechara, A. (2005). Decision making, impulse control and loss of willpower to resist drugs: a
neurocognitive perspective. Nature Neuroscience, 8, 1458–​63. DOI: 10.1038/​nn1584.
Berridge, K.C. and Robinson, T.E. (2003). Parsing reward. Trends in Neurosciences, 26, 507–​13.
DOI: 10.1016/​S0166-​2236(03)00233-​9.
Bienkowski, P., Zatorski, P., Baranowska, A., Ryglewicz, D., and Sienkiewicz-​Jarosz, H. (2010). Insular
lesions and smoking cessation after first-​ever ischemic stroke: A 3-​month follow-​up. Neuroscience
Letters, 478, 161–​4. DOI: 10.1016/​j.neulet.2010.05.008.
Blaine, S., Claus, E., Harlaar, N., and Hutchison, K. (2013). TACR1 genotypes predict fMRI response
to alcohol cues and level of alcohol dependence. Alcoholism: Clinical and Experimental Research,
37(Suppl. 1), E125–​30. DOI: 10.1111/​j.1530-​0277.2012.01923.x.
Bonthius, D.J., Solodkin, A., and Van Hoesen, G.W. (2005). Pathology of the insular cortex in
Alzheimer disease depends on cortical architecture. Journal of Neuropathology and Experimental
Neurology, 64, 910–​22.
Brasser, S.M., Castro, N, and Feretic, B. (2015). Alcohol sensory processing and its relevance for
ingestion. Physiology & Behaviour, 1(148), 65–​70.
Carpenter, C.M., Wayne, G.F., and Connolly, G.N. (2007). The role of sensory perception in
the development and targeting of tobacco products. Addiction, 102, 136–​47. DOI: 10.1111/​
j.1360-​0443.2006.01649.x.
  219

Implications for understanding addiction as choice 219

Chanraud, S., Martelli, C., Delain, F., et al. (2007). Brain morphometry and cognitive performance in
detoxified alcohol-​dependents with preserved psychosocial functioning. Neuropsychopharmacology,
32, 429–​38. DOI: 10.1038/​sj.npp.1301219.
Chase, H.W., Eickhoff, S.B., Laird, A.R., and Hogarth, L. (2011). The neural basis of drug stimulus
processing and craving: an activation likelihood estimation meta-​analysis. Biological Psychiatry, 70,
785–​93. DOI: 10.1016/​j.biopsych.2011.05.025.
Cisler, J.M., Elton, A., Kennedy, A.P., et al. (2013). Altered functional connectivity of the insular cortex
across prefrontal networks in cocaine addiction. Psychiatry Research, 213, 39–​46. DOI: 10.1016/​
j.pscychresns.2013.02.007.
Clark, L., Bechara, A., Damasio, H., et al. (2008). Differential effects of insular and ventromedial
prefrontal cortex lesions on risky decision-​making. Brain: A Journal of Neurology, 131, 1311–​22.
DOI: 10.1093/​brain/​awn066.
Claus, E.D., Blaine, S.K., Filbey, F.M., Mayer, A.R., and Hutchison, K.E. (2013). Association between
nicotine dependence severity, BOLD response to smoking cues, and functional connectivity.
Neuropsychopharmacology, 38(12), 2362–​72. DOI: 10.1038/​npp.2013.134.
Claus, E.D., Ewing, S.W.F., Filbey, F.M., Sabbineni, A., and Hutchison, K.E. (2011a).
Identifying neurobiological phenotypes associated with alcohol use disorder severity.
Neuropsychopharmacology, 36, 2086–​96. DOI: 10.1038/​npp.2011.99.
Claus, E.D., Kiehl, K.A., and Hutchison, K.E. (2011b). Neural and behavioral mechanisms of impulsive
choice in alcohol use disorder. Alcoholism: Clinical and Experimental Research, 35, 1209–​19.
DOI: 10.1111/​j.1530-​0277.2011.01455.x.
Contreras, M., Billeke, P., Vicencio, S., , et al. (2012). A role for the insular cortex in long-​term memory
for context-​evoked drug craving in rats. Neuropsychopharmacology, 37, 2101–​108. DOI: 10.1038/​
npp.2012.59.
Contreras, M., Ceric, F., and Torrealba, F. (2007). Inactivation of the interoceptive insula disrupts
drug craving and malaise induced by lithium. Science (New York), 318, 655–​58. DOI: 10.1126/​
science.1145590.
Craig, A.D. (2002). How do you feel? Interoception: the sense of the physiological condition of the
body. Nature Reviews Neuroscience, 3, 655–​66. DOI: 10.1038/​nrn894.
Craig, A.D. (2008). Interoception and emotion: a neuroanatomical perspective. In: M. Lewis,
J.M. Haviland-​Jones, and L. Feldman Barrett (eds), Handbook of Emotions (3rd Edition).
New York: Guilford Press, pp. 272–​88.
Craig, A.D. (2010). The sentient self. Brain Structure and Function, 214(5–​6), 563–​77. DOI: 10.1007/​
s00429-​010-​0248-​y.
Damasio, A. R. (1994). Descartes’ Error: Emotion, Reason, and the Human Brain, New York: Putnam.
Devito, E.E., Meda, S.A., Jiantonio, R., et al. (2013). Neural correlates of impulsivity in healthy
males and females with family histories of alcoholism. Neuropsychopharmacology. DOI: 10.1038/​
npp.2013.92.
Diehl, A., Ulmer, L., Mutschler., J , et al. (2010). Why is disulfiram superior to acamprosate in the
routine clinical setting? A retrospective long-​term study in 353 alcohol-​dependent patients. Alcohol
and Alcoholism, 45, 271–​77.
Dinur-​Klein, L., Dannon, P., Hadar, A., et al. (2014). Smoking cessation induced by deep repetitive
transcranial magnetic stimulation of the prefrontal and insular cortices: a prospective, randomized
controlled trial. Biological Psychiatry. DOI: 10.1016/​j.biopsych.2014.05.020.
Engelmann, J. M., Versace, F., Robinson, J.D., et al. (2012). Neural substrates of smoking
cue reactivity: a meta-​analysis of fMRI studies. Neuroimage, 60, 252–​62. DOI: 10.1016/​
j.neuroimage.2011.12.024.
220  

220 The role of the insula in goal-directed drug seeking and choice in addiction

Ersche, K.D., Jones, P.S., Williams, G.B., et al. (2012). Abnormal brain structure implicated in
stimulant drug addiction. Science (New York,), 335, 601–​604. DOI: 10.1126/​science.1214463.
Everitt, B.J. and Robbins, T.W. (2005). Neural systems of reinforcement for drug addiction: from
actions to habits to compulsion. Nature Neuroscience, 8, 1481–​89. DOI: 10.1038/​nn1579.
Filbey, F.M., Claus, E., Audette, A.R., et al. (2008a). Exposure to the taste of alcohol elicits activation
of the mesocorticolimbic neurocircuitry. Neuropsychopharmacology, 33, 1391–​401. DOI: 10.1038/​
sj.npp.1301513.
Filbey, F.M., Ray, L., Smolen, A., et al. (2008b). Differential neural response to alcohol priming and
alcohol taste cues is associated with DRD4 VNTR and OPRM1 genotypes. Alcoholism: Clinical and
Experimental Research, 32, 1113–​23. DOI: 10.1111/​j.1530-​0277.2008.00692.x.
Forget, B., Pushparaj, A., and Le Foll, B. (2010). Granular insular cortex inactivation as a novel
therapeutic strategy for nicotine addiction. Biological Psychiatry, 68, 265–​71. DOI: 10.1016/​
j.biopsych.2010.01.029.
Franklin, T.R., Acton, P.D., Maldjian, J.A., et al. (2002). Decreased gray matter concentration in the
insular, orbitofrontal, cingulate, and temporal cortices of cocaine patients. Biological Psychiatry, 51,
134–​42. DOI: 10.1016/​S0006-​3223(01)01269-​0.
Gardini, S. and Venneri, A. (2012). Reduced grey matter in the posterior insula as a structural
vulnerability or diathesis to addiction. Brain Research Bulletin, 87(2), 205–​11.
Gaznick, N., Tranel, D., Mcnutt, A., and Bechara, A. (2013). Basal ganglia plus insula damage yields
stronger disruption of smoking addiction than basal ganglia damage alone. Nicotine and Tobacco
Research, 16(4), 445–​53. DOI: 10.1093/​ntr/​ntt172.
Georgiadis, J.R. and Holstege, G. (2005). Human brain activation during sexual stimulation of the
penis. Journal of Comparative Neurology, 493, 33–​38. DOI: 10.1002/​cne.20735.
Goldstein, R.Z., Craig, A.D.B., Bechara, A., et al. (2009). The neurocircuitry of impaired
insight in drug addiction. Trends in Cognitive Sciences, 13, 372–​8 0. DOI: 10.1016/​
j.tics.2009.06.004.
Goldstein, R.Z., Leskovjan, A.C., Hoff, A.L., et al. (2004). Severity of neuropsychological impairment
in cocaine and alcohol addiction: association with metabolism in the prefrontal cortex.
Neuropsychologia, 42, 1447–​58. DOI: 10.1016/​j.neuropsychologia.2004.04.002.
Heather, N. (2003). Brief interventions. In: N. Heather and T. Stockwell (eds), Essential Handbook of
Treatment and Prevention of Alcohol Problems. Chichester: John Wiley, pp. 117–​38.
Hollander, J.A., Lu, Q., Cameron, M.D., Kamenecka, T.M., and Kenny, P.J. (2008). Insular hypocretin
transmission regulates nicotine reward. Proceedings of the National Academy of Sciences of the United
States of America, 105, 19480–​85. DOI: 10.1073/​pnas.0808023105.
Ishii, H., Ohara, S., Tobler, P.N., Tsutsui, K.-​I., and Iijima, T. (2012). Inactivating anterior
insular cortex reduces risk taking. Journal of Neuroscience, 32, 16031–​9. DOI: 10.1523/​
JNEUROSCI.2278-​12.2012.
Janes, A.C., Pizzagalli, D.A., Richardt, S., et al. (2010). Brain reactivity to smoking cues prior to
smoking cessation predicts ability to maintain tobacco abstinence. BPS, 67, 722–​9. DOI: 10.1016/​
j.biopsych.2009.12.034.
Janes, A.C., Smoller, J.W., David, S.P., et al. (2012). Association between CHRNA5 genetic variation at
rs16969968 and brain reactivity to smoking images in nicotine dependent women. Drug and Alcohol
Dependence, 120, 7–​13. DOI: 10.1016/​j.drugalcdep.2011.06.009.
Jentsch, J.D. and Pennington, Z.T. (2013). Reward, interrupted: Inhibitory control and its relevance to
addictions. Neuropharmacology, 76, 476–​86. DOI: 10.1016/​j.neuropharm.2013.05.022.
Kohno, M., Ghahremani, D.G., Morales, A.M. , et al. (2013). Risk-​taking behavior: dopamine D2/​D3
receptors, feedback, and frontolimbic activity. Cerebral Cortex, 25(1), 236–​45. DOI: 10.1093/​cercor/​
bht218.
  221

Implications for understanding addiction as choice 221

Koob, G.F. and Le Moal, M. (2001). Drug addiction, dysregulation of reward, and allostasis.
Neuropsychopharmacology, 24, 97–​129. DOI: 10.1016/​S0893-​133X(00)00195-​0.
Kühn, S. and Gallinat, J. (2011). Common biology of craving across legal and illegal drugs—​a
quantitative meta-​analysis of cue-​reactivity brain response. European Journal of Neuroscience, 33,
1318–​26. DOI: 10.1111/​j.1460-​9568.2010.07590.x.
Kuhnen, C.M. and Knutson, B. (2005). The neural basis of financial risk taking. Neuron, 47, 763–​70.
DOI: 10.1016/​j.neuron.2005.08.008.
Li, X., Hartwell, K.J., Owens, M., et al. (2013). Repetitive transcranial magnetic stimulation of the
dorsolateral prefrontal cortex reduces nicotine cue craving. Biological Psychiatry, 73, 714–​20.
DOI: 10.1016/​j.biopsych.2013.01.003.
Liu, X., Matochik, J.A., Cadet, J.L., and London, E.D. (1998). Smaller volume of prefrontal lobe in
polysubstance abusers: a magnetic resonance imaging study. Neuropsychopharmacology, 18, 243–​52.
DOI: 10.1016/​S0893-​133X(97)00143-​7.
Lovero, K.L., Simmons, A.N., Aron, J.L., and Paulus, M.P. (2009). Anterior insular cortex anticipates
impending stimulus significance. Neuroimage, 45, 976–​83.
Makris, N., Oscar-​Berman, M., Jaffin, S.K., et al. (2008). Decreased volume of the brain reward system
in alcoholism. Biological Psychiatry, 64, 192–​202. DOI: 10.1016/​j.biopsych.2008.01.018.
Mesulam, M.M. and Mufson, E.J. (1982). Insula of the old world monkey. III: Efferent cortical
output and comments on function. Journal of Comparative Neurology, 212, 38–​52. DOI: 10.1002/​
cne.902120104.
Miller, W.R, and Rollnick, S. (2013). Motivational Interviewing: Helping People Change (3rd ed.).
New York NY: Guilford Press.
Mizuhiki, T., Richmond, B.J. and Shidara, M. (2012). Encoding of reward expectation by
monkey anterior insular neurons. Journal of Neurophysiology, 107, 2996–​3007. DOI: 10.1152/​
jn.00282.2011.
Mohr, P., Biele, G., and Heekeren, H.R. (2010). Neural processing of risk. Journal of Neuroscience,
30(19), 6613–​19.
Morales, A.M., Lee, B., Hellemann, G., O’Neill, J., and London, E.D. (2012). Gray-​matter
volume in methamphetamine dependence: cigarette smoking and changes with abstinence
from methamphetamine. Drug and Alcohol Dependence, 125, 230–​8. DOI: 10.1016/​
j.drugalcdep.2012.02.017.
Moreno-​López, L., Catena, A., Fernández-​Serrano, M.J., et al. (2012). Trait impulsivity and prefrontal
gray matter reductions in cocaine dependent individuals. Drug and Alcohol Dependence, 125, 208–​
14. DOI: 10.1016/​j.drugalcdep.2012.02.012.
Morrison, I., Bjornsdotter, M., and Olausson, H. (2011). Vicarious responses to social touch in
posterior insular cortex are tuned to pleasant caressing speeds. Journal of Neuroscience, 31, 9554–​62.
DOI: 10.1523/​JNEUROSCI.0397-​11.2011.
Naqvi, N.H. and Bechara, A. (2005). The airway sensory impact of nicotine contributes to the
conditioned reinforcing effects of individual puffs from cigarettes. Pharmacology, Biochemistry, and
Behavior, 81, 821–​29. DOI: 10.1016/​j.pbb.2005.06.005.
Naqvi, N.H. and Bechara, A. (2006). Skin conductance responses are elicited by the airway sensory
effects of puffs from cigarettes. International Journal of Psychophysiology, 61, 77–​86. DOI: 10.1016/​
j.ijpsycho.2005.10.018.
Naqvi, N.H. and Bechara, A. (2009). The hidden island of addiction: the insula. Trends in Neurosciences,
32, 56–​67. DOI: 10.1016/​j.tins.2008.09.009.
Naqvi, N.H. and Bechara, A. (2010). The insula and drug addiction: an interoceptive view of
pleasure, urges, and decision-​making. Brain Structure and Function, 214, 435–​50. DOI: 10.1007/​
s00429-​010-​0268-​7.
222  

222 The role of the insula in goal-directed drug seeking and choice in addiction

Naqvi, N.H., Gaznick, N., Tranel, D., and Bechara, A. (2014). The insula: a critical neural substrate
for craving and drug seeking under conflict and risk. Annals of the New York Academy of Sciences.
DOI: 10.1111/​nyas.12415.
Naqvi, N.H., Rudrauf, D., Damasio, H., and Bechara, A. (2007). Damage to the insula disrupts
addiction to cigarette smoking. Science (New York), 315, 531–​34. DOI: 10.1126/​science.1135926.
Olausson, H., Lamarre, Y., Backlund, H., et al. (2002). Unmyelinated tactile afferents signal touch and
project to insular cortex. Nature Neuroscience, 5, 900–​904. DOI: 10.1038/​nn896.
Parkes, S.L. and Balleine, B.W. (2013). Incentive memory: evidence the basolateral amygdala encodes
and the insular cortex retrieves outcome values to guide choice between goal-​directed actions.
Journal of Neuroscience, 33, 8753–​63. DOI: 10.1523/​JNEUROSCI.5071-​12.2013.
Parks, G.A., Anderson, B.K., and Marlatt, G.A. (2003). Cognitive-​behavioural alcohol treatment. In: N.
Heather and T. Stockwell (eds), Essential Handbook of Treatment and Prevention of Alcohol Problems.
Chichester UK: John Wiley, pp. 69–​86.
Paulus, M.P., Lovero, K.L., Wittmann, M., and Leland, D.S. (2008). Reduced behavioral and neural
activation in stimulant users to different error rates during decision making. Biological Psychiatry,
63, 1054–​60. DOI: 10.1016/​j.biopsych.2007.09.007.
Petry, N.M. (2010). Contingency management: controversies and challenges (Editorial). Addiction, 105,
1507–​9.
Potenza, M.N., Sofuoglu, M., Carroll, K.M., and Rounsaville, B.J. (2011). Neuroscience of
behavioral and pharmacological treatments for addictions. Neuron, 69, 695–​712. DOI: 10.1016/​
j.neuron.2011.02.009.
Preuschoff, K., Quartz, S.R., and Bossaerts, P. (2008). Human insula activation reflects risk
prediction errors as well as risk. Journal of Neuroscience, 28, 2745–​52. DOI: 10.1523/​
JNEUROSCI.4286-​07.2008.
Pushparaj, A., Hamani, C., Yu, W., et al. (2012). Electrical stimulation of the insular region
attenuates nicotine-​taking and nicotine-​seeking behaviors. Neuropsychopharmacology, 38, 690–​8.
DOI: 10.1038/​npp.2012.235.
Reynolds, S.M. and Zahm, D.S. (2005). Specificity in the projections of prefrontal and insular cortex
to ventral striatopallidum and the extended amygdala. Journal of Neuroscience, 25, 11757–​67.
DOI: 10.1523/​JNEUROSCI.3432-​05.2005.
Robinson, T.E. and Berridge, K.C. (2008). The incentive sensitization theory of addiction: some current
issues. Philosophical Transactions of the Royal Society Series B: Biological Sciences, 363, 3137–​46.
DOI: 10.1016/​0166-​4328(95)00131-​C.
Rogers, R.D., Everitt, B.J., Baldacchino, A., et al. (1999). Dissociable deficits in the decision-​making
cognition of chronic amphetamine abusers, opiate abusers, patients with focal damage to prefrontal
cortex, and tryptophan-​depleted normal volunteers: evidence for monoaminergic mechanisms.
Neuropsychopharmacology, 20, 322–​39. DOI: 10.1016/​S0893-​133X(98)00091-​8.
Rose, J.E. (2006). Nicotine and nonnicotine factors in cigarette addiction. Psychopharmacology (Berl),
184, 274–​85. DOI: 10.1007/​s00213-​005-​0250-​x.
Rose, J.E., Mcclernon, F.J., Froeliger, B., and Behm, F.M. (2011). Repetitive transcranial magnetic
stimulation of the superior frontal gyrus modulates craving for cigarettes. Biological Psychiatry,
70(8), 794–​99.
Schacht, J.P., Anton, R.F., and Myrick, H. (2012). Functional neuroimaging studies of alcohol cue
reactivity: a quantitative meta-​analysis and systematic review. Addiction Biology, 18, 121–​33.
DOI: 10.1111/​j.1369-​1600.2012.00464.x.
Scott, D. and Hiroi, N. (2011). Deconstructing craving: dissociable cortical control of cue reactivity in
nicotine addiction. Biological Psychiatry, 69, 1052–​59. DOI: 10.1016/​j.biopsych.2011.01.023.
  223

Implications for understanding addiction as choice 223

Seif, T., Chang, S.-​J., Simms, J.A., et al. (2013). Cortical activation of accumbens hyperpolarization-​
active NMDARs mediates aversion-​resistant alcohol intake. Nature Neuroscience, 1–​9.
DOI: 10.1038/​nn.3445.
Sieminska, A., Buczkowski, K., Jassem, E., et al. (2008). Patterns of motivations and ways of quitting
smoking among Polish smokers: a questionnaire study. BMC Public Health, 8, 274. DOI: 10.1186/​
1471-​2458-​8-​274.
Singer, T., Seymour, B., O’Doherty, J., Kaube, H., and Dolan, R.J. (2004). Empathy for pain involves
the affective but not sensory components of pain. Science (New York), 305(5661), 1157–​62.
Small, D.M. (2010). Taste representation in the human insula. Brain Structure and Function, 214, 551–​
61. DOI: 10.1007/​s00429-​010-​0266-​9.
Standring, S. (2008). Gray’s Anatomy. Oxford: Elsevier Health Sciences.
Stefanacci, L. and Amaral, D.G. (2002). Some observations on cortical inputs to the macaque
monkey amygdala: an anterograde tracing study. Journal of Comparative Neurology, 451, 301–​23.
DOI: 10.1002/​cne.10339.
Sullivan, E.V., Müller-​Oehring, E., Pitel, A.-​L., et al. (2013). A selective insular perfusion deficit
contributes to compromised salience network connectivity in recovering alcoholic men. Biological
Psychiatry, 74, 547–​55. DOI: 10.1016/​j.biopsych.2013.02.026.
Suñer-​Soler, R., Grau, A., Gras, M.E., et al. (2011). Smoking cessation 1 year poststroke and damage to
the insular cortex. Stroke, 43, 131–​6. DOI: 10.1161/​STROKEAHA.111.630004.
Sutherland, M.T., Mchugh, M.J., Pariyadath, V., and Stein, E.A. (2012). Resting state functional
connectivity in addiction: lessons learned and a road ahead. Neuroimage, 62, 2281–​95.
DOI: 10.1016/​j.neuroimage.2012.01.117.
Tiffany, S.T. (1999). Cognitive concepts of craving. Alcohol Research and Health, 23, 215–​24.
Verdejo-​Garcia, A. and Bechara, A. (2009). A somatic marker theory of addiction. Neuropharmacology,
56(Suppl. 1), 48–​62. DOI: 10.1016/​j.neuropharm.2008.07.035.
Villafuerte, S., Heitzeg, M.M., Foley, S., et al. (2012). Impulsiveness and insula activation during
reward anticipation are associated with genetic variants in GABRA2 in a family sample enriched for
alcoholism. Molecular Psychiatry, 17, 511–​9. DOI: 10.1038/​mp.2011.33.
Vorel, S.R., Bisaga, A., Mckhann, G., and Kleber, H.D. (2007). Insula damage and quitting smoking.
Science (New York), 317, 318–​9-​author reply 318–​9. DOI: 10.1126/​science.317.5836.318c.
Wise, R.A., Wang, B., and You, Z.-​B. (2008). Cocaine serves as a peripheral interoceptive conditioned
stimulus for central glutamate and dopamine release. PloS One, 3, e2846. DOI: 10.1371/​journal.
pone.0002846.
Xue, G., Lu, Z., Levin, I.P., and Bechara, A. (2010). The impact of prior risk experiences on
subsequent risky decision-​making: the role of the insula. Neuroimage, 50, 709–​16. DOI: 10.1016/​
j.neuroimage.2009.12.097.
Yousefzadeh-​Fard, Y., Hadi-​Gharedaghi, M., Esmaeili, S., et al. (2013). Stroke modifies drug
consumption in opium addicts: role of the insula. Basic and Clinical Neuroscience, 4, 29–​36.
Zhang, X., Salmeron, B.J., Ross, T.J., et al. (2011). Factors underlying prefrontal and insula structural
alterations in smokers. Neuroimage, 54, 42–​48. DOI: 10.1016/​j.neuroimage.2010.08.008.
224  
  225

Section IV

Perspectives from behavioral
economics and cognitive
psychology
226  
  227

Chapter 13

Palpating the elephant: Current

theories of addiction in light
of hyperbolic delay discounting
George Ainslie

Abstract
Addiction appears to contradict expected utility theory and has therefore
been the subject of many re-examinations of motivation. It is variously
said to arise from and/or be maintained by conditioning, habit learning (as
distinct from the goal-directed kind), the elicitation of counterfeit reward in
the midbrain, accelerated delay discounting, hyperbolic delay discounting,
and unspecified sorts of disease or compulsion that imply addiction is not
motivated at all. Each of these models has some roots in observation but
each has problems, particularly in accounting for addictions that do not
need a neurophysiologically active agent, such as gambling or videogames.
An implication of hyperbolic delay discounting—recursive self-prediction—
adds necessary mechanisms for addiction within a motivational framework.
An addict’s “force of habit” may be motivated by what amounts to
accumulated consumption capital within an endogenous reward process. In
a recursive motivational model the addict’s impaired responsibility is more
like bankruptcy than disease.

1 Introduction
The term “addiction” invokes the metaphor of enslavement, and is thus used to describe
some force that overwhelms the other motives in a person’s life. This force must be unwel-
come; the term is not applied to spiritual rebirth, or an artist’s devotion to her work. Since
psychoactive substances are often at the root of such a force, observers often equate addic-
tion with habituation to one of these substances—​perhaps wishfully, since the possibility
of being ensnared by the workings of your ordinary motivational apparatus is unsettling,
to say the least. Nevertheless, this apparatus on its own can clearly generate overwhelming
incentives that people report being unable to resist—​to gamble, to shoplift, to seek escapist
entertainment. Of course “resistibility” is not a hard datum, even with psychoactive sub-
stances, so addiction lies at the end of a continuum rather than forming a discrete entity. The
228  

228 Current theories of addiction in light of hyperbolic delay discounting

important question is not the location of a boundary between addictions and bad habits, but
how you can have robust motives for either that fail to equilibrate with your other motives.
In discussions of human decision-​making addiction has been the elephant in the
room, the phenomenon that contradicts economists’ expected utility theory (EUT) and,
analogously, behavioral psychologists’ reward maximization, cognitive/​social psychol-
ogists’ goal pursuit (among others), and philosophers’ rational choice with all things
considered. Like the proverbial blind men we have come up with many models of this
elephant: inborn susceptibility to some temptations, entrapment by the threat of with-
drawal symptoms, weak will, aroused appetite (“hot thinking”), force of habit, decep-
tion by biochemical agents, and inconsistent evaluations of the future. These models get
recruited to one side or the other of a debate as to whether people can be held responsi-
ble for their addictive choices, which is often framed as a question of whether addiction
is a disease:

Disease: Addictive choices are involuntary, or compulsive in the sense of being beyond moti-
vation. This is a frequent self-​report by addicts. Writers have often cited William James’
quotation from an alcoholic, “Were a keg of rum in one corner of a room and were a
cannon constantly discharging balls between me and it, I could not refrain from pass-
ing before that cannon in order to get the rum” (James 1890, p. 543).1 Objectively, there
is a wide range of evidence for genetic predispositions to develop most of the familiar
addictions (Kreek et al. 2005) and for physical changes in addicts’ brains (Goldstein
and Volkow 2002).
Not a disease: Addictive choices are voluntary. Other writers have pointed out that con-
suming an addictive substance requires goal-​directed and often quite sophisticated
behavior, and furthermore that most addicts eventually stop this behavior on their
own (Heyman 2009). Some writers have gone so far as to say that addictive behavior
is a rational response to the reward offered by the activity in question, and that addicts
differ from other people only in their lower valuation of the future. Economists Gary
Becker and Kevin Murphy famously claimed, “Addictions, even strong ones, are usu-
ally rational in the sense of involving forward-​looking [utility] maximization with sta-
ble preferences” (Becker and Murphy 1988, p. 675). Researchers continue to make this
claim (e.g. Lemenze and Murray 2013).

As is often the case in longstanding debates, neither side is likely to have the whole truth.
Addictive acts are clearly motivated—​they don’t take place in front of policemen or, prob-
ably, cannons. Conversely, a failed wish to stop or limit these acts is part of their definition
(APA 2013). Unlike many real addicts, the rational addict of Becker and Murphy would
not buy treatment to help her stop. In addiction the motive for the activity breaks free
from the influence of its greater long-​term cost—​but temporarily. If such a motive were

1 Nick Heather has pointed out to me that James got the quotation from Benjamin Rush (1830). Medical
Inquiries and Observations upon the Diseases of the Mind. 4th edition. John Grigg: Philadelphia, p. 264.
  229

Parts of the elephant 229

continuous, that is, dominant at all times relative to when the choice was made, we would
have to agree with Becker and Murphy that it was just a personal taste.
The basic question is: Why does someone repeatedly do something that she expects to
regret? The answer that first comes to mind is that a chemical somehow tricks the brain;
but if this is happening we still need to know how that trick translates into conflicted moti-
vation. Furthermore, the existence of addictive patterns that do not involve a substance
implies that addiction springs from something intrinsic to the way motivation operates,
something that doesn’t depend on being unlocked by a neurophysiological turnkey. In
addition, addicts commonly continue their behavior while reporting that the cigarette
or drink or hit is no longer pleasurable. Thus three questions challenge EUT: why people
are addicted, how we can become addicted to something nonphysical, and how addictive
choices can become fixed so they persist after the pleasure is gone.

2  Parts of the elephant

Various mechanisms for addiction have been proposed. Some of them occur in everyone,
raising the question of how they are different in addicts. None addresses addiction with-
out physical reward, but I will make a suggestion about that (see Section 4.3). Several deal
with the question of persistent activity without pleasure (Sections 2.2.5–​2.2.6), but they
have limitations that call for additional explanation.

2.1  Extra rewardingness

An inherited or acquired condition that makes an addictive activity extraordinarily
rewarding may be a major factor in a person’s loss of control over it. It is common for chil-
dren of recovering alcoholics to grow up shielded from alcohol, then to plunge abruptly
into heavy drinking as soon as they first experience it. A high degree of rewardingness
may account for the power of an addiction, and discoveries of pharmacologic agents that
block this rewardingness have sometimes led to effective controls, but rewardingness
itself does not account for the conflictual nature of addiction—​why addicts continue their
activity while saying they want to stop. Steep devaluation of the future, itself hereditary
(Anokhin et al. 2011), also makes fast-​paying activities more differentially rewarding, but
by itself also fails to account for their being conflictual.

2.2  Avoiding withdrawal

The first account of addiction with a physiological basis came from the common observation
that addicts to many kinds of drug become tolerant of them and need increased consumption
to attain a given effect, preparing the way for an increasing rebound by painful physiological
“opponent processes” when the drug is stopped (Solomon and Corbit 1973). Addicts are said
to then continue a relatively unrewarding consumption rather than face withdrawal symp-
toms, returning to consumption after the opponent processes have abated in turn (Ahmed
and Koob 2005). But the opponent process pattern describes the arousal of appetites/​emo-
tions in general (Solomon and Corbit 1974). Although the threat of withdrawal may serve
230  

230 Current theories of addiction in light of hyperbolic delay discounting

to prolong binges, some opiate addicts learn to endure and even induce withdrawal so as to
cheapen their habits. Some substances, notably cocaine, do not induce physiological with-
drawal; and some non-​substance addictions do (Blaszczynski et al. 2008). Becoming trapped
by the threat of withdrawal is not itself a sufficient explanation for addiction.

2.3  Weakness of will

A weak will has always been a folk explanation of addiction, and recent experiments on
training the will have suggested it as a dimension to explore (Baumeister et al. 2006). It is
a hard hypothesis to evaluate without details of how the will operates, but the observation
that addicts often show intact willpower in other areas of their lives makes it doubtful as
a general theory. I present a more specific hypothesis in Section 4.1.

2.4  Hot thinking

Decision-​making during arousal of appetite is biased toward consumption, and has
recently been studied as hot thinking (Kahneman 2011), which is accompanied by
increased activity in the limbic and paralimbic brain areas (Bechara 2005).2 Arousal
clearly increases motivation for addictive activities, but to evaluate its role in addiction
its relationship to reward needs clarification. In particular, we need to look at whether
arousal can be imposed on a person simply by association, or whether it depends on
contingent reward.
Arousal of appetite is conventionally thought of as governed by classical conditioning
rather than by differential reward: when a stimulus is regularly followed by a motivating
experience it leads that experience to be anticipated. The first laboratory experiments on
conditioning led to a theory that behaviors could be elicited in this fashion (for instance,
clinking ice cubes leading to conditioned drinking), but careful controls revealed that
only information (stimuli, perceptions) can be conditioned (Rescorla 1988), and any
behaviors that occur are motivated by the resulting expectation (clinking ice cubes lead-
ing to an expectation of alcohol). However, it has never been clarified whether arousal of
appetite should count as a behavior or, as often seems to be assumed, as part of a (con-
ditioned) expectation. The distinction is important because expectations and behaviors
operate differently. Expectations wane when not met. Behaviors depend on contingent
reward, and thus might increase when an expectation is not met. “Conditioned” arousal
that is said to incubate in the absence of new pairings, or which occurs to mere reminders
that do not predict an unconditioned stimulus, is hard to explain unless it is a behavior.
I have argued elsewhere that arousal of appetite is better seen as a behavior, one that is
reward-​dependent but mostly not deliberate (Ainslie 2010). Appetites could be thought
of as foraging for possible rewards just as animals forage for food, and responding to
available rewards more like your livestock than like your muscles. The Roman physician
Galen already knew this, pointing out that anger was tamed like a horse, but that the

2 Emotions such as anger and the thrill of gambling lack an external consumption good but otherwise
behave like arousable appetites. I include them in the appetite category (see Ainslie 2001, pp. 65–​9).
  231

Parts of the elephant 231

“concupiscible power” (lust), like a wild boar or goat, had to be controlled by starvation
(1963 p. 47). Galen would undoubtedly have counted the craving for hard drugs as a goat.
This consideration should affect how we see the role of hot thinking in addiction, and is
developed further in Sections.4.2-​3.

2.5 Habit
Recent research into habit formation has led to the proposal that drug habits lose their
responsiveness to motivation, thus continuing after they lose their pleasurable effects
(Everitt and Robbins 2005). That is, addiction starts voluntarily but is hardened by rep-
etition into something “automatic,” beyond motivation. This proposal refers to a large
body of recent human and nonhuman research on how behavior can be governed in two
distinct systems: learning sequences of responses is “model-​free,” whereas the planning
of choices is “model-​based” (Balleine and Dickinson 1998). The two kinds of reward
learning—​habitual response sequences and planning models—​have been reported to be
subtended at least in part by different areas of the brain—​the unmapped (or model-​free)
habit system by the ventral striatum, and the model-​based planning system by the ven-
tromedial frontal lobes (Balleine and O’Doherty 2010). An animal that always finds food
in one place stops looking around and simply repeats the movements that get it there,
switching from its planning system to its habit system. Its moves can become so rou-
tine that they persist after the reward has been devalued, for instance by satiety (Yin and
Knowlton 2004)—​hence the suggestion that control by the habit system can be a model
for addictions.
The overlearning effect is not peculiar to nonhuman animals. On my drive home from
work city blocks become neighborhoods, neighborhoods become towns, and the towns
combine into a whole route, which unfolds without thought unless I interject a choice. The
existence of a habit system that operates differently from a planning system is well estab-
lished, but “mindless” would always have been a better term than “automatic” or “robotic”
for the behaviors it governs. The latter terms imply an imperviousness to contrary moti-
vation, whereas the habit system normally gives way to the planning system whenever a
choice is subject to conflicting motives. Going out of my way to the grocery store requires
only minor acts of vigilance, as does (probably) learning to find a new drug supplier.
Addictive “habits” have very little to do with mindless repetition, but on the contrary
require a high degree of flexible, goal-​directed behavior to evade suppression by a hostile
society. Furthermore, brain imaging that tracks whether a person is using model-​based
or model-​free strategies to get rewards has shown that the supposed model-​free center,
the ventral striatum, is active during both (Daw et al. 2011). The two kinds of learning
occur in parallel and are intimately connected:  “Over the course of learning, behavior
migrates from being goal-​directed to being non-​goal-​directed” (Dayan 2009). The two
systems interact from moment to moment: “People negotiate the trade-​off between the
two systems dynamically as a function of concurrent executive function demands,” for
instance when a distracting task is introduced or removed (Otto et al. 2013). Perhaps “the
model-​based system trains the model-​free system by replaying and simulating experi-
ence” (Gershman et al. 2014).
232  

232 Current theories of addiction in light of hyperbolic delay discounting

However, this difficulty does not entirely dispose of habit as a mechanism for the main-
tenance of addiction. Proponents of habit as an explanation argue that addictive sub-
stances impair the brain functions that govern switching between the systems. Ablating
the parts of a rat’s brain that call up model-​based choice may keep the rat from ever
learning to respond to changed contingencies (Robbins and Everitt 2007). Analogously,
patients who have had damage to their ventromedial prefrontal cortex do less well than
normals on changed-​contingency tasks, although the deficit has been reported to be only
partial (Fellows and Farah 2005). The reported effects of addiction on such tasks have
also been moderate, with one striking exception: cocaine addicts have been reported to
perform markedly less well than non-​addicts on changed contingency tasks, whereas
amphetamine and opiate addicts do roughly as well as non-​addicts (Ersche et al. 2008).
Ersche and colleagues speculate that the surprising difference between cocaine and the
pharmacologically similar amphetamines stems from cocaine’s unique release of a related
neurotransmitter, serotonin. Thus the habit hypothesis retains some merit for cocaine
specifically, although we might still question whether impairment measured in seconds in
a laboratory task implies inability to re-​evaluate rewards and costs over the longer term.

2.6 Counterfeit reward
It has been suggested that some addictive substances create counterfeit reward in the
brain—​that is, they deceive you into expecting reward where none will be forthcoming.
For instance, opiates have been said to “trick the system into believing that it has just
received a strong reward, which it will learn to return to” (Redish et al. 2008). But if opiates
function as rewards, they must be rewards, however unfortunate their long-​term results.
In a trickier suggestion of counterfeiting, dopaminergic agents such as cocaine can
become “wanted” on the basis of their “incentive salience,” an effect that may persist
despite not being followed by a supposedly true reward, “liked” activation of the opiate
system (Berridge and Kringelbach 2008). This effect has been produced in both non-
humans and human patients by stimulating the dopaminergic areas of their ventral
striata. Once stimulation has started, the subjects work vigorously to repeat it, but do
not show facial signs (rats) or give reports (patients) of pleasure. Moreover, the subjects
often fail to start self-​stimulation again once it has been stopped for a while. The areas
of the midbrain that are involved are at least roughly the same as those that govern the
habit (model-​free) system. Evidence from local ablations in nonhumans and dopamin-
ergic neuron loss in Parkinsonian patients suggests that reward indeed has two sepa-
rable components, wanting and liking. Liking does not lead to action unless wanting
also occurs. The incentive salience that leads to wanting has been regarded as “non-
hedonic” (Berridge 2003)—​unrelated to reward; but I have argued that this divorce is
unlikely (Ainslie 2009). The salience of a process must compete for your attention in the
same currency as other rewards, and must therefore have a reward value itself. The sharp
decline in the attractiveness of striatal stimulation with distance suggests that salience
creates a very short-​term reward, something like an itch. Addictions seem also to lose
their pleasurable component in this pattern: not a decline into frank aversiveness, but
  233

Hyperbolic discounting in a nutshell 233

a decreasing duration of pleasure or relief—​the smoker who lights cigarettes and then
stubs them out, the opiate addict whose high turns stale rapidly, the overeater who must
nibble repeatedly. In this vein, the old cartoon on my colleague’s wall keeps drawing my
attention despite no longer yielding pleasure. There may be many cases where a “want-
ing” component endures while a “liking” component fades, and these may be related to
the properties of different neurotransmitter systems. However, choice must still be deter-
mined by the final common selective factor of reward. The deterioration of pleasure into
an itch-​like pattern might well be regarded as a disease—​we call eczema a disease, after
all—​but scratching it is still a motivated behavior.

2.7  Hyperbolic delay discounting

Addiction is a predictable consequence of hyperbolic delay discounting, a universal trait.
Humans have always lived on the edge of addiction. If the boundary of addiction is placed
close to the “bad habit” range of the continuum, half the US population is reported to be
addicted (Sussman et al. 2011). Those of us who have avoided the named addictive diag-
noses are nevertheless apt to suffer from habitual overvaluation of the present moment,
as in chronic procrastination (Andreou and White 2010), overuse of credit, or unrealis-
tic future time commitment (Mullainathan and Shapir 2013). Theories of why this is so
date back at least to the theological idea of original sin, and generally point to something
innate in our makeup. An inborn hyperbolic delay discount function can explain both
our predisposition to addictions and our repeated seeking of addictive agents when their
mean pleasure value is negative. This proposal is converse to the others just reviewed—​it
reverses the figure and the ground, as it were. If a tendency to addiction is universal, the
question must be how we mostly avoid addiction. Related questions are why this avoid-
ance often fails, and why recovery is harder than avoiding the addiction in the first place.
In addition, by itself hyperbolic discounting does not account for nonphysical addictions.
Nevertheless, I will suggest that hyperbolic delay discounting provides a framework for
the other accounts of addiction.

3  Hyperbolic discounting in a nutshell

Hyperbolic discount functions describe temporary preferences for smaller, sooner (SS)
rewards over larger, later (LL) rewards during the period when an SS reward is immi-
nently available (Formula 1, Figure 13.1a); they also describe more preservation of value
at long delays than do conventional, exponential discount functions, supporting a role for
foresight in self-​control (Ainslie 2001, 2005, 2012).

Present value = Value0 / 1+ ( k × Delay ) Formula 1

where Value0 = value if immediate and k = degree of impatience (Mazur, 1987).

However, the notion of hyperbolic delay discounting has been controversial, since it
might be thought to have been heavily selected against in evolution. Further, it has not
234  

234 Current theories of addiction in light of hyperbolic delay discounting

been found in some experiments (Montague et al. 2006), as well as in mature financial
behavior, where an exponential discount curve is evident. These objections are answer-
able. As to the evolutionary objection, the attenuation of a perception in proportion to
the intensity of the stimulus for it (the Weber–​Fechner law) was a basic pattern in sen-
sory domains long before species became engaged in intertemporal choice over periods
of more than minutes. Before hominids evolved, animals did not plan for the future.
Their farsighted behaviors were based on hardwired motives, to mate or migrate or
defend territory, which pay off as the behavior is performed. Even in foraging, where
it should be adaptive to weigh a distant food patch consistently against a present one,
such consistency must not have been a strong selective factor, since nonhuman animals
do not show it (Kagel et al. 1986). Thus it is quite believable that the effect of delay on
reward evolved within the constraint of the Weber–​Fechner law, as a hyperbolic func-
tion (inverse proportion). As to people’s sometime ability to make consistent choices
over time between SS and LL rewards, the high tails of the hyperbolic discount curve
describe an incentive that can lead self-​aware individuals to approximate exponential
curves (Ainslie 1991 and Section 4.1).
Despite these answers, the hyperbolic discount function has been disruptive within behav-
ioral science, particularly in economics (Montague et al. 2006; Ainslie 2012). A proposed
compromise has been hyperboloid (quasi-​hyperbolic) functions, conventional exponential
discount curves with an extra curvature (“beta spike”) that comes from the conditioned
arousal of appetite/​emotion just before the SS reward is due (Formula 2, Figure 13.1b):

Present value = Valuea × β × δ Delay Formula 2

where Value0 = value if immediate and β has one of only two values, β = 1 when reward
is imminent or 0 < β < 1 at all other times; δ = 1—​discount rate (McClure et al. 2004).
Hyperboloid curves describe the day-​to-​day pattern of many addictions, but do not fit
those temporarily preferred behaviors that don’t entail arousal, such as procrastination;
nor do they fit behaviors that unfold over longer time periods than arousal lasts, such as
failure to save for the future, or that occur during time periods so short that preferences
for SS and LL rewards both occur during arousal, as is the case with food deliveries to

(a) (b) (c)

β
VALUE

TIME
Figure 13.1 Values of expected SS and LL rewards (a) discounted hyperbolically (b) discounted
exponentially with an added beta spike (c) discounted hyperbolically and summed with the
values of later pairs
  235

Hyperbolic delay discounting motivates recursive self-prediction 235

nonhumans with differential delays of a few seconds (e.g. Ainslie and Monterosso 2003).
Just as importantly, the beta spike does not raise the tails of the discount curves to the
levels described by hyperbolas; thus hyperboloid curves do not account for the very low
discount rates that subjects report over delays of decades (Cropper et al. 1992), or for the
increased LL choice seen when choices are bundled together (Figure 13.1c; Section 4.1).
It may also be significant that a beta spike superimposed on otherwise exponential curves
is inconsistent with proposed derivations of hyperbolic discounting from more basic
phenomena, such as an aggregation of noisy delay estimates before each choice (Kurth-​
Nelson and Redish 2012). Therefore I have argued that the underlying discount function
is purely hyperbolic, not a hyperboloid compound (Ainslie 2012). This is not to deny that
aroused appetite is a factor in addiction, but I will argue that in humans it is governed by
processes that follow from hyperbolic discounting, rather than being an exogenous factor
imposed by association (see Sections 4.2–​4.3).
The question remains of how people mostly manage to avoid addiction, and why failure
to do so makes recovery difficult. Of course, there are many ways that a society learns
to keep the chance for addictive activities at a distance or to attach penalties to them.
Beyond committing yourself to be influenced—​or bound—​by your society, there are ways
you can physically commit yourself, keep your attention away from temptations, or build
emotional resistance to them (reviewed in Ainslie 2001, pp. 73–​78). However, the inter-
nal self-​control that has become a major factor in modern societies hinges on recursive
self-​prediction.

4  Hyperbolic delay discounting motivates recursive

self-​prediction
Self-​prediction is a familiar phenomenon in behaviors that are only marginally under
voluntary control. You predict how they will develop by monitoring their current activ-
ity, and this prediction feeds back to change that development. Visceral processes such as
anger, panic, nausea, sleep (in insomniacs), and urination (in men with prostatic hyper-
trophy) are accelerated by signs that they are already happening. Darwin famously pro-
posed this phenomenon as a site of self-​control:
The free expression by outward signs of an emotion intensifies it. On the other hand, the repres-
sion, as far as this is possible, of all outward signs softens our emotions. He who gives way to violent
gestures will increase his rage; he who does not control the signs of fear will experience fear in
greater degree.

(Darwin [1872] 1979, p. 366)

This observation led him, James, and Lange mistakenly to propose that physical reflexes
originated such processes (Rolls 2005, pp. 26–​28); but the processes are nevertheless mod-
ulated by the feedback that Darwin describes. The relevance to our topic is that hyperbolic
discounting also reduces the sway of your current intentions over the future preferences
you can expect to have, making self-​prediction an important factor in many choices.
Three patterns figure in human addiction.
236  

236 Current theories of addiction in light of hyperbolic delay discounting

4.1  Intertemporalbargaining creates willpower but

also circumscribed failures
Hyperbolic discounting makes you uncertain that you will follow your own current inten-
tions, creating an incentive for recursive self-​prediction—​a topic that I have called pico-
economics (micro-​micro-​economics; Ainslie 1992). To the extent that you notice how
your current choice between an SS and LL reward predicts similar choices in the future,
you create a bundle of expectations that depend at least somewhat on the current choice—​
and which thus motivate that choice, as in Figures 13.1c and 13.2a. Seeing your current
choice as a test case creates a variant of repeated prisoner’s dilemma with your expected
future selves, and your moves in this game over time establish personal rules for when
you will count a choice of SS as a defection (evidence reviewed in Ainslie 2012). Then
the cost of eating a serving of a forbidden food, for instance, will only slightly be its effect
on your weight or health, and will mostly be its damage to the credibility of your diet.
I have presented evidence elsewhere that this example of recursive self-​prediction is the
central process of willpower (Ainslie 2001, 2005, 2012), although the name “willpower”
is sometimes given to simple response inhibition that is achieved without weighing the
alternatives.3 Such intertemporal bargaining seems to be how people learn to achieve con-
sistent choice over time despite an innate tendency to form temporary preferences for SS
rewards. However, when this bargaining fails it may be a factor that not only permits but
stabilizes addictions:
In personal rules the long tails of hyperbolic discount curves are effectively bundled
together, creating increased incentive to choose LL rewards but also to evaluate choices
in a legalistic manner. When choices are worth less in their own right than for the prec-
edents they set, choice-​making is removed from the here-​and-​now. In the extreme you
may become entangled in your rules—​rigid, emotionally unengaged, compulsive. This
development may be ultimately as undesirable as the impulses it targets.4

3 Response inhibition, as in go/​no-​go tasks, is clearly an unstable basis for resisting temptation over
time, unlike a stake of substantial self-​expectation against it. Response inhibition involves brain and
neurotransmitter activity that is at least partially distinct from the activity that accompanies choice
(Chambers et al. 2009; Luo et al. 2012). Unfortunately, many laboratory tasks designed to test the sup-
posed muscle-​like properties of will have mixed the two processes, by measuring subjects’ willingness
to continue response inhibition tasks (Kurzban et al. 2013).
4 Compulsions should be distinguished from impulses and addictions. I recommend that “compulsive”
be reserved for actions that are strictly confined by personal rules, rather than being overwhelmingly
motivated in general, or even being beyond motivation, as when people talk about “compulsive drink-
ing.” Conversely, rule-​bound, compulsive behavior is sometimes called addictive, as in workaholism,
perfectionism, anorexia nervosa, and obsessive-​compulsive personality disorder; but this usage is con-
fusing, since the motivational dynamic is quite different—​overcontrol rather than failure of control.
As a group, compulsive traits have much in common with each other and differ from substance-​and
thrill-​based activities not only in their lack of arousal but in being more consistently preferred and
integrated with your values (Ainslie 2009). Compulsions should not be called addictions or impulses,
and vice versa.
  237

Hyperbolic delay discounting motivates recursive self-prediction 237

(a) WILL (b) WILL APPETITE

Evaluate how this Evaluate how this Evaluate

choice affects choice affects likelihood of
future self-control future self-control choosing addictive
activity

Weigh
Adjust Weigh Adjust Weigh incentives for
incentives for Adjust
choice incentives for choice arousal arousal
current choice current choice

External incentives External incentives Occasion for arousal

Figure 13.2  Recursive self-prediction (a) when the current deliberate choice is a test case (b) as in
(a), but with arousal of appetite encouraged by the expectation of changing the deliberate choice

Addictions are not simply recurrent impulses, but complex compromises with your
long-​term interests that develop when you try repeatedly to resist a temptation and fail.
At first you may try to repair the damage to your expectation of self-​control by seeing
your lapse as a special case (rationalization), by staking a wider set of expectations against
this kind of lapse to create stronger resolve, or by simply repressing or denying the lapse.
However, after repeated failures you have an incentive to protect the expectations upon
which your will is based from further damage, by pulling back from attempts to control
this kind of lapse. Thus a failure of will may stabilize an addiction by defining a sort
of vice district where the law doesn’t go. Your other endeavors may continue perfectly
well, as has been the case with many figures who stayed productive despite an addiction
(Coleridge, Halsted, Erdὂs); but your choices about where to attempt to use willpower
create precedents that ramify as new situations arise. The resulting fabric of compromises
creates a disincentive to attempt a radical change. Conversely, addiction in a binge pattern
may protect personal rules by defining limited exceptions, for instance a resolution to be
considerate, perfectionistic, and monogamous that goes into abeyance when “the alcohol
is talking” (Ainslie 1999). The advice of the Anonymous organizations is to acknowledge
the futility of intertemporal bargaining ploys with respect to the index temptation (“I
am powerless over …”), a perception that actually serves to let an all-​or-​nothing stake of
expectation grow gradually (“one day at a time”) without being hedged by rationaliza-
tions. The advice to discern a higher power as taking part in your efforts may also offer
strategic protection against the abstinence violation effect (discussed in Ainslie 2004).
These are several rationales by which intertemporal bargaining may structure the com-
plex relationships that people have with their addictions.

4.2  Sudden craving is a positive feedback phenomenon

Addicts often report that reminders of addictive activity, such as watching it on TV, have
led to sudden craving followed by relapse into addiction (Stewart 2008). It is often said
that such reminders are conditioned stimuli and are thus a logical target for therapy.
But this kind of “conditioned craving” is not the same phenomenon that is observed in
238  

238 Current theories of addiction in light of hyperbolic delay discounting

laboratory experiments. Mere reminders don’t predict increased probability or imme-

diacy of the event. By contrast, when smoker subjects are given cues that are sometimes
followed by cigarette puffs, their arousal tracks the delay and probability of the puffs with
great accuracy (Field and Duka 2001). But if reminders lead to sudden arousal in the
absence of predictive information, how does it happen?
I have argued elsewhere that your consumption of an addictive good in real life is limited
not so much by its physical availability as by your willingness to consume it (Ainslie 2010).
Your entertainment of an appetite for the good is at least partially reward-​dependent, and
is often a rewarding activity itself, as witnessed by such entertainment being a venial sin
for Catholics. Even when the arousal is not rewarding per se, it is rewarded when it brings
about consumption of the addictive good. If you have made a personal rule against the
activity and are not fully confident that you will obey it, you will be apt to see arousal of
your appetite for it as a sign that you may give in, which would in turn be positively fed
back to encourage the appetite (Figure 13.2b). If you are fully confident of obeying your
rule, the appetite will stop arising, as has been reported to be the case with Orthodox Jews’
craving for cigarettes on the Sabbath (Dar et al. 2005). A corollary, that sharp arousal of
appetite by reminders should occur especially where willpower is strained—​as in con-
trolled eaters or recovering smokers—​is empirically testable, but has not been tested.
In this view, arousal is reward-​dependent, and thus a behavior. The rewards it depends
on may or may not be external. Reminders serve as occasions for it. If craving is reward-​
dependent, rather than conditioned, a practical implication is that counter-​conditioning
treatments will fail, which is in fact the case (Conklin and Tiffany 2002). Furthermore,
when we theorize that the cues associated with addictive activity are not drivers of
reflex arousal, but signals of opportunities for appetites to challenge self-​control, we are
spared having to deal with two separate selective principles for components of addictive
activities—​reward and classical conditioning, the latter having often been proposed as an
explanatory deus ex machina.

4.3  Endogenous reward is the basis of non-​substance

addictions and the “force” of addictive habits
A fundamental property of human reward has been ignored in most motivational theo-
ries: sources of reward may grow in a person without relation to any hardwired source
of value. Psychologists rarely talk about the roots of reward. When they do, they reveal
a widespread assumption—​that reward has to come from events that are innately able
to induce it. Events that do not reward by themselves are said to be sought because they
predict events that do, either by association or deduction (“secondary rewards;” see Baum
2005, pp. 77–​86). This assumption makes some sense, since otherwise you might be able to
reward yourself at will; but on the other hand, it implies explanations of common motives
that involve highly fanciful chains of association, chains that somehow do not extinguish
despite years without their primary rewards. Behavioral science allows only occasion-
ally for the possibility that the reward process often operates without being backed, as it
were, by the hard currency of hardwired rewards—​for instance in the recent goal-​setting
  239

Hyperbolic delay discounting motivates recursive self-prediction 239

literature that focuses on optimizing the risk of failure (e.g. Koch and Nafziger 2011).
Such endogenous reward is distinguished from secondary reward only by being effective
when it does not predict a primary reward—​thus amounting to a primary reward in its
own right, but one that lacks an innate inducing mechanism.
Hyperbolic delay discount functions suggest a rationale for how reward operates when
it does not entail the prediction of innate inducers. To summarize a hypothesis developed
elsewhere (Ainslie 2013; an early formulation in Ainslie 1992, pp. 228–​327): to get reward
endogenously you just have to harvest it in imagination, but doing so at will makes it
satiate quickly. To build appetite you have to define adequately infrequent occasions to
claim the reward—​make bets, ask questions, set up games. Appetite grows as it is hin-
dered, becoming potentially more and more rewarding, but hyperbolic discount curves
continually tempt you to harvest the appetite early. You will come to value those activi-
ties most where your appetite has grown large before harvesting, and to elaborate them
into complex games that control an increasing potential for reward, forming in effect
consumption capital (as used by Becker and Murphy 1988). The potential for reward that
has so accumulated lacks a general name, but might be described as hedonic importance.
Occasions for endogenous reward need to be adequately singular—​not occurring so often
that little appetite builds up—​and they need to be adequately surprising, since it is hard to
keep attention from rushing ahead to a familiar occasion and harvesting some or all of its
reward by anticipation. You are motivated to maintain the singularity of occasions by the
same recursive logic that enforces personal rules: the problem with cheating at solitaire is
that you lose your belief that future wins will be as rare as you have been expecting, and you
thus undermine the hedonic importance of the game. Similarly, saying to yourself “it’s only a
game” when your sports team is losing reduces your capacity for joy when it wins. To ignore
a bargain is to reduce the importance of finding bargains; to let an insult pass, or, in another
context, to accept casual sex, is to reduce the importance of defending your honor; to act
carelessly with your collection or hoard, is to cast doubt on its value. Most choices about
hedonic importance are less explicit, but still of the form, “I find it important only if I expect
to go on finding it important.” Moreover, “the more I have found it important, the more
important I now find it.” What starts as a hobby may become the pursuit of a holy grail—​
sighting a rare bird or astronomical event, bringing off a risky shoplift, winning a profes-
sional prize, climbing Mount Everest (an extreme example in Leamer 1999). The value of
each occasion depends on your expectation of continuing to value such occasions, just as
the value of a fiat currency depends on the expectation that it will continue to be accepted.
The consumption capital represented by the hedonic importance of an activity gives
it rewarding potential beyond its instrumental (practical) value. Crucially, the challenge
of some instrumental accomplishments makes them singular, so they can accumulate
hedonic importance on top of their instrumental value. This in turn creates incentive to
preserve your belief in their instrumental value against evidence that it may have declined,
since instrumental value is what authenticates the occasions that the task offers for endog-
enous reward. Thus hedonic goals come to parasitize instrumental ones. In the present
context, you can learn to derive at least as much reward from the events of gambling per
240  

240 Current theories of addiction in light of hyperbolic delay discounting

se as from the money you get from it—​but only to the extent that you have built up its
hedonic importance. The prize money is still indispensable, but increasingly for its role
in making the gamble itself singular, rather than for how you might expect to spend it
(discussed in Ainslie 2013b). The benchmark occasions of an addictive activity—​winning
the longshot, escaping with the goods, notching up the sexual conquest—​produce endog-
enous reward just like the successes of more benign hobbies.
Similarly, the accumulation of hedonic importance may explain the persistence of
a substance addiction after the substance has lost much of its own rewarding power.
“Scoring” drugs does not just lead to the drug effect but also the occasion of a score like
a score in hockey, which may stay effective when the drug itself has habituated. Smokers
say they use cigarettes to mark significant moments in the day, and overeaters notoriously
consume food on occasions when their actual appetite has waned. But even when the
substance isn’t pleasurable it is still necessary in order to authenticate the occasions for
endogenous reward in its pursuit, just as a gambler needs real stakes even when resigned
to eventual loss, and a kleptomaniac needs real thefts even of items she doesn’t value.
The structure of endogenous reward contingencies that serve as consumption capital is a
likely explanation of the “force” of addictive habits, more so than is the kind of mindless
habit that governs overlearned activities.

5  An intertemporal bargaining framework for addiction

Strikingly, none of the various proposals for explaining addiction contradicts any other
one. Hyperbolic delay discounting suggests how they may all be part of the same elephant:
◆ Hyperbolic delay discounting creates a universal tendency to have temporary prefer-
ences for SS rewards.
◆ Inborn factors may make some of these rewards especially tempting, but if this
increased preference were consistent rather than temporary we would call it a taste,
not an addiction.
◆ Opponent physiological processes front-​load our pleasures, so we seek them when
close to them and then seek to defer withdrawal.
◆ The credibility of personal rules, which we summarize as willpower, is effective in
reducing temporary preferences, but it is selective—​dependent on intertemporal bar-
gaining history—​rather than being a general faculty. Both failed personal rules and
overly strict personal rules create incentives to stop attempting self-​control in some
circumstances, thus stabilizing addictions.
◆ SS alternatives are strengthened by arousal of appetite, and strategies to get them may
motivate this arousal recursively so as to overcome personal rules.
◆ Behavioral inflexibility—​habit—​may be partly a neurological consequence of long
cocaine use specifically, but it is also apt to be motivated by the patterns of endogenous
reward that grow in any highly engaging activity.
  241

An intertemporal bargaining framework for addiction 241

◆ Reward that is called counterfeit is actually just very short term, resulting in low aver-
age reward levels. Differential durability of dopaminergic connections may be a factor
in why some or all addictions evolve shorter highs, but in any case this shortening can
produce the itch-​like experience of being compelled without having pleasure.
◆ To the extent that a pattern of stimuli is singular and surprising it may gather
importance—​hedonic capital—​by providing occasions for endogenous reward. Where
this reward pattern has longer-​term costs it may form a non-​substance-​based addic-
tion. Endogenous reward may also add a non-​substance component to substance-​
based addictions, fostering their persistence after the decline of substance-​based
pleasure. Strategies for getting endogenous reward also parasitize some instrumental
activities to make use of the singularity of instrumental accomplishments, creating an
incentive to misperceive some gambling-​type activities as realistic.
Addictive behavior is clearly motivated; but some factors that affect motivation can be
reasonably called diseases. I  have argued here that the motivational patterns that are
inconsistent with conventional EUT become consistent when EUT is modified to accom-
modate hyperbolic delay discounting; and that the process that emerges as crucial in
addiction and recovery is recursive self-​prediction.
In this view can we say that an addict is unable to resist temptation? Yes, if in her inter-
temporal bargaining she can see no way to bundle together enough motivation to stake
against it. The factors that give short-​term rewards an edge might reasonably be called
diseases, but the resourcelessness that follows her repeated defections in intertemporal
bargaining is more like a budgetary crisis. When the addict can’t find enough credibility to
stake against her temptations to consume, we might say that she is no longer responsible
for her choices—​but because of bankruptcy, not sickness. An addict is not insensitive to
differential motivation; her long-​range interests just cannot get adequate leverage. There
is no natural test for whether such bankruptcy “exists” or not, nor even a test for when
we should appeal to the concept. Such a concept would necessarily be culture-​bound and
would resist theoretical benchmarks, just as attempts to define legal insanity have done.
And whereas the financially bankrupt cannot discover the funds they need by a radical
restructuring of their books, addicts often discover how to reframe their choices and sud-
denly regenerate their willpower (examples in Heyman 2009, pp. 44–​64). Responsibility
in addiction will always be an elusive concept.

Acknowledgments
This chapter was presented in part at the University of Cape Town, September 11, 2014.
The material is the result of work supported with resources and the use of facilities at
the Department of Veterans Affairs Medical Center, Coatesville, PA, USA. The opinions
expressed are not those of the Department of Veterans Affairs or of the US Government.
Many of the references of which I was author or co-​author are downloadable from www.
picoeconomics.org.
242  

242 Current theories of addiction in light of hyperbolic delay discounting

References
Ahmed, S.H. and Koob, G.F. (2005). Transition to drug addiction: a negative reinforcement model
based on an allostatic decrease in reward function. Psychopharmacology, 180, 473–​90.
Ainslie, G. (1991). Derivation of “rational” economic behavior from hyperbolic discount curves.
American Economic Review, 81, 334–​40.
Ainslie, G. (1992). Picoeconomics: The Strategic Interaction of Successive Motivational States within the
Person. Cambridge: Cambridge University Press.
Ainslie, G. (1999). The dangers of willpower: A picoeconomic understanding of addiction and
dissociation. In: J. Elster and O.-​J. Skog (eds), Getting Hooked: Rationality and Addiction.
Cambridge: Cambridge University Press, pp. 65–​92.
Ainslie, G. (2001). Breakdown of Will. New York: Cambridge University Press.
Ainslie, G. (2003). Uncertainty as wealth. Behavioural Processes, 64, 369–​85.
Ainslie, G. (2004). Gods are more flexible than resolutions. Behavioral and Brain Sciences, 27, 18–​9.
Ainslie, G. (2005). Précis of Breakdown of Will. Behavioral and Brain Sciences, 28, 635–​73.
Ainslie, G. (2009). Pleasure and aversion: challenging the conventional dichotomy. Inquiry, 52, 357–​77.
Ainslie, G. (2010). The core process in addictions and other impulses: hyperbolic discounting versus
conditioning and cognitive framing. In: D. Ross, H. Kincaid, D. Spurrett, and P. Collins (eds), What
Is Addiction? Boston: MIT Press, pp. 211–​45.
Ainslie, G. (2012). Pure hyperbolic discount curves predict “eyes open” self-​control. Theory and
Decision, 73, 3–​34. DOI: 10.1007/​s11238-​011-​9272-​5
Ainslie, G. (2013a). Grasping the impalpable: The role of endogenous reward in choices, including
process addictions. Inquiry, 56, 446–​69. DOI: 10.1080/​0020174X.2013.806129. http://​www.
tandfonline.com/​eprint/​8fGTuFsnfFunYJKJ7aA7/​full. Accessed 03/​052016.
Ainslie, G. (2013b). Money as MacGuffin: a factor in gambling and other process addictions. In:
N. Levy (ed.), Addiction and Self-​Control: Perspectives from Philosophy, Psychology, and Neuroscience.
Oxford: Oxford University Press, pp. 16–​37
Ainslie, G. and Monterosso, J. (2003). Building blocks of self-​control: increased tolerance for delay with
bundled rewards. Journal of the Experimental Analysis of Behavior, 79, 83–​94.
American Psychiatric Association (2013). Diagnostic and Statistical Manual of Mental Disorders. Fifth
Edition. Washington, DC:APA Press.
Andreou, C. and White, M. (2010). The Thief of Time: Philosophical Essays on Procrastination.
New York: Oxford University Press.
Anokhin, A.P., Golosheykin, S., Grant, J.D., and Heath, A.C. (2011). Heritability of delay discounting
in adolescence: a longitudinal twin study. Behavioral Genetics, 41, 175–​83.
Balleine, B. and Dickinson, A. (1998). Goal-​directed action: contingency and incentive learning and
their cortical substrates. Neuropharmacology, 37, 407–​19.
Balleine, B.W. and O’Doherty, L.P. (2010). Human and rodent homologies in
action control: corticostriatal determinants of goal-​directed and habitual action.
Neuropsychopharmacology, 35, 48–​69.
Baum, W.M. (2005). Understanding Behaviorism (2nd ed.). Oxford: Blackwell.
Baumeister, R.F., Gailliot, M., DeWall, C.N., and Oaten, M. (2006). Self-​regulation and
personality: how interventions increase regulatory success, and how depletion moderates the effects
of traits on behavior. Journal of Personality, 74, 1773–​801.
Bechara, A. (2005). Decision making, impulse control and loss of willpower to resist drugs: a
neurocogntivie perspective. Nature Neuroscience, 8, 1458–​63.
Becker, G. and Murphy, K. (1988). A theory of rational addiction. Journal of Political Economy, 96,
675–​700.
  243

An intertemporal bargaining framework for addiction 243

Berridge, K.C. (2003). Pleasures of the brain. Brain and Cognition, 52, 106–​28.
Berridge, K.C. and Kringelbach, M.L. (2008). Affective neuroscience of pleasure: reward in humans
and animals. Psychopharmacology, 199, 457–​80.
Blaszczynski, A., Walker, M., Sharpe, L., and Nower, L. (2008). Withdrawal and tolerance phenomenon
in problem gambling. International Gambling Studies, 8, 179–​92.
Chambers, C.D., Garavan, H., and Bellgrove, M.A. (2009). Insights into the neural basis of response
inhibition from cognitive and clinical neuroscience. Neuroscience and Biobehavioral Reviews, 33,
631–​46.
Conklin, C.A. and Tiffany, S.T. (2002). Applying extinction research and theory to cue‐exposure
addiction treatments. Addiction, 97, 155–​67.
Cropper, M.L., Aydede, S.K., and Portney, P.R. (1992). Rates of time preference for saving lives.
American Economic Review, 82, 469–​72.
Dar, R., Stronguin, F., Marouani, R., Krupsky, M., and Frenk, H. (2005). Craving to smoke in orthodox
Jewish smokers who abstain on the Sabbath: A comparison to a baseline and a forced abstinence
workday. Psychopharmacology, 183, 294–​99.
Darwin, C. (1872/​1979). The Expressions of Emotions in Man and Animals. London: Julan Friedman
Publishers.
Daw, N.D., Gershman, S.J., Seymour, B., Dayan, P., and Dolan, R.J. (2011). Model-​based influences on
humans’ choices and striatal prediction errors. Neuron, 69, 1204–​15.
Dayan, P. (2009). Goal-​directed control and its antipodes. Neural Networks, 22, 213–​19.
Ersche, K.K., Roiser, J.P., Robbins, T.W., and Sahakian, B.J. (2008). Chronic cocaine but not chronic
amphetamine use is associated with perseverative responding in humans. Psychopharmacology, 197,
421–​31.
Everitt, B.J. and Robbins, T.W. (2005). Neural systems of reinforcement for drug addiction: From
actions to habits to compulsion. Nature Neruoscience, 22, 3312–​20.
Fellows, L.K. and Farah, M.J. (2005). Different underlying impairments in decision-​making following
ventromedial and dorsolateral frontal lobe damage in humans. Cerebral Cortex, 15, 58–​63.
Field, M. and Duka, T. (2001). Smoking expectancy mediates the conditioned responses to arbitrary
smoking cues. Behavioural Pharmacology, 12, 183–​94.
Galen (1963). Galen on the Passions and Errors of the Soul. Trans. P.W. Harkins. Ohio: Ohio State
University.
Gershman, S.J., Markman, A.B., and Otto, A.R. (2014). Retrospective revaluation in sequential
decision making: a tale of two systems. Journal of Experimental Psychology: General 143, 182.
Goldstein, R.Z. and Volkow, N.D. (2002). Drug addiction and its underlying neurophysiological
basis: Neuroimaging evidence for the involvement of the frontal cortex. American Journal of
Psychiatry, 159, 1642–​52
Heyman, G.M. (2009). Addiction: A Disorder of Choice. Boston: Harvard University Press.
James, W. (1890). Principles of Psychology. New York: Holt.
Kagel, J.H., Green, L., and Caraco, T. (1986). When foragers discount the future: constraint or
adaptation? Animal Behavior, 34, 271–​83.
Kahneman, D. (2011). Thinking, Fast and Slow. Farrar, Straus, and Giroux.
Koch, A.K., and Nafziger, J. (2011). Self‐regulation through Goal Setting. The Scandinavian Journal of
Economics, 113(1), 212–​227.
Kreek, M.J., Nielsen, D.A., Butelman, E.R., and LaForge, K.S. (2005). Genetic influences on
impulsivity, risk taking, stress responsivity and vulnerability to drug abuse and addiction. Nature
Neruoscience, 8, 1450–​57.
Kurth-​Nelson, Z. and Redish, A.D. (2012). Don’t let me do that!—​Models of precommitment. Frontiers
in Neuroscience, 6: 138
244  

244 Current theories of addiction in light of hyperbolic delay discounting

Kurzban, R., Duckworth, A., Kable, J.W., and Myers, J. (2013). An opportunity cost model of subjective
effort and task performance. Behavioral and Brain Sciences, 36, 661–​726.
Leamer, L. (1999). Ascent: The Spiritual and Physical Quest of Legendary Mountaineer Willi Unsoeld.
Minot, ND: Quill.
Lemenze, C., and Murray, M.P. (2013). Delay discounting and alcohol abusers: More impatient even
when not impulsive? Working paper: SSRN http://​dx.doi.org/​10.2139/​ssrn.2372634.
Luo, S., Ainslie, G., Pollini, D., Giragosian, L., and Monterosso, J.R. (2012). Moderators of the
association between brain activation and farsighted choice. Neuroimage, 59, 1469–​77. DOI: 10.1016/​
j.neuroimage.2011.08.004
Mazur, J.E. (1987). An adjusting procedure for studying delayed reinforcement. In: M.L. Commons, J.E.
Mazur, J.A. Nevin, and H. Rachlin (eds), Quantitative Analyses of Behavior V: The Effect of Delay and
of Intervening Events on Reinforcement Value. Hillsdale, NJ: Erlbaum.
McClure, S.M., Laibson, D.I., Loewenstein, G., and Cohen, J.D. (2004). The grasshopper and the
ant: separate neural systems value immediate and delayed monetary rewards. Science, 306, 503–​507.
Montague, P.R., King-​Casas, B., and Cohen, J.D. (2006). Imaging valuation models in human choice.
Annual Review of Neuroscience, 29, 417–​48.
Mullainathan, S. and Shafir, E. (2013). Scarcity: Why Having Too Little Means So Much.
New York: Henry Holt.
Otto, A.R., Gershman, S.J., Markman, A.B., and Daw, N.D. (2013). The curse of planning: Dissecting
multiple reinforcement-​learning systems by taxing the central executive. Psychological Science,
doi:10.1177/0956797612463080.
Redish, A.D., Jensen, S., and Johnson, A. (2008). A unified framework for addiction: vulnerabilities in
the decision process. Behavioral and Brain Sciences, 31, 461–​87.
Rescorla, R.A. (1988). Pavlovian conditioning: it’s not what you think it is. American Psychologist, 43,
151–​60.
Robbins, T.W. and Everitt, B.J. (2007). A role for mesencephalic dopamine in activation: Commentary
on Berridge (2006). Psychopharmacology, 191, 433–​37.
Rolls, E.T. (2005). Emotion Explained. Oxford: Oxford University Press.
Solomon, R.L. and Corbit, J.D. (1973). An opponent process theory of motivation, II: cigarette
addiction. Journal of Abnormal Psychology, 81, 158–​71. doi: 10.1037/​h0034534
Solomon, R.L. and Corbit, J.D. (1974). An opponent process theory of motivation, I: Temporall
dynamic of affect. Psychological Review, 81, 119–​45. DOI: 10.1037/​h0036128
Stewart, J. (2008). Psychological and neural mechanisms of relapse. Philosophical Transactions of the
Royal Society Series B: Biological Sciences, 363, 3147–​58.
Sussman, S., Lisha, N., and Griffiths, M. (2011). Prevalence of the addictions: A problem of the
majority or the minority? Evaluation and the Health Professions, 34, 3–​56.
Yin, H.H., and Knowlton, B.J. (2004). Contributions of striatal subregions to place and response
learning. Learning and Memory, 11, 459–​63.
  245

Chapter 14

Addiction as social choice
Howard Rachlin

Abstract
A person’s self may be viewed as the more or less extended temporal and
social patterns of that person’s overt behavior, and nothing more. In this
view, your relation with your future self is essentially a social relation to be
studied along with your relation to other people. As these temporal and
social patterns are narrow or wide, you are acting selfishly or altruistically
with respect to your future self and other people. The value of a reward (or
utility of consumption) may be quantified as a function of social discounting
(“altruism”) as well as delay discounting (“self-control”). Addiction may
thus be conceived as a kind of hyper-selfishness. This chapter draws out the
implications, in theory and practice, of this social conception of addiction.

1  Addiction and choice
Often in everyday life a contradiction arises between the value of a particular act and the
value of a pattern of acts of which the particular act is a part. For example, the particular
act of smoking a cigarette may be highly valuable relative to other particular acts avail-
able at a given moment. However, the value of a high rate of smoking over several years is
almost always very low (or highly negative) relative to the value of other available, equally
extended, behavioral patterns. By “addiction” I mean repeated and consistent choice in
such cases of the highly valued particular act.
How may addiction, as defined above, be overcome? No matter how valuable the pat-
tern, it is individual choices that have to be made. Yet, at each individual moment, the
addict prefers the addictive act to its alternative. How can a person come to choose at a
given moment what she does not prefer at that moment? Before addressing this question
let us consider the standard behavioral measure of choice between immediately more
valuable and ultimately more valuable alternatives—​the delay discount function.
The following question poses an individual choice: Which would you rather have: $1,000
five years from now or $360 right now? Stony Brook undergraduates on average said they
were indifferent between these larger-​later (LL) and smaller-​sooner (SS) rewards. Based on
answers to a series of such questions, it is possible to construct a delay discount function
(Figure 14.1a). The X-​axis represents the delay of a $1,000 reward. The Y-​axis shows the
average amount of money Stony Brook undergraduates would accept immediately as an
246  

246 Addiction as social choice

(a)

1000

$ equivalent to $1,000 at delay D

800

600

400

200

0
0 500 1000 1500 2000
Days delayed (D)

(b)
80
$ forgone to give $75 to person N

70

60

50

40

30

20

10

0
0 20 40 60 80 100
Social Distance [N]
Figure 14.1  (A) A delay discount function. The average amount of immediate money ($) equal
in value to $1,000 to be obtained after a delay (D). (B) The average amount of money for the
participant ($) equal in value to $75 to a person at a social distance, N, from the participant. The
error bars are standard errors of the mean.

even exchange for $1,000 delayed by the duration shown on the X-​axis. At all points below
the curve they preferred LL; at all points above the curve, they preferred SS. The curve
itself is the locus of points of indifference between LL and SS. The steeper the curve, the
greater the tendency to prefer SS to LL rewards. Children have steeper delay discount func-
tions than young adults; young adults have steeper delay discount functions than older
adults (Green et al. 1994). Importantly for present purposes, the delay discount functions
  247

The extended self 247

of various kinds of addicts (cigarette, heroin, cocaine, alcohol, gambling, overeating,

and others) are generally steeper than those of non-​addicts (Odum 2011). Putting aside
claims about the precise shape of the function, its generality, its parameters, etc., you may
ask: What does the function actually say about addiction? Where is it located? Where does
it come from?

2  The extended self
The great seventeenth-​century philosopher Descartes believed that nonhuman animals
have no souls, and that their behavior is governed entirely by environmental forces inter-
acting with physical mechanisms in their bodies (later called reflexes). Your dog, accord-
ing to Descartes, is a pure machine. When he strains against the leash and pulls you down
the street to get to a female dog he is in principle (Descartes would say) like a ball rolling
down a hill. Like the ball, the dog, Descartes believed, never makes deliberate choices; his
behavior is just the resultant of the outer world acting on his entirely physical nervous
system. But (Descartes believed) we humans do have souls, nonphysical entities inter-
acting with our bodies at a dimensionless point located in the center of our brains. He
believed that human behavior is the result of a conflict between our willpower (exerted by
our souls) and the same environmental forces that govern all of the behavior of our dogs.
People with strong willpower would overcome their own reflexes; people with abnormally
weak willpower would presumably be susceptible to addictions of various kinds unless
they were kept on a tight leash (so to speak) by social constraints. Unlike a dog, an alco-
holic may (according to Descartes) choose to refuse a drink; but that choice may be over-
come by physical forces, and the alcoholic may find himself doing what he did not choose
to do. From the Cartesian point of view, therefore, a delay discount function would rep-
resent the outcome of this battle between our willpower and our reflexes. A steep delay-​
discount function, such as that evinced by addicts, would signal a weak will, a strong set
of reflexes, or both. From this viewpoint, choice would be entirely an internal process; the
battle of self-​control would be a battle between inner forces (where choices are made) on
the one hand and outer forces (independent of choice) on the other.
Modern psychologists are generally physical monists; they believe that the causes of
human behavior are entirely physical; in that respect they differ from Descartes. However,
like Descartes, they view human behavior as a battle between more central forces located
in the higher regions of the brain and more peripheral forces located in lower regions of
the brain, in the peripheral nervous system, or in the viscera (Loewenstein 1996). Actual
behavior is the outcome of this battle. As Descartes might have said, an alcoholic may
choose not to drink, but may nevertheless drink. In that case, the lower forces would
have overcome the higher ones in his brain. His actual behavior would be merely the
outcome of this more central battle. Most modern psychological researchers interested in
addiction, while they may focus on behavioral observations such as those underlying the
discount function of Figure 14.1a, nevertheless view self-​control and its failure primarily
in this way—​as the outcome of a conflict located primarily in the brain. For them, the
next step after behavioral, cognitive, or social observation is to discover the neural basis
248  

248 Addiction as social choice

of the behavior. Even when observations are entirely behavioral, the investigators assume
that they are only studying peripheral indications of what is essentially a neural process.
A reader may ask: how could it be otherwise? Isn’t behavior due to muscular contrac-
tion, isn’t muscular contraction controlled by the nerves, and aren’t outgoing neural
impulses (for the most part) controlled by the brain? Yes of course. And neuroscience has
made great advances in recent years, some of them relevant to self-​control and addiction.
It is not neuroscience itself where (I believe) the problem lies; it is neuroscience as the
main path, sometimes the only path, to the solution of the problem of addiction. For me,
taking pills to cure addiction is like swallowing a spider to catch a fly. Heroin, the reader
may recall, was developed as a “non-​addictive” cure for morphine addiction, which in
turn was seen as a cure for opium addiction (Sawynok 1986). I believe that addiction is
fundamentally a behavioral and social problem; the main path to its solution is behavioral
and social. The problem with the move from behavioral and social research on addiction
to neurological research on addiction is not that neurological research is bad or irrelevant;
it is that such a move impedes the progress of behavioral and social research. You can-
not just solve the problem of behavior and then go on to the brain. One might as well say
that we can solve the problem of the brain and then go on to atomic physics. There is an
infinite amount of work to do at the behavioral level (just as there is an infinite amount of
work to do at the level of brain physiology). The direction in which behavioral research
on self-​control tells us to go is not inward, into the brain, but outward, into the social
environment where the controlling forces (the reinforcement contingencies) lie.

3  Teleological behaviorism
The approach to addiction I have been advocating is called teleological behaviorism. The fun-
damental conflict between self-​control and addiction occurs not in the brain, not inside the
organism at all, but in the reinforcement contingencies competing for control of the person’s
behavior. Figure 14.2 illustrates the kind of choice that addiction is. It is not a choice between
one particular act and another—​not a choice between having a drink and refusing the drink,
not a choice between smoking a cigarette and refusing the cigarette, but rather a choice
between a particular act (What should I do?) and an abstract pattern of acts spread out over
time (What should we do?). It is a choice between smoking the cigarette or having the drink
on the one hand and being healthy, socially accepted, and successful at one’s job on the other.

P–4 P–3 P–2 P–1 P0 P1 P2 P3 P4

What should I do?

What should we do?

Figure 14.2 An individual person at the present time (P0) in relation to a group of people or to
herself over an extended time span (P-4 to P4). In both social and temporal terms a conflict may
arise between the answers to the questions: What should I do? and “What should we do?
  249

Teleological behaviorism 249

Here is a quotation from the comedian Dick Cavett: “Once, as [Cavett and Jonathan
Miller of ‘Beyond the Fringe’] waited backstage together at the 92nd Street Y in New York
City, [Cavett] pointed disapprovingly at [Miller’s] lit cigarette. [Miller said,] ‘I know these
will kill me, I’m just not convinced that this particular one will kill me.’ ” Miller is right.
No particular cigarette can harm a person, either now or later. Only what is essentially an
abstraction—​the relationship between rate of smoking and health—​will harm him.
We all tend to focus on the particular when it comes to our own behavior. Only when,
like Cavett, we observe someone else’s behavior, or when circumstances compel us to
experience the long-​term consequences of our own behavior, are we able to feel their
force. Another way of putting Miller’s problem is to say that Miller’s behavior was under
the control of the consequences of smoking the particular cigarette whereas it should have
been under the control of the consequences of smoking at a high rate over a long period
of time. Can we say that Miller is just not a very smart person, perhaps not capable of
understanding abstract concepts? No way. Here is his Wikipedia entry:
Sir Jonathan Wolfe Miller CBE (born 21 July 1934) is a British theatre and opera director, actor,
author, television presenter, humourist, sculptor and medical doctor. Trained as a doctor in the late
1950s, he first came to prominence in the early 1960s with his role in the comedy revue Beyond the
Fringe with fellow writers and performers Peter Cook, Dudley Moore and Alan Bennett. He began
directing operas in the 1970s and has since become one of the world’s leading opera directors with
several classic productions to his credit. His best-​known production is probably his 1982 “Mafia”-​
styled Rigoletto set in 1950s Little Italy, Manhattan. In its early days he was an associate director at
the Royal National Theatre and later he ran the Old Vic Theatre. He has also become a well-​known
television personality and familiar public intellectual in both Britain and the United States.

Now let us go from the behavior of a highly sophisticated British comedian, doctor, direc-
tor, producer, etc. to that of a not very sophisticated pigeon in a Skinner-​box in my labora-
tory (Rachlin and Green 1972). If the pigeon is hungry and you offer it repeated choices
between a smaller-​sooner (SS) reward (two pellets of pigeon-​food right now) and a larger-​
later (LL) reward (four pellets delayed by as little as four seconds), the pigeon will choose
SS virtually 100% of the time (Figure 14.3a). The pigeon cannot wait a mere four seconds
for twice as much food. Its behavior is controlled by the immediate two pellets. However,
suppose we change the alternatives and offer the pigeon two paths to food (Figure 14.3b).
If the pigeon chooses the upper path, there will be a ten-​second wait, and then it will
have a second choice to make: two pellets immediately after the ten-​second wait or four
pellets delayed by a further four seconds (the same SS and LL as originally). We know
what the pigeon will do then—​it will choose SS.
If the pigeon chooses the lower path, it will also have to wait ten seconds, but then it
will have no second choice to make. It will be compelled to take LL—​four pellets delayed
by another four seconds. In other words, by choosing the lower path, the pigeon commits
itself to obtaining the larger reward at the end of the 14 seconds.
Given repeated choices between the upper and lower paths, most pigeons, most of the
time, will choose the lower and will get twice as much food as they would have gotten if
they had chosen the upper. With the two paths of Figure 14.3b, their behavior is controlled
by the higher rate of food provided by the lower path. I am not saying that pigeons that
choose the lower path are in any way smart. Pigeons are pretty dumb. It is rather I, the
250  

250 Addiction as social choice

(a) SS

Path chosen

Path rejected

LL

SS
(b)

LL

10 decond delay

LL
Figure 14.3 From an experiment with pigeon subjects by Rachlin and Green (1972). The pigeons
were choosing between smaller-sooner (SS) and larger-later (LL) rewards. When these rewards
were immediately offered, the pigeons preferred SS to LL. 14.3b. When given a choice 10 sec-
onds earlier between the upper path (leading to a second choice, the same choice as in Figure
3a) and the lower path (commitment to LL), the pigeons preferred the lower path.

experimenter, who have managed to bring their behavior under the control of its wider
consequences.

4  A proposal for developing self-​control

How then can an addict bring her own behavior under the control of its wider conse-
quences? First, and most obviously, to control her behavior she has to know what exactly
that behavior is. That is, she must make herself an expert on her own behavior. It is this
step—​self-​monitoring—​that is by far the most difficult part of self-​control. Modern com-
puter technology can make self-​monitoring easier than it used to be, but I myself prefer to
just write things down. At points in my life where I need to control my weight I keep a calorie
diary in which I write down everything I eat, its caloric content, and the sum of the calories
I eat each day. Then I make summaries each week. If I were trying to control my smoking
I would record each cigarette and the time of day I smoked it—​or each glass of scotch, each
heroin injection, each cocaine snort, each hour spent watching television or doing crossword
puzzles when I should be writing, etc. Every instance goes down in the book. There is no
denying it—​this is hard to do. For one thing, it is socially difficult. You don’t want to interrupt
  251

A proposal for developing self-control 251

a dinner party by running into the bathroom every five minutes to write down that you’ve
bitten your nails again. Nevertheless, before you even start to try to change your patterns of
eating, smoking, drinking, etc., you need to know what those patterns are.
The addict will say that by recording her behavior she is constricting her freedom (I’m
sure the pigeon, after choosing the lower path of Figure 14.3b, and being forced to obtain
LL, would accuse me, if it could speak, of restricting its freedom by withholding SS). But
in this regard it is good to remember the poet Valerie’s advice: “Be light like a bird and not
like a feather.” A bird uses the wind whereas a feather is a slave to the wind. So we must use
environmental contingencies of reinforcement and not, in the name of freedom, become
a slave to them.
This first step—​self-​monitoring—​is so important, and so difficult, that it should not be
mixed up with actual efforts at changing behavioral patterns (i.e. habits). First, addicts
must become experts on themselves. That is something our intelligence is good for. It
is there, not in some “willpower” center of our brains, where our self-​control essentially
differs from that of pigeons. We must make charts, make graphs, if that comes naturally.
But at least we should write everything down and make weekly and monthly summaries.
Sometimes this step alone, without further effort, will effect habit change. But at this point
an addict should not try in any way to change whatever habit she is trying to control.
Once an addict becomes an expert on herself, habits will be much less difficult to change.
After the addict has gained self-​observational skill, she is ready to proceed to the sec-
ond step—​to give weight to her behavior. Jonathan Miller’s problem is that each particular
cigarette weighs too little. How could he have given that particular cigarette more weight?
Let us say that Miller has already completed Step 1 and is recording each cigarette smoked
and the time it was smoked. (Note that this already gives the cigarette weight. It doesn’t just
go up in smoke but is preserved in his log.) Let us say further that the day of his encounter
with Cavett was a Monday. On that day (a “free” day) Miller smokes as much as he wants to.
He makes no effort to restrict his smoking in any way. (He is still recording each instance.)
However, on Tuesday (a “matching” day) he will smoke exactly the same number of ciga-
rettes as he did on Monday. If necessary he will sit up an extra hour on Tuesday to smoke
those two or three cigarettes to make up the total. Then on Wednesday he is free again, and
on Thursday he has to mimic Wednesday’s total. Now, when he lights a cigarette on Monday
he is in effect lighting up two cigarettes—​one for Monday, and one for Tuesday. As he keeps
to this schedule, and organizes his smoking into two-​day patterns, it should be coming
under control of the wider contingencies. Once this two-​day pattern is firmly established,
he will extend it to three days—​duplicating his Monday (“free” day) smoking on Tuesday
and Wednesday (“matching” days). Then he will extend it to four days, etc., until he reaches
seven days. Eventually, each cigarette Miller lights up on Monday will effectively be seven
cigarettes—​one for each day of the week. The weight of each cigarette will thus increase to
the point where he no longer can say, “I’m not convinced that this particular cigarette will
kill me.” At no point is he trying to reduce his smoking or exerting his willpower. Willpower
is actually bringing behavior under the control of wider (and more abstract) contingencies.
This is a power that anyone can exert who has the intelligence and is willing to invest the
effort and time. And the exercise of this power can make a smart person happy.
252  

252 Addiction as social choice

What is to prevent the smoker from breaking the pattern and impulsively smoking more
(or fewer) cigarettes on the matching day than on the free day? The answer is that nothing in
the program prevents the smoker from doing this. If the pattern is broken, the failed matching
day becomes a free day, and the sequence starts anew. Remember, on the free day, the smoker
smokes just as many cigarettes as she feels like smoking on that day. If she does that, she is
unlikely to develop strong cravings the next day when she smokes just as much as she did
yesterday. Again, the only thing the smoker is trying to do on the matching days is to match
the free days. If she truly smokes as much as she wants to on the free days, this should not be
a difficult task. If the program is working, smoking reduction should occur automatically and
effortlessly as the negative value of future smoking gains control of her free-​day behavior.
Note: There is yet a third step—​or rather a flight of steps. I have not mentioned social
support. I have not mentioned exercise. Both of these are economic substitutes for addic-
tions of various kinds. If either is lacking in an addict’s life, programs need to be estab-
lished for its institution. (I will discuss social support in more detail later.) Addiction is
not an isolated thing. It has to be regarded in the context of a complete life. To repeat: The
choice addicts make is not between performing an addictive act and not performing it, but
between performing an addictive act and performing a non-​addictive pattern of acts.

5  What is a discount function?

It is common among researchers to think of addicts as people with unusually steep delay
discount functions (Odum 2011). I will argue here that addiction is really a social phe-
nomenon except the social actors are different temporal extensions of the self. You could
ask a person whether he prefers $350 now or $1000 later. Figure  14.1a shows typical
results as delay is varied. Or you could ask a person whether he prefers $350 for himself
or $1000 for a classmate. Let us call choice of the $350 for himself a selfish choice and
choice of the $1000 for a friend an altruistic choice. We have done studies of this kind in
my laboratory in Stony Brook (Jones and Rachlin 2006). Here are the instructions we gave
to the undergraduate participants:
The following experiment asks you to imagine that you have made a list of the 100 people closest to
you in the world ranging from your dearest friend or relative at position #1 to a mere acquaintance
at #100. The person at number one would be someone you know well and is your closest friend or
relative. The person at #100 might be someone you recognize and encounter but perhaps you may
not even know their name. You do not have to physically create the list—​just imagine that you have
done so. Next you will be asked to make a series of judgments based on your preferences. On each
line you will be asked if you would prefer to receive an amount of money for yourself versus an
amount of money for the person listed. Please circle A or B for each line.

The next seven pages asked participants to make choices between relatively large amounts
of money (hypothetical regretfully) for people at various social distances and smaller
amounts for himself or herself. For each participant, at each social distance (N), we deter-
mined the amount of money for his or her self, equal in value to $75 for person-​N (the
crossover point). Figure 14.1b from Jones and Rachlin (2006) plots these crossover points
as a function of N. This is a social discount function.
  253

What is a discount function? 253

Note the similarity of Figure 14.1b to Figure 14.1a. Figure 14.1a shows how money is dis-
counted as a function of delay; Figure 14.1b shows how money is discounted as a function
of social distance. The line fitted to the points in Figure 14.1b takes the same mathematical
form (a hyperbola) as that of Figure 14.1. (The function for delay discounting is: v = V/​(1 +
kD); the function for social discounting is v = V/​(1 + k’N) where v is discounted value, V is
undiscounted value, D is delay, N is social distance, and k and k’ are constants directly pro-
portional to degree of discounting.) On the basis of delay discounting results, Rachlin and
Raineri (1992, p. 113) predicted that social discount functions would take precisely this
form. This idea (conveyed also by Figure 14.2 as well as by Figures 14.1a and 14.1b)—​that
altruism (as measured by social discounting) and self-​control (as measured by delay dis-
counting) are basically the same process—​has been expressed since then in both econom-
ics and philosophy. The economist Julian Simon (1995, pp. 375–​76) indicates what social
discounting would mean for understanding the concept of altruism:

The conceptual framework employed here obviates the age-​old question about whether an act of
giving by one individual to another should properly be labeled “altruism,” or whether instead one is
“really” being “selfish” by making oneself feel good. An individual’s discount weights vis-​à-​vis other
individuals may be considered a full description of the individual in this connection, assuming that
the individual’s behavior corresponds to his or her discounts in relation to other individuals …
Revealed preferences constitute the entire system. (Italics added.)

The concept of revealed preferences refers to an ability to construct mathematical relation-

ships from observations of behavior such as those we used to obtain the discount function
(the function “reveals” the preferences), and then to use that function to predict behavior
in some other situation.
The correspondences and parallels between social and delay discounting should serve
to demystify altruism. Whereas it may seem self-​evident that people will sacrifice some
part of their immediate good for the benefit of their temporally extended selves, and
that delay discount functions measure this tendency, it seems mysterious when people
sacrifice some part of their own good for the benefit of another person. Yet, as Simon
implies, the two forms of sacrifice are equally explicable in economic terms. A person has
common interests with other people close to him just as he has common interests with
himself at other times. As Ainslie (2001) has claimed, tradeoffs, described in terms of
hyperbolic delay discounting, may occur between P0 (oneself now) and P-​m … P0 … Pm
(oneself over a wide stretch of time) of Figure 14.2. Similarly, such tradeoffs may be
described in terms of social space. The two types of tradeoff imply that our essential selves
are not fixed by our skins but are extended beyond it both in time and in social space.
The same correspondence (between one’s self now, vis-​à-​vis one’s self at other times,
and one’s self, vis-​à-​vis one’s friends and relations) has been noted by the philosopher
Derek Parfit (1984, p. 211). Parfit makes a compelling case for what he calls reduction-
ism: “Each person’s existence just involves the existence of a brain and body, the doing
of certain deeds, the thinking of certain thoughts, the occurrence of certain experi-
ences, and so on.” Individual actions and thoughts take time and overlap with each
other giving us the illusion of a continuous self. But, aside from that overlap, there is no
254  

254 Addiction as social choice

“further thing”–​either spiritual or physical—​to a person’s self. You may have as little
in common with yourself 20 years from now as you do currently with a distant cousin.
A person’s continuous character is, according to Parfit, no more (or less) real than the
character we ascribe to groups of people such as families, clubs, firms, or nations. As
Parfit says (p. 211): “Most of us are reductionists about nations … Nations exist. Though
nations exist, a nation is not an entity that exists separately, apart from its citizens and
its territory.” If your future self is no closer to your present self than another person may
be, then there should be no essential difference between your relations to your future
self and your relations to other people. Since your concern or lack of concern for other
people involves moral issues, Parfit says, so does your concern or lack of concern for
your future self. Therefore, issues of social cooperation and altruism on the one hand
and self-​control on the other should be treated in the same way. A motorcyclist’s refusal
to wear a helmet would be, for Parfit, a moral issue, not just because other people may
have to pay a price for her brain injury, but also because her future self may have to pay
a price. To extend this to the concern of this chapter, addiction is not just a problem
for the addict but is a problem as well for society. And this is true in two ways: first,
the addict’s behavior hurts other people; second, the addict’s behavior hurts his future
self. Both are cases of overlapping, but not completely overlapping, interests. Both are
consistent with the notion of addiction as social choice.
What then is a discount function? A discount function is not an entity in a person’s head.
Looking for hyperbolic discount functions in the brain is like looking for paraboloid arcs
inside a baseball. Discount functions are rather measures of the overlap of a person’s inter-
ests with those of other persons and with those of the person at other times. These overlaps
of course will vary in extent depending on the kind of interests they are, and depending
on the person’s ongoing patterns of interaction with her future self and with other people.

6  Figure and ground in behavior

The Gestalt psychologists were a group of European trained psychologists who immi-
grated to the US before and during World War II. At that time European psychology was
heavily influenced by the philosophy of Immanuel Kant, and the application of phenome-
nology to psychology. The Gestaltists’ main focus was the psychology of perception. They
believed that perception of color, shape, size, loudness, and so forth depended not just on
particular stimuli but on the stimulus in the context of the entire perceptual field—​the
figure and its background together. The brightness of a gray spot, for instance, depends
not only on the light intensity of the spot but also, and very strongly, on the brightness
of the surrounding area. The Gestalt psychologists had a great influence on American
psychology around the middle of the twentieth century. Their influence on the psychol-
ogy of learning and motivation was less strong than their influence on the psychology of
perception. However, their most famous dictum—​the whole may be greater than the sum
of its parts (Koffka 1955), originally from Aristotle’s Metaphysics —​extends naturally to
problems of self-​control and altruism.
  255

Figure and ground in behavior 255

The teleological behavioral extension of that Gestalt dictum would say that the value of
an activity may be greater (or less) than the sum of the values of its parts. As an illustra-
tion, suppose you are driving from New York to Chicago. Your car has a CD player and
you take along some CDs to play on the trip. You like both classical and popular music,
so you take along several symphonies and several pop CDs. Suppose (perhaps contrary to
fact) that your tastes are such that:
1. Over the next 60 minutes you prefer listening to a symphony rather than to 20 three-​
minute popular songs;
2. Over the next three minutes you prefer listening to a popular song rather than to a
section of a symphony.
This is a paradigm case of a self-​control problem. The problem is that to listen to the
whole symphony (which by assumption you prefer to do) you must listen to the first
three minutes of it (which you prefer not to do). If you just do what you prefer at the
moment (assuming your preferences remain constant throughout the trip), you will drive
the whole way from New York to Chicago playing only popular songs whereas (again, by
assumption) you would have been happier if you had played only symphonies.
Similarly, an alcoholic prefers to be sober, healthy, socially accepted, and to perform
well at his job than to be drunk most of the time, unhealthy, socially rejected, and perform
poorly at his job. However, over the next few minutes, he prefers to have a drink than to
not have one. If over successive brief intervals he always does what he prefers, he will
always be drinking.
Likewise, most of us prefer to have loving families and a wide circle of entertaining and
sympathetic friends and fellow citizens who help each other out in emergencies. We prefer
to be kind to our friends and relations, and to be friendly and charitable. Yet we are too busy
at the moment to call our parents, to take our son to the ballgame, to visit a sick friend. We
throw the charity solicitation envelope unopened into the trash can, and cheat a little on our
taxes. Recall Dickens’s A Christmas Carol. The story opens with Scrooge as utterly lacking
in Christmas spirit, a selfish materialist incapable of any kind of friendship not leading to a
profitable material exchange. At the end of the story, “He became as good a friend, as good
a master, and as good a man as the good old City knew, or any other good old city, town, or
borough in the good old world.” (Dickens [1843] 2015, p. 61). The transformation comes
about through the intervention of three ghosts who show Scrooge Christmas past, Christmas
present, and Christmas yet to come. In the pattern of his life, Scrooge is able, through the
ghosts, to see the narrow spectrum of his social engagement; thus he is able to see himself.
Importantly, he is also allowed to see the pattern of the life of his clerk, Bob Cratchit; by con-
trast Scrooge cannot help but see the emptiness of his own life unless his behavior changes.
The problem, for Scrooge and the alcoholic (and Jonathan Miller), is how to make choices
over longer temporal and social spans (the whole) and avoid making choices on a case-​by-​
case basis (the parts). The reason why we find it difficult to avoid making short-​term, local
choices is that the value of the immediate and local alternative (having a drink now, getting
to where we’re going on time) is greater than that of a fraction of the longer, wider activity
256  

256 Addiction as social choice

(being sober right now, being generous right now). The reason why we should make choices
over the longer time span and social span is that the value of the longer activity (being gen-
erally sober, being generous) is greater than the sum of all the short-​term values of its parts.
Each drink refusal, each visit to a sick friend, may have no positive value in itself.
Moreover, to use behavioristic language, these pieces of the more valuable behavioral pat-
tern may never be reinforced. They are not immediately reinforced, they are not condi-
tionally reinforced, and they are not reinforced after a delay. Clearly, refusing a drink is
not immediately reinforced. It should also be clear that no single drink refusal is rein-
forced after a delay. If the alcoholic refuses a single drink, she does not wake up three
weeks later suddenly healthier and happier. To realize the value of a drink refusal, she
must put together a long string of them—​just as, to realize the value of a phone call to a
friend, you have to have a reasonably close relationship with him. You cannot just add up
the value of the parts to get the value of the whole (as economists tend to do). How would
you feel if you had watched 95 minutes of a 100-​minute movie and the projector broke?
Not, it seems safe to say, 95 percent as happy as you would have been if you had seen the
whole movie. Here is an excerpt from a review of the letters of the writer, William Styron
(Salter 2013, p. 32): “[Styron] complained always of the difficulty of writing, the torture
of it. ‘Writing for me is the hardest thing in the world.’ He loathed it, he said, every word
that he put down seemed to be sheer pain, yet it was the only thing that made him happy.”
The fact that longer behavioral patterns may have a value greater than the sum of the
values of their parts is not unique to these very broad patterns. Each broad behavioral pat-
tern is nested in still broader patterns and contains narrower patterns within it. Listening
to a symphony over an hour is nested within the pattern of listening to a mixture of sym-
phonies and popular songs for a day. At the other extreme, listening to a single verse is
nested within listening to a whole popular song. Listening to a chord sequence is nested
in listening to a verse. Even a seemingly unitary act such as a rat’s eating a food pellet is a
highly complex pattern of muscular movement.

7  Social interaction can substitute for

and displace addiction
At the beginning of this chapter I described a method for bringing your behavior under
the control of temporally wider contingencies. Now let me add to that another method,
one that is less direct. As an example of an addiction, I will again talk in terms of smoking,
but the method can be applied in a straightforward way to other addictions.
The method involves socializing with people in non-​drinking or non-​using activities.
One reason why Alcoholics Anonymous and other such groups work is that their programs
involve many elements that promote improved patterns of social cooperativeness. Be like
Scrooge after he reformed. Join clubs, take courses, call up old friends, invite people to dinner.
Build up a reputation for social cooperation. Getting that reputation is not as easy as it may
sound. You cannot just be altruistic in places and at times where others can see you. People
are hypersensitive to other people who are what you might call, situationally altruistic. For
altruism to reduce addiction, a person must develop a pattern of social cooperativeness that
  257

Social interaction can substitute for and displace addiction 257

extends throughout his life. If you make every social decision on a case-​by-​case basis, by the
time you have decided whether it’s a good idea to give that beggar a dollar, he’ll have passed
you by. (I am not saying that you should always give dollars to beggars. There may be good
reasons not to. I am just saying that if you do it, it should not be for the sake of being seen to
do it but for the sake of developing the habit of doing it.) Once you develop a pattern of social
cooperation, others will usually cooperate with you—​not always, but often. Society does play
tit for tat with us. Moreover, social support can substitute for addictive behavior in general,
and smoking in particular. To put it another way, smoking is a crutch, a kind of support that
can be discarded if the support of other people is available to take its place (Rachlin 1997).
Despite their apparent differences, cigarette smoking and social cooperation are eco-
nomic substitutes. On the basis of several studies of smoking and “social support” (the
commodity corresponding to the activity of social interaction), Fisher (1996) summa-
rized evidence for their mutual substitutability: people with more social support smoke
less than those with less social support; smoking cessation programs work better when
they are accompanied by increased social support; the sudden withdrawal of social sup-
port (e.g. by widowhood or divorce) tends to be accompanied by an increase in smoking.
Fisher stated (p. 227):
(a) Similar operations such as stressful events, personal losses, and performance chal-
lenges are likely to instigate choices for either social support or smoking;
(b) Psychological effects such as depressed mood, anxiety, or need for arousal appear to
follow the operations noted in (a) and to be associated with increased interest in both
smoking and social support;
(c) The common effects of social support and smoking appear to include anxiety reduc-
tion, mood elevation and performance enhancement.
Although smoking and social interaction may be substitutable for each other, they are
clearly different in terms of the relation between consumption and economic demand.
Nicotine generates its own antagonists in the body; therefore, like many addictive drugs,
tolerance builds up and you need greater and greater doses over time to get the same
mood-​elevating effect.
On the other hand, the more you socialize, the easier (and cheaper) it becomes. For most
children, social interaction is freely available (i.e. cheap), but eventually considerable skill
has to be acquired to maintain it (Ainsworth and Bowlby 1991). In new social circumstances
social skills need to be altered or relearned; without practice, social skills grow rusty. (Like
baseball players after a winter of inactivity, we need to redevelop our social skills after a period
of isolation.) By the time adulthood is reached, some people are clearly better than others at
getting and keeping social support (Fisher 1996). It is thus fair to say that the more social
activity is performed, the cheaper it gets; the less it is performed, the more expensive it gets.
So there is a dynamic between social activity and cigarette smoking that works as fol-
lows: to put it in economic terms, the more you smoke, the more expensive a given amount
of smoking pleasure gets (because of tolerance building up); the more you socialize, the
cheaper a given amount of social pleasure gets (because you need to practice to do it well).
If you already smoke a lot and socialize little, then both will be expensive. At that point,
258  

258 Addiction as social choice

expensive though they both are, smoking may still be less expensive than socializing. If
so, it will be easier at the moment to have a cigarette than to socialize. However, if you
increase social activity despite its expense (if you go to that party where you know only
one or two people) even while continuing to smoke, socializing will become cheaper and
cheaper. Eventually socializing will be cheaper than smoking. Since they are substitutable,
it will be much easier to stop smoking now than it was before. And you will have brought
your behavior under the control of a very valuable long-​term pattern of reinforcement.
To summarize, the concept that the value of a behavioral pattern may be greater than
the sum of the values of its parts is not just an empty slogan borrowed from Gestalt psy-
chology. It is the very basis of a teleological approach to the problem of self-​control.

Acknowledgments
Preparation of this chapter was supported by a grant from The National Institute on Drug
Abuse. Some of the material in the article is revised from Rachlin (2014).

References
Ainslie, G. (2001). Breakdown of Will. New York: Cambridge University Press.
Ainsworth, M.D.S. and Bowlby, J. (1991). An ethological approach to personality development.
American Psychologist, 46, 333–​41.
Dickens, C. ([1843] 2015). A Christmas Carol: A Ghost Story of Christmas. Sweden: Wisehouse Classics.
Fisher, E.B., Jr (1996). A behavioral-​economic perspective on the influence of social support on
cigarette smoking. In: L. Green and J.H. Kagel (eds), Advances in Behavioral Economics: Vol.
3. Substance Use and Abuse. Norwood, NJ: Ablex, pp. 207–​36.
Green, L., Fry, A.F., and Myerson, J. (1994). Discounting of delayed rewards: a lifespan comparison.
Psychological Science, 5, 33–​36.
Jones, B.A. and Rachlin, H. (2006). Social discounting. Psychological Science, 17, 283–​86.
Koffka, K. (1955). Principles of Gestalt Psychology. Oxford: Routledge and Kegan Paul.
Loewenstein, G. (1996). Out of control: visceral influences on behavior. Organizational Behavior and
Human Decision Processes, 65, 272–​92.
Odum, A.L. (2011). Delay discounting: I’m a k, you’re a k. Journal of the Experimental Analysis of
Behavior, 96, 427–​39.
Parfit, D. (1984). Reasons and Persons. Oxford: Oxford University Press.
Rachlin, H. (1997). Four teleological theories of addiction. Psychonomic Bulletin and Review, 4, 462–​73.
Rachlin, H. (2014). The Escape of the Mind. New York: Oxford University Press.
Rachlin, H. and Green, L. (1972). Commitment, choice and self-​control. Journal of the Experimental
Analysis of Behavior, 17, 15–​22.
Rachlin, H. and Raineri, A. (1992). Irrationality, impulsiveness, and selfishness as discount reversal
effects. In: G. Loewenstein and J. Elster (eds), Choice over Time. New York: Russell Sage Foundation,
pp. 93–​118.
Salter, J. (2013). Bill Styron: the ups and downs. New York Review of Books, 60, 32–​34.
Sawynok, J. (1986). The therapeutic use of heroin: a review of the pharmacological literature. Canadian
Journal of Physiological Pharmacology, 64, 1–​6.
Simon, J. (1995). Interpersonal allocation continuous with intertemporal allocation. Rationality and
Society, 7, 367–​92.
  259

Chapter 15

Nonconscious motivational
influences on cognitive processes
in addictive behaviors
W. Miles Cox
Eric Klinger
Javad S. Fadardi

Abstract
This chapter presents the motivational and goal theory of current concerns
in relation to addiction and choice. A current concern is an individual’s
motivational state from the point of becoming committed to pursuing
a particular goal until the goal is reached or the pursuit is relinquished.
During this time, the current concern guides the person’s cognitive
processes, including attention, memory, thoughts, and dreams. What is
true of goals in general is true of the goal of drinking alcohol or using
another addictive substance. We hold that the decision to use a substance
is voluntary; thus, the choice to use is subject to the person’s conscious
control. Nevertheless, implicit processes influence the decision. One of these
processes is addiction-related attentional bias. We describe (1) the research
on attentional bias and (2) interventions for helping drinkers overcome the
attentional and motivational influences on their addictive behavior.

1  A framework for cognitive processes in addictions:

the motivational goal theory of current concerns
Addicted individuals are characterized by organizing much of their daily lives
around assuring supplies of, obtaining, and consuming the subject of their addic-
tion, whether a substance such as alcohol or an activity such as gambling. The goal
theory of current concerns provides a useful framework that accounts for many of the
aspects of such addiction. The theory (summarized in Klinger and Cox 2011a) has
gradually evolved over more than four decades from an initial effort to account for
shifts in thought content (Klinger 1971) to explanations for a wide variety of affec-
tive (Klinger 1975, 1977), social and psychopathological (Klinger 1977; Klinger and
260  

260 Nonconscious motivational influences

McNelly 1969), and cognitive phenomena (e.g. Hoelscher et  al. 1981; Nikles et  al.
1998; Klinger 1978, 1990, 1996a, 1996b, 2012, 2013; Klinger et al. 1980). Because it
rests on a motivational base, it has special relevance to addictive behaviors and their
cognitive characteristics.

1.1  The nature of the goal theory of current concerns

The theory begins with the realization that what characterizes all members of the animal
kingdom during at least some stages (and usually all stages) of their lives is the imperative
of locating and consuming the substances and conditions necessary for individual and
species survival. In all but the most primitive species, these substances and conditions can
be characterized as goals.
Goals are defined here as desired endpoints—​attainment of objects or events—​that
the person is trying to achieve because he or she expects that achieving them will bring
emotional payoffs by enhancing positive affect (through appetitive goals of attaining,
keeping, or restoring something, such as a job, an achievement, or a relationship) and/​or
reducing negative affect (through aversive goals aimed at avoiding, escaping, or ridding
oneself of something valued negatively, such as an illness, a noisy roommate, or a bill
collector). The payoffs may be intrinsic to the outcome, such as a feeling of being loved,
or extrinsic in the sense of outcomes being subgoals instrumental in attaining further
outcomes that will provide the intrinsic satisfaction (e.g. arranging a date that enables a
love relationship, or pleasing one’s boss and thus protecting one’s job that enables intrin-
sically satisfying consequences, such as copious tasty food, a nice place to live, and a
good marital relationship).
To pursue a goal over more than an instant it is necessary for the brain to have a mem-
ory of it and tendencies to act on it under the right circumstances. When a dog sees a
squirrel, the dog immediately springs into action chasing the squirrel. If the squirrel
flees up a tree, the dog stands by the trunk and keeps barking. Most human goals are
more complicated than that, often requiring a sequence of actions on subgoals before
arriving at the opportunity to consummate the ultimate intrinsic goal, and often requir-
ing days, months, or years to complete the pursuit. During most of the time period
before the opportunity to consummate a particular goal its representation is absent
from consciousness. Nevertheless, the individual is likely to give cues associated with
it priority for cognitive processing. That is true for the most direct cues, as for instance
the name of a beloved person, and also more remote cues, such as allusions to similar
personal relationships.
For this to work, the brain must necessarily harbor a continuing representation of the
goal and special responsiveness to its cues. We dub this continuing representation a cur-
rent concern, defined as “the state of the organism between the time of commitment [to the
goal pursuit] and the time of consummation or disengagement” (Klinger 1975, pp. 2–​3).
The construct of current concern explicitly does not refer to the “individual’s thoughts or
actions during this state” (Klinger 1977, p. 37), which the concern underlies and potenti-
ates. Each goal is represented by its own separate current concern.
  261

A framework for cognitive processes in addictions 261

The presumption from the beginning was that the construct refers to a latent goal-​
related brain process, whose actual representation in the brain remained unspecifi-
able until recent neuroscientific advances begun to generate clues (e.g. Berkman and
Lieberman 2009; Klinger and Cox 2011a; Kouneiher et al. 2009). This latent process con-
veys processing priority to goal-​related cues in the environment and in the person’s own
ongoing stream of mentation. The processing entails attention, perceptual processing,
retention, and thought content, including both verbal thoughts and nonverbal images.
There is now ample evidence of this for waking states, in which experimental investiga-
tions using a variety of methods have shown that individuals respond selectively to goal-​
related cues with greater attention, retention, and conscious thoughts about the goal (Cox
et al. 2006; Klinger 1978, 2013; Leeman et al. 2014; Montagrin et al. 2013), as well as with
accompanying emotional responses (Bock and Klinger 1986; Klinger 1996a, 2013; Nikula
et al. 1993; Schneider 1987). Similar evidence of processing priority for goal-​related cues
has been found in dreams during sleep (Hoelscher et al. 1981; Nikles et al. 1998). The
goal-​related cues that elicit processing may be external, such as in conversations or media,
or they may be internal, part of one’s ongoing stream of thought. Their enhanced claim on
processing priority helps explain the sequencing of the content in that stream.
The cues to which people respond may have only an indirect association with the goal.
From birth onwards, people develop an extensive network of associations. Our observa-
tions indicate that response is most likely and strongest to cues most directly associated
with goals, but it also occurs regularly to more remotely associated cues.

1.2  The nonconscious nature of current concerns

The cues introduced to our participants in the thought-​and dream-​sampling studies cited
in Section 1.1 on the nature of the goal theory of current concerns occurred without
regard to participants’ ongoing mental content at the moment of stimulation, when they
would be very unlikely to be focusing on the cued goal. Yet the resulting mental content
was disproportionately related to the particular own goals adumbrated by the cues. This
supports the notion that the current concern underlying the cued goal is indeed a latent
process. In other words, it operates nonconsciously. Its products—​attended perceptions,
retrieved memories, and mental contents—​are often conscious, but these are distinct
from the underlying process.
Although not without controversy because of some failures to replicate (Cesario
2014), substantial evidence suggests that existing goal pursuits can be activated non-
consciously by experimental manipulations using priming techniques, such as by
introducing goal-​related stimuli with no explicit relation to a subsequent task and
observing their effects on the subsequent task behavior (e.g. Chartrand and Bargh
1996; Huang and Bargh 2014). Chartrand and Bargh (1996) concluded, “preconscious
processing is rich and extensive and capable of activating goals that are chronically
associated with features of the current environment. The goals then operate autono-
mously, without the need for conscious monitoring, and guide subsequent informa-
tion processing interactively with the environment” (p. 475). Moreover, it appears that,
262  

262 Nonconscious motivational influences

regardless of whether a goal pursuit is activated consciously or unconsciously, its acti-

vation makes similar demands on executive function (Huang and Bargh 2014; Marien
et al. 2012).
Together with the processing priority afforded goal-​related cues, such implicit activa-
tion of goals, such as through product placement in shops and advertising displays, can
impose a considerable burden on addicted individuals who wish to moderate or end their
consumption of alcohol. This has been amply demonstrated in relation to heavy drink-
ers (e.g. Cox et al. 2006). The specifics of the relevant alcohol findings are described in
Section 3.2 on the importance of substance-​related cognitive biases and Section 4 on
forms of cognitive bias.
At this time it is probably unknown whether it is also possible to initiate commitment
to a goal without an act of consciousness. Although that seems possible, there are grounds
for skepticism, albeit without direct contrary empirical evidence (Moskowitz and Balcetis
2014). The evidence that priming of goals through the use of subliminal stimuli or of
seemingly unrelated stimuli can influence behavior in subsequent experimental tasks
cannot distinguish between the commitment to new goals for a laboratory exercise versus
the activation of pre-​existing goals, such as to recall recent events or size up other people.
The question of nonconscious commitment thus most likely remains unresolved.

1.3  Determinants of commitments

to goals: the decision process
Life generally surrounds us with incentives, objects, and conditions that we in varying
degrees desire (appetitive or positive incentives) or desire to eliminate (aversive or nega-
tive incentives). People must keep choosing among these to turn some of them into goals.
Formal models of the decision process for these choices have been around for millen-
nia, but those still widely employed are variants of what economists call subjective utility
theory (e.g. Mongin 1997) and psychologists more often call expectancy x Value theory
(e.g. Bundorf et al. 2013; Feather 1982; Morone and Morone 2014; Van Eerde and Thierry
1996). Under this model, the transformation of an incentive into a goal is determined
primarily by two important variables: the value that the person attributes to the incen-
tive (i.e. how desirable the affective change from obtaining the incentive is expected to
be) and his or her expected likelihood of being able to achieve it within an acceptable
time frame and at acceptable cost. Theoretically, when a person is faced with needing
to choose, the incentive that has the largest product of Expectancy times Value is the
one that the individual will choose to pursue. This choice constitutes a commitment to
the goal pursuit and instates a current concern that persists until the person attains the
goal or relinquishes it if its net value and/​or expectancy diminish enough with additional
experience of the pursuit. Variants of this model have been applied in recent decades
to choices of jobs, crimes, contraceptives, family size, migration, political alternatives,
medical alternatives (summarized in Cox, Klinger, and Fadardi 2015), and, notably, binge
drinking (Quinlan et al. 2006).
  263

Disengagement from failed goal pursuits and its implications for addictive behaviors 263

Especially important for present purposes is the set of factors that determine these
choices. Expectancies and values (i.e. anticipated emotions, such as joy at goal attain-
ment and sorrow at goal abandonment) are both affected by psychoactive substances and
by a variety of other influences that have also been shown to affect addictive behaviors,
including prior mood, social support or disparagement, culturally ingrained expecta-
tions, and personality dispositions. Some of these are discussed in Section 3 on the goal
of drinking alcohol. People are often unaware of the influence these factors exert on their
decision-​making.

2  Disengagement from failed goal pursuits

and its implications for addictive behaviors
It is generally harder to let a goal go in the face of insuperable obstacles than to commit
to it. The commitment is probably undertaken in a spirit of hope and expectation; giv-
ing up is accompanied by a sequence of affectively negative events (Klinger 1975, 1977,
1993). After goal attainment, cognitive reactivity to goal-​related cues diminishes relatively
quickly (Förster et al. 2007; Moskowitz et al. 2011). In contrast, failed current concerns,
once instated, can probably not simply be terminated. Rather, as in the case of extinction
of operant responses, which is a process of inhibitory learning rather than of unlearning
the original response (Bouton 2002), ending the sway of a failed current concern is almost
certainly a process of progressive extinction of responses to cues, in this case goal-​related
cues. Just as extinguished responses can be reinstated in a variety of ways without exten-
sive relearning (e.g. Bouton 2002; Haaker et al. 2014), disengagement from goals is often
quickly reversed when circumstances become more favorable. This could happen only if
the extinguished response or abandoned goal pursuit remains represented in the brain in
a form ready to re-​emerge.
The disengagement process (incentive disengagement cycle) has been described as a
sequence of phases (Klinger 1975, 1977, 1987; Klinger and Cox 2011a) set off by an
“action crisis” (Brandstätter et al. 2013) in a goal pursuit as difficulties mount and finally
are recognized as insurmountable within acceptable costs. The phases begin with invig-
orated efforts in the goal pursuit, followed by anger sometimes accompanied by aggres-
sion, followed further by lowered mood that may range from disappointment to clinical
depression, with marked loss of interest in ordinarily attractive incentives and goals
(Klinger 1993), and ending in most cases in gradual recovery with completed disengage-
ment from the failed goal, and re-​engagement with other goals. Re-​engagement appears
to be essential to full recovery (Carver and Scheier 1998; Wrosch et al. 2013). The mid-
dle phases of this cycle are especially likely to lead to heavy substance use and other
addictions.
A common accompaniment of negative moods during the disengagement process is
rumination (e.g. Aldao et al. 2010; Koster et al. 2011), in which the person’s thought
stream is heavily preoccupied with mostly negatively toned thoughts that focus
repetitively on a limited range of personal concerns without arriving at constructive
264  

264 Nonconscious motivational influences

solutions. The current concerns theory, as in the preceding paragraph on the incen-
tive disengagement cycle, can readily account for depressive rumination—​the con-
tinual goal-​related contents of ruminative sequences serve as cues that elicit further
goal-​related contents, especially regarding goals with higher emotional charge. In
periods of negative mood, the emotional charge is likely to be negative, especially in
individuals high in trait negative affect. The effect of such rumination is generally to
deepen depression further (e.g. Marchetti et al. 2012a; Marchetti et al. 2012b; Nolen-​
Hoeksema et al. 2008).
The implications for addictive disorders, especially substance-​use disorders, are exten-
sive. People are drawn to dispel their negative affect during the disengagement cycle with
psychoactive substances and perhaps other manipulations of affect such as gambling.
During the depression phase, these become more attractive relative to alternative goals as
interest in the latter wanes. The withdrawal from nonsubstance goals leads to increased
social isolation and correspondingly leads to greater focus on substance use. Furthermore,
one can predict somewhat similar implications for people in circumstances that offer little
hope of satisfying goal attainments, with the consequence that their lives lose meaning
(Klinger 2012). These implications for addictions are discussed later in the chapter.

2.1  Areaddictive behaviors motivated or automatic?

a false dichotomy
The position of this chapter is that addictive behaviors are, like other behaviors beyond
simple reflexes, goal-​directed and voluntary (see also, e.g., Heyman 2009). Many writers
on this topic would have disagreed, arguing that addictive behaviors, or at least some of
their components, are “automatic” and beyond voluntary control. The automaticity is per-
haps most often applied to attentional biases. The term “automatic” is rarely defined but
can be taken to mean having a high probability of occurrence, if not certainty, under given
conditions, or of unconscious origin (Huang and Bargh 2014). The term “compulsive” is
also used (e.g. Feil et al. 2010), usually to describe addictive instrumental (e.g. “seeking”)
and consummatory actions (e.g. “taking”; Vanderschuren and Everitt 2005).
The image conveyed by such terms is one of mechanical rigidity. Yet the literature indi-
cates modifiability. For example, alcohol-​related behaviors termed as automatic action
tendencies can be changed through cognitive bias modification (Wiers et  al. 2011)  and
mindfulness treatment (Ostafin et al. 2012; Witkiewitz et al. 2013). Similarly, there is an
extensive literature on successful behavioral modification of obsessive-​compulsive behav-
iors, including the superiority of such treatments to pharmacological treatments (Foa
et al. 2005). Finally, there is now substantial evidence that when people change their sub-
stance use most of them do so on their own without formal treatment (Bischof et al. 2012;
Klingemann 2001; Smart 2007; Sobell et al. 2000).
The present chapter takes the position that addictions reflect processes similar to those
that characterize all striving for highly valued goals. Their high value creates the appear-
ance of automaticity. A  distinguishing feature of goal-​directedness—​indeed, one of its
operational definitions—​has long been considered to be the ability to change tactics in
  265

The goal of drinking alcohol 265

the face of obstacles, as was first strongly argued by Wolfgang Köhler (1925). At that time,
this flexibility was labeled Umweg (detour) behavior to describe new responses that cir-
cumvent obstacles to goal attainment. Detour behavior certainly characterizes the actions
of addicts who face obstacles to obtaining their substances, as for instance demonstrated
anecdotally in the wily actions of addicted nurses in stealing substances from patients and
hospital stores (Stahl 2013). Their goals may appear rigid because of the high value they
place on them, but the actions leading up to goal attainment are highly flexible.
Opposing automatization to goal directedness creates a false dichotomy. Of course,
frequent repetition of action sequences leads to integration of initially separate action
units into longer, relatively seamless action streams that require little conscious control
to unfold, as in skilled playing of a musical instrument, driving a car in traffic, or field-
ing a baseball. However, it is important to realize that this kind of response integration
does not create inflexible automatons. Even integrated action sequences are subject to
modification depending on circumstances. The guitarist may change specific fingering to
accommodate a suddenly broken string, the driver can adjust to the need to take a detour,
and the fielder’s movements will vary according to the wind, other players, or an unex-
pected bounce—​all in the service of attaining their respective goals. The integrated action
sequences remain aimed at these goals, potentially under conscious control.
As indicated in Section 1.2 on the nature of current concerns, alcohol addicts are
marked by attentional biases that steer them to alcohol cues, which in turn further whet
appetites for alcohol consumption, along with a variety of other cognitive biases, includ-
ing heightened delay discounting (discussed in Section 3.2.1 on incentive comparison
effects; see also Noël et al. 2010). Some of these are biases developed by pursuers of any
highly valued goal; others may represent ingrained individual differences in brain func-
tion or the acute effects or chronic after-​effects of substance use. They influence decision-​
making about priorities, sometimes very maladaptively, but tap variants of the kinds of
interactive neurocognitive systems active in most decision-​making (Noël et al. 2013).
Finally, it should be noted that the theory presented here has some features in common
with dual-​process models of addiction (e.g. Moss and Albery 2009; Wiers et al. 2007).
These models assume that both automatic processes (e.g. addiction-​related attentional
bias and approach tendencies) and controlled processes (i.e. deliberate responses) are
involved in addictive behaviors. We agree with this general position. The theory of cur-
rent concerns acknowledges that both kinds of processes can influence decisions about
whether or not to use an addictive substance; nevertheless, in our view, the decision is
always under the person’s voluntary control so that controlled processes can always over-
ride the automatic processes.

3  The goal of drinking alcohol

Here we depict excessive alcohol consumption as an example of an addictive behavior.
However, the main principles underlying the goal of drinking alcohol could be applied to
other kinds of substance abuse and other kinds of addictive behaviors.
266  

266 Nonconscious motivational influences

Many people, of course, do not drink alcohol. Among those who do drink, some peo-
ple drink moderately; others drink excessively, and the latter might experience various
degrees of alcohol-​related problems. Regardless of its intensity, the goal of drinking alco-
hol is formed in the same way as any other goal, consistent with value X expectancy
theory as discussed in Section 1.3 on determinants of commitments to goals. It is deter-
mined largely by (1) the value that the person attributes to drinking and (2) the person’s
expected chances of being able to derive the desired benefits from drinking versus being
able to obtain comparably satisfying benefits by other means.
This generalization regarding value X expectancy determinants of alcohol consumption
is supported, for example, by the finding in a four-​country study (Cox et al. 2002) that,
among students who had experienced negative consequences from drinking, there was
inverse correlation between adaptive motivation and amount of alcohol currently con-
sumed. In view of the factor structure of adaptive motivation (e.g. Klinger and Cox
2011b)—​with high loadings on joy anticipated from attainment of one’s goals (i.e. value)
and expected probability of attainment (i.e. expectancy)—​adaptive motivation is a rea-
sonable proxy for having satisfying and attainable goals that are alternative to drinking
alcohol. Apparently, the more that participants had such goals, the more they were able to
curb their drinking when it interfered with their attaining the goals.
There are various ways in which benefits can be derived from drinking alcohol. First,
they can be derived directly from the pharmacological effects of the alcohol. For exam-
ple, because of its effects on neurotransmitters in the brain (Tabakoff and Hoffman
2013), alcohol might serve intrinsically to reduce tension or anxiety (through the
release of gamma-​aminobutyric acid (GABA)), or it might enhance incentive salience
of reward-​related stimuli (through the release of dopamine; Berridge 2007). Second,
they can be derived indirectly—​or instrumentally, extrinsically—​because drinking
alcohol allows the person access to other valued incentives, such as approval from his
or her peers.
The various intrinsic and extrinsic determinants of drinking alcohol—​the biological,
psychological, and sociocultural variables—​have been brought together in the motiva-
tional model of alcohol use (Cox and Klinger 1988, 2011a; Cox et al. 2015). The model
shows how each of the variables that contributes to the decision to drink—​or not to
drink—​is channeled through a motivational pathway to the final decision about whether
to have a drink of alcohol on a particular occasion or not to have it. Some of the variables
that contribute to the decision are proximal to it; other variables are more distal. The most
proximal determinant of the decision is the net change in affect (enhancement of posi-
tive affect or reduction in negative affect, or both) that the person expects from drinking
alcohol compared to not drinking it. Affective change is a central motivational concept
because a desirable change in affect from its present state is the very essence of what peo-
ple are motivated to achieve. Finally, it should be emphasized that the motivational model
is a decisional model. That is, the decision to drink or not to drink is under the person’s
volitional control; nevertheless, the person might be unaware of some of the factors that
affect the decision.
  267

The goal of drinking alcohol 267

3.1  Distal determinants of drinking: past drinking experiences

The kinds of experiences that people have previously had with drinking alcohol (i.e. the
extent to which these experiences have been positive or negative) help to shape people’s
current expectations of affective changes from drinking and hence the value that they
attribute to drinking alcohol. The nature of these experiences is largely determined by
(1) each person’s constitutional predisposition and (2) the environmental influences that
affect drinking decisions.
One factor that will have affected the value of drinking alcohol is the alcohol flush
reaction (Dickson et al. 2006). It includes unpleasant reactions, such as body flushes and
nausea, which occur in people with a genetically determined deficiency in aldehyde dehy-
drogenase. Aldehyde dehydrogenase is the enzyme that normally metabolizes acetalde-
hyde, a toxic metabolite of alcohol, but acetaldehyde accumulates when people with a
deficit in the enzyme consume alcohol. The flush reaction is common among people of
Asian ancestry, but it might also occur among others. People with the reaction expe-
rience drinking alcohol as less pleasurable than other people do, and they have lower
drinking rates.
Broad sociocultural influences are another kind of determinant of the value that dif-
ferent individuals place on drinking alcohol. For instance, it is commonly known that
Italians are taught from an early age to drink in moderation, and social strictures are
placed on drinking to excess. This pattern of moderate drinking can be seen today among
Italian young people. For instance, university students in Italy have very low rates of heavy
episodic drinking, and similar low rates are seen among university students in France,
Germany, Greece, and some of the Eastern European countries (viz., Bulgaria, Hungary,
and Romania; Dantzer et al. 2006). By contrast, university students in other European
countries (viz. Belgium, Ireland, Poland, Netherlands, Slovakia) and in North America
(e.g. United States) and South America (e.g. Columbia) have high rates of heavy episodic
drinking (Dantzer et al. 2006). From a motivational perspective, these cultural influences
can be explained in terms of how they affect the value that individuals attribute to drink-
ing alcohol. That is, the value of drinking will increase to the extent that individuals model
their own drinking behavior after that of other people around and are subtly or overtly
reinforced for doing so.
Additionally, macroenvironmental influences, such as taxation and advertising, also
affect the value that individuals place on drinking alcohol. Specifically, taxes on alco-
hol make drinking less attractive, whereas advertising alcohol makes it more attractive.
In fact, it has been demonstrated that tax increases on alcohol and bans on advertis-
ing alcohol could bring about dramatic declines in alcohol-​related deaths (Hollingworth
et al. 2006).
Finally, it should be noted that substantial evidence has been gathered to demonstrate
that people’s motivation to drink alcohol—​and hence the value that they attribute to
drinking—​is related to their personality characteristics. Some personality characteristics
seem to protect people from drinking excessively; others seem to place them at risk for
268  

268 Nonconscious motivational influences

excessive drinking. Two personality characteristics have been commonly observed among
heavy drinkers that appear to contribute to problematic drinking. They are (1) behavioral
disinhibition in its various manifestations (e.g. impulsivity, reward dependence, sensation
seeking), and (2) negative emotionality (e.g. hopelessness, anxiety sensitivity). People who
are high in behavioral disinhibition value alcohol because of its rewarding properties; those
who are high in negative emotionality value it because it helps to alleviate their dysphoria.
Generalizations about behavioral disinhibition and the misuse of alcohol are, however,
somewhat tempered by recent results regarding the relationship between personality and
substance dependence. Vrieze et al. (2014) studied this relationship by testing a large sam-
ple of twins longitudinally at ages 17, 24, and 29. They found that the relationship between
the personality characteristic constraint and substance dependence was small and largely
genetic, with the genetic component declining from adolescence to adulthood. However,
the relationship appears to have been mediated by traditionalism—​the tendency to adhere
to traditional moral and social values—​rather than directly attributable to behavioral
disinhibition.

3.2  Current factors

Certain current factors, i.e. factors that have an impact at the present time, might alter
people’s expectations about the positive or negative effects that drinking alcohol will have
on their affect.
One such current factor is the immediate environmental context in which a person is
located when he or she decides whether or not to imbibe. The environment affects, for
instance, the person’s perception of how available alcohol is and the extent to which drink-
ing alcohol is approved in a particular situation (Huckle et al. 2008; Paschall et al. 2014).
If a person perceives that alcohol is readily available and that drinking it is condoned, the
affective benefits that the person expects to derive from drinking will be enhanced.
If alcohol is present, habitual drinkers might also have classically conditioned or other
learned responses to alcohol and the stimuli associated with it, and these responses might
intensify the person’s anticipation of valued changes in affect resulting from drinking (see
Field and Cox 2008; Field et al. 2009). At the same time, however, it should be noted that
various applications of classical conditioning theory (including withdrawal, appetitive,
incentive motivation, and classical operant models) which aim to understand alcohol cue
reactivity have not been well supported in empirical investigations (Drobes et al. 2001).
Alternative cognitive models to account for cue-​reactivity phenomena seem to fare better.

3.2.1 Incentive comparison effects

As discussed, the motivational model of alcohol use predicts that people will be motivated
to use alcohol or another addictive substance to the extent that they anticipate that doing
so will result in desirable effects that they want to achieve. Using an addictive substance
becomes an especially attractive way in which people can regulate their affect when they
are unable to do so by other means. For example, when drinkers are feeling anxious or
depressed, drinking alcohol is likely to temporarily alleviate their feelings of distress, and
  269

The goal of drinking alcohol 269

they are likely to turn to alcohol in order to cope with their negative feelings (Abrams
et al. 2002; Ham et al. 2007; Hussong et al. 2001; Swendsen et al. 2000).
Similarly, people who have been treated for an alcohol-​use disorder are more likely to
relapse if their life situation is stressful and they do not have meaningful incentives to replace
the alcohol (Moos et al. 1990; Zywiak et al. 2006a, 2006b); conversely, if these resources are
available to them (Moos and Moos 2007) or they are actively engaged in problem solving
(Demirbas et al. 2012), they are more likely to remain in remission. In fact, resolutions of
drinking problems might occur without treatment if drinkers allocate their resources to
achieving long-​term goals rather than to drinking alcohol (Tucker et al. 2006, 2009).
Delayed reward discounting (which is common among people with a substance-​use
disorder and an index of their impulsivity; Amlung and MacKillop 2014; MacKillop
et  al. 2010)  is one reason why incentives that could potentially serve as alternatives to
drinking alcohol might not be attractive to people with an alcohol-​use disorder. That is,
these people more than others tend to discount the value of incentives that might bring
them long-​term but not short-​term satisfaction. They prefer instead to obtain immediate
mood-​altering effects from drinking alcohol or using other drugs.
As alcohol and other substance users rely more and more on the substance as a way to reg-
ulate their affect, the attractiveness of other incentives erodes. In fact, the increasing value
that substance users attribute to their substance and the concurrent loss in value of other
incentives is reflected in the criteria that are used to define a substance use disorder. These
include: “spending a lot of time getting, using, or recovering from use of the substance”;
“not managing to do what you should at work, home or school, because of substance use”;
“continuing to use, even when it causes problems in relationships”; “giving up important
social, occupational or recreational activities because of substance use”; “using substances
again and again, even when it puts you in danger”; “continuing to use, even when you know
you have a physical or psychological problem that could have been caused or made worse
by the substance” (American Psychiatric Association 2013). In other words, the incentive
value of drinking alcohol or using another substance becomes greater than any other incen-
tive available to the person. Kalivas and Volkow (2005) have concluded that this process
is reflected physiologically in the brain as adaptation in prefrontal cells that innervate the
nucleus accumbens, which results in a decrease in the value of natural rewards, reduced
cognitive control, and increased sensitivity to substance-​related stimuli.
Using functional magnetic resonance imaging (fMRI), Ihssen et  al. (2011) assessed
the relative value of alcohol and alternative incentives among preclinical participants.
Specifically, they compared heavy drinkers’ (alcohol consumption  =  mean of 58.6 UK
units per week; one unit  =  8 grams of ethanol) and light drinkers’ (alcohol consump-
tion = mean of 4.5 UK units per week) brain reactions to various categories of pictorial
stimuli, including pictures that were alcohol-​related and those that were related to par-
ticipants’ higher-​order goals. When presented with the alcohol-​related pictures, the heavy
drinkers had stronger reactions than the light drinkers in basic emotional areas of the
insular cortex and in the reward circuitry of the ventral striatum. When presented with
the pictures related to higher-​order goals, the heavy drinkers responded more weakly
270  

270 Nonconscious motivational influences

than the light drinkers in the frontal areas of brain. Similarly, Garavan et al. (2000) used
fMRI to study cocaine users’ and comparison participants’ brain reactions to films depict-
ing individuals smoking crack cocaine and to films with explicit sexual content. In the
cocaine users, the cocaine cues activated similar brain areas as the evocative stimuli, but
the cocaine users had a weaker brain response to the explicitly sexual stimuli than the
comparison participants. The results confirm that overvaluation of alcohol or another
addictive substance and undervaluation of alternative incentives is already reflected in the
brain physiology of preclinical substance users.

3.3  Cognitive mediators

Based on their past and current experiences, people might come to value drinking alco-
hol or using another substance, and they might also expect that the valued consequences
of using the substance will occur if they use it. In this case, the person will have formed
a goal of procuring and using the substance, and the resulting current concern about
using will be reflected in each person’s cognitive processes. These will include memo-
ries of experiences about using the substance in the past; perceptions about the effects of
using the substance (arising especially from environmental and sociocultural influences);
thoughts about what will happen if the person uses; and attentional processes, such as
taking notice of stimuli related to the substance in the environment. These processes, in
turn, help to solidify each person’s intention about whether or not to use.
Even though some of the cognitive mediators are unconscious, they can still have a
strong influence on people’s decision about whether or not to use. One of the variables
that operates largely or entirely unconsciously is alcohol attentional bias. When people
who have a concern about drinking alcohol encounter stimuli that are related to alco-
hol, they selectively attend to these stimuli while simultaneously filtering out stimuli that
are not related to alcohol; in other words, they have an alcohol-​related attentional bias.
Research suggests that selective attention to alcohol fuels the motivation to drink, and it
might result in actual drinking. In fact, the magnitude of the attentional bias, as measured
with laboratory tasks, has been found to be positively associated with the amount of alco-
hol that drinkers habitually consume (see Cox et al. 2006, 2014, 2015).
The research on alcohol attentional bias and other implicit cognitive processes is pre-
sented next. The discussion describes both (1) how implicit cognitive processes are meas-
ured and (2) interventions that have been developed to overcome them.

3.4  Importance of substance-​related cognitive biases

A substance-​related cognitive bias is a bias in attention, thoughts, memories, interpreta-
tions, values (i.e. likes and dislikes), or action tendencies related to procuring or using an
addictive substance. As discussed earlier, a cognitive bias normally is implicit and auto-
matic; covertly, it might contribute to decision-​making, behavioral tendencies, or action
plans. Accordingly, in various ways substance-​related cognitive biases play an important
role in the continuation of and relapsing to addictive behaviors (Garland et  al. 2012;
Marhe et al. 2013). First, in the case of addiction to alcohol, cognitive biases can cause a
  271

Forms of cognitive bias 271

drinker to be more aware of drinking-​related cues in the environment. Second, they can
interfere with the natural flow of a drinker’s higher-​order cognitive processes, includ-
ing concentration (Waters and Green 2003) and working memory (Houston et al. 2014;
Narendran et  al. 2014; Wilcox et  al. 2014). Third, they can change the natural flow of
information processing in mesolimbic brain areas, such as the nucleus accumbens and
the amygdala (Wiers, Stelzel, Park et al. 2014). Changes in these brain mechanisms might
reduce the controlling role of reflective (cold) processes over impulsive (hot) processes
and hence lead to less controlled patterns of substance use (Pieters et al. 2014; Smith et al.
2014). When executive cognitive functions that are responsible for reflective processes
are impaired due to the adverse effects of chemical use (Fernandez-​Serrano et al. 2010),
the triggering effects of cognitive biases on drinking-​related decisions can be exacerbated
(Field et al. 2010; Loeber et al. 2009; Wiers, Stelzel, Gladwin et al. 2015). A cognitive bias,
therefore, can lead a substance user to make decisions in favour of drug use despite the
person’s conscious vows to refrain from the use.

4  Forms of cognitive bias
Attentional bias. Selective attention or attentional bias refers to a person’s tendency to focus
on and give processing priority to stimuli that are related to his or her current concerns.
Attentional bias for substance-​related stimuli has been widely studied (Coskunpinar and
Cyders 2013; Cox et al. 2006; Field and Cox 2008; Honsi et al. 2013; Leeman et al. 2014).
According to the motivational model of alcohol abuse, the saliency of substance-​related
stimuli occurs because of the relevance of the stimulus to the person’s present or previous
current concerns (e.g. Cox and Klinger 2011a). Attentional bias affects one’s perception
of substance-​related cues and hence one’s motivation to use an addictive substance (Field
and Cox 2008) because they grab and hold the person’s attention to substance-​related cues
even when substance-​unrelated cues are also present.
Most studies of attentional bias have used computer technology. A modified version of
the classic Stroop test, the addiction Stroop test (Cox et al. 2006), has been widely used
to study attentional bias among substance users and abusers. The test usually consists of
two lists of words that have been lexically matched (e.g. on number of letters, frequency of
usage, number of syllables), and each word is presented on a computer screen usually in
one of four colors. One list is substance-​related (i.e. the salient category) and the other list
is substance-​unrelated (i.e. the neutral category). The task for the participant is to ignore
the meaning of each word and to respond as quickly and accurately as possible to the
color of the font. Interference, which is a measure of the degree of the attentional bias for
the substance-​related stimuli, is calculated by subtracting each participant’s mean reaction
time to the neutral category from the mean reaction time to the salient category. In gen-
eral, the magnitude of the interference is proportional to the degree of the substance use
(Cox et al. 2006). In a recent review, Smith and Ersche (2014) concluded that the addiction
Stroop test has diagnostic potential for differentiating between recreational and addicted
drug users and for predicting treatment outcome.
272  

272 Nonconscious motivational influences

Another computerized test for measuring substance-​related attentional bias is based

on the probe-​detection paradigm, the most frequently used form of which is the (visual/​
pictorial) dot-​probe task (Lubman et al. 2000). In its simplest form, the test normally
consists of three consecutive screens: a fixation cross “+” briefly appears at the center of
the screen; a pair of pictures (one salient and one control) then briefly appears side-​by-​
side or vertically but apart from each other; finally, a dot replaces the position of one of
the pictures on the previous trial. The task for the participant is to respond to the posi-
tion of the dot as quickly as possible by pressing one of the assigned keys (e.g. right or
left arrow key). The rationale for the test is that the time spent in locating the dot when
it replaces each control picture will be longer than when it replaces the salient stimuli,
if the person has an attentional bias for the salient stimuli. The presentation time for
the visual stimuli can be either short (e.g. 200 ms) or long (2000 ms). Attentional bias
observed at shorter vsersus longer presentation times can indicate whether any atten-
tional bias that is observed occurs because the salient stimuli have captured the partici-
pant’s attention at an early stage or whether it is because the person’s attention to the
salient stimuli is being maintained (Field et al. 2004). Although there have been many
studies using the addiction Stroop test and visual-​probe task that have found attentional
bias for substance-​related stimuli (Field and Cox 2008), some studies have failed to
observe the attentional bias using these tests, probably because of procedural considera-
tions or the target group that was tested (i.e. social/​recreational users versus addicted
users) (e.g. Marks et al. 2014).
Various other methods use computer technology to detect attentional bias for substance-​
related stimuli: for example, the flicker change blindness paradigm, in which two images
alternate, has been used (Jones et  al. 2006). The two images are identical except for a
minor change between them, which the participant is required to detect. People have
been found to detect a change more readily when it is related to one of their current con-
cerns (e.g. drinking alcohol) than when it is unrelated to their current concerns (Field
et al. 2007; Jones et al. 2002, 2006).
Eye-​tracking has been also used to study substance-​related attentional bias, largely
with smokers (Baschnagel 2013; Kang et al. 2012; Kwak et al. 2007; Van Rensburg et al.
2009) and drinkers (Rose et al. 2013), but mainly in combination with other techniques,
such as the visual probe task (Marks et al. 2014) or the flicker paradigm (Hobson et al.
2013). In one study, Baschnagel (2013) used mobile eye-​tracking to study smokers’
attentional bias for smoking-​related cues in the real world. The results of these studies
show that attentional bias can be reliably detected via longer fixation times on the salient
cues due to initial orienting of attention, especially when participants’ craving is strong.
However, among social/​recreational users, the eye-​tracking studies indicate initial ori-
entation toward the salient stimuli (but not disengagement of attention from them). The
studies, however, have reported that attentional bias does not fully mediate the relation-
ship between cue reactivity and actual use of a substance (Rose et al. 2013). The limited
use of eye-​tracking in research on substance-​related attentional bias warrants further
studies, especially ones with addicted users.
  273

Forms of cognitive bias 273

4.1 Interpretation bias
The word shot might remind some people of a gun firing, but for others it elicits cognitive
processes and associations that are related to the act of drinking alcohol. Such associa-
tions are normally covert and automatic or implicit. In two recent studies (Woud et al.
2012, 2014), a novel task was used that included ambiguous open-​ended scenarios. The
task for the participants was to continue with the scenarios and complete them. They
found that heavy drinkers showed greater interpretation bias for drinking alcohol than
social drinkers did, and dependent drinkers showed greater interpretation bias than par-
ticipants who were clinically diagnosed with an anxiety or depression disorder. Moreover,
the degree of interpretation bias was positively associated with the problem drinkers’ level
of harmful drinking.

4.2 Value bias
As discussed earlier, current concerns reflect affective evaluations of goals, i.e. the value
attributed to a goal, or how much joy, sadness, or sorrow is associated with it (Klinger and
Cox 2011a). Although one can usually identify reasons for wanting to achieve a particular
goal, some aspects of a goal’s value are related to S-​R conditioning (i.e. evaluative con-
ditioning; EC) or some variant of it (Kosinski et al. 2015) that involves pairing a neutral
stimulus with a positive or negative stimulus (De Houwer et  al. 2001; Hofmann et  al.
2010), which may occur at a conscious or a nonconscious level. Despite the emphasis that
some researchers have placed on contingency awareness in EC (e.g. Field 2000), a recent
meta-​analysis by Hofmann et al. (2010) concluded that even low contingency awareness
and subliminal presentation of the unconditioned stimulus can lead to evaluative con-
ditioning of a CS. The value attributed to a goal may, therefore, be learned in both overt
and covert ways, and it can affect one’s goal-​setting and goal pursuits without the person’s
being aware of it. It is not always easy to identify the reasons for our affective associa-
tions, but when they are strong enough they might distort our memories (Kensinger and
Schacter 2005). To conclude, a goal that a person highly values is more likely to be chosen
to pursue, even though the person might not know the reasons for the goal’s desirability;
likewise, something might be assiduously avoided despite the fact the person has no clue
about the reasons for his or her distaste.

4.3  Approach–​avoidance bias
There is a subtle interaction between people’s movements and their tendency to approach
or avoid something. Naturally, people prefer to approach things that they want and to
avoid things that they do not want; this is termed executed movement (Neumann and
Strack 2000). Even when we are not responsible for executing a movement, patterns of an
object’s motion in space are linked to our emotional evaluations of the object. For exam-
ple, it is easier to positively evaluate neutral objects that approach us or move in waves
on a horizontal line (i.e. ~~~) than those that move smoothly (i.e. -​-​-​) along a horizontal
line; likewise, objects that follow a parabolic (i.e. falling) pattern of motion can acquire
274  

274 Nonconscious motivational influences

negative valence (Kosinski et al. 2015). Such subtle influences on our evaluative learning,
along with our past incentive history, determine our natural tendency to approach some
objects and to avoid others. Likewise, when trying to resolve a concern, a person might
want to approach a goal related to it (i.e. an appetitive goal) or to get rid of it (i.e. an aver-
sive goal) (Klinger and Cox 2011a). Goal expectancies may also affect people’s tendency
to approach or avoid something related to the goal. Christiansen et al. (2013) reported
that having positive alcohol expectancies can predict approach tendencies (e.g. to have
a drink). This executed movement tendency gradually acquires an automatic quality.
Evidence (Cousijn et al. 2011; Wiers, et al. 2013) shows that an automatic approach-​bias
for addiction-​related cues can contribute to the continuation of the addictive behavior.
It is evident that addicted individuals have deficits in their general cognitive perfor-
mance or executive functions (Field et al. 2010). Fadardi and Cox (2006) showed, how-
ever, that attentional bias for alcohol-​related stimuli is not an artifact of drinkers’ reduced
executive functions; nevertheless, the strength of addicted individuals’ executive func-
tions is negatively related to their cognitive bias for substance-​related stimuli. Loeber et al.
(2009) found that attentional bias for alcohol-​related stimuli was modulated by drinkers’
executive functions (i.e. attention and working memory), and other studies (Christiansen
et al. 2012; Friese et al. 2010; Smith et al. 2014) have also shown that executive function
deficits can increase abusers’ impulsive decisions to consume substances. In short, these
studies suggest that it would be important to help substance abusers to increase their gen-
eral self-​control in order for them to be better able to control their substance use.

5  Enhancing self-​control
Enhancing drinkers’ self-​control has long been a goal in the treatment of addictive behav-
iors (Marlatt 1982; Marlatt and Marques 1977; Room and Leigh 1992). Motivational
approaches, such as motivational interviewing (Rubak et al. 2005), employ techniques to
persuade substance users to progress through stages of change (i.e. pre-​contemplation,
contemplation, action, and maintenance). Systematic motivational counseling (Cox and
Klinger 2011b) uses techniques to improve substance users’ lifestyle to help them dis-
engage from substance use while they are achieving more fulfilling, substance-​unrelated
goals. Shamloo and Cox (2014) showed that (1) providing participants with hints about
how to complete two experimental tasks and (2) giving them contingent, immediate feed-
back about their performance in completing individualized goals led to an increase in
participants’ sense of control and adaptive motivation and to decreases in explicit and
implicit measures of their urge to drink alcohol. This is an interesting finding because
none of the tasks were directly related to drinking alcohol. Fadardi and colleagues (2011)
discussed the reciprocal nature of motivational and cognitive determinants of substance
abuse and how improvements in one of these determinants can bring improvements in
the other. In fact, sense of control is an important predictor of treatment outcomes in
many kinds of interventions for substance abuse, including medical interventions, cogni-
tive behavioral therapy, contingency management, mindfulness training, and cognitive
bias modification (Leeman et al. 2014).
  275

Cognitive bias modification 275

6  Cognitive bias modification

Theoretically driven and rationally based on the results of studies on different kinds of
cognitive bias in substance-​use disorders, various kinds of cognitive bias modification
(CBM) have been developed, and their effectiveness in reducing substance-​related cogni-
tive biases and improvement of treatment outcomes has been evaluated. There are various
kinds of CBM: CBM-​A refers to modification in attentional bias (e.g. Begh et al. 2013;
Fadardi and Cox 2009; Wiers, Houben et al. 2015); CBM-​I refers to techniques for chang-
ing interpretation biases related to substance-​related stimuli; and CBM-​V refers to tech-
niques for reducing positive evaluations of substance-​related stimuli or situations and
the act of using the substance. On the other hand, approach-​bias modification involves
reducing automatic tendencies to approach and obtain a substance (Eberl et  al. 2013,
2014; Sharbanee et al. 2014; Wiers, Stelzel, Gladwin et al. 2015). Although a single session
of CBM usually does not lead to reductions in attentional bias or craving and does not
help to prevent relapse (Field et al. 2009), interventions with multiple training sessions
have led to positive clinical outcomes (see Cox et al. 2015).
Most of the CBM techniques aim to strengthen substance users’ inhibitory ability when
they encounter substance-​related stimuli in order to (1) prevent allocation of their atten-
tion at an early stage of cognitive processing (i.e. while initially perceiving a stimulus) or
(2) disengage their attention at a later stage. These techniques instead promote cognitive
allocation to alternative neutral or even positive stimuli. Such a shift requires conscious
effort and frequent rehearsal, usually using a computerized program. Normally, the train-
ing procedure includes providing the trainee with a hierarchy of goals to reach through
increasingly difficult steps. The program covertly (e.g. in modified versions of the dot-​
probe task or in CBM-​I or CBM-​V techniques) or overtly (e.g. in modified versions of the
Stroop task, such as the Alcohol Attention Control Training Program (AACTP), Fadardi
and Cox 2009) reinforces controlled ways of responding to target stimuli or situations.
Thereby, the person gradually learns one of several things: (1) to ignore substance-​related
stimuli (e.g. alcohol-​related items in the environment) and simultaneously attend to
non-​alcohol-​related stimuli; (2) to interpret in a more positive, healthier, and substance-​
unrelated way situations that have been habitually associated with alcohol or another
substance; (3) to reduce the positive emotional valence associated with the act of drink-
ing alcohol or using another substance and substance-​related situations; or (4) to reject
alcohol or other substances when they are presented. As indicated, practice with these
programs both reduces the substance user’s cognitive bias and enhances his or her sense
of control over a previously habitual behavior, thereby further enhancing the motivation
to control the substance use.
Cox and colleagues (2015) evaluated whether combining systematic motivational coun-
seling and attentional training using the AACTP could have additive effects in helping
addicted drinkers reduce the amount of alcohol that they habitually drank. The study
used a 2 x 2 factorial design, and outcome was measured post-​intervention and then three
and six months later. Results showed that there was no additive effect for combining the
276  

276 Nonconscious motivational influences

two interventions. However, the attentional retraining rapidly reduced heavy drinkers’
mean weekly drinking but the effects did not endure. The motivational intervention was
effective at reducing weekly drinking; however, the effect appeared after some delay but
lasted longer than the effects of the attentional training. Clearly, more research is needed
to study the temporal course and additive effects of behavioral, CBM, and motivational
interventions for substance use and relapse prevention.

7 Summary
This chapter presents a view of addiction from the standpoint of the motivational and
goal theory of current concerns. A current concern is an individual’s motivational state
between becoming committed to pursuing a particular goal and the point at which the
goal is reached or the pursuit is relinquished. During this time, the current concern guides
the person’s cognitive processes, including attention, memories, thoughts, and dreams.
What is true of goals in general is true of the goal of drinking alcohol or using another
addictive substance. Whether or not a person decides to drink alcohol or use an addic-
tive substance will depend on the value that the person attributes to drinking or using
(i.e. the expected enhancement of positive affect or reduction in negative affect) and the
person’s expectation that the expected changes in affect will actually happen. Many fac-
tors help to determine the value that the person attributes to drinking, such as his or her
past drinking experiences, biochemical reactions to alcohol, environmental and socio-
cultural influences, and the incentives in the person’s life that might compete with trying
to obtain emotional satisfaction from drinking or using an addictive substance. When a
goal to drink or use is formed, the goal is reflected in the person’s cognitive processes (e.g.
thoughts, memories, and attention). These cognitive processes facilitate obtaining and
imbibing the addictive substance. Accordingly, interventions can help the person to coun-
teract these processes and reduce his or her use of substances. This chapter has presented
several of these interventions and the research that has been conducted to evaluate them.

Acknowledgments
The work described in this chapter was supported by Economic and Social Research
Council Grant RES-​000-​23-​9563 (awarded to W. Miles Cox and Emmanuel M. Pothos),
Economic and Social Research Council Grant RES-​000-​22-​0314 (awarded to W.  Miles
Cox and Javad Salehi Fadardi), and Economic and Social Research Council Grant RES-​
000-​23-​1269 (awarded to W. Miles Cox, Javad S. Fadardi, Steven G. Hosier, and Emmanuel
M. Pothos).

References
Abrams, K., Kushner, M.G., Medina, K.L., and Voight, A. (2002). Self-​administration of alcohol before
and after a public speaking challenge by individuals with social phobia. Psychology of Addictive
Behaviors, 16, 121–​8.
Aldao, A., Nolen-​Hoeksema, S., and Schweizer, S. (2010). Emotion-​regulation strategies across
psychopathology: A meta-​analytic review. Clinical Psychology Review, 30, 217–​37.
  277

Summary 277

American Psychiatric Association. (2013). Diagnostic and Statistical Manual of Mental Disorders (5th
ed.). Washington, DC: APA.
Amlung, M. and MacKillop, J. (2014). Clarifying the relationship between impulsive delay discounting
and nicotine dependence. Psychology of Addictive Behaviors, 28, 761–​68.
Baschnagel, J.S. (2013). Using mobile eye-​tracking to assess attention to smoking cues in a naturalized
environment. Addictive Behaviors, 38, 2837–​40. DOI: 10.1016/​j.addbeh.2013.08.005
Begh, R., Munafo, M.R., Shiffman, S., et al. (2013). Attentional bias retraining in cigarette smokers
attempting smoking cessation (ARTS): study protocol for a double blind randomised controlled
trial. BMC Public Health, 13, 1176. DOI: 10.1186/​1471-​2458-​13-​1176
Berkman, E.T. and Lieberman, M.D. (2009). The neuroscience of goal pursuit: Bridging gaps between theory
and data. In: G.B. Moskowitz and H. Grant (eds), The Psychology of Goals. New York: Guilford, pp. 98–​126.
Berridge K.C. (2007). The debate over dopamine’s role in reward: the case for incentive salience.
Psychopharmacology, 191, 391–​431.
Bischof, G., Rumph, H.-​J., and John, U. (2012). Natural recovery from addiction. In: H. Shaffer, D.A.
LaPlante, and S.E. Nelson (eds), APA Addiction Syndrome Handbook, Vol. 2: Recovery, Prevention,
and Other Issues Washington, DC: APA, pp. 133–​55.
Bock, M. and Klinger, E. (1986). Interaction of emotion and cognition in word recall. Psychological
Research, 48, 99–​106
Bouton, M.E. (2002). Context, ambiguity, and unlearning: Ssources of relapse after behavioral
extinction. Biological Psychiatry, 52, 976–​86.
Brandstätter, V., Herrmann, M., and Schüler, J. (2013). The struggle of giving up personal
goals: affective, physiological, and cognitive consequences of an action crisis. Personality and Social
Psychology Bulletin, 39, 1668–​82.
Bundorf, M.K., Mata, R., Schoenbaum, M., and Bhattacharya, J. (2013). Are prescription drug
insurance choices consistent with expected utility theory? Health Psychology, 32, 986–​94.
Carver, C.S., and Scheier, M.F. (1998). On the Self-​Regulation of Behavior. New York: Cambridge
University Press.
Cesario, J. (2014). Priming, replication, and the hardest science. Perspectives on Psychological Science,
9, 40–​48.
Chartrand, T.L. and Bargh, J.A. (1996). Automatic activation of impression formation and
memorization goals: Nonconscious goal priming reproduces effects of explicit task instructions.
Journal of Personality and Social Psychology, 71, 464–​78.
Christiansen, P., Cole, J.C., and Field, M. (2012). Ego depletion increases ad-​lib alcohol
consumption: investigating cognitive mediators and moderators. Exprimental and Clinical
Psychopharmacology, 20, 118–​28. DOI: 10.1037/​a0026623
Christiansen, P., Rose, A.K., Cole, J.C., and Field, M. (2013). A comparison of the anticipated and
pharmacological effects of alcohol on cognitive bias, executive function, craving and ad-​lib drinking.
Journal of Psychopharmacol, 27, 84–​92. DOI: 10.1177/​0269881112450787
Coskunpinar, A., and Cyders, M.A. (2013). Impulsivity and substance-​related attentional
bias: a meta-​analytic review. Drug and Alcohol Dependence, 133, 1–​14. DOI: 10.1016/​
j.drugalcdep.2013.05.008
Cousijn, J., Goudriaan, A.E., and Wiers, R.W. (2011). Reaching out towards cannabis: approach-​bias
in heavy cannabis users predicts changes in cannabis use. Addiction, 106, 1667–​74. DOI: 10.1111/​
j.1360-​0443.2011.03475.x
Cox, W.M. and Klinger, E. (1988). A motivational model of alcohol use. Journal of Abnormal
Psychology, 97, 168–​80.
Cox, W.M. and Klinger, E. (2011a). A motivational model of alcohol use: determinants of use and
change. In: W.M. Cox and E. Klinger (eds), Handbook of Motivational Counseling: Goal-​Based
278  

278 Nonconscious motivational influences

Approaches to Assessment and Intervention with Addiction and Other Problems. Chichester: Wiley-​
Blackwell, pp. 131–​58.
Cox, W.M. and Klinger, E. (2011b). Systematic motivational counseling: from motivational
assessment to motivational change. In: W.M. Cox and E. Klinger (eds), Handbook of Motivational
Counseling: Goal-​Based Approaches to Assessment and Intervention with Addiction and Other
Problems. Chichester, UK: Wiley-​Blackwell, pp. 275–​302.
Cox, W.M., Fadardi, J.S., and Pothos, E.M. (2006). The addiction-​Stroop test: theoretical considerations
and procedural recommendations. Psychological Bulletin, 132, 443–​76. DOI: 2006-​06233-​005
[pii]10.1037/​0033-​2909.132.3.443
Cox, W.M., Klinger, E., and Fadardi, J.S. (2015). The motivational basis of cognitive determinants of
addictive behaviors. Addictive Behaviors, 44, 16–​22.
Cox, W. M., Fadardi, J. S., Hosier, S. G., and Pothos, E. M. (2015). Differential effects and temporal
course of attentional and motivational training on excessive drinking. Experimental and Clinical
Psychopharmacology, 23(6), 445–454.
Cox, W.M., Fadardi, J.S., Hosier, S.G., and Pothos, E.M. (2015). Differential effects and temporal
course of attentional and motivational training on excessive drinking. Experimental and Clinical
Psychopharmacology, 23(6), 445–​54.
Cox, W.M., Schippers, G.M., Klinger, E., et al. (2002). Motivational structure and alcohol use
of university students with consistency across four nations. Journal of Studies on Alcohol, 63,
280–​85.
Dantzer, C., Wardle, J., Fuller, R., Pampalone, S.Z., and Steptoe, A. (2006). International study of
heavy drinking: attitudes and sociodemographic factors in university students. Journal of American
College Health, 55, 83–​89.
De Houwer, J., Thomas, S., and Baeyens, F. (2001). Associative learning of likes and dislikes: a review of
25 years of research on human evaluative conditioning. Psychological Bulletin, 127, 853–​69.
Demirbas, H., Ilhan, I.O., and Dogan, Y.B. (2012). Ways of problem solving as predictors of relapse in
alcohol dependent male inpatients. Addictive Behaviors, 37, 131–​4.
Dickson, P., A. James, M.R., Heath, A.C. et al. (2006). Effects of variation at the ALDH2 locus on
alcohol metabolism, sensitivity, consumption, and dependence in Europeans. Alcoholism: Clinical
and Experimental Research, 30, 1093–​100.
Drobes, D.J., Saladin, M.E., and Tiffany, S.T. (2001). Classical conditioning mechanisms in alcohol
dependence. In: N. Heather, T.J. Peters, and T. Stockwell (eds), International Handbook of Alcohol
Dependence and Problems. Chichester, UK: John Wiley, pp. 281–​97).
Eberl, C., Wiers, R.W., Pawelczack, S., Rinck, M., Becker, E.S., and Lindenmeyer, J. (2013). Approach
bias modification in alcohol dependence: do clinical effects replicate and for whom does it work
best? Developmental Cognitive Neuroscience, 4, 38–​51. DOI: 10.1016/​j.dcn.2012.11.002
Eberl, C., Wiers, R.W., Pawelczack, S., Rinck, M., Becker, E.S., and Lindenmeyer, J. (2014).
Implementation of approach bias re-​training in alcoholism-​how many sessions are needed?
Alcoholism: Clinical and Experimental Research, 38, 587–​94. DOI: 10.1111/​acer.12281
Fadardi, J.S. and Cox, W.M. (2006). Alcohol attentional bias: drinking salience or cognitive
impairment? Psychopharmacology, 185, 169–​78. DOI: DOI:10.1007/​s00213-​005-​0268-​0
Fadardi, J.S. and Cox, W.M. (2009). Reversing the sequence: reducing alcohol consumption by
overcoming alcohol attentional bias. Drug and Alcohol Dependence, 101, 137–​45. DOI: S0376-​
8716(08)00434-​1 [pii]10.1016/​j.drugalcdep.2008.11.015
Fadardi, J.S., Shamloo, Z.S., and Cox, W.M. (2011). Cognitive and motivational retraining: reciprocal
effects. In W.M. Cox and E. Klinger (eds), Handbook of Motivational Counseling: Concepts,
Approaches, and Assessment. Chichester: John Wiley, pp. 395–​411.
  279

Summary 279

Feather, N.T. (ed.) (1982). Expectations and Actions: Expectancy-​value Models in Psychology. Hillsdale,
NJ: Erlbaum.
Feil, J., Sheppard, D., Fitzgerald, P.B., Yücel, M., Lubman, D.I., and Bradshaw, J.L. (2010). Addiction,
compulsive drug seeking, and the role of frontostriatal mechanisms in regulating inhibitory control.
Neuroscience and Biobehavioral Reviews, 35, 248–​75.
Fernandez-​Serrano, M.J., Perez-​Garcia, M., Schmidt Rio-​Valle, J., and Verdejo-​Garcia, A. (2010).
Neuropsychological consequences of alcohol and drug abuse on different components of executive
functions. Journal of Psychopharmacology, 24, 1317–​32. DOI: 10.1177/​0269881109349841
Field, A.P. (2000). I like it, but I’m not sure why: can evaluative conditioning occur without conscious
awareness? Consciousness and Cognition, 9, 13–​36. DOI: 10.1006/​ccog.1999.0402
Field, M. and Cox, W.M. (2008). Attentional bias in addictive behaviors: a model of its development,
causes, and consequences. Drug and Alcohol Dependence, 97, 1–​20.
Field, M., Munafò, M.R., and Franken, I.H.A. (2009). A meta-​analytic investigation of the relationship
between attentional bias and subjective craving in substance abuse. Psychological Bulletin, 135,
589–​607.
Field, M., Duka, T., Tyler, E., and Schoenmakers, T. (2009). Attentional bias modification in tobacco
smokers. Nicotine and Tobaco Research, 11, 812–​22. DOI: ntp067 [pii]10.1093/​ntr/​ntp067
Field, M., Mogg, K., Zetteler, J., and Bradley, B.P. (2004). Attentional biases for alcohol cues in heavy
and light social drinkers: the roles of initial orienting and maintained attention. Psychopharmacology
(Berlin), 176, 88–​93.
Field, M., Wiers, R.W., Christiansen, P., Fillmore, M.T., and Verster, J.C. (2010). Acute alcohol
effects on inhibitory control and implicit cognition: implications for loss of control over
drinking. Alcoholism: Clinical and Experimental Research, 34, 1346–​52. DOI: 10.1111/​
j.1530-​0277.2010.01218.x
Field, M., Duka, T., Eastwood, B., Child, R., Santarcangelo, M., and Gayton, M. (2007). Experimental
manipulation of attentional biases in heavy drinkers: do the effects generalise? Psychopharmacology
(Berlin), 192, 593–​608. DOI: 10.1007/​s00213-​007-​0760-​9
Foa, E.B., Liebowitz, M.R., Kozak, M.J., et al. (2005). Randomized, placebo-​controlled trial of
exposure and ritual prevention, clomipramine, and their combination in the treatment of obsessive-​
compulsive disorder. American Journal of Psychiatry, 162, 151–​61.
Förster, J., Liberman, N., and Friedman, R.S. (2007). Seven principles of goal activation: A systematic
approach to distinguishing goal priming from priming of non-​goal constructs. Personality and
Social Psychology Review, 11, 211–​33.
Friese, M., Bargas-​Avila, J., Hofmann, W., and Wiers, R.W. (2010). Here’s looking at you,
bud: alcohol-​related memory structures predict eye movements for social drinkers with low
executive control. Social Psychological and Personality Science, 1, 143–​51. DOI: 10.1177/​
1948550609359945
Garavan, H., Pankiewicz, J., Bloom, A., et al. (2000). A cue-​induced cocaine craving: neuroanatomical
specificity for drug users and drug stimuli. American Journal of Psychiatry, 157, 1789–​98.
Garland, E.L., Franken, I.H., and Howard, M.O. (2012). Cue-​elicited heart rate variability and
attentional bias predict alcohol relapse following treatment. Psychopharmacology (Berlin), 222, 17–​
26. DOI: 10.1007/​s00213-​011-​2618-​4
Haaker, J., Golkar, A., Hermans, D., and Lonsdorf, T.B. (2014). A review on human reinstatement
studies: An overview and methodological challenges. Learning and Memory, 21, 424–​40.
Ham, L.S., Bonin, M., and Hope, D.A. (2007). The role of drinking motives in social anxiety and
alcohol use. Journal of Anxiety Disorders, 21, 991–​1003.
Heyman, G.M. (2009). Addiction: A Disorder of Choice. Cambridge, MA: Harvard University Press.
280  

280 Nonconscious motivational influences

Hobson, J., Bruce, G., and Butler, S.H. (2013). A flicker change blindness task employing eye tracking
reveals an association with levels of craving not consumption. Journal of Psychopharmacology, 27,
93–​97. DOI: 10.1177/​0269881112447990
Hoelscher, T.J., Klinger, E., and Barta, S.G. (1981). Incorporation of concern-​and nonconcern-​related
verbal stimuli into dream content. Journal of Abnormal Psychology, 49, 88–​91.
Hofmann, W., De Houwer, J., Perugini, M., Baeyens, F., and Crombez, G. (2010). Evaluative
conditioning in humans: a meta-​analysis. Psychological Bulletin, 136, 390–​421. DOI:10.1037/​
a0018916
Hollingworth, W., Ebel, B.E., McCarty, C.A., et al. (2006). Prevention of deaths from harmful drinking
in the United States: The potential effects of tax increases and advertising bans on young drinkers.
Journal of Studies on Alcohol, 67, 300–​308.
Honsi, A., Mentzoni, R.A., Molde, H., and Pallesen, S. (2013). Attentional bias in problem gambling: a
systematic review. Journal of Gambling Studies, 29, 359–​75. DOI: 10.1007/​s10899-​012-​9315-​z
Houston, R.J., Derrick, J.L., Leonard, K.E., Testa, M., Quigley, B.M., and Kubiak, A. (2014). Effects of
heavy drinking on executive cognitive functioning in a community sample. Addictive Behaviors, 39,
345–​49.
Huang, J.Y., and Bargh, J.A. (2014). The selfish goal: autonomously operating motivational structures as
the proximate cause of human judgment and behavior. Behavioral and Brain Sciences, 37, 121–​35.
Huckle, T., Huakau, J., Sweetsur, P., Huisman, O. and Casswell, S. (2008). Density of alcohol outlets
and teenage drinking: living in an alcogenic environment is associated with higher consumption in
a metropolitan setting. Addiction, 103, 1614–​21.
Hussong, A.M., Hicks, R.E., Levy, S.A., and Curran, P.J. (2001). Specifying the relations between affect
and heavy alcohol use among young adults. Journal of Abnormal Psychology, 110, 449–​61.
Ihssen, N., Cox, W.M., Wiggett, A., Fadardi, J.S., and Linden, D.E. J. (2011). Heavy and light drinkers’
brain reactions to alcohol, goal-​related, emotional, and neutral stimuli. Cerebral Cortex, 21,
1408–​15.
Jones, B.C., Jones, B.T., Blundell, L., and Bruce, G. (2002). Social users of alcohol and cannabis who
detect substance-​related changes in a change blindness paradigm report higher levels of use than
those detecting substance-​neutral changes. Psychopharmacology (Berlin), 165, 93–​6. DOI:10.1007/​
s00213-​002-​1264-​2
Jones, B.T., Bruce, G., Livingstone, S., and Reed, E. (2006). Alcohol-​related attentional bias in problem
drinkers with the flicker change blindness paradigm. Psychology of Addictive Behaviors, 20, 171–​7.
DOI:10.1037/​0893-​164X.20.2.171
Kalivas, P.W. and Volkow, N.D. (2005). The neural basis of addiction: A pathology of motivation and
choice. American Journal of Psychiatry, 162, 1403–​13.
Kang, O.S., Chang, D.S., Jahng, G.H., et al. (2012). Individual differences in smoking-​related cue
reactivity in smokers: an eye-​tracking and fMRI study. Progress in Neuropsychopharmacology and
Biological Psychiatry, 38, 285–​93. DOI: 10.1016/​j.pnpbp.2012.04.013
Kensinger, E.A., and Schacter, D.L. (2005). Retrieving accurate and distorted memories: neuroimaging
evidence for effects of emotion. Neuroimage, 27, 167–​77. DOI: 10.1016/​j.neuroimage.2005.03.038
Klingemann, H.K.-​H. (2001). Natural recovery from alcohol problems. In: N. Heather, T.J. Peters, and
T.R. Stockwell (eds), Handbook of Alcohol Dependence and Related Problems. New York: John Wiley,
pp. 649–​62.
Klinger, E. (1971). Structure and Functions of Fantasy. New York: John Wiley.
Klinger, E. (1975). Consequences of commitment to and disengagement from incentives. Psychological
Review, 82, 1–​25.
Klinger, E. (1977). Meaning and Void: Inner Experience and the Incentives in People’s Lives.
Minneapolis: University of Minnesota Press.
  281

Summary 281

Klinger, E. (1978). Modes of normal conscious flow. In: K.S. Pope and J.L. Singer (eds), The Stream of
Consciousness: Scientific Investigations into the Flow of Human Experience. New York: Plenum, pp. 225–​58.
Klinger, E. (1987). Current concerns and disengagement from incentives. In: F. Halisch and J. Kuhl
(eds), Motivation, Intention and Volition. Berlin: Springer, pp. 337–​47.
Klinger, E. (1990). Daydreaming. Los Angeles, CA: Tarcher (Putnam).
Klinger, E. (1993). Loss of interest. In: C.G. Costello (ed.), Symptoms of Depression. New York: Wiley,
pp. 43–​62.
Klinger, E. (1996a). Emotional influences on cognitive processing, with implications for theories
of both. In: P. Gollwitzer and J.A. Bargh (eds), The Psychology of Action: Linking Cognition and
Motivation to Behavior. New York: Guilford, pp. 168–​89).
Klinger, E. (1996b). The contents of thoughts: interference as the downside of adaptive normal
mechanisms in thought flow. In: I.G. Sarason, B.R. Sarason, and G.R. Pierce (eds), Cognitive
interference: Theories, Methods, and Findings. Hillsdale, NJ: Lawrence Erlbaum Associates, pp. 3–​23.
Klinger, E. (2012). The search for meaning in evolutionary perspective and its clinical implications.
In: P.T.P. Wong (ed.), The Human Quest for Meaning: Theories, Research, and Applications (2nd ed.).
New York: Routledge, pp. 23–​56.
Klinger, E. (2013). Goal commitments and the content of thoughts and dreams: Basic principles.
Frontiers in Psychology, 4, DOI: 10.3389/​fpsyg.2013.00415.
Klinger, E. and Cox, W.M. (2011a). Motivation and the goal theory of current concerns. In: W.M.
Cox and E. Klinger (eds), Handbook of Motivational Counseling (2nd ed.). Chichester, UK: Wiley-​
Blackwell, pp. 3–​47.
Klinger, E. and Cox, W.M. (2011b). The Motivational Structure Questionnaire, Personal Concerns
Inventory, and their variants: psychometric properties. In: W.M. Cox and E. Klinger (eds),
Handbook of Motivational Counseling (2nd ed.). Chichester, UK: Wiley-​Blackwell, pp. 205–​32.
Klinger, E., and McNelly, F.W., Jr. (1969). Fantasy need achievement and performance: a role analysis.
Psychological Review, 76, 574–​91.
Klinger, E., Barta, S.G., and Maxeiner, M.E. (1980). Motivational correlates of thought content
frequency and commitment. Journal of Personality and Social Psychology, 39, 1222–​37.
Köhler W. (1925). The Mentality of Apes. New York: Harcourt, Brace.
Kosinski, T., Chafi, A., and Rusinek, S. (2015). When motion changes liking: evaluative conditioning
with motion as unconditioned stimulus. Learning and Motivation, 49, 1–​5. DOI: 10.1016/​
j.lmot.2014.10.003
Koster, E.H. W., De Lissnyder, E., Derakshan, N., and De Raedt, R. (2011). Understanding depressive
rumination from a cognitive science perspective: the impaired disengagement hypothesis. Clinical
Psychology Review, 31, 138–​45.
Kouneiher, F., Charron, S., and Koechlin, E. (2009). Motivation and cognitive control in the human
prefrontal cortex. Nature Neuroscience, 12, 939–​45.
Kwak, S.M., Na, D.L., Kim, G., Kim, G.S., and Lee, J.H. (2007). Use of eye movement to measure
smokers’ attentional bias to smoking-​related cues. Cyberpsychology and Behavior, 10, 299–​304.
DOI: 10.1089/​cpb.2006.9953
Leeman, R.F., Bogart, D., Fucito, L.M., and Boettiger, C.A. (2014). “Killing two birds with one
stone”: alcohol use reduction interventions with potential efficacy in enhancing self-​control. Current
Addiction Reports, 1, 41–​52. DOI: 10.1007/​s40429-​013-​0008-​1
Leeman, R.F., Robinson, C.D., Waters, A.J., and Sofuoglu, M. (2014). A critical review of the literature
on attentional bias in cocaine use disorder and suggestions for future research. Experimental and
Clinical Psychopharmacology, 22, 469–​83. DOI: 10.1037/​a0037806
Paschall, M.J., Lipperman-​Kreda, S., and Grube, J.W. (2014). Effects of the local alcohol environment
on adolescents’ drinking behaviors and beliefs. Addiction, 109, 407–​16.
282  

282 Nonconscious motivational influences

Loeber, S., Vollstadt-​Klein, S., von der Goltz, C., Flor, H., Mann, K., and Kiefer, F. (2009). Attentional
bias in alcohol-​dependent patients: the role of chronicity and executive functioning. Addiction
Biology, 14, 194–​203. DOI: ADB146 [pii]10.1111/​j.1369-​1600.2009.00146.x
Lubman, D.I., Peters, L.A., Mogg, K., Bradley, B.P., and Deakin, J.F. (2000). Attentional bias for drug
cues in opiate dependence. Psychological Medecine, 30, 169–​75.
MacKillop, J., Miranda, R., Jr., Monti, P.M. et al. (2010). Alcohol demand, delayed reward discounting,
and mediation by craving in individuals with alcohol use disorders. Journal of Abnormal Psychology,
119, 106–​14.
Marchetti, I., Koster, E.H.W., and De Raedt, R. (2012a). Mindwandering heightens the accessibility of
negative relative to positive thought. Consciousness and Cognition, 21, 1517–​25.
Marchetti, I., Koster, E.H., Sonuga-​Barke, E.J., and De Raedt, R. (2012b). The default mode network
and recurrent depression: a neurobiological model of cognitive risk factors. Neuropsychology Review,
22, 229–​51.
Marhe, R., Waters, A.J., van de Wetering, B.J., and Franken, I.H. (2013). Implicit and explicit drug-​
related cognitions during detoxification treatment are associated with drug relapse: an ecological
momentary assessment study. Journal of Consulting and Clinical Psychology, 81, 1–​12. DOI:10.1037/​
a0030754
Marien, H., Custers, R., Hassin, R.R., and Aarts, H. (2012). Unconscious goal activation and
the hijacking of the executive function. Journal of Personality and Social Psychology, 103,
399–​415.
Marks, K.R., Pike, E., Stoops, W.W., and Rush, C.R. (2014). Test-​retest reliability of eye tracking
during the visual probe task in cocaine-​using adults. Drug and Alcohol Dependence, 145, 235–​7.
DOI: 10.1016/​j.drugalcdep.2014.09.784
Marks, K.R., Roberts, W., Stoops, W.W., Pike, E., Fillmore, M.T., and Rush, C.R. (2014). Fixation
time is a sensitive measure of cocaine cue attentional bias. Addiction, 109, 1501–​8. DOI: 10.1111/​
add.12635
Marlatt, G.A. (1982). Relapse prevention: a self-​control program for the treatment of addictive
behaviors. In: R.B. Stuart (ed.), Adherence, Compliance and Generalization in Behavioral Medicine.
New York: Brunner/​Mazel, pp. 329–​77.
Marlatt, G.A., and Marques, J.K. (1977). Meditation, self-​control and alcohol use. In R.B. Stuart (ed.),
Behavioral Self-​management: Strategies, Techniques and Outcomes. New York: Brunner/​Mazel, pp.
117–​53.
McGeoch, J.A. (1993). Studies in retroactive inhibition: the temporal course of the inhibitory effects of
interpolated learning. Journal of General Psychology, 9, 24–​42.
Mongin, P. (1997). Expected utility theory. In J. Davis, W. Hands, & U. Maki (Eds.), Handbook of
economic methodology (pp. 342–​350). London: Edward Elgar.
Montagrin, A., Brosch, T., and Sander, D. (2013). Goal conduciveness as a key determinant of memory
facilitation. Emotion, 13, 622–​28.
Moos, R.H. and Moos, B.S. (2007). Treated and untreated alcohol-​use disorders: Course and predictors
of remission and relapse. Evaluation Review, 31, 564–​84.
Moos, R.H., Finney, J.W., and Cronkite, R.C. (1990). Alcoholism Treatment: Context, Process, and
Outcome. New York: Oxford University Press.
Morone, A. and Morone, P. (2014). Estimating individual and group preference functionals using
experimental data. Theory and Decision, DOI: http://​dx.doi.org/​10.1007/​s11238-​014-​9431-​6
Moskowitz, G.B.,] and Balcetis, E. (2014). The conscious roots of selfless, unconscious goals. Behavioral
and Brain Sciences, 37, 151.
Moskowitz, G.B., Li, P., Ignarri, C., and Stone, J. (2011). Compensatory cognition associated with
egalitarian goals. Journal of Experimental Social Psychology, 47, 365–​70.
  283

Summary 283

Moss A.C. and Albery, I.P. (2009). A dual-​process model of the alcohol-​behavior link for social
drinking. Psychological Bulletin, 135, 516–​30.
Narendran, R., Mason, N.S., Paris, J., Himes, M.L., Douaihy, A.B., and Frankle, W.G. (2014).
Decreased prefrontal cortical dopamine transmission in alcoholism. American Journal of Psychiatry,
171, 881–​8. DOI:10.1176/​appi.ajp.2014.13121581
Neumann, R. and Strack, F. (2000). Approach and avoidance: the influence of proprioceptive and
exteroceptive cues on encoding of affective information. Journal of Personality and Social Psychology,
79, 39–​48.
Nikles, C.D.II, Brecht, D.L., Klinger, E., and Bursell, A.L. (1998). The effects of current-​concern-​and
nonconcern-​related waking suggestions on nocturnal dream content. Journal of Personality and
Social Psychology, 75, 242–​55.
Nikula, R., Klinger, E., and Larson-​Gutman, M.K. (1993). Current concerns and electrodermal
reactivity: Responses to words and thoughts. Journal of Personality, 61, 63–​84.
Noël, X., Bechara, A., Brevers, D., Verbanck, P., and Campanella, S. (2010). Alcoholism and the loss of
willpower: A neurocognitive perspective. Journal of Psychophysiology, 24, 240–​8.
Noël, X., Brevers, D., and Bechara, A. (2013). A triadic neurocognitive approach to addiction for
clinical interventions. Frontiers in Psychiatry, 4 DOI:http://​dx.doi.org/​10.3389/​fpsyt.2013.00179
Nolen-​Hoeksema, S., Wisco, B.E., and Lyubomirsky, S. (2008). Rethinking rumination. Perspectives on
Psychological Science, 3, 400–​24.
Ostafin, B.D., Bauer, C., and Myxter, P. (2012). Mindfulness decouples the relation between automatic
alcohol motivation and heavy drinking. Journal of Social and Clinical Psychology, 31, 729–​45.
Paschall, M.J., Lipperman-​Kreda, S., and Grube, J.W. (2014). Effects of the performance: a role
analysis. Psychological Review, 76, 574–​91.
Pieters, S., Burk, W.J., Van der Vorst, H., Engels, R.C., and Wiers, R.W. (2014). Impulsive and reflective
processes related to alcohol use in young adolescents. Frontiers in Psychiatry, 5, 56. DOI: 10.3389/​
fpsyt.2014.00056
Quinlan, S., Jaccard, J., and Blanton, H. (2006). A decision theoretic and prototype conceptualization
of possible selves: Implications for the prediction of risk behavior. Journal of Personality, 74,
600–​30.
Room, R. and Leigh, B.C. (1992). Self-​control concerns and drinking loss of control in general and
clinical populations. Journal of Studies on Alcohol, 53, 590–​93.
Rose, A.K., Brown, K., Field, M., and Hogarth, L. (2013). The contributions of value-​based decision-​
making and attentional bias to alcohol-​seeking following devaluation. Addiction, 108, 1241–​9.
DOI: 10.1111/​add.12152
Rubak, S., Sandbaek, A., Lauritzen, T., and Christensen, B. (2005). Motivational interviewing: a
systematic review and meta-​analysis. British Journal of General Practice, 55, 305–​12.
Schneider,W.(1987). Ablenkung und Handlungskontrolle:EineKognitiv-​MotivationalePerspektive.
Unpublished Diploma thesis,University of Bielefeld, Bielefeld, Germany.
Shamloo, Z.S., and Cox, W.M. (2014). Information-​enhancement and goal setting techniques for
increasing adaptive motivation and decreasing urges to drink alcohol. Addictive Behaviors, 39,
1205–​13. DOI: 10.1016/​j.addbeh.2014.03.023
Sharbanee, J.M., Hu, L., Stritzke, W.G., Wiers, R.W., Rinck, M., and MacLeod, C. (2014). The effect
of approach/​avoidance training on alcohol consumption is mediated by change in alcohol action
tendency. PLoS One, 9, e85855. DOI: 10.1371/​journal.pone.0085855
Smart, R.G. (2007). Natural recovery or recovery without treatment from alcohol and drug problems
as seen from survey data. In: H. Klingemann and L.C. Sobell (eds), Promoting Self-​change from
Addictive Behaviors: Practical Implications for Policy, Prevention, and Treatment New York: Springer
Science + Business Media, pp. 59–​71.
284  

284 Nonconscious motivational influences

Smith, D.G. and Ersche, K.D. (2014). Using a drug-​word Stroop task to differentiate recreational from
dependent drug use. CNS Spectrums, 19, 247–​55. DOI: 10.1017/​S1092852914000133
Smith, D.G., Simon Jones, P., Bullmore, E.T., Robbins, T.W., and Ersche, K.D. (2014). Enhanced
orbitofrontal cortex function and lack of attentional bias to cocaine cues in recreational stimulant
users. Biological Psychiatry, 75, 124–​31. DOI: 10.1016/​j.biopsych.2013.05.019
Sobell, L., C., Ellingstad, T.P., and Sobell, M.B. (2000). Natural recovery from alcohol and drug
problems: methodological review of the research with suggestions for future directions. Addiction,
95, 749–​64.
Stahl, B. (2013). Addicted nurses keep licenses. StarTribune, 32: A31, A38–​9.
Swendson, J.D., Tennen, H., Carney, M.A., et al. (2000). Mood and alcohol consumption: an
experience sampling test of the self-​medication hypothesis. Journal of Abnormal Psychology,
109(2), 198–​204.
Tabakoff, B. and Hoffman, P.L. (2013). The neurobiology of alcohol consumption and alcoholism: An
integrative history. Pharmacology, Biochemistry and Behavior, 113, 20–​37.
Tucker, J.A., Roth, D.L., Vignolo, M.J., and Westfall, A.O. (2009). A behavioral economic reward index
predicts drinking resolutions: Moderation revisited and compared with other outcomes. Journal of
Consulting and Clinical Psychology, 77, 219–​28.
Tucker, J.A., Vuchinich, R.E., Black, B.C., and Rippens, P.D. (2006). Significance of a behavioral
economic index of reward value in predicting drinking problem resolution. Journal of Consulting
and Clinical Psychology, 74, 317–​26.
Van Eerde, W. and Thierry, H. (1996). Vroom’s expectancy models and work-​related criteria: a meta-​
analysis. Journal of Applied Psychology, 81, 575–​86.
Van Rensburg, K.J., Taylor, A., and Hodgson, T. (2009). The effects of acute exercise on attentional
bias towards smoking-​related stimuli during temporary abstinence from smoking. Addiction, 104,
1910–​7. DOI:10.1111/​j.1360-​0443.2009.02692.x
Vanderschuren, L.J.M.J., and Everitt, B.J. (2005). Behavioral and neural mechanisms of compulsive
drug seeking. European Journal of Pharmacology, 526, 77–​88.
Vrieze, S.I., Vaidyanathan, U., Hicks, B.M., Iacono, W.G., McGue, M. (2014). The role of constraint
in the development of nicotine, marijuana, and alcohol dependence in young adulthood. Behavior
Genetics, 44, 14–​24.
Waters, H. and Green, M.W. (2003). A demonstration of attentional bias, using a novel dual task
paradigm, towards clinically salient material in recovering alcohol abuse patients? Psychological
Medicine, 33, 491–​8.
Wiers, R.W., Bartholow, B.D., van den Wildenberg, E., et al. (2007). Automatic and controlled
processes and the development of addictive behaviors in adolescents: a review and a model.
Pharmacology, Biochemistry and Behavior, 86, 263–​83.
Wiers, C.E., Kuhn, S., Javadi, A.H. et al. (2013). Automatic approach bias towards smoking cues is
present in smokers but not in ex-​smokers. Psychopharmacology (Berlin), 229, 187–​97. DOI: 10.1007/​
s00213-​013-​3098-​5
Wiers, R.W., Eberl, C., Rinck, M., Becker, E.S., and Lindenmeyer, J. (2011). Retraining automatic
action tendencies changes alcoholic patients’ approach bias for alcohol and improves treatment
outcome. Psychological Science, 22, 490–​7.
Wiers, R.W., Houben, K., Fadardi, J.S., van Beek, P., Rhemtulla, M., and Cox, W.M. (2015). Alcohol
Cognitive Bias Modification training for problem drinkers over the web. Addictive Behaviors, 40,
21–​6. DOI: 10.1016/​j.addbeh.2014.08.010
Wiers, C.E., Stelzel, C., Park, S.Q., et al. (2014). Neural correlates of alcohol-​approach bias in alcohol
addiction: the spirit is willing but the flesh is weak for spirits. Neuropsychopharmacology, 39,
688–​97. DOI: 10.1038/​npp.2013.252
  285

Summary 285

Wiers, C.E., Stelzel, C., Gladwin, T.E., et al. (2015). Effects of cognitive bias modification training on
neural alcohol cue reactivity in alcohol dependence. American Journal of Psychiatry, 72, 335–​43.
DOI: 10.1176/​appi.ajp.2014.13111495
Wilcox, C.E., Dekonenko, C.J., Mayer, A.R., Bogenschutz, M.P., and Turner, J.A. (2014). Cognitive
control in alcohol use disorder: deficits and clinical relevance. Reviews in the Neurosciences, 25,
1–​24. DOI:10.1515/​revneuro-​2013-​0054
Witkiewitz, K., Lustyk, M.K., and Bowen, S. (2013). Retraining the addicted brain: a review of
hypothesized neurobiological mechanisms of mindfulness-​based relapse prevention. Psychology of
Addictive Behaviors, 27, 351–​65.
Woud, M.L., Fitzgerald, D.A., Wiers, R.W., Rinck, M., and Becker, E.S. (2012). “Getting into the
spirit”: alcohol-​related interpretation bias in heavy-​drinking students. Psychology of Addictive
Behaviors, 26, 627–​32. DOI: 10.1037/​a0029025
Woud, M.L., Pawelczak, S., Rinck, M., et al. (2014). Alcohol-​related interpretation bias in alcohol-​
dependent patients. Alcoholism: Clinical and Experimental Research, 38, 1151–​9. DOI: 10.1111/​
acer.12334
Wrosch, C., Scheier, M.F., and Miller, G.E. (2013). Goal adjustment capacities, subjective well‐being,
and physical health. Social and Personality Psychology Compass, 7, 847–​60.
Zywiak, W.H., Stout, R.L., Longabaugh, R., et al. (2006a). Relapse-​onset factors in Project
MATCH: The Relapse Questionnaire. Journal of Substance Abuse Treatment, 31, 341–​5.
Zywiak, W.H., Stout, R.L., Trefry, W.B., et al. (2006b). Alcohol relapse repetition, gender, and predictive
validity. Journal of Substance Abuse Treatment, 30, 349–​53.
286

Chapter 16

Self-​regulation, controlled processes,

and the treatment of addiction
Andrew J. Vonasch
Heather M. Maranges
Roy F. Baumeister

Abstract
A common view is that a core feature of addiction is loss of conscious
control. We present an alternative view in which we emphasize the role
of conscious control and self-regulation in both sustaining and ending an
addiction. In our model, addictive behaviors occur because addicts self-
regulate successfully to fulfill problematic desires. Therefore, the most crucial
step in treatment is that the addict consciously decides to stop fulfilling
those desires. The addict must then self-regulate to quit. We describe the
self-regulatory steps an addict must take to quit using drugs (set clear
standards, maintain motivation to quit, maintain self-efficacy to quit, train
willpower, and avoid depleting willpower through avoiding too many
concurrent uses) and to avoid relapse (make lifestyle changes to minimize
temptation, and avoid losing motivation after a lapse).

1 Introduction
Addictive behavior is often tragic for the addict and his or her loved ones. Addicts behave
in ways that seem irrational and incompatible with conscious, deliberate choice. After all,
no-​one thinking clearly would choose to be an addict. The traditional scientific perspec-
tive on addiction has been to treat it as a disease of the mind brought on by the consump-
tion of drugs (e.g. Jellinek 1960). The implication is that the drugs take over the mind,
causing the addict to lose control over his own choices and actions. This perspective is
also consistent with popular methods of treating addiction. For example, Alcoholics
Anonymous asks addicts to submit to a higher power and to acknowledge their power-
lessness over alcohol. Thus, the traditional perspective is that unconscious, automatic,
and uncontrolled processes determine addictive behaviors. From this perspective there
is little room for consciously controlled processes such as choice, preference, willpower,
and future planning.
  287

Self-regulation theory 287

In this chapter, however, we explore a complementary view. In reading the scientific

literature, we have found that consciously controlled processes are indeed very impor-
tant for explaining addictive behavior (Baumeister and Vonasch 2015; Baumeister
et al. 2015). Automatic processes surely do contribute to addictive behaviors, but so do
consciously controlled processes. Addicts use willpower and planning both to use and
to stop using drugs (Baumeister and Vonasch 2015; Baumeister et al. 2015). Addicts
consciously choose to use drugs because they like the subjective experiences the drugs
provide (Heyman 2009). Unconscious, automatic processes are nonetheless important
because they influence how motivated people are to use drugs. People who strongly
enjoy drugs (or who dislike the feeling of not using drugs because they suffer with-
drawal symptoms or because they are self-​medicating) are presumably much more
vulnerable to addiction than people who dislike drug experiences. While automatic
processes often determine people’s motivations to use drugs, people consciously con-
trol when they actually use drugs. For example, even a person “hopelessly” addicted to
crack will usually exert self-​control to avoid lighting up until he is beyond the sight of
a police officer. In our view, the fact that addicts use self-​control in this and other ways
offers hope that they can control their behavior to cease their problematic use.
Elsewhere (i.e. Baumeister and Vonasch 2015; Baumeister et  al. 2015), we have
reviewed the literature and concluded that addicts use conscious processes to regulate
their addictive behaviors in each stage of addiction, i.e. the stage before regular use,
the stage of regular use, the stage of quitting or reducing use, and the stage of main-
taining sobriety (or reduced levels of use). In this chapter, we only have space to sum-
marize the major findings from those reviews and their implications for treatment. We
first introduce the reader to the most up-​to-​date scientific theories about consciously
controlled processes, especially willpower (called self-​regulation in the scientific lit-
erature). We then apply these theories to explain addictive behaviors. We focus in this
chapter on the controlled processes involved in the quitting and sobriety maintenance
stages of addiction. Based on our understanding of these controlled processes, we offer
some treatment suggestions to aid addicts in quitting and in maintaining sobriety or
reduced levels of drug use. Many of the treatment suggestions we make will be famil-
iar to readers, but we think it may be useful to think about these treatments from the
perspective that many addictive behaviors are controlled rather than out of control.

2  Self-​regulation theory
Self-​regulation refers to the ability to override one’s impulses and exert top-​down control
over the self ’s thoughts, emotions, impulses, or actions (Baumeister et  al. 1994). When
Sally has an urge to smoke a cigarette and decides to wait until her lunch break, she is self-​
regulating the timing of her behavior. If she decides to distract herself from her urges by
thinking about something else, she is regulating her thoughts (and desires). Self-​regulation
usually involves overriding and inhibiting an automatic, intuitive thought, emotion,
impulse, or action (Baumeister et al. 1994). However, self-​regulation can also involve doing
288  

288 Self-regulation, controlled processes, and the treatment of addiction

something rather than nothing when the automatic impulse is to be passive (Vohs et al.
2014). For example, when Brad is feeling lazy but decides to get off the couch and go to
the gym anyway, he is self-​regulating his behavior. A person is self-​regulating whenever he
changes his behavior to do something other than what his initial intuitions would motivate.
Self-​regulation by definition involves a directed change in behavior, not just a random
different behavior. People self-​regulate in order to bring their behavior in line with stand-
ards. Standards can take many forms, including ideals, norms, values, goals, peer expecta-
tions, and desires to behave consistently with one’s past self. These may be set by others.
For example, Sally waited to smoke a cigarette until her lunch break because of a rule
established by her employer or government stipulating that employees are not allowed
to smoke except on designated breaks. In addition, however, many standards are self-​
imposed. For example, Stan is attempting to curtail his drinking by setting a three-​drink
limit for himself each night.
Thus, standards help to guide people’s behavior. To know when to self-​regulate, people
compare their behavior to a standard and adjust their behavior until it meets the standard
(Carver and Scheier 1981, 1982). Clear standards aid self-​regulation because they make
it easy to know when behavior is sufficiently adjusted. Clear, explicit standards constitute
what some call “bright lines,” referring to the fact that one cannot miss that one is crossing
the line (Ainslie 2001, p. 94 et seq.). Zero drinks is a bright-​line rule, because consump-
tion of any amount is a clear violation. Trying to drink less is not a bright line because
there are many different ways of calculating “less.” This is likely a reason why many treat-
ment programs, including Alcoholics Anonymous, require clients to remain completely
abstinent. When standards are unclear or conflicting, self-​regulation is hampered because
it is difficult to know when self-​regulation has been successful.
To successfully adhere to a standard you must attend to both the standard and your
behavior. Although it may seem as though you are always aware of your own behavior,
psychologists have shown that people are not always self-​aware enough to regulate their
behavior successfully. Several factors can impair the self ’s ability to monitor behavior
and thereby weaken self-​regulation, including intense emotions, distractions found in
social settings like bars and parties, and intoxication. People often find it difficult to keep
track of their participation in a busy evening’s events even when sober, let alone while
drunk or high. Studies have shown that alcohol reduces self-​awareness (Hull 1981), so
it is especially difficult to monitor your own behavior when you have been drinking. As
a result, alcohol users have impaired self-​regulation across a variety of domains: they
eat more, smoke more, spend more money, leave overly large tips, are less modest, and
drink more alcohol (Baumeister et al. 1994). For a person trying to quit alcohol or other
drugs, drinking alcohol disrupts the monitoring process and makes it even harder to
know when to stop. This can result in a vicious cycle of increasing consumption.
If you know your desired standard of behavior and you know that you are currently
not meeting that standard, you may still fail to achieve the standard because of a lack of
willpower. One perspective on self-​regulation that has engaged our laboratory is the lim-
ited resource model of self-​regulation (Baumeister et al. 1994, 1998, 2007; Muraven and
  289

Self-regulation theory 289

Baumeister 2000). The core idea is that the self ’s ability to regulate behavior is limited by
a resource that becomes depleted with effort. In other words, willpower is limited.
Many studies have been conducted testing this idea using a dual-​task paradigm in
which participants who engaged in self-​regulation in a first task performed worse on any
subsequent task that also required self-​regulation. For example, participants in one study
(Baumeister et al. 1998) were randomly assigned to either restrain themselves from eating
a plate of freshly baked cookies and instead taste test a plate of radishes, or to indulge in
the cookies and leave the radishes alone. Of course, it is much easier to enjoy the cookies
than to restrain oneself from eating them and instead eat bitter radishes. In the second
part of the experiment, participants were given a puzzle to solve that actually had no cor-
rect answer (i.e. the puzzle was impossible). Experimenters timed how long the partici-
pants took before giving up. Participants who had depleted their willpower by restraining
themselves from eating the cookies gave up sooner on the puzzle than participants who
had been allowed to eat the cookies. We call the depletion of willpower “ego depletion,” in
recognition of its similarity to an old Freudian theory of the operations of the ego.
Recent theoretical developments have added nuances to the theory (Baumeister and
Vohs 2014). It is now clear that ego depletion is not caused by running out of the limited
resource. Rather, depleted people restrict their usage of the resource in order to conserve
it for the future (Muraven et al. 2006). Even depleted people can use willpower, but they
often choose to conserve their energy instead (Beedie and Lane 2012). This is good news
for people struggling to avoid addictive behaviors. Even when depleted, people can exert
willpower if they are motivated enough. However, exerting willpower feels very effort-
ful and every feeling and desire tends to feel more extreme when people are depleted
(Vohs et  al. 2014). One should not count on strong motivation to overcome strong
desires forever. People who are severely depleted may ultimately exhaust their resources
(Baumeister and Vohs 2014). Because any act of self-​regulation can deplete the resource,
a person who is attempting to quit drugs should avoid exerting self-​regulation for other
purposes (such as dieting, or completing a difficult assignment at work) while quitting.
There is some evidence that glucose in the bloodstream (a chemical used for energy
by the brain and body) relates to the limited resource (Gailliot et al. 2007; Gailliot and
Baumeister 2007). Diversion of glucose from the brain to other body parts can lead to
symptoms resembling depletion. For example, during the late luteal, or premenstrual,
phase of the menstrual cycle, many women experience symptoms of premenstrual syn-
drome (PMS) as the ovaries increase their metabolic expenditures and use more glucose
(Webb 1981, 1986). This may leave less glucose available for self-​regulation, implicated
by PMS symptoms’ resemblance to ego depletion (Gailliot et  al. 2010). Women’s use
of nicotine, caffeine, alcohol, and nonprescribed drugs increases during the late luteal
phase (Mello et al. 1987; Marks et al. 1994; Snively et al. 2000). Alcohol also reduces
blood glucose, which may partially explain why alcohol impairs self-​regulation.
Willpower can be trained, much like strengthening a muscle (Muraven et  al. 1999).
Thus, in the long term exercising willpower strengthens it, but in the short term exercis-
ing willpower depletes it. Training willpower has been shown to increase the success-
ful quitting rates of smokers. For example, one study assigned participants to practice
290  

290 Self-regulation, controlled processes, and the treatment of addiction

squeezing a hand grip for as long as they could twice a day or to resist sweets all day for
two weeks before attempting to quit smoking (Muraven 2010). Those who practiced these
habits—​which required self-​regulation to monitor and change thoughts and behavior—​
were more successful at quitting smoking than participants who performed easy and brief
math problems or wrote down their urges in a diary, which did not require self-​control.

3 Addiction
Although we do not have space to fully describe the theory of controlled processes within
addiction (see Baumeister et al. 2015), here we briefly outline the main arguments. The
word “addiction” originally referred to a strong desire for something, but now it more
often refers to a problematic desire (Orford 2001). The stereotypical view of addiction
is that it is characterized by an overwhelming, perhaps uncontrollable and irresistible,
craving. This view is advocated by some scholars, such as Leshner (1997, 1999), who view
addiction as a brain disease. We do not dispute that addiction, like all mental phenomena,
is mediated by brain processes. What we take issue with is the idea that, because addiction
changes the brain, it does so exclusively through uncontrolled processes that bypass the
self. The available evidence supports the idea that the conscious self plays an integral role
in establishing a pattern of addictive behavior.
Addiction is not just a brain state but a pattern of addictive behavior. At any one point
in time, an addict may face a desire to use. A  recent study by Hofmann et  al. (2012)
found that desires for alcohol and tobacco among regular drinkers and smokers (not all
of whom were clinically addicted) were actually weaker on average than all other desires,
and they were less likely to be rated as “irresistible.” Desires to drink and smoke were
quite frequent, however, and that may be a key to understanding addiction. It is not that
one must use a tremendous amount of willpower to overcome an overwhelming urge to
smoke, but that one must use a little bit of willpower many times each day. Viewed in that
way, it is hardly surprising that many people fail to suppress those urges every single time.
What differentiates an addict from a casual user is the degree of problematic use.
Addictive behavior is characterized by a series of dysfunctional, impulsive choices to use
drugs despite substantial negative consequences (Redish et al. 2008; Heyman 2009). As
we said earlier, it is likely that no one ever decides to become a drug addict. But people
do choose to use drugs each time they use them (aside from rare cases in which people
are coerced to use drugs). Most of these everyday decisions to use drugs are guided by
short-​term, pleasure-​seeking motivations. People use drugs because drugs make them
feel good (or because drugs prevent them from feeling bad). Using drugs once usually has
few negative consequences, but using drugs repeatedly can be seriously detrimental to
health, finances, occupational performance, and social relationships.
Some theorists have promoted the brain disease theory of addiction by pointing to
evidence of brain changes linked to long-​term addiction (Leshner 1997, 1999; Robinson
and Berridge 2003). However, these changes do not indicate that people lose the ability
to choose and control their actions. As critics such as Schaler (2000) and Heyman (2009)
have pointed out, the changes are mainly localized in the so-​called desire centers of the
  291

Deciding to fight against addiction 291

brain rather than the motor control centers. Thus, when addiction changes the brain, it
changes the patterns of wanting, not the ability to control one’s movements. The addict
does not lose control; rather, the addict develops a persistent desire to use drugs for their
pleasurable effects and to avoid displeasure from not using drugs.
Addiction does not begin or end overnight. The life course of addiction and addictive
behaviors can be subdivided into four heuristic stages (Baumeister and Vonasch 2015;
Baumeister et  al. 2015). These stages are only guidelines to help think about addictive
behaviors, rather than rigidly differentiated steps. For example, it may not usually be pos-
sible to specify the exact moment at which a person becomes “addicted.” The four stages
are pre-​addiction, regular use, fighting against addiction, and fighting against relapse.
We will not discuss the pre-​addictive or regular use stages further, except to emphasize
the importance of the addict’s decision to cease regular use and fight against addiction.
The remainder of the chapter focuses on the decision to fight against addiction, the self-​
regulatory processes involved in fighting against addiction, the self-​regulatory processes
involved in fighting against relapse, and treatment suggestions for fighting against both
addiction and relapse.

4  Deciding to fight against addiction

Given our view that many addictive behaviors are the result of controlled but unwise and
problematic choices, treating addicts primarily involves helping them make better choices.
The first step for an addict is to resolve to reduce or eliminate her use of the problematic
substance or substances. The literature is very clear that an addict is most successful in quit-
ting when the addict decides herself (albeit sometimes with encouragement from others)
to cease problematic use (e.g. Chapman and Mackenzie 2010; Miller and Rollnick 1991;
Liskow et al. 1990). Coerced sobriety rarely lasts long beyond the coercion (Freedberg and
Johnston 1978; Gallant et al. 1973; Rosenberg and Liftik 1976). When self-​motivated, how-
ever, many people can quit or reduce consumption to non-​problematic levels. Many people
who at one time abused substances now continue to use those substances at rates deemed
not problematic (Heather and Robertson 1981; Zinberg 1984).

4.1  Motivations to quit

Most addicts (including those who meet the clinical diagnostic standards laid out in the
Diagnostic and Statistical Manual of Mental Disorders) quit without seeking formal treat-
ment (Chapman and Mackenzie 2010; Heyman 2009). For example, since the Surgeon
General’s first report linking smoking to cancer, many millions of people quit smoking, and
90 percent of them did so without professional help or treatment (American Cancer Society
1986). Far more people were addicted at some point in life than are currently addicted,
which indicates that most people do stop their addictive behaviors (Compton et al. 2007).
For many people, a lifestyle change precedes quitting. For example, during the Vietnam
War, 20 percent of American soldiers in Vietnam met clinical standards for narcotics addic-
tion. But upon returning home, the addiction rate among Vietnam veterans fell to only
1 percent, mostly without formal treatment (Robins et al. 1975). Many aspects of military
292  

292 Self-regulation, controlled processes, and the treatment of addiction

life in Vietnam may have been conducive to heroin use: combat, stress, uncertainty, danger,
boredom, tolerant norms, and the easy availability of the drug. Most of these diminished or
vanished upon returning home to the United States. Most addicts apparently decided that
abusing narcotics would interfere with work, family, and the other trappings of civilian life.
Other lifestyle changes are less dramatic than returning from war but nonetheless are
associated with reduced addiction rates. Many people quit addiction in connection with
major pragmatic and symbolic life changes, including becoming a parent, getting mar-
ried, or having a brush with death (Russell and Davies 2009). Events like these catalyze
not only the decision to quit but also immediate attempts to do so, as opposed to plan-
ning and waiting to quit. Smokers who try to quit immediately after deciding to quit are
more likely to remain abstinent from smoking (for at least six months) than those who
delay their plan (West and Sohal 2006). In general, people abuse drugs more when they
do not have other good uses of their time, and people use drugs less when they have other
important uses of their time (Heyman 2009).
The central role of personal enjoyment suggests that voluntary, self-​serving behavior
is an important component of addiction. Although it may be that some addicts use only
from fear of withdrawal or the discomfort that accompanies abstinence (e.g. Kristeller
1994; Naqvi and Bechara 2010), enjoyment is in general a good predictor of contin-
ued use. On the whole, the more addicts like their drug, the less likely they are to quit.
People who expect alcohol to be especially pleasurable have low rates of success in quit-
ting (Brown 1985). Likewise, about half of smokers who know the habit is bad for them
reported continuing to smoke because of the enjoyment derived from it and the other half
because it relieved stress (Fidler and West 2009). As mentioned before, positive feelings
or avoidance of negative feelings play a large role in continued drug use. For example, one
study found that individuals who had the most positive emotional response to smoking
after a night of abstinence were most likely to relapse (Strong et al. 2011). This positive
feedback undermines the necessary initial desire to quit, the catalyst to use self-​regulation
to alter relevant thoughts and behaviors.

4.2  Beliefs about addiction

Because addiction is about choice more than physiological dependence, addicts’ beliefs
about addiction affect their decisions to continue abusing drugs (Heather et  al. 1982,
1983). Many addicts deny they have a problem with drugs because they believe they can
overcome any deleterious effects of drugs. The problem is that addicts do not always know
when their addiction becomes a problem for themselves and for other people. The impli-
cations of people’s lack of realization of their problem can be tragic. In 1995, Shannon
Hoon, lead singer of the talented band Blind Melon, sang about his drug use, claiming that
“I smoke a lot of grass, and I pop a lot of pills, but I’ve never done nothing that my spirit
couldn’t kill.”1 Ironically, Hoon died of an overdose not long after recording those lyrics.

1 The lyrics were slightly altered from the original version of “The Pusher” by Steppenwolf to emphasize
Hoon’s purported ability to withstand the effects of the drugs.
  293

Self-regulation in fighting against addiction 293

It may be difficult to persuade an addict to quit using drugs insofar as drugs are pleas-
urable. Observers may readily see the negative consequences, such as the expense or the
danger, but addicts may not adequately consider those costs. Another eminent musician,
Layne Staley (vocalist for Alice in Chains), much beloved by some of the present authors,
sang about his decision to use drugs, claiming a seemingly rational basis for his addic-
tion: “What’s my drug of choice? Well what have you got? I don’t go broke, and I do it a
lot.” Staley also died of a drug overdose. Ultimately, addiction can be a costly and even
fatal disorder, but many addicts presumably do not acknowledge the danger. Hence two
important first steps toward recovery are acknowledging the problem and deciding to
stop abusing drugs.
One of the most effective treatments for addiction is “motivational interviewing”
(Miller and Rollnick 1991). This is a very simple treatment that focuses on helping the
addict decide to quit and giving him resources to help guide him to a successful treat-
ment plan when he does decide to quit. Motivational interviewing can take as little as
fifteen minutes—​just enough time to discuss the pros and cons of continued drug use
and to help the addict reach his own conclusion that he must quit drugs. Obviously, a
brief interview contains no opportunity to reduce an addict’s drug tolerance or address
any physiological problems. Nonetheless, two meta-​analyses of the scientific studies on
motivational interviewing have shown it to be highly effective in improving successful
quit rates (Burke et  al. 2003; Rubak et  al. 2005). Motivational interviewing, like other
motivational treatments, is effective because it changes people’s minds about addiction
and helps them decide to fight against addiction.

5  Self-​regulation in fighting against addiction

Once an addict has decided to fight against addiction, he or she must self-​regulate to
succeed. In this section, we will apply the theories of self-​regulation described earlier to
explain how addicts succeed or fail in attempting to rein in their problematic drug use.
Once a person has been addicted to a substance and used it regularly for some time, the
person’s use may have become habitual. Therefore, self-​regulation in the fighting against
addiction phase is primarily used to override the bad habits developed in the regular
use phase.

5.1 Standards
To change one’s habits, one must first know which habits one wishes to change, and how.
Thus, as we mentioned before, self-​regulation begins with determining clear standards for
behavior. Many drug treatment programs establish very strict behavioral standards, most
commonly complete abstinence. There are costs and benefits to this approach. The main
benefit of such bright lines is that they are eminently clear and unambiguous standards—​
and clear standards are extremely helpful for successful self-​regulation. But there are also
two possible risks. First, a person may be less likely to decide to fight against addiction if
the standard appears unattainable or miserable. A person may decide it is not even worth
294  

294 Self-regulation, controlled processes, and the treatment of addiction

trying to reduce his drinking or drug use if one must maintain complete abstinence to
meet the standard. It may seem too painful, or he may simply enjoy drinking too much
to give it up entirely. Adopting less stringent, but still clear, standards may help avoid this
risk. For example, an alcoholic might be willing to adhere to a one-​drink limit, thereby
dramatically reducing his drinking without having to give it up entirely.
The second risk is that setting a very stringent standard means that one has failed as
soon as there is any violation, which can be severely demotivating (Marlatt and Gordon
1985). For instance, an addict may begin a program for quitting smoking in which com-
plete abstinence is required. At some point, she may have a lapse of judgment or willpower
and have a single cigarette. Once she has broken the rule that smoking is forbidden, she
may think to herself, “What the hell! I might as well enjoy myself tonight. I’ll quit again
tomorrow.” She may then proceed to smoke an entire pack. This phenomenon is known
informally to researchers as the “what the hell effect” and more formally as the abstinence
violation effect (Cochran and Tesser 1996; Herman and Polivy 1983). When people vio-
late a standard even by a little bit, they may decide to overindulge because they have failed
already, and (to them) a large failure is not qualitatively different from a small failure.
This is a major downside to setting a standard of complete abstinence and believing that a
single violation of that means you “fell off the wagon.” As violations, a sip and a three-​day
binge count the same. But, of course, a binge is much more damaging than a sip.

5.2  Motivations for quitting

For an addict to quit successfully in the long term, the addict usually must be internally
motivated to adhere to the standards. Treatments in which addicts are coerced into quit-
ting (an external motivation) may be successful in forcing the addict to quit but once they
are complete, addicts tend to begin using again because they enjoy using and they did not
make a conscious decision (internal motivation) to quit (Freedberg and Johnston 1978;
Gallant et al. 1973). When drinkers are persuaded to quit to avoid punishment, they tend
to quit until the threat of punishment is removed (Rosenberg and Liftik 1976). Celebrity
addicts cycle through numerous treatment programs in which they are sober for sev-
eral weeks, but shortly afterwards begin drinking again. In our framework, these relapses
occur because the treatment programs offer strong external motivations to not drink dur-
ing the program, but once the celebrity returns to the glamour of Hollywood, both her
internal and external motivations change, so it is rewarding and enjoyable to have a drink.
Externally imposed standards can be effective, but usually only in the short run. An
example of an externally imposed standard is contingency management therapy (e.g.
Petry et al. 2000). Contingency management entails rewarding people for good behavior
(abstinence, in most cases), or punishing people for bad behavior (i.e. using the abused
substance). In the most effective contingency management programs, addicts receive a
small reward (usually money) for each sequential day of abstinence. The rewards increase
daily, such that on the first day you may only receive $1, but on the second day you receive
$2, and so on. By the tenth day, you might really want to smoke a cigarette, but doing
so would not only cost you $10 today, it would also mean starting over at $1 tomorrow,
  295

Self-regulation in fighting against addiction 295

instead of the $11 you could earn by remaining sober. Of course, these programs cannot
last indefinitely: they cost a substantial amount of money, even though the per diem pay-
ment eventually plateaus. The trouble is that these programs are most effective only as
long as the payments continue. Many smokers who had successfully stopped smoking for
18 continuous months during a contingency management program began smoking again
once the payments stopped (Volpp et al. 2009).
Other external incentives have been tried, and are effective, though temporarily. For
example, one study offered abstinent smokers a trip to Hawaii (King et al. 1987). Again,
the incentive increased quit rates, but they were short-​lived. Some people do quit long-​
term for external reasons, as long as the long-​term reasons remain reasons. For example,
almost 100  percent of physicians who wrote a resignation letter in case of a potential
failed drug test remained abstinent for the duration of their career (Crowley 1984).
Internal motivation to quit appears to be essential for long-​term success. Addicts who
have decided for themselves that they want to quit and that they intend to quit are indeed
more likely to quit than those who merely perceive that society or their peers think
they ought to quit (Smit et  al. 2011). Again, most addicts quit on their own without
any formal counseling or therapy (Chapman and Mackenzie 2010). We believe this is
because the main reason for prolonged addiction is that, although addicts may want
to quit, they do not want to abstain from using on any particular occasion. This does
not mean that therapy is unhelpful or unnecessary. As we stated before, therapies like
motivational interviewing are highly effective at helping people recognize their problem
and motivating them to solve it earlier than they would have on their own (Miller and
Rollnick 1991).

5.3  Self-​efficacy
People are unlikely to attempt to quit if they believe it will not be possible to quit. To
quit successfully, an addict must believe it is possible to stop using. In technical terms,
the addict must possess self-​efficacy or, more precisely, the belief that it is possible to
quit. Several studies have shown that smokers with high self-​efficacy were more likely to
quit (and less likely to relapse) than smokers with low confidence in their ability to quit
(e.g. DiClemente 1981; Garvey et  al. 1992). Higher self-​efficacy is also associated with
greater rates of quitting alcohol (e.g. Solomon and Annis 1990). Presumably, self-​efficacy
is important because if you do not believe you are capable of quitting (either on your own
or with the help of a particular treatment program), you would have little reason to put in
effort. At the first sign of possible failure, rather than buckling down, you might give up.
However, self-​efficacy is also important for another reason.
Having high self-​efficacy may actually reduce the temptation to smoke (DiClemente
et al. 1985). People who believe they cannot quit (i.e. have low self-​efficacy) report being
very tempted to smoke. Among people who actually do quit, however, there was only a
minimal correlation between temptation strength and self-​efficacy. In other words, peo-
ple who did quit often did so despite being very tempted to use again. Therefore, having
strong temptations did not make it impossible to quit. But believing that it is impossible
296  

296 Self-regulation, controlled processes, and the treatment of addiction

to quit may actually make it harder to quit because it increases the temptation to give in
to one’s urges.

5.4 Willpower
Controlling one’s behavior is a form of self-​control and, according to the limited resource
theory, self-​control depends on willpower. Many sources of evidence confirm that will-
power is one potent key to stopping addictive behavior. For one thing, people who report
having higher overall trait self-​control are both less likely to become addicted and more
likely to successfully quit than people with low trait self-​control. People with high trait
levels of self-​control are more likely to successfully quit smoking than people with lower
self-​control (Brandon et al. 2003). Furthermore, low self-​control behaviors like failing to
use seatbelts are more common among habitual smokers than nonsmokers (Remington
et al. 1985). People whose willpower has been depleted also tend to drink more alcohol
than less-​depleted people (Christiansen et al. 2012).
People high in trait self-​control are not necessarily better at resisting a temptation that
is staring them in the face. Rather, people high in trait self-​control tend to avoid danger-
ous situations in which they might be tempted (Fishbach and Shah 2006). Avoiding the
need to use willpower is a better strategy than counting on the strength of one’s willpower
to hold up in the face of temptation.
Because willpower relies on a limited resource, addicts have a particularly difficult time
fighting addiction when they are also self-​regulating other aspects of their behavior. For
example, people who attempted to diet and quit smoking at the same time were less suc-
cessful at quitting smoking than people who focused their mental energy on quitting
smoking and did not attempt to restrict their diet (Patten and Martin, 1996). People who
lead particularly stressful and busy schedules are more likely to fail when attempting to
restrict their drinking than people with less busy and depleting schedules (Muraven et al.
2005). Other experiments suggest this is not simply a case of busy people drinking more.
People who were randomly assigned as part of an experiment to do a depleting task (as
opposed to an easy task that did not require self-​regulation) also drank more than they
intended to (Muraven et al. 2002). Although we are not aware of any experiments testing
the effects of depletion on the use of hard drugs (such studies may not have been done,
for ethical and practical reasons), the theory would suggest that addicts with depleted
willpower would also be more likely to fail to restrain themselves from using hard drugs.
Depletion may be of particular concern for smokers because smoking itself reduces the
depletion effect (Heckman et al. 2012). Smokers may therefore rely in part on cigarettes
to help them regulate themselves throughout their day and in order to reduce depletion
from other domains in their life. Therefore, when quitting smoking, even more so than
when quitting other drugs, it may be especially important to avoid stress and other deplet-
ing situations.
Recall that willpower can be trained, helping addicts to quit. The smokers who were
asked to do simple self-​control exercises (avoiding sweets or performing daily handgrip
exercises) for two weeks before quitting smoking more than doubled (from 12 percent
  297

Self-regulation for fighting against relapse 297

to 27 percent) their rate of success in quitting smoking for at least one month (Muraven
2010). Quitting smoking is clearly a difficult thing to do, but training willpower may be a
great new tool to help people quit.
Based on the limited resource theory of self-​regulation, we recommend addicts train
their willpower by performing willpower exercises before attempting to quit (see Muraven
2010). Timing is critical. Training willpower during quitting may deplete one’s willpower
and make it even harder to quit. Willpower exercises should be performed daily for at
least two weeks before quitting. We also recommend quitting during a period of reduced
stress, when willpower reserves will be at their highest. Quit during vacation, or a slow
period at work. Try to keep a positive mood, as this alleviates depletion (Tice et al. 2007).
If possible, addicts should think about how quitting affirms one of their own core val-
ues, as this also alleviates depletion (Schmeichel and Vohs 2009). When tempted, addicts
should think in terms of the long-​term consequences of quitting (they will be healthier,
more productive at work, have a better family life, etc.) rather than the immediate pleas-
ure they could have by drinking, smoking, or otherwise consuming their drug of choice.

6  Self-​regulation for fighting against relapse

When a person has successfully self-​regulated to reduce drug use to reasonable levels,
the next stage is avoiding relapse by maintaining reasonable levels of use (or complete
abstinence) for the duration of the person’s life. From the perspective of a physiological
addiction, this should become easier over time as the addict’s brain reconfigures itself to
adjust to life without its drug. From a self-​regulatory perspective of addictive behavior,
the risks of relapse persist for different reasons. Primarily, former addicts are at risk of
relapse because they are fully aware of the pleasurable effects of the drug. Even decades
after quitting alcohol, for instance, an addict might seek the pleasure of a drink because
he remembers how good alcohol used to make him feel. (In this perspective, use of the
drug is itself self-​regulation, because the person uses it to adjust his or her mood and
emotional state.) The more pleasurable the addict remembers the drug to have been, the
more likely he is to begin using again, even long after the quitting. For example, the more
positive people expect alcohol to be, the less likely they are to maintain sobriety after quit-
ting (Brown 1985). Pleasurable memories of drug use present risks that all former drug
users face.
As when fighting to quit drugs in the first place, maintaining sobriety requires self-​
regulation. Most of our advice for quitting also applies to avoiding relapse. Maintaining
clear behavioral standards remains key. Former addicts must maintain clear rules about
what behaviors they allow themselves and which behaviors violate their standards.
Efficacy remains important. People must believe they are capable of remaining sober. And
people must maintain willpower to avoid temptation.
The long-​term nature of fighting against relapse offers special challenges for preserv-
ing willpower. One should not think of fighting addiction as doing mental battle against
a single powerful urge. Maintaining sobriety is more like swatting away a particularly
persistent mosquito. It is not very difficult to swat the mosquito any one time, but one
298  

298 Self-regulation, controlled processes, and the treatment of addiction

must persist in swatting it every time, or risk relapse. Like a mosquito, addicts’ urges to
use come frequently, but do decrease in strength and frequency over time (Hofmann et al.
2012). It would be easy after several months or years of sobriety to become complacent
and assume further addictive behaviors are not a risk. Because willpower can become
depleted, however, whenever an opportunity to use arises, the addict must hope that his
willpower stores are adequate to resist. No studies have been done to test whether training
willpower is effective in the long term. It is likely that continuing to train one’s willpower
would maintain one’s willpower strength, but this idea should be accepted only tentatively.

6.1  Avoiding temptation

The best strategy for avoiding self-​control failure is to avoid tempting situations. For most
addicts this will involve self-​regulating to reorganize important parts of one’s life, includ-
ing reprograming habits and changing social groups to avoid other drug users.
Addicts who resist making the required life changes may find it very difficult to quit.
Imagine trying to abstain from drinking yet going to the bar with your old drinking bud-
dies every Friday. Every week you would be tempted to drink, and a single lapse in judg-
ment could end your sobriety. It would be much easier to avoid the bars altogether. Any
time you put yourself in a situation in which you habitually used drugs, your automatic
inclination would be to use, and you would have to control and restrain yourself. Making
lifestyle changes that remove an addict from temptation is a strong strategy, and research
shows that it is highly effective (Marlatt and Donovan 2005). People who change their
lifestyles so that they do not face frequent temptations are far more likely to remain absti-
nent. From our self-​regulatory perspective on addiction, this is because it is easier to
avoid situations that require willpower than to use willpower to resist temptations.

6.2  Avoiding relapse from a single failure

Many people assume that an addict who has failed to abstain one time has relapsed. It is
true that many people who, after quitting, smoke a single cigarette, or have a single drink,
tend to “fall off the wagon” and relapse into regular use (Garvey et al. 1992). However,
there is no evidence that this occurs because that single cigarette or drink takes over the
person’s motor cortex and forces him to smoke or drink in excess. More likely, the belief
that having a single cigarette or drink irreparably ruins one’s sobriety causes the addict to
think “what the hell” and go on a binge.
One clever study tested these two competing ideas to explain why smokers who smoke a
single cigarette tend to relapse (Juliano et al. 2006). A group of regular smokers abstained
for four days and then were randomly assigned to one of three conditions for the fifth
day of the study. On that fifth day, participants were either given five real cigarettes, five
denicotinized cigarettes, or no cigarettes to smoke. The researchers then offered monetary
rewards to participants if they did not smoke for an additional six days. The key result
was that participants who had smoked either type of cigarette on day five were more
likely to relapse within the next six days than participants who did not smoke on day
five. The study showed that the nicotine did not cause relapse, because participants who
  299

CONCLUSION 299

smoked cigarettes with the nicotine removed were just as likely to relapse as participants
who smoked cigarettes with nicotine. A more likely explanation is that participants who
believed they had already “ruined” their streak of abstinence were more likely to give up
and stop fighting against relapse.

7  Treatment suggestions
In this section we briefly summarize treatment suggestions that are consistent with or that
arise from the self-​regulation model of addictive behaviors. Self-​regulation requires moti-
vation, clear behavioral standards, and the use of willpower to adhere to those standards.
Continued successful self-​regulation entails avoiding relapse. Therapeutic techniques that
increase motivation to avoid substance abuse, provide clear behavioral standards, and
strengthen and maintain willpower to adhere to those standards will likely aid in con-
trolling addictive behaviors. Maintaining motivation, standards, and willpower are also
needed to avoid relapse, though a complementary strategy is to avoid situations in which
willpower will be needed. Realistically, it may be impossible to completely avoid situa-
tions in which one might be tempted. For example, an alcoholic probably cannot avoid
all grocery stores in which alcohol is sold. Therefore, we recommend a combination of
maintaining motivation, adhering to clear standards, strengthening willpower, and avoid-
ing situations in which willpower must be used to control addictive behaviors.

8 Conclusion
In this chapter, we have briefly summarized the self-​regulatory theory of addiction
(Baumeister and Vonasch, 2015; Baumeister et  al. 2015). This theory holds that self-​
regulatory processes are fundamental to understanding addictive behaviors. We sum-
marized the role of self-​regulation in quitting addiction and preventing relapse. We also
offered treatment suggestions for helping addicts quit (or reduce usage to reasonable lev-
els) and for helping them prevent relapse.

References
Ainslie, G. (2001). Breakdown of Will. Cambridge: Cambridge University Press.
American Cancer Society (1986). Cancer facts and figures. Available at: http://​www.cancer.org/​
research/​cancerfactsstatistics/​cancerfactsfigures2010/​index. Accessed 05/​005/​2016.
Baumeister, R.F. and Vohs, K.D. (2014). Strength model of self-​regulation as limited
resource: assessment, controversies, update. Manuscript submitted for publication.
Baumeister, R.F. and Vonasch, A.J. (2015). Uses of self-​regulation to facilitate and restrain addictive
behavior. Addictive Behaviors, 44, 3–​8.
Baumeister, R.F., Heatherton, T.F., and Tice, D.M. (1994). Losing Control: How and Why People Fail at
Self-​regulation. Waltham, MA: Academic Press.
Baumeister, R.F., Vohs, K.D., and Tice, D.M. (2007). The strength model of self-​control. Current
Directions in Psychological Science, 16, 351–​5.
Baumeister, R.F., Vonasch, A.J., and Vohs, K.D. (2015). Self-​regulation and addictive behaviors.
Manuscript in preparation.
300  

300 Self-regulation, controlled processes, and the treatment of addiction

Baumeister, R.F., Bratslavsky, E., Muraven, M., and Tice, D.M. (1998). Ego depletion: is the active self a
limited resource? Journal of Personality and Social Psychology, 74, 1252–​6.
Beedie, C.J. and Lane, A.M. (2012). The role of glucose in self-​control another look at the evidence and
an alternative conceptualization. Personality and Social Psychology Review, 16, 143–​53.
Brandon, T.H., Herzog, T.A., Juliano, L.M., Irvin, J.E., Lazev, A.B., and Simmons, V.N. (2003).
Pretreatment task persistence predicts smoking cessation outcome. Journal of Abnormal Psychology,
112, 448–​56.
Brown, S.A. (1985). Reinforcement expectancies and alcoholism treatment outcome after a one-​year
follow-​up. Journal of Studies on Alcohol and Drugs, 46, 304–​308.
Burke, B.L., Arkowitz, H., and Menchola, M. (2003). The efficacy of motivational interviewing: a meta-​
analysis of controlled clinical trials. Journal of Consulting and Clinical Psychology, 71, 843–​61.
Carver, C.S. and Scheier, M.F. (1981). Attention and Self-​regulation. New York: Springer-​Verlag.
Carver, C.S. and Scheier, M.F. (1982). Control theory: A useful conceptual framework for personality–​
social, clinical, and health psychology. Psychological Bulletin, 92, 111–​35.
Chapman, S. and MacKenzie, R. (2010). The global research neglect of unassisted smoking
cessation: causes and consequences. PLoS Medicine, 7, e1000216.
Christiansen, P., Cole, J.C., and Field, M. (2012). Ego depletion increases ad-​lib alcohol
consumption: Investigating cognitive mediators and moderators. Experimental and Clinical
Psychopharmacology, 20, 118–​28.
Cochran, W. and Tesser, A. (1996). The “what the hell” effect: some effects of goal proximity and goal
framing on performance. In: L.L. Martin and A. Tesser (eds), Striving and Feeling: Interactions
Among Goals, Affect, and Self-​Regulation. Hillsdale, NJ: Erlbaum, pp. 99–​120.
Compton, W.M., Thomas, Y.F., Stinson, F.S., and Grant, B.F. (2007). Prevalence, correlates, disability,
and comorbidity of DSM-​IV drug abuse and dependence in the United States: results from the
national epidemiologic survey on alcohol and related conditions. Archives of General Psychiatry, 64,
566–​76.
Crowley, T.J. (1984). Contingency contracting treatment of drug-​abusing physicians, nurses, and
dentists. NIDA Research Monographs, 46, 68–​83.
DiClemente, C.C. (1981). Self-​efficacy and smoking cessation maintenance: a preliminary report.
Cognitive Therapy and Research, 5, 175–​87.
DiClemente, C.C., Prochaska, J.O., and Gibertini, M. (1985). Self-​efficacy and the stages of self-​change
of smoking. Cognitive Therapy and Research, 9, 181–​200.
Fidler, J.A. and West, R. (2009). Self-​perceived smoking motives and their correlates in a general
population sample. Nicotine and Tobacco Research, 11, 1182–​8.
Fishbach, A. and Shah, J.Y. (2006). Self-​control in action: implicit dispositions toward goals and away
from temptations. Journal of Personality and Social Psychology, 90, 820–​32.
Freedberg, E.J. and Johnston, W.E. (1978). Effects of various sources of coercion on outcome of
treatment of alcoholism. Psychological Reports, 43, 1271–​78.
Gailliot, M.T. and Baumeister, R.F. (2007). The physiology of willpower: linking blood glucose to self-​
control. Personality and Social Psychology Review, 11, 303–​27.
Gailliot, M.T., Baumeister, R.F., DeWall, C.N., et al. (2007). Self-​control relies on glucose as a limited
energy source: willpower is more than a metaphor. Journal of Personality and Social Psychology, 92,
325–​36.
Gailliot, M.T., Hildebrandt, B., Eckel, L.A., and Baumeister, R.F. (2010). A theory of limited metabolic
energy and premenstrual syndrome symptoms: increased metabolic demands during the luteal
phase divert metabolic resources from and impair self-​control. Review of General Psychology, 14,
269–​82.
  301

Conclusion 301

Gallant, D.M., Bishop, M.P., Mouledoux, A., Faulkner, M.A., Brisolara, A., and Swanson, W.A. (1973).
The revolving-​door alcoholic: an impasse in the treatment of the chronic alcoholic. Archives of
General Psychiatry, 28, 633–​5.
Garvey, A.J., Bliss, R.E., Hitchcock, J.L., Heinold, J.W., and Rosner, B. (1992). Predictors of smoking
relapse among self-​quitters: a report from the Normative Aging Study. Addictive Behaviors, 17,
367–​77.
Heather, N. and Robertson, I. (1981). Controlled Drinking. London: Methuen.
Heather, N., Rollnick, S., and Winton, M. (1983). A comparison of objective and subjective measures
of alcohol dependence as predictors of relapse following treatment. British Journal of Clinical
Psychology, 22, 11–​17.
Heather, N., Winton, M. and Rollnick, S. (1982). An empirical test of “a cultural delusion of alcoholics.”
Psychological Reports, 50, 379–​82.
Heckman, B.W., Ditre, J.W., and Brandon, T.H. (2012). The restorative effects of smoking upon self-​
control resources: a negative reinforcement pathway. Journal of Abnormal Psychology, 121, 244–​49.
Herman, C.P. and Polivy, J. (1983). A boundary model for the regulation of eating. Psychiatric Annals,
13, 918–​27.
Heyman, G.M. (2009). Addiction: A Disorder of Choice. Cambridge, MA: Harvard University Press.
Hofmann, W., Vohs, K.D., and Baumeister, R.F. (2012). What people desire, feel conflicted about, and
try to resist in everyday life. Psychological Science, 23, 582–​88.
Hull, J.G. (1981). A self-​awareness model of the causes and effects of alcohol consumption. Journal of
Abnormal Psychology, 90, 586–​600.
Jellinek, E.M. (1960). The Disease Concept of Alcoholism. New Haven, CT: Hillhouse Press.
Juliano, L.M., Donny, E.C., Houtsmuller, E.J., and Stitzer, M.L. (2006). Experimental evidence for
a causal relationship between smoking lapse and relapse. Journal of Abnormal Psychology, 115,
166–​73.
King, A.C., Flora, J.A., Fortmann, S.P., and Taylor, C.B. (1987). Smokers’ challenge: immediate and
long-​term findings of a community smoking cessation contest. American Journal of Public Health,
77, 1340–​41.
Kristeller, J.L. (1994). Treatment of hard-​core, high-​risk smokers using FDA approved pharmaceutical
agents: An oral health team perspective. Health Values: The Journal of Health Behavior, Education &
Promotion, 18, 25–​32
Leshner, A.I. (1997). Addiction is a brain disease, and it matters. Science, 278, 45–​47.
Leshner, A.I. (1999). Science-​based views of drug addiction and its treatment. Journal of the American
Medical Association, 282, 1314–​16.
Liskow, B., Nickel, E., Tunley, N., Powell, B., and Penick, E.E. (1990). Alcoholics’ attitudes toward and
experiences with disulfiram. American Journal of Drug and Alcohol Abuse, 16, 147–​60.
Marks, J.L., Hair, C.S., Klock, S.C., Ginsburg, B.E., and Pomerleau, C.S. (1994). Effects of menstrual
phase on intake of nicotine, caffeine, and alcohol and nonprescribed drugs in women with late luteal
phase dysphoric disorder. Journal of Substance Abuse, 6, 235–​43.
Marlatt, G.A., and Donovan, D.M. (eds). (2005). Relapse Prevention: Maintenance Strategies in the
Treatment of Addictive Behaviors (2nd ed.). New York: Guilford Press.
Marlatt, G. and Gordon J. (eds) (1985). Relapse Prevention: Maintenance Strategies in the Treatment of
Addictive Behaviors. New York: Guilford.
Mello, N.K., Mendelson, J.H., and Palmieri, S.L. (1987). Cigarette smoking by women: interactions
with alcohol use. Psychopharmacology, 93, 8–​15.
Miller, W.R. and Rollnick, S. (1991). Motivational Interviewing: Preparing People to Change Addictive
Behavior. New York: Guilford Press.
302  

302 Self-regulation, controlled processes, and the treatment of addiction

Muraven, M. (2010). Practicing self-​control lowers the risk of smoking lapse. Psychology of Addictive
Behaviors, 24, 446–​52.
Muraven, M. and Baumeister, R.F. (2000). Self-​regulation and depletion of limited resources: Does self-​
control resemble a muscle?. Psychological Bulletin, 126, 247–​59.
Muraven, M. and Shmueli, D. (2006). The self-​control costs of fighting the temptation to drink.
Psychology of Addictive Behaviors, 20, 154–​60.
Muraven, M., Baumeister, R.F., and Tice, D.M. (1999). Longitudinal improvement of self-​regulation
through practice: Building self-​control strength through repeated exercise. Journal of Social
Psychology, 139, 446–​57.
Muraven, M., Collins, R.L., and Neinhaus, K. (2002). Self-​control and alcohol restraint: an initial
application of the self-​control strength model. Psychology of Addictive Behaviors, 16, 113–​20.
Muraven, M., Collins, R.L., Shiffman, S., and Paty, J.A. (2005). Daily fluctuations in self-​control
demands and alcohol intake. Psychology of Addictive Behaviors, 19, 140–​47.
Muraven, M., Shmueli, D., and Burkley, E. (2006). Conserving self-​control strength. Journal of
Personality and Social Psychology, 91, 524–​37.
Naqvi, N.H. and Bechara, A. (2010). The insula and drug addiction: an interoceptive view of pleasure,
urges, and decision-​making. Brain Structure and Function, 214, 435–​50.
Orford, J. (2001). Addiction as excessive appetite. Addiction, 96, 15–​31.
Patten, C.A. and Martin, J.E. (1996). Does nicotine withdrawal affect smoking cessation? Clinical and
theoretical issues. Annals of Behavioral Medicine, 18, 190–​200.
Petry, N.M., Martin, B., Cooney, J.L., and Kranzler, H.R. (2000). Give them prizes and they will
come: Contingency management for treatment of alcohol dependence. Journal of Consulting and
Clinical Psychology, 68, 250–​7.
Redish, A.D., Jensen, S., and Johnson, A. (2008). A unified framework for addiction: vulnerabilities in
the decision process. Behavioral and Brain Sciences, 31, 415–​37.
Remington, P.L., Forman, M.R., Gentry, E.M., Marks, J.S., Hogelin, G.C., and Trowbridge, F.L. (1985).
Current smoking trends in the United States: the 1981-​1983 behavioral risk factor surveys. Journal
of the American Medical Association, 253, 2975–​78.
Robins, L.N., Helzer, J.E., and Davis, D.H. (1975). Narcotic use in Southeast Asia and afterward: An
interview study of 898 Vietnam returnees. Archives of General Psychiatry, 32, 955–​61.
Robinson T.E, and Berridge K.C. (2003). Addiction. Annual Review of Psychology, 54, 25–​53.
Rosenberg, C.M. and Liftik, J. (1976). Use of coercion in the outpatient treatment of alcoholism.
Journal of Studies on Alcohol and Drugs, 37, 58–​65.
Rubak, S., Sandbæk, A., Lauritzen, T., and Christensen, B. (2005). Motivational interviewing: a
systematic review and meta-​analysis. British Journal of General Practice, 55, 305–​12.
Russell, C. and Davies, J.B. (2009). Empirical, logical and philosophical arguments against cigarette
smoking as a pharmacologically compelled act. Current Psychology, 28, 147–​68.
Schaler, J.A. (2000). Addiction Is a Choice. Chicago: Open Court Publishing.
Schmeichel, B.J. and Vohs, K. (2009). Self-​affirmation and self-​control: affirming core values
counteracts ego depletion. Journal of Personality and Social Psychology, 96, 770–​82.
Smit, E.S., Fidler, J.A., and West, R. (2011). The role of desire, duty and intention in predicting attempts
to quit smoking. Addiction, 106, 844–​51.
Snively, T.A., Ahijevych, K.L., Bernhard, L.A., and Wewers, M.E. (2000). Smoking behavior, dysphoric
states and the menstrual cycle: Results from single smoking sessions and the natural environment.
Psychoneuroendocrinology, 25, 677–​91.
Solomon, K.E. and Annis, H.M. (1990). Outcome and efficacy expectancy in the prediction of
post‐treatment drinking behaviour. British Journal of Addiction, 85, 659–​65.
  303

Conclusion 303

Strong, D.R., Leventhal, A.M., Evatt, D.P., et al. (2011). Positive reactions to tobacco predict relapse
after cessation. Journal of Abnormal Psychology, 120, 999–​1005.
Tice, D.M., Baumeister, R.F., Shmueli, D., and Muraven, M. (2007). Restoring the self: Positive affect
helps improve self-​regulation following ego depletion. Journal of Experimental Social Psychology, 43,
379–​84.
Volpp, K.G., Troxel, A.B., Pauly, M.V., et al. (2009). A randomized, controlled trial of financial
incentives for smoking cessation. New England Journal of Medicine, 360, 699–​709.
Vohs, K.D., Baumeister, R.F., Vonasch, A.J., Pocheptsova, A., and Dhar, R. (2014). Self control
resource depletion impairs active initiative and thus produces passivity. Unpublished manuscript.
Webb, P. (1981). Increased levels of energy exchange in women after ovulation. The Physiologist, 24, 43.
Webb, P. (1986). 24-​hour energy expenditure and the menstrual cycle. American Journal of Clinical
Nutrition, 44, 614–​9.
West, R., and Sohal, T. (2006). “Catastrophic” pathways to smoking cessation: findings from national
survey. British Medical Journal, 332, 458–​60.
Zinberg, N.E. (1984). Drugl Set and Setting: the Basis for Controlled Intoxicant Use. New Haven CT: Yale
University Press.
304  
  305

Section V

Implications for treatment,
prevention, and
public health
306  
  307

Chapter 17

The blindfold of addiction

Beth Burgess

Abstract
Addiction can be likened to a state of temporary blindness, wherein addicts
are prevented from seeing how to make better choices by the deep-seated
biological and psychological mechanisms that contribute to the formation
of an addictive disorder. This chapter aims to provide readers with a deep
appreciation of the physical and mental implications of an addiction, and
offers advice on how best to understand and help someone whose decision-
making is impaired by their disorder. The author argues that an addict’s choice
to recover can only be realized if the treatment they receive is tailored to their
problem and targets the susceptibilities, circumstances, cognitive and visceral
biases, and co-occurring mental health concerns that seem to combine to
cause addiction. The author offers specific tools and strategies to help remove
the blindfold of addiction so that addicts can clearly see the path to recovery
and become capable of making self-respecting, constructive choices.

1 Introduction
A blindfolded man often stumbles where others can tread safely. He can’t see what lies
before him; he is unaware of the dangers he may be facing; he is ignorant of the best way
to proceed. With his eyes shrouded in darkness, a man will feel vulnerable and lost. He
may accidentally choose a path that will take him to rocky terrains that tear at his feet, to
quicksands threatening to suck him down, into overgrown forests that confuse and bewil-
der him. If a false step is taken, he may be scared of turning back. But we can agree that,
despite his impairment, a blindfolded man still chooses which steps to take.
The best way of leading a blindfolded man out of difficulty is by guiding him gently,
encouraging him not to be scared, and helping him to navigate when he’s not sure
which way to go; it’s by explaining how the blindfold limits him, and teaching him to
use his other senses and skills to compensate for the lack of sight. A good guide will
thoroughly educate him about the layout of the terrain, so that he can proceed safely
on his own.
Addicts are essentially blindfolded individuals and are often unaware of how blind they
are. They make choices to take substances without fully understanding the biochemical
308  

308 The blindfold of addiction

implications it may have for them—​and without foreseeing the long-​term consequences.
They may struggle to find their strengths, feeling confused, helpless, and cut off from
solutions. They may try to forge ahead, not seeing or understanding why their choices
lead them into danger. Addicts add to their own difficulties because, once intoxicated,
they are less sure-​footed and less able to make sound judgements. And the further they go
down a dangerous path, the longer the way back will be.
But people who become addicts didn’t put the blindfold on themselves. They didn’t
choose the circumstances which predisposed them to problems and they are not respon-
sible for a lack of practical education about addiction. What they are ultimately responsi-
ble for is choosing to ask for help when they realize they are struggling. To do that, they
must be willing to accept that they’ve taken a wrong turn and have the courage to take a
radical new direction, despite their fears. They must be willing to become educated about
their limitations and to make new choices to enable them to recover.
An addict may choose to take the arm of someone who they believe knows the terri-
tory better than they do, whether that is an addiction worker, a therapist, or a sponsor.
Clearly, it is essential that their chosen guide themselves knows the correct path to take to
lead the addict out of their difficulties. If an addict doesn’t get effective support, they may
as well have no choice. The addict will be back to blind choice at best, and feel even more
hopeless at worst.
I hope this chapter will be an enlightening read for anyone who wants to understand
the territory of addiction more deeply, and the pragmatics of working with it, so they can
better serve their clientèle. In this text, I will investigate the choices addicts make, why
they make them, and how the choice-​making process is influenced. I will present a clear
way of understanding and classifying addiction, so that service providers can ensure they
are offering effective treatment choices to their clients. I will also suggest ways of combat-
ing impaired decision-​making and of helping clients to correct unhelpful thoughts and
feelings that lead to negative choices.
I start from the position that addicts make dysfunctional, but understandable, choices
in the course of becoming addicted and sustaining their addiction. When the addict asks
for help to recover, they have already initiated a powerful shift towards more construc-
tive choices. This chapter aims to help treatment providers build on that critical choice by
offering advice, tools, and treatments that enable a successful recovery.

2  Defining addiction
When referring to addiction, my language will mainly pertain to substance addictions,
purely for the sake of concision. Process addictions—​such as gambling—​may also be
considered by the reader, but will not be referred to as often. When defining addiction,
I mean a “sessional” pattern of compulsive behavior, which is part of a larger long-​term
pattern. For me, addiction is defined by how a person thinks and behaves during a series
of individual sessions of substance taking. People who always set out with the genuine
intention of drinking small amounts, and consistently drink until they finish the bottle or
  309

Defining addiction 309

pass out, are addicts—​even if they do not have alcohol between sessions. Someone who
genuinely intends to have one cookie and cannot help but finish the entire packet every
time is an addict—​even if they don’t have sugar on a regular basis.
It is what happens when the substance is actually consumed that defines addiction.
A severe addict will experience overwhelming urges to continue using a substance after
having had one “hit” of it, and a feeling of discomfort tantamount to bereavement if they
are stopped from having their substance again once the impact of the initial hit starts
to subside. An addict may experience cravings and obsessive thoughts about substances
outside of a session, but not all do, especially once they begin to work on recovery.
Abstinence does not necessarily mean someone is not an addict; it just means they have
ceased being in active addiction. If a fully-​fledged addict returns to taking their substance,
they will ultimately be unable to control their intake because they will experience the
same compulsion again.
Regular or heavy use of substances alone does not define addiction. There are people
who have substance misuse problems that are not of an addictive nature. People may,
for example, drink unsafe amounts of alcohol or take illegal drugs without developing
an addiction. Physical dependency—​when the body adapts to the constant presence of
a substance, causing withdrawal symptoms when it is discontinued—​does not necessar-
ily signify addiction either. Binge drinkers partying hard at the weekend might experi-
ence withdrawal shakes upon stopping without experiencing the compulsive feelings that
accompany addiction. As a side note, I view smoking cigarettes as a matter of dependence
more than addiction for most smokers. Humans only need to consume small amounts of
nicotine to encounter withdrawal upon stopping (Benowitz and Henningfield 1994), so
anyone who starts recreationally smoking several cigarettes a day is likely to find it diffi-
cult to quit. The true nicotine addicts are those who almost want to light another cigarette
before their current one is even finished.
Both addicts and non-​addicts may misuse substances temporarily to have fun, to
mask pain, to deal with stress, grief, boredom, or because of peer pressure. If these
behaviors never become accompanied by the compelling urge to continue a session,
then I wouldn’t define a person as addicted, even though the level of substance use may
be high. There are usually other clues that show a person is not (yet) an addict, such as a
steady pattern of substance misuse that doesn’t escalate. If someone can end a “session”
after a desired amount of a substance without the maddening urge for more, they aren’t
addicted.
Because addiction involves a certain neurological state in the body, it needs to be treated
differently from nonaddicted substance use, however heavy. Unfortunately, professionals
dealing with substance misuse problems often fail to differentiate between the two issues,
leading to serious complications for the client. If inappropriate or misleading advice is
given, because neurological conditions have not been taken into account, the client may
have unrealistic expectations and stumble down a false path. By giving clients advice tai-
lored to their specific problem, and which takes into account how their brain works, they
will be able to make informed choices for recovery.
310  

310 The blindfold of addiction

3  The addictive personality type

If substances were the definitive factors in addiction, then everyone who ever dab-
bled in drugs or drank alcohol would become addicted. Just 14 percent of people who
drink develop alcohol dependence (Anthony et al. 1994), and among those most man-
age to escape long-​term addiction (Grant et. al. 2004). Only a quarter of people who
use heroin become dependent, despite it being considered one of the most addictive and
dependence-​inducing drugs (Nutt et al. 2007). And while substance abuse has risen over
the years, dependence has not (Grant et. al. 2004). There must, then, be more to addiction
than the chemicals themselves.
Indeed, research shows that the “addictive personality” is a genuine phenomenon with
biological roots and behavioral implications. Unacknowledged by the “addiction = moral
failure” brigade, this susceptibility is already in place before an individual ever takes a drink
or drug. If your biological parent is addicted to drugs or alcohol, you are at a higher risk
of becoming addicted, irrespective of whether you’ve ever had contact with them or their
addiction (Kendler et al. 2012; Goodwin et al. 1974). A review by Schuckit (2009) showed
that a variety of heritable genes can contribute to an addictive personality, including those
causing a physically higher tolerance to substances and those which drive impulsivity.
The predisposition to impulsivity, sensation seeking, and addiction has been linked
with a lower number of D2 striatal dopamine receptors in the brain, which is apparent
before an addict ever touches a drug (Dalley et al. 2007). Individuals starting out with
fewer D2s are vulnerable to dopaminergic dysfunction when they ingest substances that
provoke a dopamine response (Buckholtz et al. 2010). Lower levels of D2 reception are
associated with exaggerated biological reactions to dopamine-​releasing substances like
food, alcohol, and drugs, or dopamine-​producing processes like gambling and sex. These
exaggerated responses result in cravings for more. At the same time, a low number of D2
receptors has been shown to cause a propensity to not learn from negative feedback and
to repeat mistakes (Klein et al. 2007). We can already see a pathway to addiction in this
combination of genetic effects alone.
These findings are only the start of what we are learning about biological predisposi-
tion. There appear to be inherited risk factors for alcoholism, for example, in the lev-
els and activity of certain neurochemicals in the body, including GABA and serotonin
(Ratsma et al. 2002). There may be many other genetic variants that make addiction more
likely to develop, some of which may only be activated by particular environments. In
short, there is evidence of a biological predisposition to addiction, which helps us begin
to understand why some people get addicted and others don’t.

4  Why humans try substances

Not everyone with an addictive predisposition will turn out to be an intractable addict.
Addiction cannot occur without initially taking a substance, and there are particular fac-
tors that make use or abuse of substances more likely. Variables in local, familial, and
peer environments can play a role in either protecting people with addictive personalities
  311

“Deciding” to become addicted 311

against substance misuse or making the risk higher (Hicks et al. 2013; Dick et al. 2013;
Kendler et al. 2012). Emotional and psychological factors seem to be critical predictors of
addiction, which we will see in more detail later in this chapter. For example, Dick et al.
(2013, pp. 2108–​17) found that it was not just children with highly sociable, sensation-​
seeking personalities who were drawn towards underage drinking; youngsters with con-
sistently low emotional stability and behavioral problems before the age of five were also
likely to be drinking by mid-​adolescence.
Some may say that addiction is a choice, regardless of the genetic, environmental, and
circumstantial factors that influence it, since people are rarely physically forced to take
their first hit. But humans are designed to make mistakes, especially when young. We
learn how to walk, talk, and interact with our environments precisely through trial and
error. People usually experiment with substances before they have developed full exec-
utive functioning, which only completes in the 20s (De Luca and Leventer 2008). The
immaturity of the brain makes young people both prone to risky behavior like substance
taking and vulnerable to the neurological impact of those substances. If starting out with
a predisposition to addiction, teenagers toying with substances might unwittingly be
making a mistake that will unfortunately prove fatal later on.
Human instincts also drive people to choose to do things which either cause pleasure or
avoid pain (Freud ([1895] 1954). When someone has their first experiences of intoxicat-
ing or pleasurable substances, they are likely to find that negative feelings, anxieties, and
stresses diminish and sociability increases. These effects will feel particularly beneficial to
those who either have emotional difficulties or enjoy new sensations. Little will the people
with a compromised dopamine system know that they have taken something which could
potentially harm them more than other people. Little will they suspect that, over time,
this good experience will become a horrible trap that they can’t escape from. All they may
feel is increased wellbeing and relief. All they may know is: “this feels better—​let’s do some
more of it.”
Without being born with a fully developed brain packed with skills and knowledge, it is
difficult to make rational, healthy choices from the beginning. No-​one knows or decides
their genetic make-​up or which other risk factors may have personal significance for
them. Addicts cannot, therefore, be condemned for initially experimenting with the same
chemicals that people less prone to addiction do with impunity. This argument becomes
increasingly compelling when you consider that some substances that predisposed people
become addicted to are legal and available on every high street. Indeed, drinking alcohol
is largely encouraged and endorsed by society and, in certain cultures, having a hangover
is even greeted with a wry grin and viewed as a badge of honor. Alcohol, caffeine, and
sugar are commonly used to socialize or to celebrate special occasions, and advertising
for these products is inescapable.

5  “Deciding” to become addicted

Despite predisposition, no-​one starts out as hopelessly addicted to alcohol or any other
substance. Addiction is a progressive disorder, meaning that the longer someone takes a
312  

312 The blindfold of addiction

particular “poison,” the less control they will have over their use. Anyone who uses sub-
stances heavily will find it takes a toll on them after some time, so people who develop
a severe addiction must be, consciously or unconsciously, choosing to suffer the conse-
quences of their use, including the risk of dependence.
Before the addictive process has firmly taken hold, an individual will have clues that
their substance use is damaging and unhealthy. They may experience worsening hango-
vers, come-​downs, or criticism from others. The substance-​misusing person is now at a
crossroads of choice. They can continue down the path to addiction or branch off into
early recovery. The individual knows which path is healthier and some choose to recover
early on. There are people who become teetotal without ever plumbing the desperate
depths of alcoholism and others who grow out of drug use. These people weigh up the
stakes and choose to respect their body and promote their self-​interest over the long term.
But others do the opposite; they deny they have a problem and continue on their not-​so-​
merry way. How, then, might this decision to submit to the risk of addiction come about?
Likewise, why would anyone ever risk doing drugs which are known to be particularly
addictive, such as crack or heroin?
Human beings are generally poor at forecasting what their feelings and behavior will
be in different circumstances (Wilson and Gilbert 2003). Moreover, they tend to under-
estimate the strength of visceral forces on them in the future. In the case of addiction,
people may underestimate the power of a craving and choose to take substances in the
miscalculated belief that they will not get hooked or will be able to resist the pull of
addiction later on (Loewenstein 1999). Lower D2 receptor availability leads to a dimin-
ished ability to stop and think through the consequences of one’s actions (Buckholtz
et al. 2010), showing that people predisposed to addiction may be at a higher risk of
miscalculation. Add to this the failure to learn from negative feedback (Klein et  al.
2007), and we can see how a biological predisposition to addiction exerts its influence
on maladaptive choices.
Another cognitive problem that addicts suffer from is the inability to wait for larger
rewards, instead favoring smaller, immediate ones. The process is called delay discount-
ing, and all human beings are somewhat subject to this motivational bias since we were
originally designed to survive and reproduce quickly, rather than to consider our future
happiness. But addicts seem to fall into this trap in the most extreme way, placing greater
worth on the short-​term benefits of abusing substances versus the long-​term benefits of
not doing so. One particular study (Mitchell et al. 2005) showed that abstinent alcoholics
chose larger, delayed rewards much less often than non-​alcoholic controls. The experi-
ment tested volunteers with between one day and 17 years of abstinence and found that a
longer time abstinent did not necessarily reduce delay discounting. Significantly, among
those tested, the more biological family members they had with alcohol problems, the
more they themselves had difficulty waiting for rewards. This suggests that a propensity
for excessive delay discounting is heritable and occurs even in abstinence. There are also
other factors that mean choice making is impaired in addicts (see Holton and Berridge,
Chapter 9, this volume).
  313

Emotional “choosing” 313

6  Emotional “choosing”
Emotional and psychological factors also affect people’s choices. Someone who decides to
go on destroying their body and their life with a substance has a very different mindset
from someone who acknowledges the issue and abstains.
I would argue that the person who chooses to continue down the path to addiction
is partially making that choice because they suffer from chronically low self-​esteem or
other mental health difficulties. Indeed, individuals who are at higher risk of developing
addictions share certain genetic traits with those who are at a higher risk of develop-
ing certain psychological disorders (Schuckit 2009). Substance misuse, for these people,
is less about seeking pleasure than it is about—​consciously or unconsciously—​trying to
avoid extremely painful feelings. The level of pain these individuals suffer from negative
consequences to health, relationships, finances, or anything else, does not match up to the
level of pain they already feel deep inside—​and that they experience in all too stark reality
when not using substances.
In my experience as a therapist, mental health problems are a result of unhappy child-
hood experiences more often than they are purely a direct result of addiction. Something
quite negative usually precipitates any really hardcore use of substances. I have never had
a client for whom everything bar the addiction was rosy. My clients come to me with anxi-
ety disorders, feelings of worthlessness, and all manner of unfortunate emotional states,
which have mostly been apparent before the onset of substance misuse. If people choose
to self-​medicate with substances (Khantzian 1990), we can understand their solution as
a product of damaged self-​esteem. People who saw themselves as more worthy might
go to a health professional to deal with their emotional problems instead of deciding to
self-​medicate by using alcohol and drugs. Self-​destructive choices reflect the turmoil that
exists within the individual.
Mental health disorders and childhood difficulties muddy the waters when it comes to
the matter of addiction and choice because these life experiences can influence biology—​
and the D2 receptors are implicated yet again. Studies show that there is a lower avail-
ability of the D2 receptors in people experiencing obsessive compulsive disorder, social
phobia (Schneier et al. 2008), and depression. One study found that depression and anxi-
ety were pronounced in very young children with genetic coding for fewer D2 receptors
(Hayden et al. 2010), showing that these patterns can precede substance use. These pre-
schoolers also received less supportive parenting and expressed more negative emotions
when interacting with their parents. The research indicated that these factors were bound
up in a vicious cycle. Where the problem starts is difficult to unpick, since parents will
both be subject to the genes that influence their own personality and behavior and will
have passed some of them on to the child.
To my mind, difficult early life experiences are the final piece of the puzzle when it
comes to severe addiction. Without a pervasive, implicit “unhappiness” caused by early
trauma and stress, people prone to addiction might be able to stop their self-​destruction
before it becomes entrenched (see Burgess, Chapter  22, this volume, for much more
314  

314 The blindfold of addiction

on the links between mental health, trauma, and addiction). Full addiction is formed
from a mix of susceptible genes, early stress or trauma, and heavy substance use. The
substance use itself often becomes part of the picture precisely because of the other
elements.

7  When choice comes too late

At some stage, even for those who are self-​medicating, the pain of using substances will
outweigh the mental pain of not doing so. As a result of substance misuse, individu-
als may suffer severe consequences, from social exile to unemployment, illness, poverty,
and homelessness. But even when events become dire, some people still don’t stop using
addictive substances. Sadly, by the time the most terrible long-​term consequences are
acknowledged and experienced, it may be too late, biologically, to turn back from addic-
tion because use of addictive substances changes the brain.
The human body will always try to self-​right its functions (Sterling and Eyer 1988). The
more a substance which causes chemical changes in the body is taken, the more the brain
fights back. Eventually, the body decreases the availability of D2 receptors, meaning more
of a substance is needed just to keep the body functioning. But it’s not just the difficulty of
initially withdrawing from a substance that drives addicts to continue to use. Due to the
continual self-​righting required when someone’s body relies on substances, the brain’s natu-
ral reward functions become chronically impaired (Nestler et al. 2001) and may take on an
abnormal new set point for reward (Koob and Le Moal 2001). Normal dopamine-​tickling
experiences like eating, socializing, and sex, no longer “do it” for the addicted individual.
In addition, the first taste of something that unexpectedly fills an individual with well-
being will be recorded in the brain as something of high value (Robinson and Berridge
1993), making them crave that experience and seek it out repeatedly. This phenomenon
can happen even when a substance is unhealthy and can be an unconscious drive. Anyone
who still yearns for low-​quality comfort food that they ate as a kid will be familiar with this
experience. For someone with a compromised dopamine system, drugs or alcohol may
be indelibly marked as deserving of a special place in the psyche because of their earlier
surprisingly pleasurable effects—​irrespective of an addict’s desperate desire to stop using
later on. The neural system also becomes so hypersensitive to mental triggers associated
with their substance that addicts can unconsciously crave something they have come to
consciously despise whenever their environment reminds them of their particular poison.
Symptomatically, all this leaves addicts in a chronic state of stress when they don’t have
their substance, and life may seem unbearable without it. When cravings surface, an addict
will feel a compulsion to kill the cravings as quickly as possible, setting aside rational
thought while a biological need must be served. And it is, by now, a biological need, as the
brain has an altered set point and a permanent unconscious marker making the addict
feel they need to use. What this looks like in practice is people taking substances despite
all the horrors they cause. The decision to stop has now become very difficult indeed, as
both cravings and withdrawal symptoms punish an addict if they make that choice.
  315

IDENTIFYING SPECIFIC TYPES OF SUBSTANCE MISUSE 315

8  The choice to quit

Addicts can choose to cease addictive behavior, no matter how difficult it is to go against
the neural grain. Alcoholics do turn up in A&E in full alcohol withdrawal, having finally
chosen to put the bottle down. Drug addicts do pack themselves off to detox (or cold turkey
it on the sofa) hoping to leave their old life forever. No matter how low someone’s self-​
esteem or how serious their mental health issues, at some point an addict will not be able
to bear their addiction any longer. Whether mentally, physically, financially, or socially, the
pain will be too much for them to handle. This is the classic case of “reaching rock bottom.”
The addict may have been at the crossroads of choice many times before, and either
refused to ask for directions or been met by an ineffective guide. But at some point, they
are likely to sink down at the crossroads and start pleading for help to recover, looking
anywhere and everywhere for someone to point them to the right path. It may not be easy
to swallow the heady mixture of shame and pride to do so, but it becomes possible at a
critical point.
It is important to consider how the decision to ask for help might come about, since
it may mark the turning point in someone’s life. It cannot be pure desperation. After all,
many addicts have dragged themselves along rockbottom for some time. The choice to ask
for help seems to be marked by finally having some insight into the problem: for example,
an addict starting to realize some of the specific patterns to which they are falling prey.
An individual need not understand the scientific factors to know that their life and brain
are out of control. At this stage, the addict is connecting the dots, perhaps realizing that
they are no longer using substances to self-​medicate their old problems as much as they
are using substances to block out the problems that substances themselves are causing.
The light of insight, or a change of perspective, could be a powerful predictor of asking for
help. Research has shown that higher levels of insight into addiction help addicts com-
mit to change, and that those with greater insight enjoy a higher likelihood of prolonged
abstinence (Jung et al. 2011).

9  Identifying specific types of substance misuse

Deciding to ask for help is the most crucial choice an addict can make. This is a criti-
cal skylight for change—​and whenever that choice is made, there must be effective
help there to meet it. Indeed, how the cry for help is met can monumentally help or
hinder the recovery process. No addict must be condemned to a life of suffering and
the same vicious cycle forever. People can recover, but they must be given the right
tools and information to enable them to make positive, effective choices. They need a
guide that gently nudges, encourages, and enlightens them so they can find the road
to recovery. This all starts with a treatment provider clearly understanding, and help-
ing a substance misuser to understand, what sort of problem they have and where it
may come from.
There are different treatment implications when it comes to people who are substance
misusers and those who are actually addicted. It may also be necessary to treat “habitual”
316  

316 The blindfold of addiction

addicts (like many smokers) differently from those with congenitally compromised neu-
rochemical or receptor levels which influence their behavior. These differentiations will
help service providers avoid methods which might be contraindicated in addiction and
will help to ensure treatment is more effective. If an addict has started out with a predis-
position due to their dopamine receptors, for example, they will be unable to ever take
their substance safely again. Their brain will find it impossible to normalize their dopa-
mine responses. They may also need interventions to help them with impulse control, for
example, more so than other candidates.
Obviously, when a client comes into a service, there is unlikely to be a handy brain scan-
ner lying around to indicate just how addicted their neurobiology is. But we can uncover
the pattern of addiction from a client’s history, behaviors, and experiences. Whenever
I first meet a client who displays a problem with substances, I ask them two questions
to find out whether they are most likely addicted by nature, addicted by habit, or if the
problem is temporary misuse:
(A) Is there a history of addictive or compulsive behaviors in your biological family?
I explain that this does not have to mean alcoholism or drug addiction; the addictive nature
may show up as a problem controlling sugar, for example, or in other compulsive behaviors.
(B Once you’ve had one “hit,” do you find yourself (more often than not) unable to stop
a session of substance taking?
I sometimes use examples here to illustrate e.g. Can you leave an unfinished drink on
the table or does that feel very difficult? If you can only have one hit, do you feel agitated,
irritable, or bereft? Have you ever missed crucial events (job interviews, court cases, etc.),
or arrived hungover or intoxicated due to being unable to stop using substances?
The key below shows how a client’s answers may help shed light on the type of problem
they are likely to have:
(A) YES (B) YES
If an individual comes from a biological line of addicts, and already has trouble
controlling their own behavior, they most likely have a congenital predisposition to
becoming addicted and are, in fact, an addict.
(A) YES (B) NO
If a client can genuinely stop a session of substance use, they aren’t a full-​blown
addict; but family members with addictive tendencies always raise a red flag. Since
addiction is progressive, someone with a likely biological predisposition who
answers “no” to (B) may well say “yes” to the same question later down the line.
(A) NO (B) YES
This person is likely to have a habitual addiction or dependence without an addictive
personality.
(A) NO (B) NO
With neither family members displaying addictive behaviors nor an overwhelming
compulsion to continue using a substance after the first hit, the client has a tempo-
rary substance misuse problem, not an addiction.
  317

PSYCHO-​EDUCATION 317

10  Providing effective help and treatment

Services must be very careful not to set anyone up to fail, whether they are an addict or
someone with a different substance misuse problem. Clients may have their chance to
recover completely jeopardized if they are presented with false choices. Guilt, shame, and
low self-​esteem perpetuate addiction and substance misuse. Not only will inappropriate
treatments not work, they are likely to add to that person’s sense of shame and drive them
right back to their bottle, pipe, or syringe.
The advice offered to the client must fit their substance misuse profile. For example,
giving a “drink diary” to an addict is pointless and often harms their self-​esteem, because
it implies that sessional control is achievable with enough effort. Addicts will either put
a “0” in their daily column or an excessive amount, because once the first hit has been
taken, control is unachievable. Non-​addicted drinkers, on the other hand, may find such
an exercise helpful. Advising a person who is not an addict that they can never drink again
may be counter-​productive and make them feel hopeless. But telling an addict that they
could control their use if only they tried is even worse, because it is nigh-​on impossible
for them to do so. Their biochemistry makes it so compelling for them to continue tak-
ing a substance once they’ve started that it would be setting them up for frustration and
failure to insist they try to moderate their use. Addicts must be advised that abstinence is
the way forward for them. Not only should they be told this, but addiction workers should
also explain why. It is not enough to say that substances are damaging their health and
wellbeing—​they know that. It is not enough to tell them they should stop drinking—​they
may take this to mean they should cut down or moderate. Explain to them that they need
to be abstinent because, scientifically, it’s the only viable way to recover.
Knowing how various difficulties can lead to and perpetuate addiction, it is possible
to formulate precise treatments to help clients with their handicaps. Effective addiction
treatment should incorporate psycho-​education, training to curb impulsivity, manage
stress, and counter cognitive and visceral biases that contribute to poor choices, plus
facilities to heal mental health and self-​esteem problems. By offering treatment which
deals with all contributing factors, addicts can finally be empowered to be able to realisti-
cally choose recovery. What follows are suggested tools and best practices to help free an
addict of their blindfold effectively.

11  Psycho-​education
People who do not understand addiction may feel confounded, frustrated, and angry
about the behavior of addicts. They see an act that they can do with impunity and can’t
understand how another person can lose control doing the same thing. But addicts rarely
understand addiction either and often feel increasingly baffled and hopeless about their
inability to stop drinking or taking drugs. An addict may curse their lack of willpower or
their poor decisions without realizing that their own addiction-​stricken brain is foiling
them and tricking them into making self-​defeating choices. It’s usually not an addict’s
willpower that needs work so much as their awareness. Educating addicts about addic-
tion allows them to peek out from under the blindfold and see how the land really lies.
318  

318 The blindfold of addiction

Without insight into their condition, addiction remains an invisible enemy and an addict
won’t know how to arm themselves against it.
When the brain is addicted, it will do everything it can to make an addict continue
taking substances, including besieging the addict with extreme cravings and irrational
thoughts about using. It is simply the brain trying to return to what it now thinks is a
normal state (i.e. being awash with substances) (Sterling and Eyer 1988). In an addicted
brain, operant and neural conditioning has been established, meaning that many situa-
tions will create an urge to use a substance (See 2002; Robinson and Berridge 1993). In
short, an addict’s brain is so wired to expect its usual poison that it will become extremely
perturbed if that substance is not forthcoming when and where it normally would be.
This is where the brain steps in and starts trying to persuade and manipulate the stub-
bornly abstinent addict into using one more time. It will transmit compelling thoughts
about drinking or taking drugs, such as: “Just one hit will be OK,” “No-​one’s here—​I could
probably get away with it,” or “I’ve managed a week without drinking—​I deserve a reward.
A drink would be nice.” Educating an addict about their brain’s subversive agenda can
stop them unwittingly falling prey to irrational thoughts that might otherwise sabotage
their recovery efforts.
The more an addict understands how addiction is sustained and the “sessional”
nature of it, the more empowered they will be to break the vicious cycle they are in.
Entrenched addicts’ cravings demand relief as strongly as extreme hunger, thirst, or
pain. But the craving for their substance only becomes completely insurmountable after
the first hit of a session is taken. It is indescribably hard to stop after that initial taste; an
addict’s neurobiology will kick and scream against stopping, no matter what that person
consciously wants. For a crack user, cravings to take more may occur only seconds or
minutes into starting a session, an alcoholic’s may start minutes to hours after initial
ingestion, and a heroin users might begin several hours after the first hit, as the buzz
starts to wear off and withdrawal looms. An addict’s potential ability to reject their drug
becomes operable and realistic only when there is none of that dopaminergic substance
in their system. Out-​of-​session cravings feel unpleasant but they are not unbeatable. If
an addict caves in to the craving to start a session, they may feel as if they are satisfy-
ing it; in reality, they are just replacing that craving with an unbearable, overwhelming
craving to take more—​which is why they can’t stop after the first hit. Continually taking
substances maintains the cravings forever. Abstaining is the only way out of the trap;
never taking the first hit breaks the cycle of addiction. It is essential to teach clients
methods for dealing with their cravings when clean and sober, as well as to advise them
that these will generally become less urgent and frequent after a sustained period of
abstinence and recovery work.
Knowing that addicts often fall for short-​term gains to the detriment of their long-​term
aims, helping addicts overcome this propensity might help them deal with cravings, too.
Delay discounting in addicts can be improved by training their working memory (Bickel
et al. 2010). Since delay discounting perpetuates addiction, using memory games to intro-
duce psycho-​education topics might be an effective way of killing two birds with one stone.
  319

Building self-esteem 319

Addiction science also helps us forewarn addicts about cross-​addiction. It is only fair to
advise addicts that they may also need to steer clear of certain things to which they are not
currently addicted. Plenty of alcoholics give up alcohol only to find themselves floored by
a sugar or gambling addiction. Cross-​addiction may be explained by a protein that accu-
mulates in the brain, rendering addicts more prone to other addictions. DeltaFosB is pro-
duced when there is too much dopamine in the brain; so, when dopaminergic substances
are used repeatedly, it reaches a critical mass. The protein appears to flip a genetic switch,
sensitizing the addict to dopamine-​producing experiences and making them crave them
(Nestler et al. 2001). It is likely that this protein and others, such as BDNF (Wang et al.
2013), may be responsible for more permanent changes in an addict’s neurology, causing
long-​term cravings and addictive behaviors. Quite simply, the message is that if you’ve
ever been addicted to substances, it’s wise to avoid most intoxicants. This sort of education
can empower addicts to make informed choices, so they can avoid falling into another
addictive trap.

12  Building self-​esteem

Whenever someone contacts me for help, I presume that they start from a place of con-
fusion, not a place of weakness. Addicts are not weak people, and this is precisely what
I tell them. If low self-​esteem has made someone choose self-​destructive behaviors like
addiction, then building up an addict’s feelings of self-​worth is a good way to cement the
path to recovery. How the client is treated is all-​important in this process. Respecting
them, acknowledging their best qualities, showing genuine interest in their progress,
and delight in their successes, can all help the client develop better self-​esteem. Using
positive language and techniques from solution-​focused therapy (de Shazer 1985) can
guide addicts to recognize their strengths and see a brighter future. I encourage clients
to take on an attitude of self-​compassion, curiosity, and objectivity, and deter them from
counter-​productive self-​shaming. The emphasis should always be on moving forward
and finding answers.
Some addicts believe that it is impossible to break the cycle and live a happy life with-
out their substance. It is a vital task for all those who work with addicted clients to help
them see how real and rewarding recovery is. I think it’s a shame that addiction workers
are sometimes told not to disclose their own recovery to clients. It’s a waste of that very
powerful tool, hope. Sometimes hope is all that keeps an addict trying to beat their addic-
tion. Role models of recovery can be highly motivating and inspirational. If the addict can
first start to see that recovery is possible and desirable, they’ll already feel they have more
choice than they ever did before.
Early on in the process of recovery, there may still be trips and stumbles, but the addict
who has genuinely committed to recovery will have the firm resolve to never give up.
It is vital to avoid the shame that drives people back into addictions, so mistakes must
be framed as part of the learning process. Something I  have found very effective with
my own clients is drawing a mental line when they have made the decision to recover.
Everything before the line was as a result of the client’s misunderstanding about addiction.
320  

320 The blindfold of addiction

They couldn’t help listening to, and falling for, their brain’s unhelpful neural demands.
Everything after the line is a learning zone, where mistakes may still be made, but will be
learned from rather than repeated. The line represents commitment, choice, and crossing
over from shame into integrity, which further builds self-​esteem.

13  Countering cognitive and visceral biases

Practicing forward thinking can help counter the myopia and impulsivity from which
addicts often suffer. Asking clients to imagine life in five years’ time is a positive way to
start any change work, as it forces them to orientate themselves towards a longer-​term
view. Get them to make a visual image of what their existence might be like if they were
to continue their addiction and then contrast that with a mental picture of them free of
their addiction. Ask your client to really see, feel, and hear how things are different down
each path, including what they will lose and gain, what they might be thinking and doing,
the things others may say to them, and how it will feel inside. Making a vivid picture of
the paths in front of them helps clients to play the long game and realize the benefits of
choosing recovery.
The “Play the Tape Forward” exercise in my book The Happy Addict (Burgess 2013,
pp. 141‒43) may also be helpful; it neatly combines lessons about immediate cause and
effect as well as longer-​term chain reactions. It’s useful to do when cravings occur and
essentially involves listing the chain of consequences that might arise from choosing
either to use or stay abstinent. Clients are invited to stop and contemplate the impact of
their choice on the coming days, weeks, months, and years. Encourage them to consider
not only the potential physical, legal, financial, or emotional impact, but also to imagine
how they will feel about themselves, the opportunities that might open or close as a result
of their choice, and the fact that their decision can either add strength to the cycle of crav-
ing or weaken it.

14  Stress management and cognitive-​behavioral change

When beginning recovery work, it is essential that clients have a realistic crisis plan.
Developing ways of managing stress and discomfort without taking substances is vital.
Asking clients to mentally rehearse strategies for dealing with hard-​to-​handle emotions
or cravings can help them make effective choices when in difficulty. One of the most
valuable things you can tell someone suffering from addiction is that an emotion is not
an instruction to act. I have said this very thing to many addicts and—​after an initial look
of bewilderment—​it is as if a light has suddenly shone through the blindfold. In many
ways, addictive behavior becomes a visceral knee-​jerk response. The addict is so used
to responding to tough mental states by using substances that they don’t even under-
stand there is a different choice they can make. Allowing someone to see that they do
have choice in how they behave—​even when they feel awful—​can be very enlightening
for them.
  321

Mental health 321

Correcting unhelpful thinking patterns can permanently decrease the occurrence of

negative emotions. If an addiction has taken someone to awful places where they have
been treated badly, they may have picked up very negative ways of looking at the world.
Some addicts might have seen the darker side of life for a long time or developed nega-
tive thinking styles based on childhood experiences. It may take time to change old
thinking patterns since all past thinking and behavior is wired in the brain. If a certain
thought pattern is repeated often enough, this process ends up becoming extremely
efficient and a specific neural pathway is formed (See 2002). If an addict has learned
to think negatively, this response may feel just as compelling and automatic to them as
using a substance.
Dialectical behavioral therapy (DBT), developed by Linehan (1993), can teach addicts
to choose more helpful ways of thinking and behaving. DBT helps clients build toler-
ance to difficult emotions, to control impulses, deal with stress, make behavioral changes,
and to be more balanced in their thoughts. DBT’s advantage over the more well-​known
cognitive-​behaviour therapy is that it is a more gentle and validating therapy, which
means it is suitable for people suffering from low self-​esteem and, in my experience,
those struggling with ambivalence. DBT also draws on Buddhist concepts that address
issues that addicts commonly wrestle with. Useful Buddhist teachings for addicts include
acceptance, non-​attachment, self-​compassion, non-​judgment, equanimity, and mindful-
ness. A review of mindfulness studies showed that it can alleviate problems with alcohol,
cocaine, amphetamines, marijuana, nicotine, and opiates. There is evidence that practic-
ing mindfulness can reduce cravings and help addicts develop greater awareness of the
impact of their actions and choices (Chiesa and Serretti 2014).

15 Mental health
The issue of substance misuse as self-​medication highlights why effective mental
health treatments are so important—​because there are people who arrive at the cross-
roads of choice and do try to choose the healthier path, only to be let down. Some
people try to seek help for mental health issues or substance-​related problems and
they are not helped effectively. They may then wind up going further down that path
of addiction.
I encounter many clients who have been turned away from mental health services
because they are still using substances. Substances are blamed for blurring the picture
and addicts are told they must stop drinking or taking drugs before anyone can take
their emotional symptoms seriously or give them any help for their issues. And vice
versa—​mental health patients are not welcomed with open arms at substance misuse
services because they sometimes present behavioral challenges. While mental ill health
can be worsened by alcohol and drugs, my clients all tell similar stories. The distressing
feelings came before they ever picked up a drink or a drug. If we can consider the fact
that these two issues are intertwined, rather than separate, more effective treatments can
be devised.
322  

322 The blindfold of addiction

Whenever I work with an addict, a vital part of the treatment is releasing the pain
of any traumatic experiences that have led to low self-​esteem, painful feelings, dis-
ordered thinking, and unhelpful behavior—​in summary, mental health symptoms.
Helping an addict to heal their emotional pain is just as important as dealing with the
actual mechanism of addiction, because it removes that need to self-​medicate. When
an addict no longer has unbearable feelings from the past haunting them, they are
much more likely to commit to sobriety and recovery. They no longer want to press
the self-​destruct button.
In my experience, everything from addiction to anxiety to eating disorders to
depression can be resolved or vastly improved by dealing with past traumas. I recom-
mend therapies which actively help people to process past traumas healthily over those
which focus on continuously talking about the problem. If recovery is about choice,
and choice is brought about by empowerment, addicts should be encouraged to be
active in ending their own difficulties rather than reflecting on the pain over and over
again. Helping someone to free themselves from the perpetual loop of pain allows
them to not only see that things can be different, but that they have the power to make
them so.

16 Conclusion
Addiction is a disorder that occurs because faulty, if understandable, choices are made
against a backdrop of predisposing factors. Since these factors are beyond the potential
addict’s control, it makes sense for society to help counter them as much as possible. Clear
addiction education should be provided in schools to help young people understand how
addiction develops and what it looks and feels like. Children should be taught how to
discern and value longer-​term rewards and should be supported with their mental and
emotional health as early as possible.
Insight prompts people to ask for help, especially when they are desperate. This natu-
rally makes a case for promoting more practical addiction education in settings where
addicts may find themselves at rockbottom. Hospitals, prisons, police stations, and
drop-​in units should train workers to truly understand addiction, know how to differ-
entiate it from substance misuse, and be able to explain it clearly to suspected sufferers.
Recovery is the choice to right what has gone wrong—​but, with all the will in the world,
the blindfolded addict isn’t likely to be able to change without assistance. When an addict
does ask for help, they’re asking to be empowered and enlightened; so that instead of
struggling to recover, battling against all their biological forces, their cognitive limita-
tions, and their painful feelings, they can more happily and easily make the right choices.
For an addict to be able to fully capitalize on their decision to get help and recover, they
need support that makes their choice realizable and fruitful. This should include precise
classification of their problem, appropriate education, and holistic treatment that covers
all their issues. Only then can they hope to circumvent the blindfold; and only then can
they choose, and keep choosing, recovery.
  323

Conclusion 323

References
Anthony, J.C., Warner, L.A., and Kessler, R.C. (1994). Comparative epidemiology of dependence
on tobacco, alcohol, controlled substances, and inhalants: basic findings from the National
Comorbidity Survey. Experimental and Clinical Psychopharmacology, 2, 244‒68.
Benowitz, N.L. and Henningfield, J.E. (1994). Establishing a nicotine threshold for addiction—​the
implications for tobacco regulation. New England Journal of Medicine, 331, 123–​25.
Bickel, W.K., Yi, R., Landes, R.D., Hill, P.F., and Baxter, C. (2011). Remember the future: working
memory training decreases delay discounting among stimulant addicts. Biological Psychiatry, 69,
260–​65.
Buckholtz, J.W., Treadway, M.T., Cowan, R.L., et al. (2010). Dopaminergic network differences in
human impulsivity. Science, 329, 532.
Burgess, B. (2013). The Happy Addict: How to Be Happy in Recovery From Alcoholism or Drug Addiction.
London: Eightball Publishing.
Chiesa, A. and Serretti, A. (2014). Are mindfulness-​based interventions effective for substance use
disorders? A systematic review of the evidence. Substance Use & Misuse, 49, 492–​512.
Dalley, J.W., Fryer, T.D., Brichard, L., et al. (2007). Nucleus accumbens D2/​3 receptors predict trait
impulsivity and cocaine reinforcement. Science, 315, 267‒70.
De Luca, C.R. and Leventer, R.J. (2008). Developmental trajectories of executive functions across the
lifespan. In: P. Anderson, V. Anderson, and R. Jacobs (eds), Executive Functions and the Frontal
Lobes: A Lifespan Perspective. Washington, DC: Taylor & Francis, pp. 3–​21.
De Shazer, S. (1985). Keys to Solution in Brief Therapy. New York: W.W. Norton.
Dick, D.M., Aliev, F., Latendresse, S.J., et al. (2013). Adolescent alcohol use is predicted by
childhood temperament factors before age 5, with mediation through personality and peers.
Alcoholism: Clinical and Experimental Research, 37, 108–​17.
Freud, S. (1954 [1895]). A project for a scientific psychology. In: M. Bonaparte, A. Freud, and K. Ernst
(eds), E. Mosbacher and J. Strachey (trans.), The Origins of Psycho-​Analysis: Letters to Wilhelm Fliess,
Drafts And Notes, 1887–​1902. New York: Basic Books, pp. 347‒445.
Goodwin, D.W., Schulsinger, F., Møller, N., Hermansen, L., Winokur, G., and Guze, S.B. (1974).
Drinking problems in adopted and nonadopted sons of alcoholics. Archives of General Psychiatry,
31, 164‒69.
Grant, B.F., Dawson, D.A., Stinson, F.S., Chou, S.P., Dufour, M.C., and Pickering, R.P. (2004). The 12-​
month prevalence and trends in DSM-​IV alcohol abuse and dependence: United States, 1991‒1992
and 2001‒2002. Drug and Alcohol Dependence, 74, 223–​34.
Hayden, E.P., Klein, D.N., Dougherty, L.R., et al. (2010). The dopamine D2 receptor gene and
depressive and anxious symptoms in childhood: associations and evidence for gene-​environment
correlation and gene-​environment interaction. Psychiatric Genetics, 20, 304‒10.
Hicks, B.M., Johnson, W., Durbin, C.E., Blonigen, D.M., Iacono, W.G., et al. (2013). Gene–​
environment correlation in the development of adolescent substance abuse: selection effects of
child personality and mediation via contextual risk factors. Development and Psychopathology, 25,
119–​32.
Kendler, K.S., Sundquist, K., Ohlsson, H., Palmér, K., Maes, H., Winkleby, M.A., et al. (2012). Genetic
and familial environmental influences on the risk for drug abuse: a national Swedish adoption study.
Archives of General Psychiatry, 69, 690‒97.
Khantzian, E.J. (1990). Self-​regulation and self-​medication factors in alcoholism and the addictions.
Similarities and differences. Recent Developments in Alcoholism, 8, 255‒71
Klein, T.A., Neumann, J., Reuter, M., Hennig, J., von Cramon, D.Y., and Ullsperger, M. (2007).
Genetically determined differences in learning from errors. Science, 318, 462‒65.
324  

324 The blindfold of addiction

Koob, M. and Le Moal, M. (2001). Drug addiction, dysregulation of reward, and allostasis.
Neuropsychopharmacology, 24, 97–​129.
Jung J.G., Kim J.S., Kim G.J., Oh M.K., and Kim S.S. (2011). Brief insight-​enhancement intervention
among patients with alcohol dependence. Journal of Korean Medical Science, 26, 11‒16.
Linehan, M.M. (1993). Skills Training Manual for Treating Borderline Personality Disorder.
New York: Guilford.
Loewenstein, G. (1999). A visceral account of addiction. In: J. Elster and O.J. Skog (eds), Getting
Hooked: Rationality and Addiction. Cambridge: Cambridge University Press, pp. 235–​64.
Mitchell, J.M., Fields, H.L., D’Esposito, M., and Boettiger, C.A. (2005). Impulsive responding in
alcoholics. Alcoholism: Clinical and Experimental Research, 29, 158–​69.
Nestler, E.J., Barrot, M. and Self, D.W. (2001). DeltaFosB: a sustained molecular switch for addiction.
Proceedings of the National Academy of Sciences USA, 98, 42–​6.
Nutt, D., King, L.A., Saulsbury, W., and Blakemore, C. (2007). Development of a rational scale to assess
the harm of drugs of potential misuse. Lancet, 369, 47‒53.
Ratsma, J.E., Van Der Stelt, O., and Gunning, W.B. (2002). Neurochemical markers of alcoholism
vulnerability in humans. Alcohol and Alcoholism, 37, 523‒33.
Robinson, T.E. and Berridge K.C. (1993). The neural basis of drug craving: an incentive-​sensitization
theory of addiction. Brain Research Reviews, 18, 247‒91.
Schneier, F.R., Martinez, D., Abi-​Dargham, A., et al. (2008). Striatal dopamine D(2) receptor
availability in OCD with and without comorbid social anxiety disorder: preliminary findings.
Depression and Anxiety, 25, 1‒7.
Schuckit, M.A. (2009). An overview of genetic influences in alcoholism. Journal of Substance Abuse
Treatment, 36, S5–​S14.
See, R.E. (2002). Neural substrates of conditioned-​cued relapse to drug-​seeking behavior. Pharmacology
Biochemistry and Behavior, 71, 517–​29.
Sterling, P. and Eyer, J. (1988). Allostasis: a new paradigm to explain arousal pathology. In: S. Fisher and
J.T. Reason (eds), Handbook of Life Stress, Cognition, and Health. New York: Wiley, pp. 629‒49.
Wang, J., Fanous, S., Terwilliger, E.F., Bass, C.E., Hammer, R.P., Jr, and Nikulina, E.M. (2013).
BDNF over-​expression in the ventral tegmental area prolongs social defeat stress-​induced cross-​
sensitization to amphetamine and increases DeltaFosB expression in mesocorticolimbic regions of
rats. Neuropsychopharmacology, 38, 286–​96.
Wilson, T.D. and Gilbert, D.T. (2003). Affective forecasting. Advances in Experimental Social
Psychology, 35, 345–​411.
  325

Chapter 18

Behavioral economics as a framework

for brief motivational interventions
to reduce addictive behaviors
James G. Murphy
Ashley A. Dennhardt
Ali M. Yurasek

Abstract
Laboratory research guided by behavioral economics suggests that
reinforcement from substance-free activities influences rates of alcohol
consumption, and that addiction may be related to a sharp discounting
of delayed rewards. These behavioral economic mechanisms have also
shown consistent relations to alcohol and drug consumption in naturalistic
studies, and have informed the development of efficacious interventions
such as contingency management. A key next step in increasing the public
health impact of these basic reinforcement mechanisms is to understand
their specific relations to the development of addiction and their potential
to be manipulated in brief interventions. Behavioral economic laboratory
research suggests that increasing the salience of delayed outcomes and
the extent to which the behavior leading to those rewards or punishers is
viewed as part of a coherent pattern can reduce impulsive choices. This
chapter discusses a novel brief intervention approach—the Substance Free
Activity Session—that attempts to reduce alcohol and drug consumption
by targeting these behavioral economic mechanisms of change. It will also
discuss behavioral economic approaches to quantifying addiction severity or
potential, including demand curve and monetary and behavioral allocation
based measures of substance reinforcing efficacy.

1 Overview
The overall goal of this book is to describe and discuss a radical change in how addiction
should be conceptualized, with implications for improved scientific understanding and
a more rational response to problems of addiction in society at large. We applaud this
goal and eagerly agreed to contribute a chapter because the assessment and intervention
approaches we will describe are a direct extension of the “addiction as choice” perspective
326  

326 Behavioral economics as a framework for brief motivational interventions

(Vuchinich and Tucker 1988; Murphy et  al. 2005, 2012; Bickel et  al. 2014; Tucker et  al.
2002, 2009, 2012).
First, we aim to describe in practical terms some novel indices of addiction severity
(or addiction potential) that measure individual differences in the degree of economic
demand for drugs and in the relative value of drugs compared to alternatives. These rein-
forcing efficacy measures are viewed as a complement to traditional severity indices that
focus on measuring consumption patterns or measures of drug use problems or depend-
ence symptoms, and have shown predictive utility in identifying the likelihood of suc-
cessful change following an intervention or an attempt to change without intervention.
Second, we will describe in detail a novel brief motivational intervention approach
(called the Substance-​Free Activity Session: SFAS) that is based on behavioral economic
theory and the assumption that addiction reflects a potentially modifiable choice process.
The approach is intended as a complement to standard brief motivational interviews for
alcohol and drug use and focuses on increasing future orientation and engagement in
patterns of goal-​directed alternatives to drinking and drug use.

2  Reinforcement pathology model of addiction

Before proceeding, in keeping with the theme of the book, we should note that our
view of addiction is consistent with the reinforcement pathology model proposed by
Bickel and colleagues (2011, p. 413; 2014, p. 650). That model is an extension of behav-
ioral economic and behavioral theories of choice approaches (Vuchinich and Heather
2003)  and assumes that the value a person places on a substance is a function of the
benefit/​cost ratio of substance use in relation to the benefit/​cost ratios of other avail-
able activities. Addiction is understood as a continuous phenomenon (rather than an
all-​or-​nothing disease or physiological state) that is defined as a pattern of fairly consist-
ent preference for the reinforcing commodity (e.g. a drug or food item, or activity such
as gambling, sex, or Internet use) relative to other activities. Reinforcement patholo-
gies such as alcohol or drug addiction are presumed to result from ongoing patterns of
interactions between endogenous (e.g. physiologically mediated subjective response to
drugs, elevated stress or arousal, elevated delay discounting) and environmental/​con-
textual factors such as low availability of alternatives and ready access and low price of
the reinforcing commodity, social contexts that reinforce use of the commodity, as well
as life events that cause stress or other negative affect. The process is self-​perpetuating
(i.e., recursive) because repeated use of many addictive commodities will have direct
negative effects on the availability of alternatives, which will in turn increase the rela-
tive degree of preference for the drug. For example, as Rachlin (2000, pp. 82) described
in his Relative Addiction model, social activity is an important potential substitute for
drug use but “because frequent drug use often erodes social support”—​substance users
may avoid or alienate friends or family who are not using—​continued heavy drug use
directly undermines what would be the most viable alternative to using drugs. Although
Rachlin described the particular relevance of deficits in social support as a contributor
  327

Assessment strategies 327

to a vicious cycle of increasing relative valuation of addictive commodities, behavioral

economic research suggests that levels of substance use generally show inverse relations
with a variety of different alternative reinforcers.
It is important to note that we do not view choice models of addiction as implying
that substance use is volitional, or that individuals make a conscious choice to become
addicted. Instead, we think the unique implication of choice or behavioral economic
models of addiction is to emphasize that patterns of addictive behavior are sensitive to
the influence of the environment, and in particular the relative availability and price of
drugs and of alternative reinforcers. In other words, these models assume that patterns of
choices between drugs and alternatives adhere to basic behavioral (learning theory) and
economic principles.
The choice or reinforcement pathology models have two clear implications for assess-
ment and treatment that will be discussed in the remainder of this chapter—​first, that the
assessment of substance abuse or addiction severity, and response to treatment, should
include measurement of the relative valuation of drug-​related and drug-​free rewards;
second, that treatment should attempt to reduce the overvaluation of current relative
to future rewards and increase engagement in regular patterns of behavior that lead to
delayed reinforcement.

3  Assessment strategies: behavioral economic

approaches to quantifying risk for addiction
and response to treatment
Behavioral economic theory uses the term reinforcing efficacy (RE), which in laboratory
settings is quantified by the amount of behavior (e.g. lever presses, time) allocated to
gain access to the reinforcer, to describe the relative level of preference for a reinforcer
such as alcohol (Bickel et al. 2014; Heinz et al. 2012; Hursh and Silberberg 2008). The
reinforcing efficacy (also referred to as reward value or relative reinforcing efficacy) of a
given drug is theorized to both be a product of the direct reinforcing effects of the drug
and individual differences in decision making (e.g. delay discounting). Although RE is a
relatively stable individual difference variable, it has also been shown to vary as a function
of experiential and environmental states (e.g. craving, negative affect, the availability of
alternative reinforcers: Amlung and MacKillop 2014; Bickel et al. 2014; Rousseau et al.
2011) and even substance use interventions (Dennhardt et al. 2015; Murphy et al., 2015).
There is some evidence that elevated alcohol reinforcing efficacy is associated with indices
of substance use problems and dependence (Heinz et al. 2012; Skidmore et al. 2014) and
predicts increasing drinking trajectories over time (Tucker et al. 2009, 2012), as well as
poor response to intervention (MacKillop and Murphy 2007; Murphy et al. 2005).
There are three general self-​report approaches to measuring RE that are all derived
from behavioral economic laboratory drug self-​administration models: purchase tasks,
reinforcement survey instruments, and measures of discretionary monetary expenditures
towards alcohol/​drugs (Heinz et al. 2012; Murphy et al. 2009; Tucker et al. 2009).
 328
Person A Person A Omax
(Higher Demand)
10 Intensity Person B
$10.00
CONSUMPTION

Elasticity Person A

EXPENDITURE
Person B $1.00
(Lower Demand)
1

Person B
$0.10

Breakpoint $0.01
0.1
$0.01 $0.10 $1.00 $10.00 $0.01 $0.10 $1.00 $10.00
PRICE PRICE
Figure 18.1  Alcohol demand (left panel) and expenditure (right panel) curves. Intensity = consumption at zero cost. Breakpoint = the price
that suppresses consumption to zero. Elasticity = aggregated price sensitivity, Omax = maximum total expenditure on alcohol. Person A
exhibits higher alcohol demand and expenditures compared to Person B, reflecting a higher value for alcohol as a reinforcer (reinforcing
efficacy)

 
  329

Assessment strategies 329

3.1  Drug and alcohol demand

Alcohol (Murphy and MacKillop 2006) and drug (Collins et al. 2014; Bruner and Johnson
2014; Jacobs and Bickel 1999; Mackillop et al. 2008) purchase tasks estimate reward value
by generating demand curves that plot consumption as a function of price and identify
how much someone would consume given unrestricted (free) access to alcohol/​drugs, how
much money they would spend on alcohol/​drugs, and the extent to which their consump-
tion level is price sensitive (Figure 18.1). The data required to generate these measures can
be obtained in a few minutes by asking approximately 15 questions about hypothetical
drug/​alcohol purchases across a range of prices. These scales can also be modified to pro-
vide information on contextual influences on demand (e.g. how many drinks would you
purchase if you had a class the next day? Skidmore and Murphy 2011). Demand curves
yield multiple indices that include consumption at minimum price (i.e. intensity), the price
that reduces consumption to zero (i.e. breakpoint), maximum expenditure (Omax), and the
aggregate slope of the demand curve (i.e. elasticity). Hypothetical alcohol purchase tasks
have been most widely investigated and yield reliable and valid individual difference meas-
ures of reinforcing efficacy that are correlated with lab-​based alcohol consumption and
a variety of collateral indices of alcohol problem severity, including alcohol use disorder
symptoms and craving (Skidmore et al. 2014; Mackillop et al. 2010). Alcohol demand has
also been shown to increase acutely in response to experimentally induced elevations in
craving (MacKillop et al. 2010) and stress (Amlung and MacKillop 2014), and to decrease
acutely following administration of the anti-​craving medication, naltrexone (Bujarski
et  al. 2012). In one study, elevated demand at baseline predicted greater levels of typi-
cal weekly drinking six months following a brief intervention, even after controlling for
baseline drinking and gender (MacKillop and Murphy 2007). Thus, even within samples
of regular drinkers there are individual differences in demand that may reflect the propen-
sity for alcohol problems and drinking patterns that are relatively impervious to standard
brief intervention approaches. Further, there is some evidence that reductions in demand
are detectable immediately following treatment, and that the degree to which treatment
reduces demand predicts long-​term drinking outcomes (Dennhardt et al. 2015).

3.2  Proportionate substance-​related reinforcement

Proportionate substance-​related reinforcement is defined as the relative allocation of time
and enjoyment that is associated with substance use compared to substance-​free behav-
ior (Correia et al. 1998; Skidmore et al. 2014). Reinforcement survey instruments define
reinforcement as the product of activity frequency and enjoyment ratings, and addiction
researchers have modified these measures to differentiate and quantify substance-​related
and substance-​free reinforcement (Correia et al. 1998, 2003). These self-​report measures
can be completed in approximately 5–​10 minutes in clinical or research settings. High
proportionate substance-​related reinforcement is theorized to be an early indicator of dis-
proportionate reliance on substance-​related reinforcement compared to alternative (non-​
drug) reinforcers (Murphy et al. 2007, 2012). As such, it may predict the likelihood of
330  

330 Behavioral economics as a framework for brief motivational interventions

subsequent escalation of substance misuse and a lower probability of maintaining healthy

drinking patterns.
There is considerable evidence that proportionate substance-​related reinforcement
plays an important role in alcohol and other drug use. Numerous experimental stud-
ies have shown that high rates of substance use are most likely in contexts devoid of
substance-​free sources of reinforcement, that substance abuse is associated with dimin-
ished dopamine response to naturally occurring substance-​free rewards such as food and
erotic stimuli (Garavan et al. 2000; Koob 2006; Volkow et al. 2003), and that substance
use will generally decrease if access to alternative reinforcers is increased (Carroll et al.
2009; Higgins et al. 2004). One longitudinal study found that the presence of alternative
reinforcers negatively predicted the development of smoking over a four-​year period (age
18-​22) (Audrain-​McGovern et al. 2011). Another longitudinal study with heroin depend-
ent adults found that subjective valence ratings of pleasant (drug-​free) pictures negatively
predicted future heroin use after controlling for baseline craving and heroin use (Lubman
et al. 2009). In emerging adult populations, there is evidence that heavy drinkers report
less reinforcement from non-​drug activities compared to matched controls (Correia
et al. 1998, 2003), and that illicit drug use is associated with incrementally lower levels of
substance-​free reinforcement relative to heavy drinking (Meshesha et al. 2015). Murphy
et al. (2005, p. 97) found that a matching law-​based measure (Herrnstein 1970) of the RE
of alcohol predicted brief intervention outcomes; participants who reported a smaller
proportion of their total activity participation and enjoyment (reinforcement) from sub-
stance use at baseline reported lower levels of follow-​up drinking, even after controlling
for their baseline drinking level.
Researchers have also conceptualized obesity as a reinforcement pathology reflecting
an overreliance on the short-​term reinforcement associated with food relative to food-​
free reinforcement (Epstein et  al. 2007). In a finding that closely parallels the alcohol
treatment outcome study summarized above, a recent longitudinal study found that
changes in proportionate food-​related reinforcement (i.e. relative activity participation
and enjoyment related to eating compared to other activities) predicted future weight loss
among overweight or obese adults who were participating in weight loss treatment, and
that the influence of proportionate food-​related reinforcement was independent of initial
body mass index values (Buscemi et al. 2014).

3.3  Alcohol-​Savings Discretionary Expenditure Index (ASDE)

The ASDE measures the difference between relative discretionary expenditures towards
alcohol and relative discretionary expenditures towards savings. Savings is used in order
to capture relative preference for alcohol (an immediate reinforcer) versus the delayed
reinforcement associated with monetary savings. Larger ASDE values reflect greater valu-
ation of alcohol relative to savings and have been shown to predict relapse following alco-
hol problem resolution (Tucker et al. 2009, 2012), and poor treatment response (Worley
et al. 2015), even after controlling for other significant measures of problem severity. The
ASDE, like the demand curve and proportionate substance-​related reinforcement indices,
may provide a clinically relevant index of alcohol value that is associated with greater
  331

A brief behavioral economic intervention approach to reduce alcohol and drug use 331

problem severity and poor treatment response. Because young adults typically do not allo-
cate significant amounts of money to savings, this measure has been modified in two stud-
ies with college students to reflect Relative Total Discretionary Expenditures on Alcohol
rather than discretionary expenditures on alcohol relative to savings (RDEA:  Murphy
et  al. 2009), and this metric has shown significant relations with other alcohol reward
metrics and with alcohol problems (Skidmore et al. 2014).

3.4  Clinical utility of behavioral economic assessment indices

These indices, when used in conjunction with other risk factors such as elevated con-
sumption and dependence symptoms, may provide a comprehensive model of substance
use severity that may prove useful in identifying individuals who are most at risk for
escalating substance abuse severity and in need of intervention services. All three of these
self-​report measurement approaches can be completed in clinical and clinical research
settings to provide unique information on severity and need for treatment, to identify
specific clinical concerns that might be a focus in treatment (e.g. elevated demand, low
rates of drug-​free reinforcement), and to measure response to treatment. Skidmore et al.
(2014) suggested potentially unique applications of the various RE indices as screening
and outcome measures in clinical contexts. For example, Intensity and Omax could be
especially useful clinical screening measures for risky drug use as they can be measured
with the very brief Alcohol or Drug Purchase Task (intensity with a single item asking
about maximum consumption at price = 0). Proportionate substance-​related reinforce-
ment could provide useful information on the specific need for a treatment that increases
substance-​free activities (Murphy et  al. 2012), and changes in the reinforcing efficacy
indices over time could be monitored as a secondary outcome measure or indicator of
a need for additional treatment. As noted earlier, there is evidence that change in these
metrics following treatment predicts drinking outcomes.

4  A brief behavioral economic intervention approach

to reduce alcohol and drug use: the Substance-​Free
Activity Session (SFAS)
Laboratory research guided by behavioral economics suggests that reinforcement from
drug-​free activities influences rates of addictive behaviors, and that addiction may be
related to a sharp discounting of delayed rewards (Ainslie, 1975). These behavioral eco-
nomic mechanisms have also shown consistent relations to addictive behavior patterns
in naturalistic studies, and have informed the development of efficacious interventions
such as Contingency Management. A  key next step in increasing the public health
impact of these basic reinforcement mechanisms is to understand their specific relations
to the development of addiction and their potential to be manipulated in brief interven-
tions. Behavioral economic laboratory research suggests that increasing the salience
of delayed outcomes and the extent to which the behavior leading to those rewards or
punishers is viewed as part of a coherent pattern can reduce impulsive choices (Logue
2000). The remainder of this chapter will discuss novel brief intervention approaches
332  

332 Behavioral economics as a framework for brief motivational interventions

that attempt to reduce addictive behaviors by targeting these behavioral economic

mechanisms of change.
The SFAS is a single session intervention that supplements a standard alcohol or drug-​
focused motivational interview. It uses well-​validated principles of motivational inter-
viewing (MI) (Miller and Rollnick 2002) to target the behavioral economic mechanisms
of substance-​free reinforcement and delayed reward discounting. As noted above, these
mechanisms originated in theoretical and laboratory work within the areas of behav-
ioral economics and operant psychology, but have more recently been “translated” into
variables that can be measured in naturalistic or clinical contexts (Mackillop et al. 2010;
Murphy et al. 2009, 2012; Murphy and Dennhardt 2016; Tucker et al. 2009). The SFAS
can be best understood as a direct application of MI to target increased engagement
in substance-​free activities, but it also integrates elements from cognitive-​behavioral
therapy/​community reinforcement for addiction (Carroll et al. 2012), behavioral acti-
vation (Lejuez et al. 2011), and motivational counseling (Cox and Klinger 2011). The
SFAS is distinguished from the latter approaches by its brevity and appropriateness for
non-​treatment-​seeking populations, inclusion of personalized feedback, and explicit
emphasis on enhancing patterns of future-​oriented and goal-​directed behavior.

4.1  Substance-​free reinforcement.

Some studies suggest that it may be possible to increase substance-​free activities using a
brief intervention approach, and that change in these alternative behaviors may precipi-
tate change in substance use. An experimental study that did not include a substance-​use
intervention found that college students who were instructed to increase drug-​free behav-
iors (exercise or creative activities) reported doing so, and also reduced alcohol use com-
pared to control participants (Correia et al. 2005). Additionally, Murphy et al. (2005, p. 98)
found that heavy drinking college students who reduced their drinking following a brief
intervention showed increased reinforcement from substance-​free activities at follow-​up,
and specifically increased academic activity. Other research with college students also
suggests that engagement in academic, athletic, or volunteer activities is inversely related
to substance use (Meshesha et al. 2015; Vaughan et al. 2009, Wechsler et al. 1997), and
that even among heavy drinkers, enjoyment ratings for substance-​free recreational activi-
ties are positively related to motivation to change drinking (Murphy et al. 2007). Thus,
substance users who have a number of enjoyable alternatives to substance use may be
more likely to reduce their consumption following an intervention, a finding that mirrors
research with adult substance users who change without formal treatment (Tucker et al.
2009). Together, these findings highlight the need for brief intervention approaches that
increase alternative activities among non-​treatment-​seeking substance abusers.

4.2  Delayed reward discounting

Individuals who use drugs may under-​engage in the constructive alternatives to drug use
identified above because the benefits of these activities are generally delayed. Delayed
reward discounting (DRD) refers to the level of decrease in subjective value associated
with reward delay. Although the value of all rewards decreases as their receipt is delayed,
  333

Substance-Free Activity Session (SFAS) treatment elements 333

there are substantial individual differences in the degree that delayed rewards are dis-
counted, and this discounting phenomenon may be a core feature of substance abuse
(Madden and Bickel 2009). Whereas substance use generally provides immediate rein-
forcement (e.g. anxiety reduction, euphoria, social facilitation), many substance-​free
activities (e.g. attending class, studying) do not (Murphy et al. 2006). For example, col-
lege students who sharply discount the value of delayed academic and career outcomes
may be less likely to engage in the behaviors necessary for success in these domains (e.g.
studying, internships), and may instead allocate their behavior towards immediately rein-
forcing activities such as using drugs. Indeed, numerous studies have demonstrated that
substance abusers discount the value of delayed rewards more steeply than control par-
ticipants (Bickel et al., 2014; MacKillop et al. 2010).
Delay discounting may be a particularly relevant risk factor for university students, in
light of the fact that many of the academic and career-​related activities that would presum-
ably compete with alcohol and drug use are associated with delayed and uncertain out-
comes (e.g. graduation, admission to graduate school, career success). Behavioral economic
laboratory research suggests that increasing the salience of delayed outcomes and the extent
to which the behavior leading to those rewards or punishers is viewed as part of a coherent
pattern can reduce impulsive choices (Monterosso and Ainslie 1999). Hofmeyr and col-
leagues (2011, p. 406), for example, found that forcing smokers to think of their decisions or
behaviors (smoking) as a series of “bundles” or patterns decreased discounting.
One clinical implication of this research is that, short of creating immediate and pow-
erful alternatives to substance use through intensive contingency management (Higgins
et al. 2004), community reinforcement (Carroll et al. 2012), or cognitive rehabilitation
approaches (Bickel et  al. 2011; Houben et  al. 2011), interventions should attempt to
encourage substances abusers to view their day-​to-​day behavior as comprising patterns
leading towards long-​term outcomes. The personalized drug and alcohol feedback
included in motivational interviewing approaches may help to accomplish this perspec-
tive shift; drug use decisions are aggregated to form meaningful tallies, like instances
of drug use per week, money spent on drug use during a month/​year, and rates of
drug use relative to peers (Dimeff et al. 1999). However, a key and unique implication
of behavioral economic theory is that interventions should attempt to aggregate more
global day-​to-​day decisions and activities (both substance-​related and substance-​free)
into cohesive patterns that have implications for long-​term substance-​free rewards.

5  Substance-​Free Activity Session (SFAS)

treatment elements
5.1  Session
introduction, clarification of goals,
and congruence with substance use patterns
To date, the SFAS has been used in conjunction with a standard brief motivational inter-
view for alcohol and drug use (BMI; Dimeff et al. 1999). The SFAS session usually occurs
a week after the BMI session and begins with a brief follow-​up of the previous session to
clarify any questions the student may have and to check on progress with their substance
334  

334 Behavioral economics as a framework for brief motivational interventions

use goals if they were motivated to change their use. The clinician next introduces the SFAS
session, describing the purpose as a way to help the student get the most out of their college
experience through clarifying their goals and deciding how to best organize their time.
After introducing the SFAS, the purpose of the first part of the session is to explore the
student’s values by engaging them in a discussion about why they chose to attend college.
Students describe a variety of motives for attending college that range from socializing and
having “the college experience” to fulfilling their parents’ expectations. However, almost
all students begin college with goals of completing a degree to gain entry into their profes-
sion of choice or, for students who are undecided about their specific career aspirations,
to simply have more opportunities and potential for future financial stability. Consistent
with the underlying motivational interviewing framework (Miller and Rollnick, 2013), all
ideas are accepted and explored, and the clinician attempts to understand and highlight
the student’s specific motivations for attending college.
Students are also asked about their future plans and specific goals for college. At this
point, after asking the student to carefully articulate their reasons for attending college
and their college goals, the clinician asks about the relationship between the student’s
substance use and their goals for college and their career. It is important not to assume
that alcohol or drugs are interfering with the student’s goals and that, if they are, the clini-
cian remains nonjudgmental, letting the student make judgments regarding the harmful-
ness of their use. If a student denies that their substance use has an impact on their goals,
the clinician may then ask the student how they have been able to maintain this balance
between drinking and academics, and also to ask if they anticipate future incompatibil-
ity between drinking/​drug use and college or career success. Many students who deny
current academic impairment related to substance use acknowledge that continued or
increasing use would likely cause future impairment as their academic or career demands
increase. This both reinforces positive behaviors (e.g. avoiding drinking on week nights)
and encourages the student to think more in depth about the role of alcohol or drugs in
accomplishing their goals for college on beyond. Many students express that they expect
they will need to reduce their drinking in the future to accomplish their college goals.

5.2  Feedback elements

The next section of the SFAS involves presenting the student with feedback on the finan-
cial benefits of college, the requirements necessary to reach their earlier stated goals, how
they spend their time, and substance-​free activities they might enjoy. Many of the ele-
ments in this section are personalized and the feedback is created from information the
student provided in the assessment prior to the intervention. This section also includes an
exercise to help students view their activities in terms of current value (how enjoyable it is
currently) and future value (how might the activity help them to meet long-​term goals).
These specific components are described in detail below.
5.2.1 Graduating from college
In this section, the clinician provides the student with facts about college graduation
rates and the financial benefits associated with graduating from and maintaining a high
  335

Substance-Free Activity Session (SFAS) treatment elements 335

grade-​point average (GPA). Students are provided with figures that illustrate the differ-
ences in hourly wage and yearly income between those who obtain a high school diploma
and those who earn a college degree and a figure showing income differences based on
college GPA (obtaining a college degree and a higher grade point average are associated
with higher income in the US). One purpose of providing this information is to attempt to
enhance the value of future rewards by making the long-​term rewards of a college degree
and achieving a high GPA more salient. Students are also provided with university gradu-
ation statistics, which indicate that only about half of US students who start a four-​year
degree go on to graduate. This is often surprising to students and can reinforce the notion
that completing a college degree is not a certainty and will require considerable effort, but
ultimately yield significant financial rewards.

5.2.2 Intended career requirements

As mentioned above, students will often discuss their career goals in the first section of
the intervention. In this section of the feedback, students are provided with a summary of
the requirements for entering their chosen career path or what they need to accomplish
to achieve the aforementioned goals. If a student is unsure or has not chosen a major, this
section focuses on helping them do so. For those with an identified career path, informa-
tion is provided on the specific educational requirements (e.g. graduate or professional
school), acceptance standards (e.g. average GPA for students admitted into Law School),
and licensure requirements if applicable (e.g. in many cases a drinking and driving or
drug arrest might preclude future licensure).

5.2.3 Personalized career-​related activities

Students are also provided with a list of institution-​specific extracurricular activities that
are consistent with their interests and career goals. If the student expresses interest in get-
ting involved in any of the activities, more specific information is discussed such as how
to join and the time/​location of the activity. Students are also asked about potential barri-
ers to getting involved and the clinician may engage in problem solving with the student if
appropriate. Students may think of ways to allow for time to participate in extracurricular
activities by cutting back on the number of nights spent drinking or by employing better
time management skills or study habits. Maintaining a collaborative and motivational
interviewing stance is especially important in this section to avoid making the student
feel like they are agreeing to participant in activities only to placate the clinician. Rather,
as with other sections of this intervention, the students’ interests and goals should guide
the content and focus.

5.2.4 Time allocation

In this section, students are provided with personalized feedback about how they spend
their time across a variety of domains (see Figure 18.2). The clinician guides the student
through the personalized feedback, explores their reaction, and facilitates the student’s
examination of how the current time allocation fits with or is discrepant from his or her
ideal time allocation. Many times the student recognizes that their recent time alloca-
tion is not conducive to making progress towards their long-​term goals and from here
336  

336 Behavioral economics as a framework for brief motivational interventions

16
Family
14 Studying
Exercising
12 Drinking
Extracurricular
10
Working
Class
8
Internet
6 Internships/Community
Drug Use
4

0
Figure 18.2  Time allocation feedback element

the clinician is able to help the student think of how they might restructure their time to
reduce the discrepancy. To further increase motivation, students are asked about the ben-
efits of making such changes and how the changes might impact their progress towards
their goals. Students are also asked to discuss potential barriers to implementing the
goals, and the clinician offers to help them with problem solving to address those barriers.
In addition to showing the student how they spend their time, students receive a figure
summarizing sample-​level associations between grade point average and average time
spent studying, attending class, and drinking. This figure reflects that, on average, uni-
versity students who spend more time in class and studying earn higher grades, and that
student who spend more time drinking earn lower grades. Given the earlier information
presented that highlighted the association between GPA and future income, the student
is able to identify that spending more time studying and less time drinking could lead to
greater future income. If the student is interested, some general time management strate-
gies can also be reviewed in this section.

5.2.5 Weighing the immediate and future benefit of activities

This module is designed to increase the student’s awareness of the immediate versus
delayed values of the activities they engage in. Students are asked to identify activities
that have a high “current value”; things they really enjoy in the moment. Students will
often name activities like drinking with friends, watching TV or surfing the Internet, or
playing video games. Students are also asked to think of three activities that have high
“future value,” meaning that engaging in them now will lead to benefits in the future. To
help the students conceptualize this idea, the clinician provides an example such as physi-
cal activity/​exercise: exercising is high in future value because exercising now is likely to
lead to health benefits in the future. Students often mention studying, going to class, and
participating in an academically focused extracurricular activity as activities that have
  337

Substance-Free Activity Session (SFAS) treatment elements 337

high future value. These activities are written along the X-​axis of a chart. Students are
then asked to rate the current value of all of the activities on a scale from 1 to 10 and the
clinician uses a marker to create bars that reflect these values. Predictably, many students
rate the activities they identified as having high current value very highly (8–​10) and
often the high future value activities (e.g. studying) very low in current value. Students
are then asked to do the same thing, but now rating all activities on their level of future
value. Activities high in current value, such as drinking, are typically rated fairly low in
future value, but occasionally a student will rate drinking as high future value, and point
out that drinking with friends can solidify relationships that they will maintain through-
out their lifetimes. Consistent with motivational interviewing, the clinician validates this
perception and focuses the conversation around other activities with high future value.
The clinician may also ask the student about ways other than drinking that he/​she could
strengthen social relationships.
Once all activities have been rated, student can see the visible representation of the
relative current and future values for each activity. This activity frequently prompts stu-
dents to discuss how they would like to increase the amount of time spent on activities
that have high future value, and also express the difficulties they experience in allotting
the time for these high future value activities. In these cases the clinician often discusses
the concept of self-​control, and validates that it is indeed difficult to choose to engage in
activities that may not be as enjoyable in the moment but will help them meet their goals.
The clinician also provides the student with strategies for improving self-​control such as
making specific commitments, enlisting social support, plotting or graphing outcomes/​
progress, maintaining rest/​nutrition, reminding oneself of goals and values, and utilizing
self-​reinforcement for goal progress. The concept is framed as thinking about “what they
want now” versus “what they want most” and how to behave in a way that will be most
likely to give them what they want most in the future.
5.2.6 Coping with stress and depression.
The SFAS intervention also includes an optional module on coping with stress and
depression. This is included because these symptoms can interfere with goal pursuit and
increase the reinforcing efficacy of alcohol (Murphy et  al. 2013). If a student indicates
elevations on the Depression, Anxiety, and Stress Scale (DASS; Lovibond and Lovibond
1995), they are provided with a list of the symptoms they endorsed. The clinician provides
empathic support related to the stressful nature of college and inquires about the student’s
preferred coping strategies. The clinician also inquires about the role of alcohol/​drugs and
stress and, if the student is interested, often suggests alternative coping strategies such as
behavioral activation, relaxation, or problem-​solving approaches.
5.2.7 Substance-​free recreational activities
The final feedback section of the feedback for the SFAS intervention focuses on substance-​
free activities. The student is provided with a list of the substance-​free activities they
reported enjoying in the assessment the week earlier. They are also provided with others
they did not list, but that are related to general interests they identified or to their major.
The clinician engages the student in a discussion about these activities (e.g. what they like
338  

338 Behavioral economics as a framework for brief motivational interventions

about the activity, what new activities they would like to try in the near future). In addi-
tion to hobbies and social/​leisure activities, many students mention wanting to engage
in activities that were discussed during the career feedback portion such as volunteering
or completing an internship experience. To increase motivation to participate in these
activities, the students are asked about the benefits of participating and again about the
barriers in order to help problem-​solve any issues that may prevent the student from
engaging in these activities. Research suggests that students who increase their engage-
ment in substance-​free activities often reduce their drinking even without any specific
drinking intervention (Correia et al. 2005).

5.3  Summary and goal setting

By this point in the SFAS intervention, the student has been provided with a large amount
of information and may lose sight of the central themes of the session. Thus, the clinician
summarizes and attempts to integrate the significant content from the session (e.g. the
student’s college goals, how these goals have been impacted by alcohol use, any specific
feedback elements that seemed to resonate with the student.) This often elicits further
reflection from the student on one or more of the sections and serves as a springboard to
the goal-​setting exercise that concludes the session. In this concluding segment the main
college goals that had been set in the early part of the session are revisited, sometimes
revised, and are supplemented with small, short-​term goals that the student can begin to
work on immediately or in the very next semester. For example, if the student had set a
goal to graduate with a GPA of 3.5, a smaller short-​term goal might be to study a speci-
fied number of hours daily. It is important with the short-​term goals that they follow the
SMART guidelines (Rubin 2002), that is, that they are specific, measureable, attainable,
realistic, and time-​based. Similar to other sections of the SFAS session, students will be
asked about benefits of achieving these goals, barriers to achieving them and about how
they will know if their plan is working. The purpose of this is to help the students once
they leave the session and throughout college, be able to identify barriers, problem solve
around them, and to self-​monitor their progress.

6 Outcomes
Murphy and colleagues developed the SFAS session as a brief approach to enhance engage-
ment in future-​oriented substance-​free activities that might “compete” with drinking. After
drafting a treatment manual, the authors solicited feedback from heavy drinking college
students in a series of focus groups, followed by an open pilot trial (N = 14; Murphy et al.
2012b) that demonstrated reductions in drinking. Participants in the initial randomized
controlled trial (Murphy et al. 2012) were 82 college freshman (50 percent female) who
reported two or more past-​month heavy drinking episodes. In comparison to an alcohol
BMI plus a relaxation training active control session, the alcohol BMI + SFAS condition
was associated with significantly greater reductions in alcohol related problems on the
Young Adult Alcohol Consequences Questionnaire (YAACQ; Read et  al. 2006)  at both
one-​month and six-​month follow-​up assessments. The BMI + SFAS showed large effect
  339

Outcomes 339

18

16

14

12
Alcohol Problems

10

0
Pre-Intervention 1-Month Post- 6-Months Post-
Intervention Intervention

Relaxation SFAS

Figure 18.3  Changes in past-month number of alcohol problems from baseline to follow-up

by intervention condition

size reductions in alcohol problems at one month and largely maintained that reduction
at the six-​month follow-​up, whereas the BMI + Relaxation condition showed no change
in problems at one month and a small effect size change at six months (see Figure 18.3).
This effect was partially mediated by an increase in protective behavioral strategies, such
as using a designated driver, avoiding drinking games, and leaving a bar/​party at a predeter-
mined time. Additionally, students in the BMI plus SFAS condition who reported lower lev-
els of substance-​free reinforcement or higher symptoms of depression at baseline reported
greater reductions in heavy drinking compared to participants in the BMI + Relaxation con-
trol condition. Finally, participants reported increases in two of the intended mechanisms of
change, namely consideration of future consequences (Strathman et al. 1994) and evening
studying. These findings suggest that incorporating a single session focused on increasing
engagement in alternative activities can enhance the effects of standard BMIs.
A second randomized controlled trial attempted to replicate and extend the Murphy
et al. (2012, pp. 6–​7) by evaluating an abbreviated version of the alcohol BMI + SFAS that
were administered back-​to-​back in a single hour. This session was compared to a similarly
timed alcohol BMI + Education session control (Yurasek et al. 2014). This study adapted
the BMI and SFAS sessions to address both drug and alcohol use. Unlike the original
Murphy et al study, all participants reduced their alcohol consumption and problems at
follow-​up, and there was no significant advantage for the BMI + SFAS. However, par-
ticipants in the BMI + SFAS condition used marijuana on significantly fewer days at the
340  

340 Behavioral economics as a framework for brief motivational interventions

16.00

14.00

12.00
Marijuana Use Days

10.00

8.00

6.00

4.00

2.00

0.00
Pre-Intervention 1-Month Post- 6-Months Post-
Intervention Intervention

SFAS Education

Figure 18.4  Change in number of days using marijuana by condition

six-​month follow-​up compared to those in the BMI + ED (p = 0.023; see Figure 18.4). It
is possible that compressing the administration of the BMI + SFAS sessions into a single
hour reduced efficacy for alcohol outcomes relative to its original two hour (separated
by a week) administration, or that alcohol education is a more viable supplement to a
standard MI than is relaxation. We are currently exploring a two-​session version of the
BMI + SFAS that also included a booster phone call in an ongoing multisite clinical trial.

7  Potential future applications and extensions

7.1  Behavioral economic assessment approaches
The studies reviewed above provide support for the use of behavioral economic measure-
ment of the relative valuation of drug-​related and drug-​free rewards in applied clinical
settings. The demand curve, monetary expenditure, and relative behavioral allocation
indices of substance reinforcing efficacy can provide information on an individual’s rela-
tive valuation of drugs that might complement existing measures of substance use pat-
terns and problems. Relative valuation of drugs might be especially useful for individuals
taking part in harm-​reduction interventions as a way to predict the degree of control
over substance use and the likelihood of clinically relevant increases in use. Importantly,
these indices are relatively brief and appear to both predict response to intervention and
change following an intervention that eventually reduced substance use (Dennhardt et al.
2015; Worley et al. 2015). A recently completed study in our laboratory found that both
reductions in demand (measured immediately after a brief intervention) and reductions
  341

Potential future applications and extensions 341

in proportional alcohol expenditures (during the month after a brief intervention) pre-
dicted reductions in drinking in multivariate models that controlled for baseline drinking
level (Murphy et al., unpublished data). Furthermore, change in these reinforcing efficacy
indices was a more robust predictor than baseline levels of the indices.
Future research should continue to validate user-​friendly approaches to measuring
reinforcing efficacy among substance abusers presenting for treatment, and in high-​risk
populations as part of brief intervention screening efforts. Given the focus on measur-
ing ongoing patterns of substance-​related and substance-​free activities and expenditures,
ongoing prospective assessment using smartphones or other technology would be an
ideal way to capture dynamic changes in substance reinforcing efficacy that might indi-
cate a need for tailored intervention elements.

7.2  Behavioral economic brief intervention approaches

The SFAS is an initial effort to incorporate behavioral economic elements into the general
structure of a motivational interviewing-​based brief intervention, and a controlled pilot trial
suggests that it may increase the efficacy of popular brief alcohol interventions (Murphy
et al. 2012). Specifically, the SFAS extends traditional brief interventions by attempting to
increase the substance abuser’s commitment to important goals, engagement in substance-​
free activities, and general degree of future orientation. These goals can be accomplished
even without specific motivation to reduce drinking or drug use, but the theoretical assump-
tion is that increases in future orientation and engagement in substance-​free activities will
indirectly increase the likelihood of reductions in alcohol and drug use (e.g. Correia et al.
2005). To date, our two completed and two ongoing SFAS trials have focused on college
students. This is a high-​risk group in terms of alcohol and drug use, and it also ideally suited
for the SFAS given the general incompatibility between successful completion of college and
career-​development goals and heavy substance use. Young adulthood is also an ideal time
for an intervention that does not require help seeking or motivation to change, and that
focuses on developing goals and increasing future orientation.
Future research should investigate the SFAS as an adjunct to substance-​use inter-
ventions with other populations. Empirically supported interventions such as com-
munity reinforcement therapy or contingency management (Higgins et  al. 2004)  are
examples of intensive intervention approaches that target some of the same theoreti-
cal mechanisms as the SFAS (e.g. increasing substance-​free reinforcement, reducing
impulsive decision-​making). Individuals with severe substance abuse and or cognitive
impairment may require these intensive approaches, but the SFAS may be a more viable
approach with individuals with mild to moderate substance use problems. In addition
to college students, we suspect that there are a number of other potential populations
of substance abusers who might benefit from the SFAS. In particular, young adults who
are not college students might also benefit from an approach that helps them to address
drinking/​drug use in the context of developing a greater consideration of the future and
identifying patterns of goal-​directed substance-​free activities. Similarly, military veter-
ans are a high-​risk group that might also lack viable alternatives to drinking and require
342  

342 Behavioral economics as a framework for brief motivational interventions

an approach that attempts to specifically address this issue (McDevitt-​Murphy et  al.
2014). The SFAS may be especially helpful for individuals with psychiatric comorbidity,
which is often associated with elevated alcohol reward value (Murphy et al. 2013) and
diminished engagement in rewarding alternatives to substance use (Murphy et  al.
2012). Finally, individuals who are entering retirement, unemployed, recently divorced
or separated, or transitioning from a controlled environment such as a prison or reha-
bilitation facility may benefit from an approach such as the SFAS. When incorporated
outside of educational settings the SFAS could encourage individuals to identify life
goals and corresponding patterns of behavior that might facilitate progress towards
those goals, discuss the extent to which the goals and associated behavior patterns are
congruent with drinking/​drug use, provide personalized information to facilitate goal
progress and enhance the salience of the benefits of goal achievement, and encourage
commitment to feasible and specific short-​term goals to facilitate attainment of the
long-​term goals. Depending on the population, goals might include increasing physical
fitness, developing a particular skill or hobby, improving social or family relationships,
or pursuing a volunteer, vocational, or educational goal. The clinician should attempt
to focus the session as much as possible on goals that are generally incompatible with
heavy substance use.

References
Amlung, M.T., Acker, J., Stojek, M.K., Murphy, J.G., and MacKillop, J. (2012). Is talk “cheap?” an initial
investigation of the equivalence of alcohol purchase task performance for hypothetical and actual
rewards. Alcoholism: Clinical and Experimental Research, 36, 716–​24.
Amlung, M, and MacKillop, J. (2014). Clarifying the relationship between impulsive delay discounting
and nicotine dependence. Psychology of Addictive Behaviors, 28, 761–​68.
Ainslie, G. (1975). Specious reward: A behavioral theory of impulsiveness and impulse control.
Psychological Bulletin, 82, 463–​96.
Audrain-​McGovern, J., Rodriguez, D., Rodgers, K., and Cuevas, J. (2011). Declining alternative
reinforcers link depression to young adult smoking. Addiction, 106, 178–​87.
Bickel, W.K., Jarmolowicz, D.P., Mueller, E.T., and Gatchalian, K.M. (2011). The behavioral economics
and neuroeconomics of reinforcer pathologies: implications for etiology and treatment of addiction.
Current Psychiatry Reports, 13, 406–​15.
Bickel, W.K., Johnson, M.W., Koffarnus, M.N., MacKillop, J., and Murphy, J.G. (2014). The behavioral
economics of substance use disorders: reinforcement pathologies and their repair. Annual Review of
Clinical Psychology, 13, 20.1–​20.37.
Bruner, N, and Johnson, M. (2014). Demand curves for hypothetical cocaine in cocaine-​dependent
Individuals. Psychopharmacology, 231, 889–​97.
Buscemi, J., Murphy, J.G., Berlin, K. S., and Raynor, H.A. (2014). A behavioral economic analysis
of changes in food-​related and food-​free reinforcement during weight loss treatment. Journal of
Consulting and Clinical Psychology, 82(4), 659.
Bujarski, S., MacKillop, J., and Ray, L.A. (2012). A behavioral economic approach to understanding
alcoholism pharmacotherapy mechanisms: evidence for naltrexone in Asian Americans.
Experimental Clinical Psychopharmacology, 20, 181–​90.
  343

Potential future applications and extensions 343

Carroll, M.E., Anker, J.J., and Perry, J.L. (2009). Modeling risk factors of nicotine and other drug abuse
in the preclinical laboratory. Drug and Alcohol Dependence, 104, S70–​8.
Carroll, K.M., Nich, C., LaPaglia, D.M., Peters, E.N., Easton, C.J., and Petry, N.M. (2012). Combining
cognitive behavioral therapy and contingency management to enhance their effects in treating
cannabis dependence: less can be more, more or less. Addiction, 107, 1650–​59. doi:10.1111/​
j.1360-​0443.2012.03877.x
Collins, R., Vincent, P., Yu, J., Liu, L., and Epstein, L. (2014). A behavioral economic approach to
assessing demand for marijuana. Experimental and Clinical Psychopharmacology, 22, 211–​21.
Correia, C.J., Benson, T.A., and Carey, K.B. (2005). Decreased substance use following increases in
alternative behaviors: a preliminary investigation. Addictive Behaviors, 30, 19–​27.
Correia, C.J., Carey, K.B., and Borsari, B. (2002). Measuring substance-​free and substance related
reinforcement in the natural environment. Psychology of Addictive Behaviors, 16, 28.
Correia, C.J., Carey, K.B., Simons, J., and Borsari, B.E. (2003). Relationships between binge drinking
and substance-​free reinforcement in a sample of college students: a preliminary investigation.
Addictive Behaviors, 28, 361–​68.
Correia, C.J., Simons, J., Carey, K.B., and Borsari B.E. (1998). Predicting drug use: application of
behavioral theories of choice. Addictive Behaviors, 23, 705–​709.
Cox, W.M. and Klinger, E. (eds) (2011). Handbook of Motivational Counseling: Goal-​based Approaches
to Assessment and Intervention with Addiction and Other Problems. London: John Wiley and Sons.
Dennhardt, A.A. and Murphy, J.G. (2011). Associations between depression, distress tolerance, delay
discounting, and alcohol-​related problems in European American and African American college
students. Psychology of Addictive Behaviors, 25, 595.
Dennhardt, A.A., Yurasek, A.M., and Murphy, J.G. (2015). Change in delay discounting and substance
reward value following a brief alcohol and drug use intervention. Journal of the Experimental
Analysis of Behavior, 103(1), 125–​40.
Dimeff, L.A., Baer, J.S., Kivlahan, D.R., and Marlatt, G. A. (1999). Brief Alcohol Screening and
Intervention for College Students (BASICS): A Harm Reduction Approach. New York: Guilford Press.
Epstein, L., Temple, J., Neaderhiser, B., et al. (2007). Food reinforcement, the dopamine D2 receptor
genotype, and energy intake in obese and nonobese humans. Behavioral Neuroscience, 121, 877–​86.
Garavan, H., Pankiewicz., J, Bloom, A., et al. (2000). Cue-​induced cocaine craving: neuroanatomical
specificity for drug users and drug stimuli. American Journal of Psychiatry, 157, 11789–​98.
Heather, N. and Vuchinich, R.E. (eds) (2003). Choice, Behavioural Economics and Addiction.
Oxford: Elsevier.
Heinz, A.J., Lilje, T.C., Kassel, J.D., and de Wit, H. (2012). Quantifying reinforcement value and
demand for psychoactive substances in humans. Current Drug Abuse Reviews, 5, 257–​72.
Herrnstein, R.J. (1970). On law of effect. Journal of the Experimental Analysis of Behavior, 13, 243–​66.
Higgins, S.T., Heil, S.H., and Lussier, J. (2004). Clinical implications of reinforcement as a determinant
of Substance Use Disorders. Annual Review of Psychology, 55, 431–​61.
Hofmeyr, A., Ainslie, G., Charlton, R., and Ross, D. (2011). The relationship between addiction and
reward bundling: an experiment comparing smokers and non‐smokers. Addiction, 106, 402–​409.
Houben, K., Wiers, R.W., and Jansen, A. (2011). Getting a grip on drinking behavior: training working
memory to reduce alcohol abuse. Psychological Science, 22, 968–​75.
Hursh, S, and Silberberg, A. (2008). Economic demand and essential value. Psychological Review, 115,
186–​98.
Jacobs, E. and Bickel, W. (1999). Modeling drug consumption in the clinic using simulation
procedures: demand for heroin and cigarettes in opioid-​dependent outpatients. Experimental and
Clinical Psychopharmacology, 7, 412–​26.
344  

344 Behavioral economics as a framework for brief motivational interventions

Koob, G.F. (2006). The neurobiology of addiction: a neuroadaptational view relevant for diagnosis.
Addiction, 101, 23–​30.
Lejuez, C.W., Hopko, D.R., Acierno, R., Daughters, S.B., and Pagoto, S.L. (2011). Ten year revision
of the brief behavioral activation treatment for depression: Revised treatment manual. Behavior
Modification, 35, 111–​61. DOI:10.1177/​0145445510390929
Logue, A.W. (2000). Self-​control and health behaviour. In: W.K. Bickel and R.E. Vuchinich (eds),
Reframing Health Behavior Change with Behavioral Economics. Mahwah: Lawrence Erlbaum
Associates, pp. 167–​92.
Lovibond, S.H. and Lovibond, P.F. (1995). Manual for the Depression Anxiety Stress Scales (2nd ed.).
Sydney: Psychology Foundation.
Lubman, D.I., Yucel, M., Kettle, J.W., et al. (2009). Responsiveness to drug cues and natural
rewards in opiate addiction: associations with later heroin use. Archives of General Psychiatry,
66, 205–​12.
Madden, G.J. and Bickel, W.K. (eds) (2009). Impulsivity: The Behavioral and Neurological Science of
Discounting. Washington, DC: APA.
Mackillop, J. and Murphy, J. (2007). A behavioral economic measure of demand for alcohol predicts
brief intervention outcomes. Drug and Alcohol Dependence, 89, 227–​33.
MacKillop, J., Murphy, J.G., Ray, L.A., et al. (2008). Further validation of a cigarette purchase task
for assessing the relative reinforcing efficacy of nicotine in college smokers. Experimental Clinical
Psychopharmacology, 16, 57–​65.
MacKillop, J., Miranda, R.R., Monti, P.M., et al. (2010). Alcohol demand, delayed reward discounting,
and craving in relation to drinking and alcohol use disorders. Journal of Abnormal Psychology, 119,
106–​14.
McDevitt-​Murphy, M.E., Murphy, J.G., Williams, J.L., et al. (2014). Randomized controlled trial of two
brief alcohol interventions for oef/​oif veterans. Journal of Consulting and Clinical Psychology, 82,
562–​68.
Meshesha, L.Z., Dennhardt, A.D., and Murphy, J.G. (2015). Polysubstance use is associated with
deficits in substance-​free reinforcement in college students. Journal of Studies on Alcohol and Drugs,
76, 106–​16.
Miller, W.R. and Rollnick, S. (2002). Motivational Interviewing: Preparing People to Change.
New York: Guilford.
Miller, W.R. and Rollnick, S. (2013). Motivational interviewing: Helping People Change.
New York: Guilford.
Monterosso, J. and Ainslie, G. (1999). Beyond discounting: possible experimental models of impulse
control. Psychopharmacology, 146, 339–​47.
Murphy, J., Barnett, N., Goldstein, A., and Colby, S. (2007). Gender moderates the relationship
between substance-​free activity enjoyment and alcohol use. Psychology of Addictive Behaviors, 21,
261–​65.
Murphy, J.G. and Dennhardt, A.A. (2016). The behavioral economics of young adult substance abuse.
Preventive Medicine. DOI: 10.1016/​j.ypmed.2016.04.022
Murphy, J.G. and MacKillop, J. (2006). Relative reinforcing efficacy of alcohol among college student
drinkers. Experimental and Clinical Psychopharmacology, 14, 219–​27.
Murphy, J.G., Correia, C.J., Colby, S.M., and Vuchinich, R.E. (2005). Using behavioral theories of
choice to predict drinking outcomes following a brief intervention. Experimental and Clinical
Psychopharmacology, 13, 93–​101.
Murphy, J.G., MacKillop, J., Skidmore, J.R., and Pederson, A.A. (2009). Reliability and validity of a
demand curve measure of alcohol reinforcement. Experimental Clinical Psychopharmacology. 17,
396–​404.
  345

Potential future applications and extensions 345

Murphy, J.G., Dennhardt, A.A., Skidmore, J.R., et al. (2012). A randomized controlled trial of a
behavioral economic supplement to brief motivational interventions for college drinking. Journal of
Consulting and Clinical Psychology, 80(5), 876–886 .
Murphy, J.G., Yurasek, A.M., Dennhardt, A.A., et al. (2013). Symptoms of depression and PTSD are
associated with elevated alcohol demand. Drug and Alcohol Dependence, 127(1), 129–​36.
Murphy, J.G., Dennhardt, A.D., Yurasek, A.M., et al. (2015). Behavioral economic predictors of brief
alcohol intervention outcomes. Journal of Consulting and Clinical Psychology, 83(6), 1033–​43. http://​
dx.doi.org/​10.1037/​ccp0000032
Rachlin, H. (2000), The lonely addict. In: W.K. Bickel and R.E. Vuchinich (eds), Reframing Health
Behavior Change with Behavioral Economics. Mahwah, NJ: Lawrence Erlbaum Associates, pp. 145–​65.
Read, J.P., Kahler, C.W., Strong, D.R., and Colder, C.R. (2006). Development and preliminary validation of
the young adult alcohol consequences questionnaire. Journal of Studies on Alcohol, 67(1), 169–​177.
Rousseau, G., Irons, J., and Correia, C. (2011). The reinforcing value of alcohol in a drinking to cope
paradigm. Drug and Alcohol Dependence, 118, 1–​4.
Rubin, R.S. (2002). Will the real SMART goals please stand up? Industrial Organizational Psychologist,
39, 26–​27.
Skidmore, J.R. and Murphy, J.G. (2011). The effect of drink price and next-​day responsibilities.
On college student drinking: a behavioral economic analysis. Psychology of Addictive Behaviors,
25, 57–​68.
Skidmore, J.R., Murphy, J.G., and Martens, M.P. (2014). Behavioral economic measures of alcohol
reward value as problem severity indicators in college students. Experimental and Clinical
Psychopharmacology, 22, 198–​210.
Strathman, A., Gleicher, F., Boninger, D.S., and Edwards, C. (1994). The consideration of future
consequences: weighing immediate and distant outcomes of behavior. Journal of Personality and
Social Psychology, 66, 742–​52.
Tucker, J.A., Vuchinich, R.E., and Rippens, P.D. (2002). Predicting natural resolution of alcohol-​
related problems: A prospective behavioral economic analysis. Experimental and Clinical
Psychopharmacology, 10, 248–​57.
Tucker, J.A., Roth, D.L., Vignolo, M.J., and Westfall, A.O. (2009). A behavioral economic reward index
predicts drinking resolutions: Moderation revisited and compared with other outcomes. Journal of
Consulting and Clinical Psychology, 77, 219–​28.
Tucker, J., Roth, D., Huang, J, Crawford, M., and Simpson, C. (2012). Effects of interactive voice
response self-​monitoring on natural resolution of drinking problems: utilization and behavioral
economic factors. Journal of Studies in Alcohol and Drugs, 73, 686–​98.
Vaughan, E., Corbin, W., and Fromme, K. (2009). Academic and social motives and drinking behavior.
Psychology of Addictive Behaviors, 23, 564–​76.
Volkow, N.D., Wang, G.J., Ma, Y., et al. (2003). Expectation enhances the regional brain metabolic and
the reinforcing effects of stimulants in cocaine abusers. Journal of Neuroscience, 23, 11461–​68.
Vuchinich, R.E. and Tucker, J.A. (1988). Contributions from behavioral theories of choice to an analysis
of alcohol abuse. Journal of Abnormal Psychology, 97, 181–​95.
Wechsler, H., Davenport, A., Dowdall, G., Grossman, S., and Zanakos, S. (1997). Binge drinking,
tobacco, and illicit drug use and involvement in college athletics: a survey of students at 140
American colleges. Journal of American College Health, 45, 195–​200.
Worley, M.J., Shoptaw, S.J., Bickel, W.K., and Ling, W. (2015). Using behavioral economics to predict
opioid use during prescription opioid dependence treatment. Drug and Alcohol Dependence,
148, 62–​68.
Yurasek, A.M., Dennhardt, A.A., and Murphy, J.G. (2014). A randomized controlled trial of a
behavioral economic intervention for substance abuse. Manuscript submitted for publication.
346  

Chapter 19

Role of choice biases and choice

architecture in behavioral economic
strategies to reduce addictive
behaviors
Jalie A. Tucker
Susan D. Chandler
JeeWon Cheong

Abstract
Human investment activities are vulnerable to delay discounting and a
range of other common choice biases. This chapter summarizes conceptual
work and research on choice biases and discusses implications for individual
and public health strategies to reduce addictive behaviors, with emphasis
on public health. Principles of population science, prevention, and public
health practice are summarized to explicate the basis for an integrated
intervention strategy, informed by research on human choice behavior,
which spans clinical, community, healthcare system, and policy interventions.
Interventions may remediate choice biases (e.g. seek to reduce delay
discounting) or manipulate the architecture of choice by framing options
to help people choose in their best interests (e.g. make the more beneficial
choice the default option). Choice architecture strategies implemented
within healthcare systems and communities have greater potential for
population impact than individual clinical treatments, and what mix of
options may maximize population benefits remains to be determined.

1 Introduction
Behavioral economic research has demonstrated robustly that investment activities in
humans and animals are vulnerable to delay discounting and a range of other common
choice biases (Ainslie 1975; Rachlin et al. 1981; Madden and Bickel 2010). This chapter
summarizes conceptual work and research concerned with choice biases and discusses
implications for individual and public health strategies to reduce addictive behaviors,
with emphasis on the latter domain. Strategies range from supply and demand reduction
  347

Public health foundations of addictive behavior change strategies 347

interventions focused on the problem commodity or behavior to newer approaches that

seek to remediate choice biases or manipulate the “architecture of choice” by framing
options to help people make choices in their best interests (Thaler and Sunstein 2008; see
section 4). In addition to having utility for individual interventions, choice architecture
strategies implemented in community, healthcare, financial, and other organizational
venues have potential for reaching larger numbers of people and having greater popula-
tion impact, which is a central public health goal. This broadened scope of intervention
strategies is guided by a population perspective on addictive behaviors, which character-
izes problem severity in the population as lying along a continuum from none to mild/​
moderate to severe and seeks a spectrum of interventions that collectively addresses the
problem severity continuum (Institute of Medicine 1990; Tucker and Simpson 2011).
Historically, approaches to changing addictive behavior emphasized individual
approaches epitomized by clinical treatments, and evidence-​based treatments are now
well established, including for substance use and other addictive disorders (McGovern
and Carroll 2003; Miller 2009). However, consensus is emerging that further advances
in reducing the harms and costs of addictive behaviors will require augmenting clinical
interventions with public health and policy interventions that target social, economic,
and other structural factors involved in problem genesis, maintenance, and resolution
(Humphreys and McLellan 2011; Tucker and Simpson 2011; cf. Prado et al. 2013; see sec-
tion 2). The next generation of interventions seeks to combine key elements of individual-​
level interventions with dissemination concepts and strategies pioneered in public health.
Theoretical systems that provide organizing concepts and principles applicable at mul-
tiple levels are essential, and behavioral economics provides a useful transdisciplinary
framework for this endeavor (US National Institutes of Health (US NIH) 2010). These
approaches are not mutually exclusive, and development of multi-​component strategies
and policies built on behavioral economic theory and research is now within reach.
The chapter is organized as follows: We first describe basic principles of population sci-
ence, prevention, and public health practice and explicate why a broadened scope of behav-
ior change strategies is needed that extends beyond clinical treatments for diagnosable
disorders (see section 2). Second, we summarize evidence for systematic biases in human
choice behavior that can guide development of interventions aimed at promoting sound
choices over time that benefit health, finances, and other important areas of human activ-
ity (see section 3). Third, we connect these findings to existing and potential intervention
strategies for substance-​related problems spanning clinical, community, healthcare systems,
and policy interventions (see section 4). The chapter ends with consideration of strengths
and limitations of current work and future directions for research and practice that rest on
behavioral economic and choice architecture concepts and findings (see section 5).

2  Public health foundations of addictive

behavior change strategies
We rely on an expanded population perspective of addictive behaviors, originally
articulated in the influential 1990 US Institute of Medicine (IOM) book, Broadening
348  

348 Role of choice biases and choice architecture in behavioral economic strategies

Substance Use
none
moderate

heavy
abuse
dependence

Prevention Telehealth Outpatient Inpatient

and brief treatment treatment
interventions
severe
substantial

moderate
mild
none Problems
Figure 19.1 Population perspective on substance use, problems, and interventions.
Adapted with permission from Institute of Medicine, Broadening the Base of Treatment for Alcohol Problems, 1990,
by the National Academy of Sciences, Courtesy of the National Academies Press, Washington, D.C.

the Base of Treatment for Alcohol Problems. As shown in Figure 19.1, the severity of
dimensions related to addictive behaviors (e.g. substance use practices and harms and
costs of use, including physical dependence) are characterized as lying along con-
tinua ranging from mild to moderate to severe. The degree of severity on the multiple
dimensions is usually similar, but not necessarily identical (e.g. one may use substances
excessively with minimal negative consequences or, conversely, serious consequences
may result from minimal use). The nature of the configuration determines interven-
tion targets and goals.
Because most harms and costs are attributable to the population segment with mild to
moderate problems, intervening effectively with this segment has the greatest potential
for improving population health outcomes (IOM 1990). Individuals with serious to severe
problems—​i.e., those who fulfill clinical diagnostic criteria for substance use dependence or
are labelled addicts or alcoholics in the vernacular—​are a more visible, but minority popula-
tion segment. Clinical treatments target and deliver intensive intervention resources to this
segment, whereas public health approaches emphasize developing a spectrum of interven-
tions that collectively address target behaviors and outcomes of varying severity (Humphreys
and Tucker 2002; IOM 1990; Springer and Phillips 2006; Tucker and Simpson 2011). Clinical
treatments are an important component of services, but less intensive and more prevention-​
oriented services are appropriate for the majority with less serious problems.
These relationships between population distributions of problem severity and interven-
tion needs are shown in Figure 19.2, which is adapted from another IOM report (1994)
concerned with preventing mental health and substance use disorders. The report rec-
ommended a three-​pronged approach that includes (1) universal prevention that targets
the general population without regard to risk status; (2) selective prevention that targets
  349

Public health foundations of addictive behavior change strategies 349

Prevention Strategies Based Treatment Sector Services for

on population Risk Distribution Substance Use Disorders

Relapse
prevention,
Substance medication,
Use behavioral
Disorder monitoring,
mutual help

Residential/day Detoxification
Moderate treatment, outpatient/
risk telehealth inpatient

Low/no risk

Universal Selective Indicated Prevention Acute Longterm Aftercare

Figure 19.2  Intervention sprectrum for prevention and treatment of substance misuse.

Adapted with permission from Institute of Medicine, Reducing Risks for Mental Disorders: Frontiers for Prevention
Intervention Research, 1994, by the National Academy of Sciences, Courtesy of the National Academies Press,
Washington, D.C.

identifiable subpopulations with risk factors; and (3) indicated prevention that targets high-​
risk persons with symptoms or early onset of a given disorder (cf. Springer and Phillips
2006). The first two categories typically involve briefer, less costly interventions that can
be disseminated over large population segments or key risk groups, whereas the third cat-
egory involves screening, early case finding, clinical treatment, aftercare, and relapse pre-
vention. The former interventions are relatively less efficacious at the individual level, but
their aggregate impact on population health may be greater because they reach many more
people (Tucker and Grimley 2011). Indeed, in some cases, a targeted selective intervention
with a sizeable at-​risk subpopulation (e.g. injection drug users, sexually active emerging
adults) may be an optimal intervention choice from a cost–​benefit perspective.
Effective prevention programs require information about population distributions
and determinants of incidence and prevalence of a given disorder; developmental tra-
jectories of disorders, including risk periods and remission patterns; and risk and pro-
tective factors that span individual, social, economic, and other structural variables,
including barriers to services. A basic principle in allocating resources across the popu-
lation is sensitivity to the “prevention paradox” (Midanik 2001). Because population
disease burden is usually due to the majority population segment with low to moderate
risk/​problems, as is the case for substance-​related problems, the greatest impact on
population health will come from allocating most resources to the middle of the distri-
bution rather than to the tails.
In the substance abuse area, however, services have tended to focus on the distribution
tails in the form of universal prevention delivered without regard to audience characteris-
tics, needs, or preferences (e.g. Project DARE (Ennett et al. 1994); Project STAR (Kaminski
et al. 2002)), or clinical treatment for the minority with serious to severe problems. Even
350  

350 Role of choice biases and choice architecture in behavioral economic strategies

in public health practice, risk-​based interventions remain the dominant approach (Frank
and Jepson 2013; Rose 1985). Selective prevention with high-​risk, susceptible groups
exemplifies this approach. This contrasts with population-​based strategies that intervene
“upstream” by targeting determinants of disease incidence that may variously involve envi-
ronmental, regulatory, healthcare systems or policy levels (Prado et al. 2013; Rose 1985).
In principle, this strategy has greater potential for population reach and impact, even if
individual benefits are smaller. Another advantage is that a population-​based strategy does
not require triaging individuals or risk groups into different interventions, which depends
on effective screening and decision algorithms for referral and intervention.
In summary, modern prevention for addictive and other health-​related behaviors is
grounded in a population perspective and builds on knowledge and applications for indi-
vidual behavior change. As summarized next, research on behavioral choice has identi-
fied a number of biases in human decision-​making that are sufficiently common to be
viewed as normative and hold promise for guiding improved interventions for addictive
behaviors that span clinical and public health strategies.

3  Normative biases in human decision-​making

Biased choice means that patterns of choice deviate from those predicted by expected
utility in economic analyses. Real people do not choose “rationally” in line with what
classic economic models predict, and their “irrational” deviations show reliable regulari-
ties. Early experimental work in operant psychology was instrumental in challenging
and refining economic models of intertemporal choice (e.g. Ainslie 1975)  and laid the
empirical foundation for behavioral economics, which is a merger of operant psychology
and consumer demand theory in microeconomics (Rachlin et al. 1981). Now an estab-
lished transdisciplinary field, modern behavioral economics offers organizing concepts,
common terminology, and diverse methodologies that have advanced knowledge and
application in many areas (US NIH 2010), including its original applied field of addic-
tive behaviors (Vuchinich and Heather 2003; Becker and Murphy 1988). Findings on
normative choice biases, described next, are particularly relevant for understanding and
intervening with addictive behaviors, the hallmark of which is a chronic preference for
short-​term rewards that are associated with delayed negative consequences.
Delay discounting. Four decades of behavioral economic research has repeatedly dem-
onstrated that the choice behavior of humans and animals is vulnerable to delay discount-
ing: shorter term outcomes are weighed more heavily than delayed outcomes, even when
the latter have higher overall value (Ainslie 1975; MacKillop et  al. 2011; Madden and
Bickel 2010). Contrary to classic economic theories that assume a constant rate of dis-
counting regardless of the intertemporal intervals of choice, behavioral economic studies
have consistently found that the form of the discount curve is hyperbolic or hyperbola-​
like. This means that the value of rewards changes dynamically as a function of time to
availability (Ainslie 1975; Chapman and Johnson 1995). Value increases sharply when
reward availability is imminent and can result in abrupt preference reversals away from
later rewards of higher value in favor of sooner rewards of lower value.
  351

Normative biases in human decision-making 351

Of greatest relevance to the present chapter is the large body of research indicating
higher discount rates among individuals exhibiting addictive behaviors, particularly
those who meet clinical diagnostic criteria for substance use or other addictive disorders.
Research showing higher discounting among persons engaged in substance misuse, over-
eating, gambling, and related risk behaviors (e.g. risky sex, poor financial planning) has
been reviewed narratively (Bickel and Marsch 2001; cf. Madden and Bickel 2010) and in
a recent meta-​analysis (MacKillop et al. 2011). A shared feature is chronic insensitivity to
control of current behavior by valuable, delayed outcomes in favor of repeated choices of
less valuable short-​term outcomes with delayed harmful effects. However, because most
relevant studies were cross-​sectional case-​control comparisons and were not longitudinal,
etiological relationships remain ambiguous concerning whether higher discounting pre-
cedes the onset of addictive behaviors, whether engaging in addictive behaviors results in
higher discounting, or some combination of the two (MacKillop et al. 2011).
Discount rates show variability both across and within individuals over time, and indi-
vidual discount rates should not be viewed as an immutable trait as in personality con-
ceptions and measures of impulsivity (MacKillop et al. 2011; Odum and Bauman 2010;
Tucker et  al. 2010). For example, steeper discounting is associated with lower income
and education and younger age (Green et al. 1996; Jaroni et al. 2004; Reimers et al. 2009),
but has unreliable associations with gender and personality measures of impulsivity (e.g.
Epstein et al. 2003; Swann et al. 2002). Moreover, heterogeneity exists among individuals
with a given addictive behavior pattern. Among such individuals, relatively lower dis-
counting or greater sensitivity to delayed rewards predicts positive outcomes of behavior
change attempts (e.g. MacKillop and Kahler 2009; MacKillop and Murphy 2007; Tucker
et al. 2009; Washio et al. 2011).
Discount rates also vary as a function of the context of choice and can be altered by
changes in that context. A robust generalization from experimental work is that prefer-
ence for a given activity depends on the other options available in the choice context
and on the constraints on access to them (e.g. price, time, effort to obtain; Rachlin et al.
1981). Thus, preference for addictive behaviors such as substance misuse can be reduced
by increasing the direct constraints on access to those behaviors and by enriching the
environment with higher-​valued alternatives that do not involve addictive behaviors
(Vuchinich and Heather 2003). These contextual manipulations are effective within cer-
tain boundary conditions (e.g. Odum and Baumann 2010)  and underlie classic supply
reduction and demand reduction drug control strategies, respectively.
Other choice biases. Research reviewed elsewhere (Odum and Baumann 2010; Tucker
et al. 2010) supports a range of additional choice biases, as follows, that have relevance for
designing interventions to improve health or addiction-​related outcomes:
◆ Sign effect: Positive outcomes are discounted more than negative outcomes in inter-
temporal choices that involve various gains and losses in health, financial, and envi-
ronmental/​ecological outcomes (e.g. Baker et al. 2003; Hardisty and Weber 2009).
◆ Magnitude effect: Smaller rewards are discounted more than larger rewards regardless
of other parameters such as delay to receipt (e.g. Baker et al. 2003; Hardisty et al. 2013).
352  

352 Role of choice biases and choice architecture in behavioral economic strategies

◆ Sequence effect: Sequences of outcomes that end in gains are preferred to those that end
in losses, even when the overall utilities of the sequences are the same; for example,
holding overall health constant during a given interval, sequences during the interval
that involved improving health were preferred over stable health or sequences that
involved declining health (Chapman 1996; Richardson et al. 2011).
◆ Delay duration or dynamic inconsistency effect: Discount rates are higher for proximal
future intervals compared to more distant intervals, including outcomes many years
in the future; for example, hypothetical public outcomes over very long delays (30–​
900 years) were discounted less than those occurring over shorter delays (1–​30 years;
Chapman 2001).
◆ Domain effect: Discount rates vary over commodity classes such as monetary, social,
and health-​related outcomes for reasons other than overall utility; for example, per-
sonal health outcomes were discounted differently than financial or environmental
(e.g. air quality) outcomes (Hardisty and Weber 2009).
Some biases have fairly straightforward implications for behavior change interventions
(Tucker et al. 2010; see section 4). For instance, sign and magnitude effects suggest plac-
ing emphasis on promoting changes in behavior that will lead to longer durations of full
health or greater health improvements, rather than on preventing behaviors associated
with declining health. Sequence effects suggest the value of contextualizing choices as
cohesive, temporally extended series of events instead of as choices involving brief dis-
crete events. Duration effects are particularly pertinent to real-​world healthcare system
and policy decisions, which can involve the allocation of present resources toward uncer-
tain outcomes years or decades into the future.
Intervention implications of other findings are more complex, including evidence that
deviation from expected utility varies across commodity domains; depend on whether
outcomes involve gains or losses; and may be affected by personal characteristics, includ-
ing addiction-​related and health status (Odum and Baumann 2010; Tucker et al. 2010;
see also section 4). For example, Weatherly and Terrell (2011) found that discount rates
differed among commodity domains that involved either tangible/​economically consum-
able goods (e.g. money, cigarettes, retirement income) or potential for personal or societal
betterment (e.g. ideal body image or dating partner, education legislation, medical treat-
ment). In a study that included smokers, problem drinkers, and normal controls, Bickel
et al. (2012) found that problem drinkers exhibited “altruism” by showing lower discount-
ing in a reward condition that involved social groups compared to an individual choice
condition, whereas smokers evidenced higher discounting across conditions. In a similar
vein, Richardson et al. (2011) found that respondents valued ameliorating ill health of
others, especially for severe conditions.
Observed variation across commodity domains deserves further study because real-​
world choices involving health and addictive behaviors typically involve multiple domains
(Bickel and Vuchinich 2000; Vuchinich and Heather 2003). Unlike choice scenarios in
experimental preparations between small reward amounts available now versus larger
amounts of the same reward available later, addiction-​related choices involve choosing
  353

Implications for improving intervention strategies 353

between engaging in qualitatively different behaviors. For example, a recovering sub-

stance misuser repeatedly chooses between engaging in drug use now, which will lead to
later negative consequences, and engaging in sobriety-​supportive behaviors (e.g. attend-
ing mutual help groups), which will lead to valuable benefits in areas of life-​functioning
disrupted by substance misuse. Better understanding of domain effects is basic to improv-
ing interventions for addictive behaviors.
Another feature of real-​world choices is that they often involve a mix of potential
gains and losses. Research on choices involving losses lags behind research on choices
involving gains and is revealing inconsistencies. Whereas studies of intertemporal
choices involving rewards showed orderly delay discounting effects, studies of choices
involving losses, pain, punishment, or other emotional distress demonstrated a “pre-
sent” bias in favor of experiencing the punishment or loss sooner rather than later (e.g.
Hardisty et  al. 2013; Harris 2012; Story et  al. 2013). Furthermore, greater heteroge-
neity in individual time preferences is common in choices involving losses compared
to rewards, and time preferences for losses and rewards are not necessarily correlated
(Harris 2012).
Another under-​investigated aspect of choices involving loss and punishment concerns
the role of negative affective and health states (Metcalfe and Mischel 1999). People gener-
ally discount future health outcomes more steeply when they are in “hot” emotional or
deprived states (craving a drug, hungry) than when they are in “cold” states (Chapman
2005; Loewenstein 2005). “Dread” of future pain also appears subject to discounting pro-
cesses (Story et al. 2013). In addition, individuals in poor states of health tend to discount
health outcomes less steeply (van der Pol and Cairns 2001), and persons living in poorer
countries tend to prefer immediate but inferior health outcomes compared to persons
living in wealthier nations (Robberstad and Cairns 2007). These relationships pose chal-
lenges for creating decision-​making contexts that promote optimal pro-​health choices
when persons are in poor health or pain and have limited income.
In summary, human behavioral choice shows reliable regularities that involve devia-
tions in utility as predicted by rational economic theory. These biases are common and
need to be taken into account when devising interventions to promote beneficial deci-
sions that involve real-​world outcomes. Behavioral economic research has advanced
beyond early experimental preparations involving hypothetical choices of different
monetary rewards available sooner vs. later to modeling complex choices among quali-
tatively different reinforcers. Evidence of interactions among multiple parameters of
intertemporal choice is accumulating (e.g. domain and sequence effects, asymmetries
between choices involving gains and losses), but the research is limited (Baker et  al.
2003; Hardisty et al. 2013; see section 5). This limitation is a general one and not specific
to the health literature.

4  Implications for improving intervention strategies

Established intervention strategies. Interventions for substance-​related problems have
matured sufficiently to comprise a spectrum of strategies that increasingly map onto the
354  

354 Role of choice biases and choice architecture in behavioral economic strategies

epidemiology of substance-​related problems. Although not mutually exclusive, interven-

tion strategies can be placed in one of three general categories:
◆ Drug supply reduction strategies seek to reduce drug-​related problems by constraining
drug availability through border control, drug interdiction, and policing; imposition of
criminal penalties for possession and trafficking; increasing other costs (e.g. through taxa-
tion, liquor outlet restrictions; see Anderson et al. 2009; Cook and Moore 2002); and more
recently monitoring physician prescribing practices (Morgan et al. 2013). This costly strat-
egy, dominant in the US, has likely reached maximum effectiveness and has had unin-
tended negative consequences (black/​grey market development, differential negative
impact on African-​American males; Pacula et al. 2014; Warren et al. 2011; Zawilska 2011).
◆ Drug demand reduction strategies include enriching the environment with positive
activities that compete with drug use; expanding prevention and early intervention
(Agerwala and McCance-​Katz 2012); and increasing treatment capacity and ease of
access (McKellar et  al. 2012). This arena is where opportunities remain to improve
outcomes and population health. Evidence-​based treatments (Miller 2009) serve the
minority with serious dependence, but treatment slots are limited, access is often “high
threshold” with long waiting times, and treatment can be costly, inconvenient, and
stigmatizing (IOM 1998). Environmental enrichment involving structural interven-
tions is in its infancy (Solinas et al. 2010).
◆ Harm reduction strategies aim to reduce negative consequences and costs of substance
misuse, but not necessarily use per se (MacCoun and Reuters 2001; Marlatt et al. 1997).
Treatment rather than incarceration is offered for nonviolent drug use and posses-
sion, including providing opiate pharmaco-​substitutes for the severely dependent.
Successful tactics include reduction of transmission of diseases associated with drug
use (e.g. HIV/​AIDS, hepatitis, sepsis) through clean needle exchanges (Des Jarlais
et al. 1996) and reduction of fatal opiate overdoses by making opiate-​antagonists avail-
able to first responders (Davis et al. 2014) and passing Good Samaritan laws (Network
for Public Health Law 2014) that protect opiate users who aid overdose victims. The
approach has worked well in several European countries, Australia, and Canada, but
remains at odds with U.S. drug control policies (MacCoun and Reuters 2001).
Each distinct strategy relies on somewhat different “program theories” about the control-
ling variables of substance misuse that can be framed using established behavior change
principles (Skinner 1953): (1) behavior is controlled by its consequences; (2) punishment
temporarily suppresses behavior and then only if it is swift and consistent; (3) incentivizing
positive change works better than punishing negative acts and outcomes; and (4) contin-
gent incentives made available on variable schedules are most effective for maintaining
change. Supply reduction strategies rely heavily on punishment, harm reduction strategies
rely heavily on incentivizing change while minimizing punishment, and demand reduction
strategies often use both rewarding and punishing contingencies delivered on intermittent
schedules to initiate and maintain behavior change (see Budney and Higgins 1998).
Research on behavioral choice adds another important consideration to this list of
behavioral principles. Because human choice is fraught with irrational elements that result
  355

Implications for improving intervention strategies 355

in deviations from expected utility, these newer findings, when combined with established
change principles, offer novel ways to improve intervention strategies and outcomes.
Specifically, choice architecture strategies use these findings to craft decision-​making
contexts that enhance the probability of beneficial choices and reduce the probability of
poor choices (e.g., Loewenstein et al. 2007). Although addiction-​related applications are
in their infancy, they hold promise because choice biases such as discounting are accentu-
ated in persons exhibiting addictive behaviors or disorders.
Within the incipient applied literature, two general approaches can be identified. The
first approach seeks to remediate the choice biases or provide incentives for better choices
(e.g. Bickel et  al. 2011; Budney and Higgins 1998). The second approach accepts that
biased choices are normative and structures health messages, choices, interventions,
and contexts in ways that use the biases to promote good choices and outcomes (e.g.
Loewenstein et al. 2007; Ortendahl and Fries 2005; Thaler and Sunstein 2008). The former
strategy is exemplified by behavior modification and psychotherapeutic interventions,
whereas the latter strategy is found in some behavioral economic and public health inter-
ventions (Marlatt et al. 1997; Tucker et al. 2010).
Remediating choice biases. The first strategy aimed at reducing the negative effects of
biased choices, typically delay discounting, is better established and includes the follow-
ing intervention approaches and targets:
◆ Incentivize better choices and outcomes by offering tangible incentives or privileges. For
example, offer incentives for abstinence (e.g. Budney and Higgins 1998; Petry et  al.
2000, 2012; Schumacher et al. 2007) or for engaging in therapeutic activities that sup-
port pro-​health behaviors and environments that promote the value of long-​term
rewards of sobriety (e.g. Bickel and Marsch 2001; Lei et al. 2012; Petry et al. 2012).
◆ Help individuals take a longer view of their personal cost-​benefit situation to promote
choices of health-​related options likely to yield valuable benefits longer term (Logue 2000;
Lash et al. 2013). This is integral to motivational interviewing techniques (Miller and
Rollnick 2012), which highlight longer views of the future to help shift behavior toward
self-​controlled sober patterns. Self-​monitoring of addictive behaviors similarly helps
link discrete daily choices to longer-​term behavior patterns that are less immediately
preferred, but have greater long-​term value (Rachlin 1995; Simpson and Vuchinich
2000; Tucker et al. 2012).
◆ Intervene directly to reduce the choice bias. Bickel and colleagues (Bickel et  al. 2011;
Radu et al. 2011; cf. Black and Rosen 2011) have shown that working memory training
lowered discount rates among stimulant users in treatment, although effects on drug
use outcomes were not reported. Exposure to contingency management procedures
similarly decreased delay discounting on laboratory tasks in smokers (Mueller et al.
2009; Yi et al. 2008) and opioid-​dependent persons (Landes et al. 2012). These find-
ings indicated that impulsive choice styles can be changed, but whether such changes
reduce addictive behaviors remains unclear and merits investigation.
◆ Use normative feedback to create a new framework for choices involving socially mean-
ingful behavior. For example, college students tend to overestimate the extent of
356  

356 Role of choice biases and choice architecture in behavioral economic strategies

drinking and other risk behaviors among their peers and may underestimate their
own risks. Providing accurate information through assessment of personal risk taking
and feedback on relevant group norms can promote desirable reductions in drinking
and other risk taking (e.g. Moreira et  al. 2009; Perkins 2003). Such framing effects
can change maladaptive behaviors by altering the surrounding social contexts within
which choices occur.
Using choice biases to promote healthy decisions and outcomes. This strategy accepts that
biased choices are normative and, rather than intervening to change them, focuses on struc-
turing health messages, choices, interventions, and contexts in ways that use the biases to
promote good choices and outcomes. The approach has firm traction in financial retire-
ment and savings programs (Beshears et al. 2006; Thaler and Benartzi 2004). A prototypical
approach involves changing the default choices regarding savings, such that participation
in a plan is automatic unless explicitly declined, as opposed to requiring affirmative opt-​in.
Although less well developed in addiction and other health-​relevant applications, the
approach is promising, as illustrated by the following examples:
◆ Positive outcomes are discounted more than negative outcomes, and outcomes in
proximal future intervals are discounted more than those in distant intervals. These
sign and duration effects can be exploited beneficially by using contingency contract-
ing to make the negative consequences of unhealthy behaviors more immediate and
certain (Logue 2000; Lash et  al. 2013). Similarly, health messages can be framed as
long-​term future losses rather than short-​term gains or losses (e.g. Lazaro et al. 2001;
Ortendahl and Fries 2005).
◆ Outcome sequences of similar overall utility that end in gains are preferred to those
that end in losses. Sequencing effects can be used therapeutically by linking pro-​health
behaviors in a concrete manner that highlights accruing benefits (e.g. by charting, self-​
monitoring; Logue 2000; Simpson and Vuchinich 2000; Tucker et al. 2012).
◆ Persons with more serious substance-​related problems tend to have steeper discount
rates and are more responsive to shorter-​term contingencies. This bias can be taken
advantage of to promote recovery by providing “treatment on demand” with mini-
mal delays in order to increase utilization when substance misusers experience moti-
vational shifts toward recovery-​seeking behaviors (Marlatt et  al. 1997; Tucker and
Davison 2000). Such shifts are common but often transitory, and long treatment wait-
ing times fail to capitalize on these “teachable moments.”
◆ Discounting of private or personal outcomes appears to be higher than discounting
of social outcomes (Bickel et  al. 2012), but when choices are made over long inter-
vals, personal and social outcomes tend to be discounted similarly (Lazaro et  al.
2001). These findings suggest ways to arrange choices that involve different domains
and timeframes to promote pro-​health choices and benefits for individuals and larger
groups. Moreover, public decision-​making about funding health and social programs
appeared to be influenced by many of these same biases, including sign, duration,
magnitude, and framing effects (Tucker et al. 2010; West et al. 2003).
  357

Future directions for drug-control strategies and research 357

More generally, Loewenstein et al. (2007) argued that a desirable feature of the second
strategy is that it can benefit persons with more extreme choice biases without overly con-
straining the choices of more rational decision-​makers. Termed asymmetric paternalism,
the strategy helps those susceptible to choice biases make better choices without otherwise
infringing on freedom of choice for less biased decision-​makers. A prototypical example
based on discounting is changing the order of food presentation in a cafeteria line so that
healthy foods, not desserts, are presented first. This does not change the overall avail-
ability of desserts, but the switch helps overweight people with limited self-​control make
better food choices. Another example involves “opt-​out” HIV testing policies. Originally,
persons had to “opt in” and give consent for HIV testing. This resulted in suboptimal test-
ing rates, especially in resource-​poor, high-​prevalence areas (World Health Organization
(WHO) 2004). Testing guidelines (Branson et al. 2006; WHO/​UNAIDS 2007) now rec-
ommend that persons who make contact with healthcare systems be informed that they
will be tested unless they explicitly request to opt-​out. The overall opportunity for testing
is unchanged, but opt-​out testing has significantly improved testing rates during ante-​
natal care (Kennedy et al. 2013) and in STI clinics (Heijman 2009). A randomized HIV
prevention trial with young adults in sub-​Saharan Africa (Baisley et al. 2012) found sig-
nificantly higher testing uptake among ten opt-​out communities (90.6 percent) compared
to ten opt-​in communities (60.5 percent).

5  Future directions for drug-​control strategies

and research
Obviously, various approaches that take choice behavior into account are not mutu-
ally exclusive, and what mix of options works best remains to be determined. This may
include incorporating relevant features into existing supply, demand, and harm reduction
interventions and developing newer approaches that seek to remediate choice biases or
manipulate the architecture of choice by framing options to help people make choices in
their best interests.
One of the most exciting opportunities offered by choice architecture is the potential
for scalability, or the capacity of systems to reach and serve expanding numbers of people
and thus increase impact on population health (Hollands et al. 2013; Thaler and Sunstein
2008). Although intensive, individual treatment can be highly efficacious, it is costly, high-​
threshold, and often received late in a disease process when many harms and costs have
already been incurred and may not be completely remediable. Thus, reach is inherently
limited (Tucker and Grimley 2011). Changes in default options (e.g. treatment on demand,
opt-​out HIV testing) can increase opportunities to make pro-​health decisions and influence
determinants in ways that decrease population incidence of clinical disorders. If properly
crafted, they are suitable for implementation at healthcare systems, policy, regulatory, and
legal levels with greater reach than clinical treatments or high-​risk prevention programs.
Although growing rapidly, the transdisciplinary behavioral economic research base
remains uneven in scope, methodology, level of analysis, and quality. Some of the more
358  

358 Role of choice biases and choice architecture in behavioral economic strategies

intriguing findings with applied significance involve domain effects among different com-
modity classes. Because disruption in multiple areas of life-​health functioning is a central
feature of substance misuse, understanding domain effects is basic to developing real-​
world choice contexts and interventions to promote pro-​health decisions. However, the
choice preparations used in human research are almost exclusively hypothetical and often
require participants to imagine and judge scenarios that they have not experienced in
real life. Although a modest literature shows that choice patterns of hypothetical and real
money choices are similar, comparable evidence is lacking with respect to social, health,
and other valuable non-​tangible commodities. Furthermore, research on whether per-
formance on laboratory discounting tasks predicts real-​world health behaviors is limited
with mixed results (Story et al. 2014). Thus, domain effects remain poorly understood,
and it is unclear whether they may reflect intrinsic functional differences between com-
modities, methodological differences in assessing health and other commodity prefer-
ences, or both (Chapman 2002; Tucker et al. 2010).
Another consideration in developing health and addiction-​related interventions is to
avoid assuming that expected utility is always maximized by reducing delay discounting
and by orienting choices around future outcomes that are almost always probabilistic.
Sensitivity to the changing constraints on and availability of different commodities in
choice contexts is basic to wise decision-​making, and the temporal units over which util-
ity is maximized will be variable.
More generally, there is continuing debate about theoretical accounts of discounting
and other choice biases (e.g. Hardisty et al. 2013; Story et al. 2014) and how to conceptual-
ize and organize the evidence base for choice architecture interventions involving health
outcomes (e.g. Hollands et al. 2013). Although beyond the scope of this chapter, some of
these issues have applied significance. For example, Hollands et al. (2013) argued that cur-
rent evidence for scalability and population impact of choice architecture interventions
is very limited, despite endorsement of the approach among European policy-​makers.
However, the scope of their literature review was restricted to simple manipulations of
small “micro-​environments” such as restaurants and elevators that have no grounding
in behavioral economic research on behavioral choice and choice biases. In contrast, this
forms the foundation of asymmetric paternalism (Loewenstein et al. 2007), which appears
to provide a superior framework for applications to addictive behaviors given the promi-
nent and accentuated role of choice biases including delay discounting in such disorders.
Finally, intriguing evidence is emerging about the role of language and culture in inter-
temporal choice. In an international study of temporally distributed economic and health
behaviors (Chen 2013), speakers of languages with little or no grammatical distinction
between present and future (e.g. Mandarin Chinese) were more likely to engage in future-​
oriented behaviors than speakers of languages that strongly distinguish present and future
(e.g. English). Speakers of weak future-​time reference (FTR) languages were more likely
to save, be physically active, and use condoms; accumulated more wealth by retirement;
and were less likely to smoke or be obese. Effects ranged from individual financial prac-
tices to national savings rates. Although it remains ambiguous whether language is a
  359

Future directions for drug-control strategies and research 359

causal variable or whether this FTR distinction reflects other cultural variables that influ-
ence savings and health behaviors, languages that bring the future into the present were
associated with more optimal choice patterns than those strongly demarcating the future
from the present. Western-​dominated delay discounting research explicitly demarcates
past, present, and future outcomes, and the Chen (2013) findings raise questions about
fundamental orienting assumptions.
In conclusion, the development of coordinated multi-​component intervention strate-
gies built on behavioral economic theory and research is now within reach. In addition
to guiding numerous individual-​level interventions, behavioral economics generally and
choice architecture strategies specifically can be implemented within healthcare systems
and communities and at a policy level. A scientifically grounded range of services based
on concepts and findings from behavioral economics can benefit larger numbers of peo-
ple, which is a central orienting assumption and goal in public health.

Acknowledgments
Manuscript preparation was supported in part by US National Institutes of Health/​National
Institute on Alcohol Abuse and Alcoholism grant no. 1 R01 AA017880-​01A1 and by coop-
erative agreement #U48-​DP001915 from the Centers for Disease Control and Prevention.

References
Agerwala, S.M. and McCance-​Katz, E.F. (2012). Integrating screening, brief intervention, and referral
to treatment (SBIRT) into clinical practice settings: a brief review. Journal of Psychoactive Drugs, 44,
307–​17.
Ainslie, G. (1975). Specious reward: a behavioral theory of impulsiveness. Psychological Bulletin, 82,
463–​96.
Anderson, P., Chisohlm, D., and Fuhr, D.C. (2009). Effectiveness and cost-​effectiveness of policies and
programs to reduce the harm caused by alcohol. Lancet, 373, 2234–​46.
Baisley, K., Doyle, A.M. and Changalucha, J., et al. (2012). Uptake of voluntary counselling and testing
among young people participating in an HIV prevention trial: comparison of opt-​out and opt-​in
strategies. PLoS One, 7(7), e42108.
Baker, F., Johnson, M.W., and Bickel, W.K. (2003). Delay discounting in current and never-​before
cigarette smokers: similarities and differences across commodity, sign, and magnitude. Journal of
Abnormal Psychology, 112, 382–​92.
Becker, G. and Murphy, J.M. (1988). A theory of rational addiction. Journal of Political Economy, 4,
675–​700.
Beshears, J., Choi, J.J., Laibson, D., and Madrian, B.C. (2006). The importance of default options
for retirement savings outcomes: evidence from the United States. In: J. Brown, J. Liebman, and
D.A. Wise (eds), (2006). Social Security Policy in a Changing Environment. Chicago: University of
Chicago Press, pp. 167–​95.
Bickel, W.K., Jarmolowicz, D.P., Mueller, E.T., Franck, C.T., Carrin, C., and Gatchalian, K.M.
(2012). Altruism in time: social discounting differentiates smokers from problem drinkers.
Psychopharmacology, 224, 109–​20.
Bickel, W.K. and Marsch, L.A. (2001). Toward a behavioral economic understanding of drug
dependence: delay discounting processes. Addiction, 96, 73–​86.
360  

360 Role of choice biases and choice architecture in behavioral economic strategies

Bickel, W.K. and Vuchinich, R.E. (eds) (2000). Reframing Health Behavior Change with Behavioral
Economics. Mahwah, NJ: Lawrence Erlbaum Associates.
Bickel, W.K., Yi, R., Landes, R.D., Hill, P.F., and Baxter, C. (2011). Remember the future: working
memory training decreases delay discounting among stimulant addicts. Biological Psychiatry, 69,
260–​65.
Black, A.C. and Rosen, M.I. (2011). A money management-​based substance use treatment increases
valuation of future rewards. Addictive Behaviors, 36, 125–​28.
Branson, B.M., Handsfield, H.H., Lampe, M.A., et al. (2006). Revised recommendations for HIV
testing of adults, adolescents, and pregnant women in health-​care settings. Morbidity and Mortality
Weekly Report, 55(R14), 1–​17.
Budney, A.J. and Higgins, S.T. (1998). National Institute on Drug Abuse Therapy Manuals for Drug
Addiction: Manual No. 2: A Community Reinforcement Plus Vouchers Approach: Treating Cocaine
Addiction. NIH Pub. No. 98-​4309. Rockville, MD: National Institute on Drug Abuse.
Chapman, G.B. (1996). Temporal discounting and utility for health and money. Journal of Experimental
Psychology: Learning, Memory, and Cognition, 22, 771–​91.
Chapman, G.B. (2001). Time preferences for the very long term. Acta Psychologica, 108, 95–​116.
Chapman, G.B. (2002). Your money or your health: time preferences and trading money for health.
Medical Decision Making, 22, 410–​16.
Chapman, G.B. (2005). Short-​term cost for long-​term benefit: time preference and cancer control.
Health Psychology, 24, S41–​8.
Chapman, G.B. and Johnson, E.J. (1995). Preference reversals in monetary and life-​expectancy
evaluations. Organizational Behavior and Human Decision Processes, 62, 300–​17.
Chen, M.K. (2013). The effect of language on economic behavior: evidence from savings rates, health
behaviors, and retirement assets. American Economic Review, 103(2), 690–​731.
Cook, P.J. and Moore, M.J. (2002). The economics of alcohol abuse and alcohol-​control policies. Health
Affairs, 21, 120–​33.
Davis, C.S., Ruiz, S., Glynn, P., Picariello, G., and Walley, A.Y. (2014). Expanded access to naloxone
among firefighters, police officers, and emergency medical technicians in Massachusetts. American
Journal of Public Health, 104: e7–​9.
Des Jarlais, D.C., Marmor, M., Paone, D., et al. (1996). HIV incidence among injecting drug users in
New York City syringe-​exchange programs. Lancet, 348, 987–​91.
Ennett, S.T., Tobler, N.S. and Ringwalt, C.L., et al. (1994). How effective is drug abuse resistance
education? A meta-​analysis of Project DARE outcome evaluations. American Journal of Public
Health, 84, 1394–​401.
Epstein, L.H., Richards, J.B., Saad, F., Paluch, R., Roerman, J., and Lerman, C. (2003). Comparison of two
measures of delay discounting in smokers. Experimental and Clinical Psychopharmacology, 11, 131–​38.
Frank, J. and Jepson, R. (2013). High-​risk versus population prevention strategies for NCDs: Geoffrey
Rose revisited in the twenty-​first century. In: D.V. McQueen (ed.), Global Handbook on
Noncommunicable Diseases and Health Promotion. New York: Springer, pp. 3–​19.
Green, L., Myerson, J., Lictman, D., Rosen, S., and Fry, A. (1996).Temporal discounting in choice
between delayed rewards: the role of age and income. Psychology of Aging, 11, 79–​84.
Hardisty, D.J. and Weber, E.U. (2009). Discounting future green: money versus the environment.
Journal of Experimental Psychology: General, 138, 239–​340.
Hardisty, D.J., Appelt, K.C., and Weber, E.U. (2013). Good or bad, we want it now: fixed-​cost present
bias for gains and losses explains magnitude asymmetries in intertemporal choice. Journal of
Behavioral Decision Making, 26, 348–​61.
Harris, C.R. (2012). Feelings of dread and intertemporal choice. Journal of Behavioral Decision Making,
25, 13–​28.
  361

Future directions for drug-control strategies and research 361

Heijman, R.L. J., Stolte, I.G., Thiesbrummel, H.F.J. et al. (2009). Opting out increases HIV testing in a
large sexually transmitted infections clinic. Sexually Transmitted Infections, 85, 249–​55.
Hollands, G.J., Shemilt, I., Maerteau, T.M., et al. (2013). Altering micro-​environments to change
population health behavior: towards an evidence base for choice architecture interventions. BMC
Public Health, 13, 1218.
Humphreys, K. and McLellan, A.T. (2011). A policy-​oriented review of strategies for improving the
outcomes of services for substance use disorder patients. Addiction, 106, 2058–​66.
Humphreys, K. and Tucker, J.A. (2002). Toward more responsive and effective intervention systems for
alcohol-​related problems. Addiction, 97, 126–​32.
Institute of Medicine (1990). Broadening the Base of Treatment for Alcohol Problems. Washington,
D.C.: Institute of Medicine.
Institute of Medicine (1994). Reducing Risks for Mental Disorders: Frontiers for Prevention Intervention
Research. Washington, D.C.: Institute of Medicine.
Institute of Medicine Committee on Community-​Based Drug Treatment (1998). Bridging the
Gap between Practice and Research: Forging Partnerships with Community-​Based Drug and
Alcohol Treatment, S. Lamb, M.R. Greenlick, and D. McCarty (eds). Washington, D.C.: National
Academies Press.
Jaroni, J., Wright, S., Lerman, C., and Epstein, L.H. (2004). Relationship between education and delay
discounting in smokers. Addictive Behaviors, 29, 1171–​76.
Kaminski, R.A., Stormshak, E.A., Good, R.H.3rd, and Goodman, M.R. (2002). Prevention of
substance abuse with rural Head Start children and families: results of Project STAR. Psychology of
Addictive Behaviors, 16(4 Suppl.), S11–​S26.
Kennedy, C.E., Fonner, V.A., Sweat, M.D., Okero, F.A., Baggaley, R., and O’Reilly, K.R. (2013).
Provider-​initiated HIV testing and counselling in low-​and middle-​income countries: a systematic
review. AIDS and Behavior, 17, 1571–​90.
Landes, R.D., Christensen, D.R., and Bickel, W.K. (2012). Delay discounting decreases in those completing
treatment for opioid dependence. Experimental and Clinical Psychopharmacology, 20, 302–​309.
Lash, S.J., Burden, J.L., Parker, J.D., Stephens, R.S., and Budney, A.J. (2013.) Contracting, prompting and
reinforcing substance use disorder continuing care. Journal of Substance Abuse Treatment, 44, 449–​56.
Lazaro, A., Barberan, R., and Rubio, E. (2001). Private and social time preferences for health and
money: an empirical estimation. Health Economics, 10, 351–​56.
Lei, H., Nahum-​Shani, I., Lunch, K., Oslin, D., and Murphy, S.A. (2012). A “SMART” design for
building individualized treatment sequences. Annual Review of Clinical Psychology, 8, 21–​48.
Loewenstein, G. (2005). Hot–​cold empathy gaps and medical decision making. Health Psychology, 24,
S49–​S56.
Loewenstein G., Brennan T., and Volpp, K. (2007). Asymmetric paternalism to improve health
behaviors. Journal of the American Medical Association, 298, 2415–​17.
Logue, A.W. (2000). Self-​control and health behavior. In: W.K. Bickel and R.E. Vuchinich (eds),
Reframing Health Behavior Change with Behavioral Economics. Mahwah, NJ: Lawrence Erlbaum
Associates, pp. 167–​92.
MacKillop, J., Few, L.R., Ray, L.A., Sweet, L.H., and Munafo, M.R. (2011). Delayed reward discounting
and addictive behavior: a meta-​analysis. Psychopharmacology, 216, 305–​21.
MacKillop, J. and Kahler, C.W. (2009). Delayed reward discounting predicts smoking cessation
treatment outcome in heavy drinkers. Drug and Alcohol Dependence, 104, 197–​203.
MacKillop, J. and Murphy, J.G. (2007). A behavioral economic measure of demand for alcohol predicts
brief intervention outcomes. Drug and Alcohol Dependence, 89, 227–​33.
Madden, G.J. and Bickel, W.K. (eds) (2010). Impulsivity: The Behavioral and Neurological Science of
Discounting. Washington, D.C.: APA Books.
362  

362 Role of choice biases and choice architecture in behavioral economic strategies

Marlatt, G.A., Tucker, J.A., Donovan, D.M., and Vuchinich, R.E. (1997). Help-​seeking by substance
abusers: the role of harm reduction and behavioral economic approaches to facilitate treatment
entry and retention. In: L.S. Onken, J.D. Blaine, and J.J. Boren (eds), Beyond the Therapeutic
Alliance: Keeping the Drug Dependent Individual in Treatment NIDA Research Monograph No. 165.
Rockville, MD: National Institute on Drug Abuse, pp. 44–​84.
MacCoun, R.J. and Reuters, P. (2001). Drug War Heresies: Learning from Other Vices, Times, and Places.
Cambridge: Cambridge University Press.
McGovern, M.P. and Carroll, K.M. (2003). Evidence-​based treatments for substance use disorders.
Psychiatric Clinics of North America, 26, 991–​1010.
McKellar, J., Austin, J., and Moos, R. (2012). Building the first step: a review of low-​intensity
interventions for stepped care. Addiction Science & Clinical Practice, 7, 26.
Metcalfe, J. and Mischel, W. (1999). A hot/​cool-​system analysis of delay of gratification: Dynamics of
willpower. Psychological Review, 106, 3–​19.
Midanik, L.T. (2001). Where should alcohol problems prevention be targeted? Comments on Gmel
et al. Addiction, 96, 511.
Miller, P.W. (2009). Evidence-​Based Addiction Treatment. Waltham, MA: Academic Press.
Miller, W.R. and Rollnick, S. (2012). Motivational Interviewing: Helping People Change (3rd ed.).
New York: Guilford.
Morgan, L., Weaver, M., Sayeed, Z., and Orr, R. (2013). The use of prescription monitoring programs
to reduce opioid diversion and improve patient safety. Journal of Pain and Palliative Care
Pharmacotherapy, 27, 4–​9.
Moreira, T., Smith, L., and Foxcroft, D. (2009) Social norms interventions to reduce alcohol misuse
in University or College students. Cochrane Database of Systematic Reviews, 3 (CD006748).
doi: 10.1002/​14651858.CD006748.pub2
Mueller, E.T., Landes, R., Kowal, B., et al. (2009). Delay of smoking gratification as a laboratory model
of relapse: effects of incentives for not smoking, and relationship to measures of executive function.
Behavioral Pharmacology, 20, 461–​73.
Network for Public Health Law (2014). Legal interventions to reduce overdose mortality: naloxone
access and overdose good Samaritan laws. [Online] (Updated May 2014). https://​www.
networkforphl.org/​_​asset/​qz5pvn/​network-​naloxone-​10-​4.pdf. Accessed 20/​09/​2014.
Odum, A.L. and Baumann, A.A. (2010). Delay discounting: state and trait variable. In G.J. Madden and
W.K. Bickel (eds), Impulsivity: The Behavioral and Neurological Science of Discounting. Washington,
D.C.: APA Books.
Ortendahl, M. and Fries, J.F. (2005). Faming health messages based on anomalies in time preference.
Medical Science Monitor, 11, RA253–​6.
Pacula, R.L., Kilmer, B., Wagenaar, A.C., Chaloupka, F.J., and Caulkins, J.P. (2014). Developing public
health regulations for marijuana: lessons from alcohol and tobacco. American Journal of Public
Health, 104, 1021–​28.
Perkins, H.W. (eds) (2003). The Social Norms Approach to Preventing School and College Age Substance
Abuse. New York: Jossey-​Bass.
Petry, N.M., Martin, B., Cooney, J.L., and Kranzler, H.R. (2000). Give them prizes, and they will
come: contingency management for treatment of alcohol dependence. Journal of Consulting and
Clinical Psychology, 68, 250–​57.
Petry, N.M., Barry, D., Alessi, S.M., Rounsaville, B.J., and Carroll, K.M. (2012). A randomized trial
adapting contingency management targets based on initial abstinence status of cocaine-​dependent
patients. Journal of Consulting and Clinical Psychology, 80, 276–​85.
Prado, G., Lightfoot, M., and Brown C.H. (2013). Macro-​level approaches to HIV prevention among ethnic
minority youth: state of the science, opportunities, and challenges. American Psychologist, 68(4), 286–​99.
  363

Future directions for drug-control strategies and research 363

Rachlin, H. (1995). Self-​control: beyond commitment. Behavioral and Brain Sciences, 18, 109–​59.
Rachlin, H., Battalio, R., Kagel, J., and Green, L. (1981). Maximization theory in behavioral
psychology. Behavioral and Brain Sciences, 4, 371–​88.
Radu, P.T., Yi, R., Bickel, W.K., Gross, J.J., and McClure, S.M. (2011). A mechanism for reducing delay
discounting by altering temporal attention. Journal of the Experimental Analysis of Behavior, 96,
363–​85.
Reimers, S., Maylor, E.A., Stewart, N., and Chater, N. (2009). Associations between a one-​shot delay
discounting measure and age, income, education and real-​world impulsive behavior. Personality and
Individual Differences, 47, 973–​78.
Richardson, J.R. J., McKie, J., Peacock, S.J., and Iezzi, A. (2011). Severity as an independent determinant
of the social value of a health service. European Journal of Health Economics, 12, 163–​74.
Robberstad, B. and Cairns, J. (2007). Time preferences for health in northern Tanzania: an empirical
analysis of alternative discounting methods. Pharmacoeconomics, 25, 73–​88.
Rose, G. (1985). Sick individuals and sick populations. International Journal of Epidemiology, 14, 32–​38.
Schumacher, J.E., Milby, J.B., Wallace, D., et al. (2007). Meta-​analysis of day treatment and
contingency-​management dismantling research: Birmingham Homeless Cocaine Studies (1990–​
2006). Journal of Consulting and Clinical Psychology, 75, 823–​28.
Simpson, C.A. and Vuchinich, R.E. (2000). Temporal discounting in the value of objects of
choice: discounting, behavior patterns, and the value of health behavior. In: W.K. Bickel and
R.E. Vuchinich (eds), Reframing Health Behavior Change with Behavioral Economics. Mahwah,
NJ: Lawrence Erlbaum Associates, pp. 193–​218.
Skinner, B.F. (1953). Science and Human Behavior. New York: Free Press.
Solinas, M., Thiriet, N., Chauvet, C., and Jaber, M. (2010). Prevention and treatment of drug addiction
by environmental enrichment. Progress in Neurobiology, 92, 572–​92.
Springer, F. and Phillips, J.L. (2006). The IOM model: A tool for prevention planning and
implementation. Prevention Tactics, 8, 1–​7.
Story, G.W., Vlaev, I., Seumour, B., et al. (2013). Dread and the disvalue of future pain. PLOS Biology, 9,
e1003335.
Story, G.W., Vlaev, I., Seymour, B., et al. (2014). Does temporal discounting explain unhealthy
behavior? A systematic review and reinforcement learning perspective. Frontiers in Behavioral
Neuroscience, 8, 76. DOI: 10.3389/​fnbeh.2014.00076.
Swann, A.C., Bjork, J.M., Moeller, F.G., and Dougherty, E.M. (2002). Two models of
impulsivity: relationships to personality traits and psychopathology. Biological Psychiatry, 15, 988–​94.
Thaler, R.H. and Benartzi, S. (2004). Save More Tomorrow™: using behavioral economics to increase
employee saving. Journal of Political Economy, 112 (S1), S164–​S87.
Thaler, R.H. and Sunstein, C.R. (2008). Nudge: Improving Decisions about Health, Wealth, and
Happiness. New York: Penguin Books.
Tucker, J.A. and Davison, J.W. (2000). Waiting to see the doctor: the role of time constraints in the
utilization of health and behavioral health services. In: W.K. Bickel and R.E. Vuchinich (eds),
Reframing Health Behavior Change with Behavioral Economics. Mahwah, NJ: Lawrence Erlbaum
Associates, pp. 219–​64.
Tucker, J.A. and Grimley, D.M. (2011). Public Health Tools for Practicing Psychologists. Ashland,
OH: Hogrefe & Huber.
Tucker, J.A. and Simpson, C.A. (2011). The recovery spectrum: from self-​change to seeking treatment.
Alcohol Research & Health, 33, 371–​79.
Tucker, J.A., Simpson, C.A., and Khodneva, Y. (2010). The role of time and delay in health decision
making. In: G.J. Madden and W.K. Bickel (eds), Impulsivity: The Behavioral and Neurological Science
of Discounting. Washington, D.C.: APA Books, pp. 243–​72.
364  

364 Role of choice biases and choice architecture in behavioral economic strategies

Tucker, J.A., Roth, D.L., Vignolo, M., and Westfall, A.O. (2009). A behavioral economic reward index
predicts drinking resolutions: moderation re-​visited and compared with other outcomes. Journal of
Consulting and Clinical Psychology, 77, 219–​28.
Tucker, J.A., Roth, D.L. Huang, J., Crawford, M.S., and Simpson, C.A. (2012). Effects of IVR self-​
monitoring on natural resolution of drinking problems: utilization and behavioral economic factors.
Journal of Studies on Alcohol and Drugs, 73, 686–​98.
US National Institutes of Health (2010). Science of Behavior Change Meeting Report. http://​nihroadmap.
nih.gov/​documents/​SOBC_​Meeting_​Summary_​2009.pdf. Accessed 26/​09/​2014.
van der Pol, M.M. and Cairns, J.A. (2001). Estimating time preferences for health using discrete choice
experiments. Social Science and Medicine, 52, 1459–​70.
Vuchinich, R.E. and Heather, N. (2003). Choice, Behavioral Economics and Addiction.
Oxford: Pergamon.
Warren, P., Chiricos, T., and Bales W. (2011). The imprisonment penalty for young black and Hispanic
males: a crime-​specific analysis. Journal of Research in Crime and Delinquency, 49, 56–​80.
Washio, Y., Higgins, S.T., Heil, S.H., et al. (2011). Delay discounting associated with treatment response
among cocaine-​dependent outpatients. Experimental and Clinical Psychopharmacology, 19, 243–​48.
Weatherly, J.N. and Terrell, H.K. (2011). Delay discounting of different commodities II: Confirmatory
analyses. Journal of General Psychology, 138, 35–​48.
West, R.R., McNabb, R., Thompson, A.G. H., Sheldon, T.A., and Evans, J.G. (2003). Estimating
implied rates of discount in healthcare decision-​making. Health Technology Assessment, 7, 1–​60.
World Health Organization (2004). Investing in a comprehensive health sector response to HIV/​
AIDS—​scaling up treatment and accelerating prevention. Geneva, Switzerland: WHO.
World Health Organization and Joint United Nations Programme on HIV/​AIDS (2007). Guidance on
provider-​initiated HIV testing and counselling in health facilities. Geneva, Switzerland: UNAIDS.
Yi, R., Johnson, M.W., Giordano, L.A., Landes, R.D., Badger, G., and Bickel, W.K. (2008). The effects
of reduced cigarette smoking on discounting future rewards: an initial evaluation. Psychological
Record, 58, 163–​74.
Zawilska, J.B. (2011). “Legal highs”—​New players in the old drama. Current Drug Abuse Reviews, 4,
122–​30.
  365

Chapter 20

How an addict’s power of choice

is lost and can be regained
Gabriel Segal

Abstract
“[M]ost alcoholics … have lost the power of choice in drink. Our so-called
will power becomes practically non-existent” (Alcoholics Anonymous). In
active addiction, when an addict who is trying to control or cease using
attempts to choose to abstain, but fails, he chooses against his own will.
He has lost the power to choose as he wants. Often, when an abstinent
addict relapses, the relapse is preceded by a cognitive dysfunction that
selectively disables his willpower in relation to his substance. These
modes of disempowerment in choice making are correctly explained
by a specific disease theory of addiction, articulated in the literature of
Alcoholics Anonymous, and subsequently vindicated by contemporary
cognitive/affective neuroscience. The cognitive dysfunction that precedes
relapse is caused by stress. Twelve-step programs are effective at relapse
prevention and are so because they are comprehensive stress-reduction and
management programs.

1 Introduction
In 2001–​2005 the National Epidemiologic Survey on Alcohol and Related Conditions
surveyed 43,093 individuals. They found that about two ​thirds of those meeting DSM-​IV
criteria for alcohol-​dependence at some point in their lives had a relatively brief course
of excessive drinking (for at most a few years, around the emotional turbulence of teens
and early twenties or around midlife) and exhibited relatively mild dependency. By con-
trast, the remaining third suffered from severe and chronic dependency. Usually they
exhibited signs of dependency in their early teens, and their symptoms gradually but
continually worsened over years and decades. Most patients in treatment programs and
members of Alcoholics Anonymous (AA) fall into this group (Willenbring et al. 2009).
I will assume that these results generalize reasonably well to alcoholics and other sub-
stance addicts. In this chapter, I will consider only this type of “severe” addict. I will set
aside the question whether those who meet criteria for substance dependence for a rela-
tively short period suffer from an ephemeral version of the same underlying condition as
severe, chronic addicts, or are merely alike in superficial respects and just present some
366  

366 How an addict’s power of choice is lost and can be regained

of the same symptoms resulting from a different underlying neuropsychological cause.

I  will also set aside the question of which, if any, behavioral addictions (to sex, gam-
bling, work, shopping, etc.) fall into the same category as severe substance addiction, and
which are merely superficially similar. In this chapter, I will use the terms “addict(ion)”
as short for “severe, chronic addict(ion),” except where I state otherwise.
I will argue that addiction consists in a specific type of impairment in the subjects’
choice-​making systems, one that was very well described in AA’s foundational literature.
In active addiction, addicts have, in a certain specific sense, “lost the power of choice”
when it comes to their favoured substance (Alcoholics Anonymous 1939, p.  24). This
does not mean that when an addict or alcoholic takes drugs or drinks, their behavior is
unintentional, or beyond their control in the manner of a reflex knee-​jerk, the shaking of
a Parkinson’s sufferer, a sleepwalker, or someone under hypnosis (though on rare occa-
sions it can be, or at least can seem to be, something like any of those). Nor does it mean
that they could not do otherwise if they chose to do otherwise, and stuck to that choice.
So, for example, an alcoholic suffering from the most severe cravings, with a glass of their
favourite tipple right before them, would likely refrain from drinking, at least for a while,
if a gun were placed to their head. To that extent at least, their drinking is intentional
action, under their control, and the result of a choice to drink rather than to refrain. In
that sense, and that sense only, they have the power of choice over their using.
But in another and more interesting sense, they lack the power of choice. There are two
dimensions to this. The first is that, in many cases, while they could refrain from using if
they were so to choose, they lack the power to choose to refrain. The second dimension is
that, often, after a choice to remain abstinent has been made, that choice gets overturned
for no good reason, and the addict ends up using. In those cases, the addict’s choice lacks
the power to cause the chosen course of action. Both of these dimensions involve ways in
which the addicts choose against their own wills.
In relation to the first dimension, even if the addict very much wants to choose to
refrain and wants not to choose to use, they choose to use anyway. Thus they lack the
power to choose as they would like to choose. In these cases, their action is motivated by
a desire that they would rather not have and in spite of a desire that fits better with their
overall goals in life and that better represents their wills. Acting on a desire that one would
rather not have is acting against one’s will in the sense of Frankfurt (1971).
Addicts also often choose against their own wills in a second and slightly different
sense. In these cases, the addict knows that they are acting against their own best interests.
They know that, while they use because they want to use, or feel an urge to use, they have,
all things considered, much better reasons for refraining than they have for using. Yet still
they choose to use. Choosing to do something while believing that, all things considered,
one has better reasons for choosing not to do it, is choosing against one’s own will in the
sense of Davidson (1980). (See Heather and Segal (2015) for further discussion of the
relevance of Davidson’s work to understanding addiction, and see Holton and Berridge
(this volume) for detailed discussion of compulsive aspects of addictive choice making.)
The second dimension is that, often, when an addict chooses not to use, they do not
stick to that choice. In such cases, the addict changes their mind for no good reason. It
  367

The nature of the impairment 367

is not that they acquire new information, or rationally re-​evaluate the pros and cons of
using. Rather, their resolve to refrain, their willpower, gets overpowered by the urge to
use, or simply evaporates. Such choices are made against the subject’s own will in yet a
third sense, the sense of Holton (2009). These cases are of particular interest and impor-
tance to long-​term recovery maintenance, as I will explain below.
The problem of addiction should then be easy to solve: the addict merely has to choose to
refrain from using, and stick to that choice. Simple. But it is notoriously difficult for addicts
to do this, no matter how hard or how often they try, with or without help from doc-
tors, psychiatrists, counselors, and therapists. Many die while still trying to implement this
seemingly simple solution. The reason for this is that the addict’s choice-​making circuits,
specific affective and cognitive systems in the addict’s mind-​brain, are severely impaired.
In section 2, I will describe the exact nature of the impairment, drawing on the articu-
lated “disease” model provided by AA, and (as I  shall argue) vindicated by contempo-
rary cognitive-​affective neuroscience, thanks to the “incentive-​sensitization” theory of
Terry Robinson and Kent Berridge (see, e.g., Robinson and Berridge 1993 and Holton
and Berridge, this volume). In section 3, I will show how 12-​step programs based on AA’s
original version provide a comprehensive solution to the problem.

2  The nature of the impairment

2.1  AA’s “disease” theory
Alcoholics Anonymous (1939) offers an articulated tripartite disease theory of the nature
of alcoholism. Alcoholism consists in what they call a “physical allergy” and a “mental
obsession” and it involves a “spiritual malady.” I believe this account is correct in essence
and detail, as I will explain.

2.1.1 The “physical allergy”

The so-​called “physical allergy” consists simply in the phenomenon of the ingestions of
alcohol setting off cravings for more. So, when an alcoholic has a drink, instead of satis-
fying the desire, the drink precipitates an even greater one. This is the main reason why
alcoholics typically end up drinking far more than they want to or intend to when they
have their first drink of the day. The condition is not literally an allergy in the contempo-
rary medical sense. But it is like an allergy in that it is an oversensitivity of the system to
a substance. Like a genuine allergy, it is a kind of sensitization. From a scientific point of
view, the “physical allergy” is a priming effect of the drug stimulating the mesocorticolim-
bic system. (For details, see the discussion of the appetizing effect, “l’appetit vient en man-
geant” (appetite comes with eating), described in Berridge 2001, p. 239.) All the familiar
addictive drugs physiologically stimulate dopamine receptors independently of any pleas-
urable or intoxicating effects they have, and priming is one result of this.1 However, the
power and causal significance of priming varies from substance to substance. While, for

1 Not all addictive substances directly stimulate dopamine receptors. Sugar does not, but can lead to
sensitization and addiction.
368  

368 How an addict’s power of choice is lost and can be regained

example, ingestions of alcohol and cocaine quickly cause cravings for more, the priming
effect of injected heroin is relatively small compared to its satisfying sedative effects.

2.1.2 The “mental obsession”

I will use the term “obsession” to cover two phenomena. People in active addiction are
obsessed with their drug in the normal sense of “obsession.” Thoughts of using take up
inordinate amounts of their time and have huge hold over them. This is one reason why
controlled drinking or using regimes tend not to work for addicts. They may be able to
keep within set limits for a few weeks or months. But during such periods, continued use
perpetuates the obsession and sooner or later control is overcome.
The AA literature tends to use the term “mental obsession” also to refer to the specific
fact that, in alcoholics, the idea that somehow it will be possible to “drink like a gentleman”
keeps returning. Someone who has suffered from active addiction for a number of years will
have considerable experience of their own failures to limit their use. Over and over again
they will have set out with an intention to use some relatively small amount, and ended up
using many times more. This often occurs in a single day or evening. But it is also common
in the longer term. For example, an addict will go on holiday for two weeks, taking with
them what they think will be an adequate supply. Within one day the supply is used up, and
they are off to score some more. Or they will use one evening with the idea of not doing so
again for the foreseeable future, only to repeat the same routine a day or a week later.
The idea that this lack of ability to limit their intake will somehow wear off, and that it
will possible to “partake like a gentleman,” has a very firm foothold in the addict’s mind.
In the first instance, it typically takes addicts years to accept that they have any kind of
problem at all. This is often the case even when the problem has become obvious to those
close to them, who have told them about it in no uncertain terms. Thus the addict repeats
over and over the basic, stock error of starting out believing that they will use within
reasonable limits and ending up using much more. But the obsession goes far beyond
just that. A key aspect of addiction is that, even after the addict has come to observe and
accept that they have very severe problems with controlled use, the obsession that this
time it will be OK, this time they will just have the one, returns. Even if they have become
convinced, after years of experience of total failure, that controlled use is in no way a
realistic option for them, that conviction later evaporates and the idea that this time they
will have just that one comes back and takes control. And succumbing to the temptation
to have just that one often has catastrophic consequences.
It is not hard to see how having just that one often precipitates a relapse for an addict who
has previously decided to stay abstinent. In order to break abstinence and have just one, the
individual must already be in a state of mind in which their previous resolve is overcome by
an urge to use. For whatever reason, taking the drug seems more attractive than maintain-
ing abstinence. After the first use, the motivation for not using decreases while the motiva-
tion for using increases. The former occurs because abstinence has already been broken
and there seems to be less to lose by having a second after the first than there was by having
a first after abstinence. The latter occurs for various reasons. Unpleasant come-​down and
withdrawal effects kick in and these can be relieved by a hair of the dog. Moreover, the
  369

The nature of the impairment 369

very fact that the addict has used again causes stress and distress, which further increase
the desire for drug-​induced relief. And so the cycle continues. In many cases, the situation
is worsened because the first use leads to an immediate binge. That “just one” precipitates
cravings. And the intoxicating effect of the drug itself lowers the subject’s capacity to take
long-​term rational considerations into account, and enhances devil-​may-​care thinking. So
the plan of sticking to one flies out of the window and a binge results.
The abstinent addict’s obsession, their tendency suddenly, for no adequate reason, to
think it will be safe to have just the one, is a well-​attested phenomenon, well-​described
and illustrated with examples on pages 35–​43 of Alcoholics Anonymous (1939). “Alf,”
who I described in Segal (2013a), provides a good example. Alf had spent nine weeks as
an inpatient at a 12-​step rehabilitation clinic and another 28 days as a day-​patient. He was
a member of AA and had read pages 35–​43 of Alcoholics Anonymous (1939). He knew
from experience of his own last years of very severe drinking that the chances of having
just a small amount to drink and then stopping were close to zero. In his later years of
drinking, in spite of every effort, he had never managed to control his intake. If he had
one drink, he would finish the bottle and buy another one. He had heard over and over
again that this was the lot of the alcoholic: abstinence, or alcoholic chaos, with no third
option. And he was completely aware that he was an alcoholic of that sort, which was why
he was abstinent.
Yet he drank again. What happened was that he became worried about money. He
decided that he would need to look over his bank accounts and come to terms with his
economic situation. He said to himself that after looking at the accounts, he would reward
himself with a big drink. And it would calm him down. It simply did not occur to him that
there was any risk involved. He would have a large drink to calm down and then return to
abstinence. The information that he was an alcoholic and that a drink risked relapse simply
remained absent from his consciousness and had no impact on his planning. He remained
in this state for three whole days! There was a three-​day gap between the moment he
decided to look at his bank accounts and have a drink, and the moment he actually did
it. Throughout that period, it never crossed his mind that the plan was crazy. Alf had the
drink. He then continued to drink pure spirits throughout his waking hours, non-​stop,
for 11 days, until he was rescued by a neighbor who managed to arrange a detox for him.
Alf had previously been told by his doctor that his drinking was putting his health at
grave risk, and that if he continued it might well kill him. His psychiatrist had told him
that permanent brain damage would be a likely outcome of any further drinking at all.
Thus his “forgetting” that taking a drink would be taking a serious risk was like someone
forgetting that playing Russian roulette might result in getting shot in the head—​a very
odd thing to “forget” indeed!
This is the “insanity” referred to in Step Two of AA’s Twelve Steps: an addict who knew
perfectly well that one use is likely to lead to a relapse suddenly finds persuasive a totally
feeble reason for having just one, and forgets all about the likelihood of dire consequences.
The addict’s willpower is “practically non-​existent” (Alcoholics Anonymous 1939, p. 24,
my emphasis), because his reasons for not using cease to have any rational effect on his
cognition, and his will to abstain simply disappears.
370  

370 How an addict’s power of choice is lost and can be regained

To understand what explains this kind of major cognitive breakdown it is necessary first
to understand the neurology that appears to underlie addiction: incentive sensitization of
the mesocorticolimbic (dopamine) system.

2.2  Incentive sensitization

The mesocorticolimbic system is an evolutionarily primitive neural system that under-
lies reward motivation and interacts with learning. If a mammal does something that
brings it more pleasure than expected, the system records this. The next time the animal
is exposed to the same cue, its motivation to obtain the reward is increased. Moreover,
the cue itself becomes an object of interest to the animal. Similar cues grab its atten-
tion and become attractive in their own right. So, for example, when a puppy hears the
sound of a can being opened in the kitchen, its ears prick up, it races eagerly towards the
sound, and then expends considerable effort on trying to get fed. In the terminology of
Robinson and Berridge the cue takes on high “incentive value” and produces a powerful
“wanting” for the reward. “Wanting” is an unconscious motivational state, an uncon-
scious drive for action. In humans, “wantings” often cause conscious urges, desires, crav-
ings or feelings of need
According to Robinson and Berridge (e.g. 1993), what happens when someone becomes
addicted is that the mesocorticolimbic system becomes hypersensitive to cues. The sys-
tem easily detects and reacts strongly to them, which reactions are the neural basis of
powerful “wantings.” Further, “wantings” become dissociated from the actual value of the
reward: the system overestimates the benefits of the action. And they can also become
dissociated from the consciously expected value of the action. This is why many addicts
continue to use drugs even when they no longer enjoy or expect to enjoy the effects—​
indeed, even when they expect to feel very ill as a result of ingestion. (For example, about
50 percent of alcoholics who have been fitted with an Antabuse implant still drink, know-
ing it will make them very sick.)
It seems that a chief evolutionary function of the mesocorticolimbic system is to make
sure that the animal gets enough food and water to survive. That explains why addicts can
feel that they urgently need their drug: a cognitive system within their brains is telling them
precisely that. If an addict becomes unsure that they will be able to use, they often experi-
ence great fear. They may even panic. This fear can be far greater than would be justified by
the risks of withdrawal and may occur in the absence of any conscious belief that serious
consequences could result from deprivation. Powerful “wantings” are well described as
feelings of faux-​starvation. It is very much as if the addict unconsciously believes they are
starving: an unconscious cognitive system in their head is, in effect, telling them so.
As said above, the incentive-​sensitization theory explains the “physical allergy,” the
tendency of an addict’s ingestion of a drug to cause cravings for more. It also explains
the addict’s obsession with using. Over time, the addict’s system comes to associate
more and more cues with rewarding use. Crucially, these cues include thoughts and
feelings in the addict’s mind: addicts use to alleviate boredom, to celebrate good times,
to deal with unpleasant emotions. After many years of conditioning, the addict is in a
  371

The nature of the impairment 371

more or less constant state of “wanting”: hence the more or less constant desire or urge
to use.2

2.3  “Wantings,” willpower and denial

Often, addicts’ “wantings” are persistent and powerful enough to overcome their will-
power very easily. Addictive using often involves acting against the will in the three spe-
cific senses mentioned above: Frankfurt’s, Davidson’s, and Holton’s. It should be remarked
that, although all these kinds of case are sometimes described in terms of “weakness of
the will,” the addict’s will may actually not be weak. It is merely weaker than the “want-
ing.” The pathologically swollen “wantings” of a sensitized dopamine system are often
enormously powerful and/​or persistent and would easily shatter the strongest of wills as a
hurricane might shatter a wooden shed.
In the Frankfurt, Davidson, and Holton types of case, the addict may remain fully aware
of the reasons they have for not using and of the rationally compelling nature of those rea-
sons. They are acting in ways that they themselves would tend to regard as irrational and
contrary to their own best interests. But when they are in denial, as addicts often are, they
fail to grasp the harsh realities of their using and its consequences. Narcotics Anonymous
tells us: “Denial is the part of the disease that makes it difficult, if not impossible, for us
to acknowledge reality … We denied that we had a problem with drugs, regardless of all
evidence to the contrary. We lied to ourselves, believing that we could use again success-
fully” (Narcotics Anonymous 1993, p. 7).
Alf, mentioned above, provides a good illustration of the phenomenon of addictive
denial. Alf first became attracted to the idea of taking drugs at a young age, around seven
or eight. Drugs seemed exciting, different, rebellious and a route to some kind of alterna-
tive lifestyle or reality. Alf saw plenty of publicity about the dangers of drugs. He heard
that one could become an “addict,” that it could ruin one’s life, and even end it. But none
of that had the slightest rational impact on his plans. He never felt at risk. He duly started
inhaling solvents at about ten years of age and became a daily user of marijuana at 14. This
pattern of use caused him to become addicted, in the strongest sense of the term, that is,
to experience cravings and obsession, within a year or so. He cross-​addicted to alcohol at
about twenty. And his extreme denial persisted for decades. He failed to acknowledge any
kind of problem, in spite of drinking to excess nearly every day and in spite of very clear
warnings from those close to him. He was also aware that he easily passed every diagnos-
tic test for alcoholism. But he did not see this as a cause for concern. He had no view about

2 While there is good empirical evidence of dopamine sensitization in rats, it has been observed in
humans in only a few studies (e.g. Boileau et al. 2006), and some still question whether it causes mes-
olimbic hyper-​reactivity in humans. The question arises because addicts often appear to have decreased
dopamine receptivity (e.g. Volkow et al. 2004). This may be a negative-​feedback consequence of drug
use that masks hyper-​reactivity, or an artifact of neuro-​imaging technology. Alternatively, continued
excessive use in these cases might be explained by other factors, including withdrawal, tolerance, and
perhaps feelings of decreased enjoyment of other activities.
372  

372 How an addict’s power of choice is lost and can be regained

what kind of condition alcoholism is or whether he really was an alcoholic, even though
he met the diagnostic criteria. The denial persisted until his early forties when his doctor
told him that if he carried on drinking as he was, he would quickly experience very severe
health problems and likely an early death.
He then went to see a counsellor who had the view that there is no such thing as alco-
holism, and so, no such disease. She had the view that controlled drinking is not only a
viable option for problem drinkers, but is usually the safest one. Alf of course did not
want to quit drinking entirely, and was delighted at the prospect of learning to drink like
a gentleman. It took him seven years of abject failure at this experiment to see that that
this was not a viable option.
The basic case of denial may be an extreme case of “cognitive dissonance” (bending of
beliefs under emotional pressure) or of more specific types of failure of self-​awareness
(Goldstein et  al. 2009)  or both. But an equally important case is that of the cognitive
breakdown that often precedes relapse, the case where willpower is not so much overpow-
ered as “practically non-​existent.” And the key factor in all this, I will argue, is: “stress.”

2.4 “Stress”
I will use the term “stress,” in inverted commas, as a semi-​technical term to cover what
the subject would consciously recognize as feelings of stress (or their cause), stress that
is not consciously felt but that might be picked up on by questionnaires or family mem-
bers, and the specific phenomenon of happy stress or reward stress (Peciña et al. 2006).
People tend to drink to celebrate. On reflection, this is a bit strange: why drink if you are
already happy? This is explicable if rewards cause stress, and alcohol relieves that stress.
And smokers like to smoke after sex (the expression “mandatory post-​coital cigarette”
is sometime used). All these forms of “stress” appear to correlate well with the release of
a hormone called “corticotropin-​releasing factor” (CRF) in the central nervous system.
Drugs are effective for quick, short-​term relief of stress. When people start using drugs
in their early teens to handle stress caused by emotional turbulence, they come to rely
on them for that purpose, instead of learning how to process emotions in mature ways.
As this is happening, the dopamine system gets increasingly attuned to “stress” as a cue
for use, hence as a trigger for “wanting.” Moreover, withdrawal states are themselves very
stressful. They and the “stress” they cause are relieved by hairs of the dog. Hence the asso-
ciation between “stress” and “wanting” is strengthened by that mechanism too. “stress”
thus becomes an extremely powerful cue for “wanting.”
Secondly, when someone is “stressed” CRF is actually released inside the nucleus
accumbens shell, part of the dopamine system (Walsh et al. 2014). This has the effect of
enhancing the system’s sensitivity to cues, increasing its reactivity and amplifying posi-
tive motivation for reward-​seeking behaviours (Peciña et al. 2006). If you like chocolate
biscuits, then your tendency to binge when presented with a packet will increase when
you are “stressed”. CRF released inside the nucleus accumbens shell therefore increases an
addict’s “wantings” in general, and specifically will increase the power of the “wanting”
response to the cue that “stress” provides.
  373

Twelve-step programs 373

Thirdly, “stress” has the effect of adversely affecting higher cognitive functions and mem-
ory. “Stressed” individuals tend temporarily to “forget” things, not to think straight, and to
act instinctively, in a dopamine-​driven fashion (e.g. Arnsten 2009; Li and Sinha 2007).
A plausible hypothesis, then, is that this triple-​whammy effect of “stress” causes the
spectacular memory failures and insane thinking that immediately precede relapse in the
addict who has previously been fully aware of the grave dangers of that first use. “Stress”
provides a powerful cue for “wanting.” CRF is released inside the nucleus accumbens and
sends the dopamine system into overdrive, thus further empowering the “wanting.” And
“stress” undermines the memory and clear thinking that would be required for the addict
to keep their head and not fall victim to the obsession.
Suggestive further support for the hypothesis is provided by the phenomenon of cross-​
sensitization between drugs and “stress” in animals. Animals previously exposed to “stress”
may become sensitized to drugs and animals sensitized by drugs may become hypersen-
sitive to “stress” (Robinson and Berridge 2003). Cross-​sensitization from drugs to stress
thus probably also occurs in human addicts, and this may enhance the effect of “stress”
on the dopamine system and add an extra whammy to the equation. This makes it even
more probable that “stress” can push even the resolutely abstinent addict off the wagon.
The hypothesis predicts that a good “stress”-​reduction and management program would
work for relapse prevention. I will now argue that this prediction is correct by showing
first that 12-​step programs work and, second, that they are, indeed, “stress”-​reduction and
management programs.

3  Twelve-​step programs
3.1  Efficacy of 12-​step programs
There has been much public controversy concerning the efficacy of AA and other 12-​step
programs or fellowships. In fact, the evidence that AA and 12-​step treatments are among
the most effective treatments for alcoholism is strong. Evidence is available from a variety
of different types of study, including randomized controlled trials (Kelly and Yeterian
2012; Kelly and Beresin 2014). Further, AA surveys routinely find that over 70 percent of
members report over one year’s continuous sobriety (e.g. AA 2011, AA 2015). And two
scientific studies also found sobriety rates of 70 percent or over among members who
attended meetings regularly (Kaskutas 2009).
But these studies concern only the relationship between membership or regular attend-
ance of meetings and sobriety. In fact, attending meetings is only a small part of a full 12-​
step program. Meetings serve to remind members of the risk of relapse, as an opportunity
to share information about how they use 12-​step principles to stay sane and sober, to put
their problems in perspective and to share them. Many individuals attend meetings but
do not do the full program. And these people are much more vulnerable to relapse than
those who do the whole program thoroughly.
No scientific studies have yet been done on abstinence rates among those who have
completed all 12 steps and continued to practice the program in the recommended way.
374  

374 How an addict’s power of choice is lost and can be regained

However, I have done extensive, though informal and unscientific, research into the ques-
tion. I  have polled many thousands of addicts (15,000–​20,000 subjects, mainly on the
Internet, in Facebook and Google groups) asking: “Have you done the Twelve Steps thor-
oughly and then relapsed while still doing them thoroughly? Or do you know of any-
one meeting those conditions?” My sample included many disaffected ex-​members of
Anonymous fellowships, who were not satisfied with and indeed are very opposed to the
programs and the fellowships.
The result was robust. Only six of the thousands offered a claim to the effect that they are
or knew someone who had done the Steps and relapsed while still doing them. Of those
six, three quickly changed their minds after a short discussion about how they had been
working on each specific Step prior to their relapse. One person claimed to know someone
who had relapsed while still doing the Steps, but was unable to provide any details except to
say that her acquaintance has major problems concentrating on any task at all. Two other
claims remain. One individual admits that he cannot remember what he was doing in the
days before he relapsed. The other does appear, prima facie, to have done and continued to
do the Steps as specified in Alcoholics Anonymous (1939). An anomaly in her case is that
she is an atheist who regards her higher power as nature, yet her sponsor was guiding her
through the Steps featuring a supernatural God as the higher power. It seems that this mis-
match undermined the efficacy of the program. Whether this can be put down to a failure
with a specific Step remains an open question that is still under discussion.
I believe that these informal studies provide good anecdotal evidence that 12-​step pro-
grams are 99–​100 percent effective, if followed thoroughly. This might seem incredible.
But skepticism may be alleviated if one thinks of the program as analogous to diet and
exercise: a strict regime of diet and exercise is more-or-less guaranteed to control weight.
However, such regimes only work for people who follow them, and this is not easy to do.
People who try to follow the regime and do not keep their weight down fail, not because
the regime doesn’t work but because they are not following it.3
In any event, even if the success rate among those who thoroughly work the program
is less than my informal survey suggests, it is high enough to indicate that what they are
doing significantly reduces the chances of relapse. The remainder of this chapter will be
concerned with an account of how the programs work to deal with “stress.”4

3.2  A “spiritual” solution to a “spiritual malady”

In classical AA terms, the result of the Twelve Steps is to have something called a “spiritual
awakening.” Co-​founder and chief architect of the Twelve Steps and the fellowship of AA,
Bill Wilson, believed that freedom from the obsession came from good conscious contact

3 If you remain skeptical, I would suggest that you search for subjects who have done all Twelve Steps
thoroughly and relapsed while still working on them thoroughly. If you find a candidate, then I would
ask you to ask them if they wouldn’t mind talking to me and put us in touch, as their story would likely
provide valuable information concerning the causes of relapse.
4 What follows is drawn from Segal (2013b). Some of the text is repeated verbatim.
  375

Twelve-step programs 375

with God, whom he thought of as a loving, all-​powerful, creative intelligence, universal

mind, or spirit of the universe. However, he realized that not everyone who had a success-
ful recovery through the Steps understood the process in religious or supernatural terms.
In Appendix II of Alcoholics Anonymous (1939) (added in 1941) the “spiritual” awaken-
ing is described in terms of a personality change, and a profound alteration in one’s reac-
tion to life. It is presented as involving an “inner resource” which members identify with
“their own conception of a Power greater than themselves.” (Only) AA’s “more religious
members” call this awareness of a Power greater than themselves “God-​consciousness.”
In the actual Steps this “Power” is referred to as “God as we understood Him,” rather
than just “God.” The words “as we understood Him” feature thanks to the insistence of
non-​religious members of Wilson’s group (Alcoholics Anonymous 1984, p. 199; 1980).
They meant the expression “God as we understood Him” to have a very liberal interpreta-
tion, along the lines of “anything you like (other than yourself).” So we are left with the
idea that developing a relationship to something other than one’s self, along the lines pre-
sented in the Steps, can tap an inner resource and bring about a personality change and
a profound alteration in one’s reaction to life of a kind that are sufficient to bring about
recovery from alcoholism. Looking at the Steps that way makes it relatively easy to see
how they work in entirely naturalistic terms, with no appeal to any supernatural force.
The “spiritual” solution relieves the obsession by providing a remedy for the third aspect
of the disease of addiction, which AA calls the “spiritual malady.” Symptoms of “spiritual”
malady are irritability, restlessness, and discontent due to anger, fear, resentment, frustra-
tion, dissatisfaction with life, low self-​esteem, wounded pride, guilt, jealousy, existential
angst, etc. These all stem from unmanaged, misdirected and/​or overactive instincts, unsat-
isfied, creating emotional turbulence and disturbance to peace of mind (the psychology
of addiction and recovery as understood in terms of misdirected instincts is beautifully
articulated in Alcoholics Anonymous 1952).
Wilson describes the root of the problem thus: “Selfishness—​self-​centredness! That, we
think, is the root of our troubles” (Alcoholics Anonymous 1939, p. 62). “Spiritual” malady
results from three specific kinds of selfishness: self-​seeking, self-​absorption, and self-​will.
Self-​seeking is placing the satisfaction of one’s own desires as the end goal of actions and
interactions with others.
Self-​absorption is focusing attention on oneself and one’s feelings. The self-​absorbed
individual thinks of other things in terms of how they do or can affect him-​or herself.
They have little interest in how other things are just in themselves. For example, when
thinking of the past or future, they do not focus on general or other-​centered facts, such
us who won or will win an election. Rather, they mentally represent scenarios centered on
themselves, like plays in which they are the lead character. And when thinking of present
matters, they tend to focus on their own feelings and how these can be improved, rather
than on what is going on in the external world.
Self-​will is stubborn, excessive adherence to trying to have things as one thinks they
should be. Self-​willed individuals are not inclined to take advice, or listen carefully to
the suggestions of others about what they should do. They do not tend to accept moral
376  

376 How an addict’s power of choice is lost and can be regained

or other correction. They think they know best what is right, at least for them, and are
inclined to try too hard to bring it to be. They do their own thing, their way. And they are
apt to get very frustrated and stressed if their will is thwarted.
A person can be in many ways unselfish yet still be self-​seeking, self-​willed, and self-​
absorbed. They may be relatively thoughtful and generous. But they treat others well
because, at the end of the day, it suits them. They do not in any meaningful sense treat
others as being as important as they are, or place the needs of others level with their own.
They may act morally, doing what is right. But they do not do it because it is right but
because that is how they feel more comfortable.
Drug abuse of its nature involves self-​absorption, self-​seeking, and self-​will. The selfish
individual is concerned with their own feelings, so strives to adjust the world to make
them feel better. As a life strategy, this fails spectacularly because reality rarely bends to
one’s will. Other people rarely feel and behave just as one wants and the world does not
readily provide everything one desires. As the strategy fails, the selfish individual may
turn to drugs as a convenient fix for their feelings. With such substances, the individual
can become God of their own feelings and so rule the realm that matters to them most.
They can do things their own way to have things their own way. And if they learn to do
this early in life, as most addicts do, they never learn to deal with their feelings properly.
They become used to getting high or anaesthetised and substance use becomes their chief
or only coping strategy. If they have vulnerable dopamine systems, these will become
sensitized and they will become addicts.
Addicts are not necessarily born more selfish than anyone else.5 They start out using
drugs to alter their feelings. And, for whatever reason, their dopamine systems, the seat
of instinctive drives, become disordered and oversensitive to drug cues. When sensitized
and deprived of drug-​induced stimulation, it induces a state of “faux starvation” in the
poor sufferer and renders them desperate.
The phenomenon of addiction—​rooted in the malfunctioning dopamine system—​just
brings into sharp relief a kind of maladjustment that is widespread. “Spiritual” malady is
not confined to addiction and can lead to problematic behaviors of many different kinds.
But it often gives rise to behavioral “addictions” to romance, exercise, and all manner of
other things, from obsessively looking at property prices to watching soap operas on TV.
All of these are used to escape from disturbing feelings, and most involve a desperate and
misplaced attempt to satisfy some instinctive need: money for food and material posses-
sions, or love or power or the esteem of others.
These activities do not involve emotionally satisfying interactions with people or things
and so fail to satisfy the instinctive drives that cause them. The self-​willed individual then
just keeps on trying and failing with the same strategies, and “addictive” behavior results.
Self-​centeredness can lead to discontent and “stress” in all the important areas of life.
It leads to relationship problems as the individual tries to cause the partner to feel and

5 Converging evidence suggests that the dopamine systems of youths who are predisposed to addiction
are often particularly highly responsive to emotionally salient stimuli (Leyton and Vezina 2014). This
is unsurprising if addiction itself is neurally based on incentive sensitization.
  377

Twelve-step programs 377

act in ways that will make them feel good, instead of treating the other and their feelings
as ends-​in-​themselves. By focusing on their own feelings, the selfish person is hampered
in their ability to listen to others properly and learn from them. Their sensitivity to and
awareness of the feelings of others is impaired and so they cannot be truly close to them.
The same applies to an individual’s relationship with their physical environment. The
selfish person tends to look at the environment in terms of its affordances for pleasure, ask-
ing themselves, what can it do for me? This general approach detracts from the individual’s
capacity to appreciate it for what it is. A selfish person in a beautiful forest, for example, will
tend, maybe subconsciously, to think along the lines of, can I find food, take some wood or
use it in some other way? If the forest provides no such affordances, then a selfish person
will fail to see a point to it. They will focus attention on themselves and not on the beauty
of the ponds and trees around them. A selfless person is able to interact with the forest
cognitively and emotionally in a quite different way: they can simply appreciate its beauty
and take joy in being there. Or they can tend to it just to make the world a better place.
“Spirituality” as it figures in 12-​step recovery from addiction can be understood simply
as the opposite of self-​absorption, self-​seeking, and self-​will. The “spiritually” fit person
treats their feelings not as objects, ends-​in-​themselves to be manipulated away from dis-
tress and towards pleasure, but as modes of awareness of and contact with other people
and elements of the world around them. The feelings of a “spiritually” fit person connect
them to the rest of the universe in a way that allows them to interact with its denizens in
meaningful and satisfying ways. The “spiritually” fit person humbly accepts others and
the rest of the world as they are, and goes with the flow. The “spiritually” fit person tries
to change their mind to fit the word, and not the other way around. That is the central
principle of the philosophies of Taoism, Buddhism, and Stoicism and it lies also at the
core of 12-​step recovery. Alcoholics Anonymous (1952, p. 48) expresses it this way: “we
needed to change ourselves to meet conditions, whatever they were” (my emphasis). Since
the causes of “stress” are not worldly situations but one’s psychological reactions to them,
this strategy works to alleviate it.
“Spirituality” opens an individual’s mind to the world as it is, brings connectedness to
it and allows them to live serenely and contentedly in it. I will conclude with a descrip-
tion of the Steps and a sketch of how they work to treat “spiritual” malady, keep turbulent
emotions at bay, leaving the addict sufficiently free of “stress” to be free from the obsession
and risk of relapse.

3.3  How the Steps work

Here are the Twelve Steps of Narcotics Anonymous:
1 We admitted that we were powerless over our addiction, that our lives had become
unmanageable.
2 We came to believe that a Power greater than ourselves could restore us to sanity.
3 We made a decision to turn our will and our lives over to the care of God as we
understood Him.
4 We made a searching and fearless moral inventory of ourselves.
378  

378 How an addict’s power of choice is lost and can be regained

5 We admitted to God, to ourselves, and to another human being the exact nature of our
wrongs.
6 We were entirely ready to have God remove all these defects of character.
7 We humbly asked Him to remove our shortcomings.
8 We made a list of all persons we had harmed, and became willing to make amends to
them all.
9 We made direct amends to such people wherever possible, except when to do so
would injure them or others.
10 We continued to take personal inventory, and when we were wrong promptly
admitted it.
11 We sought through prayer and meditation to improve our conscious contact with
God as we understood Him, praying only for knowledge of His will for us and the
power to carry that out.
12 Having had a spiritual awakening as a result of these steps, we tried to carry this mes-
sage to addicts, and to practice these principles in all our affairs.
And here is an example of a secular version of the kind sometimes used by fellowship
members who find it easier to work without a concept of God:6
1 We admitted we were powerless over our addiction—​that our lives had become
unmanageable.
2 Came to believe that a power greater than ourselves could restore us to sanity.
3 Made a decision to turn our will and our lives over to nature and go with its flow.
4 Made a searching and fearless moral inventory of ourselves.
5 Admitted to ourselves and to another human being the exact nature of our wrongs.
6 Were entirely ready to give up all these defects of character.
7 Adopted a practice of humility so that our shortcomings might be overcome.
8 Made a list of all persons we had harmed, and became willing to make amends to
them all.
9 Made direct amends to such people wherever possible, except when to do so would
injure them or others.
10 Continued to take personal inventory and when we were wrong promptly admitted it.
11 Sought through reflection and meditation to improve our conscious contact with
nature and its inhabitants, seeking only for knowledge of right thought and action.
12 Having had a spiritual awakening as the result of these steps, we tried to carry this
message to other addicts, and to practice these principles in all our affairs.

6 See Roger (2013) for a collection of secular sets of steps.

  379

Twelve-step programs 379

Here is a brief explanation of how the program works. At Step One, the addict gets the
ideas of the “allergy” and obsession firmly ensconced in their head. They come to rec-
ognize that indulging in a single use brings a grave risk and that unless they take action
to change things, the addiction will get the better of them again, insanity will result and
they will slip and then relapse. They recognize that they have not been able to manage
their own thoughts, emotions and behavior and that if things are to change, they need
assistance. At Step Two, they come to believe that assistance is available, that something
more powerful than them can relieve the obsession. The key aspect of Step Three is just
the decision to do the remainder of the Steps with the idea of placing trust in the program
and fellowship, rather than in their own thinking, which has hitherto proved catastrophic
in relation to their substance. In relation to staying abstinent, the addict decides to do
things according to the dictates of something other than their own will. Strictly speaking,
that is all that is required for taking Step Three.
However, the idea of turning one’s will and life over to the care of a higher power is
key to the whole program. It is important for the addict to stop trying to control every-
thing: other people and real-​world situations. And it is important that they stop trying
to control their own emotions and thoughts in the very crude and unhealthy ways they
have been used to: taking drugs, escapism, denial. Instead, they will turn management of
themselves over to the Steps and whatever further higher power they may choose, such
as God, their group or fellowship, good orderly direction, human goodness, love, nature,
truth, the Tao or the Dharma way. Addicts need to give up running on self-​will, they need
to learn to do the right thing, then let go and leave the rest up to nature. Taking Step Three
is the beginning of this new approach to life.
It is at Step Four that the real work begins. On Step Four, the addict makes a list of
all instances where their character defects (or “maladjustments”; Alcoholics Anonymous
1952, p. 50) have caused them emotional disturbance. These will be instances where such
things as resentment, fear, envy, pride, self-​pity, jealousy, frustrated greed or guilt arising
from harms done have caused trouble. In the case of resentment, for example, they learn
to recognize which aspects of their own character are adversely affected by the actions of
others that they find offensive, whether it is wounded pride, fear of loss of status or finan-
cial security or whatever. They learn to look at what actions of their own caused others
to act aggressively towards them. They learn to forgive others, and themselves, to strive
to give up on blaming others and self, to learn from their mistakes, put the past behind
them, and move on.
On Step Five, they share all this with someone else. This is an element of most talking
therapies. It takes the sting out of troubling feelings and helps the addict feel accepted by
the rest of the world, and so helps the addict accept and live with themselves as they really
are, warts and all. This realistic acceptance of their defects clears the way for the addict to
try to become less defective—​a better person.
On Steps Six and Seven, they pause to reorient their approach to life. These Steps rein-
force the addict’s realistic attitudes towards themselves and their place in the world. Step
380  

380 How an addict’s power of choice is lost and can be regained

Six involves willingness to keep striving to improve one’s character, to keep working on
one’s maladjustments, hence to keep working on recovery. Step Seven places humility
at the center of recovery. Maintaining humility helps the addict not to expect too much
out of life, not to want more than they need, not to feel deprived, jealous, or resentful,
not to fear failure or success but to feel grateful for being alive and for whatever lot they
have. Humility keeps a person grounded and guards against overexcitement and reward
“stress.” Humility also underpins motivation to do service to others. And it forces the
under-​confident addict to take responsibility for their current and future actions and
reactions to life. Steps Six and Seven help the addict avoid taking anything for granted
and getting smug. They help the addict keep in mind that the work of recovery is never
finished. Occasionally, an addict will relapse after years of abstinence, while apparently
still working on the program. They go to meetings, have sponsees, meditate, maybe pray,
and do some inventory. Yet they relapse. In these cases, it is typically neglect of Six and
Seven that is the culprit. They get complacent, cease using inventory to work on char-
acter development, become ungrateful for their lot, become discontent, irritable, and
“stressed” out.
Steps Eight and Nine are self-​explanatory:  the addict does whatever they can to set
right wrongs they have done, without causing more harm. Their conscience thus becomes
clear, and they become more able to accept who and what they are and to live happily with
themselves.
Steps Ten, Eleven, and Twelve comprise a program of long-​term recovery maintenance.
Step Ten involves keeping a daily inventory and making sure to set right any wrongs as
they arise. It includes keeping Steps Six and Seven in mind: the addict maintains aware-
ness of their own ongoing defects and keeps in place the aim of improving. The inventory
also includes a credit section: pats on the back for things done well. This helps maintain
self-​esteem and happiness. Step Eleven involves taking time, each day, to consider right
thoughts, feelings and action (or God’s will):  the addicts train themselves to try to do
what is right rather than what they want, and to keep their thoughts and feelings in order,
measured and humble. By “meditation” the founders of AA had in mind the study of
religious texts. But these days most practitioners do some form of mindful meditation
instead of or in addition to that. Step Twelve involves practicing the principles of humility
and right action in all aspects of life, and using one’s experience of recovery to help other
addicts. Being part of a group that saves lives, and actively participating in this process,
works wonders for a person’s self-​esteem, connection with humanity and feeling that their
own life is meaningful and worth living. Working with addicts also keeps the reality of the
risk and dangers of relapse in clear view, and maintains a sense of gratitude for the addict’s
own recovery and relative wellbeing.
Steps Ten, Eleven, and Twelve work very well together to keep a recovering addict in
good emotional shape. If any circumstance disturbs their serenity, they can do an inven-
tory and share this with their sponsor to process the turbulent emotions, meditate to
calm down and get grounded, and then go and help someone in dire need, to take their
  381

Twelve-step programs 381

mind off their own problems, put things in perspective, and do something really useful
for humanity.7

References
Alcoholics Anonymous (1939). Alcoholics Anonymous. New York: Alcoholics Anonymous World
Services.
Alcoholics Anonymous (1952). The Twelve Steps and Twelve Traditions. New York: Alcoholics
Anonymous World Services.
Alcoholics Anonymous (1980). A Newcomer Asks. New York: Alcoholics Anonymous World Services.
Alcoholics Anonymous (1984). Pass It On. New York: Alcoholics Anonymous World Services.
Alcoholics Anonymous (2011). Membership Survey. New York: Alcoholics Anonymous World Services.
Alcoholics Anonymous (2015). Membership Survey. York: Alcoholics Anonymous (Great Britain) Ltd.
Arnsten, A.F.T. (2009). Stress signalling pathways that impair prefrontal cortex structure and function.
Nature Reviews Neuroscience, 10, 410–​22.
Berridge, K. (2001). The neural basis of drug craving: an incentive-​sensitization theory of addiction.
In: Medin, D. (ed), The Psychology of Learning and Motivation: Advances in Research and Theory, 40,
pp. 223–​78.
Boileau, I., Dagher, A., Leyton, M., et al. (2006). Modeling sensitization to stimulants in humans: an
[11C]raclopride/​positron emission tomography study in healthy men. Archives of General
Psychiatry, 63, 1386–​95.
Davidson, D. (1980). How is weakness of the will possible? In: D. Davidson (ed.), Essays on Actions and
Events. Oxford: Clarendon Press, pp. 21–​42.
Frankfurt, H.G. (1971). Freedom of the will and the concept of a person. Journal of Philosophy,
68, 5–​20.
Heather, N. and Segal, G. (2015). Is addiction a myth? Donald Davidson’s solution to the problem of
akrasia says not. International Journal of Drug and Alcohol Research ISSN 1925-​7066. http://​www.
ijadr.org/​index.php/​ijadr/​article/​view/​195. Accessed 05/​05/​2016.
Goldstein, R. Z., Craig, A. D., Bechara, A., et al. (2009). The neurocircuitry of impaired insight in drug
addiction. Trends in Cognitive Science, 13(9), 372–​80.
Holton, R. (2009). Willing, Wanting, Waiting. Oxford: Oxford University Press.
Kaskutas, L. A. (2009). Alcoholics Anonymous effectiveness: faith meets science. Journal of Addictive
Diseases, 28, 145–​57.
Kelly J.F. and Yeterian J.D. (2012). Empirical awakening: the new science on mutual help and
implications for cost containment under health care reform. Journal of Substance Abuse, 33, 85–​91.
Kelly, J. and Beresin, G. (2014). In defense of 12 Steps: what science really tells us about addiction.
Wbur’s Common Health Reform snd Reality, April 12. http://​commonhealth.wbur.org/​2014/​04/​
defense-​12-​step-​addiction. Accessed 05/​05/​2016.
Leyton, M. and Vezina, P. (2014). Dopamine ups and downs in vulnerability to addictions: a
neurodevelopmental model. Trends in Pharmacological Sciences, 35, 268–​76.
Li, C.-​S.R. and Sinha, R. (2007). Inhibitory control and emotional stress regulation: neuroimaging
evidence for frontal-​limbic dysfunction in psycho-​stimulant addiction. Neuroscience and
Biobehavioral Reviews, 32, 581–​97.

7 Many thanks for most helpful discussion and comments to Kent Berridge, Patrick Butlin, Thomas
Crowther, Daniel Friesner, Rita Goldstein, Nick Heather, Edmund Henden, John Kelly, Louise King,
Marc Lewis, and Alistair Sweet.
382  

382 How an addict’s power of choice is lost and can be regained

Narcotics Anonymous (1993). It Works How and Why: The Twelve Steps and Twelve Traditions of
Narcotics Anonymous. CA: Narcotics Anonymous World Services.
Peciña, S., Schulkin, J., and Berridge, K.C. (2006). Nucleus accumbens corticotropin-​releasing factor
increases cue-​triggered motivation for sucrose reward: paradoxical positive incentive effects in
stress? BMC Biology, 4, 8.
Robinson, T.E. and Berridge, K. C. (1993). The neural basis of drug craving: an incentive-​sensitization
theory of addiction. Brain Research Reviews, 18, 247–​91.
Roger, C. (2013). The Little Book: A Collection of Alternative 12 Steps. Canada: AA Agnostica.
Segal, G. (2013a). Alcoholism, disease and insanity. Philosophy, Psychology and Psychiatry, 20, 297–​315.
Segal, G. (2013b). Twelve Steps to Psychological Good Health and Serenity: A Guide. Grosvenor
House: Guilford.
Volkow, N.D., Fowler J.S., Wang G.-​J., and Swanson, J.M. (2004). Dopamine in drug abuse and
addiction: results from imaging studies and treatment implications. Molecular Psychiatry, 9, 557–​69.
Walsh, J. J., Friedman, A. K., Sun, H., et al. (2014). Stress and CRF gate neural activation of BDNF in
the mesolimbic reward pathway. Nature Neuroscience, 17(1), 27–​29.
Willenbring, M. L., Massey, S.H., and Gardner, M.B. (2009). Helping patients who drink too much: an
evidence-​based guide for primary care physicians. American Family Physician, 80, 44–​50.
  383

Section VI

Implications for the public

understanding of addiction
and for legal responsibility
for addictive behaviour
384
  385

Chapter 21

What addicts can teach us about

addiction: A natural history approach
Gene M. Heyman
Verna Mims

Abstract
According to the spokespersons for US federal health institutes, addiction
is a chronic, relapsing disease, yet every US survey of the prevalence and
correlates of psychiatric disorders shows that the majority of those who
meet the DSM criteria for addiction are in remission. For illegal drugs,
remission typically occurs by age 30, whereas for cigarettes and alcohol
dependence typically persists well into the 40s. Contrary to conventional
wisdom, the evidence says that remission after age 30 is stable. Since most
addicts do not seek treatment, logic implies that everyday experiences
must promote positive change in addicts. In support of this logic, research
shows that the correlates of remission include economic pressures, family
obligations, and values, such as the desire to be a better person. In light of
the research on how drug dependence ends, interventions should focus on
factors that encourage addicts to make positive non-drug choices.

1  Introduction: goals and approach

Our long-​term goal is to change the conversation about addiction, replacing “Beyond a
reasonable doubt … addiction is a disease of the brain,” (NIDA 2010) to “Addicts retain
the capacity to quit drugs. How do we get them to quit sooner or prevent addiction in the
first place?” Our short-​term goal is to present the research that establishes this view of
addiction. The studies show that addiction to illegal drugs typically ends after a few years,
usually without the benefit of interventions, and that the correlates of remission include
many of the factors that influence everyday choices, such as economic pressures, fam-
ily pressures, the desire to be a better person, and the desire to lead a more meaningful
life. No disease or psychiatric disorder has this profile. As these findings become better
known, we expect that they will persuade researchers, clinicians, the public, and addicts
themselves that addiction does not entail compulsive drug use. Indeed, research presented
in this chapter reveals that most addicts already know this, or, put more accurately, they
behave as if they know this. They quit drugs, the conditions that persuade them to do so
are not necessarily extraordinary and, in the case of illegal drugs, they typically quit drugs
386  

386 What addicts can teach us about addiction: a natural history approach

at clinically significant levels by about age 30. This is what all the major epidemiological
surveys have shown, including those sponsored by the agencies that continue to claim
that addiction is a chronic disease (e.g. the National Institute on Drug Abuse: NIDA). The
results reviewed in this chapter are among the most reliable in behavioral science. If given
the attention they deserve, they promise to change received opinion regarding the nature
of addiction—​which, in turn, promises to change addiction.
This chapter is based on research that describes how addicts behave, including epidemio-
logical studies, treatment interventions, experiments, and historical accounts of the effects
of changes in drug availability. Although the findings sensibly reinforce one another, they
are not well known. NIDA directors and spokespersons, journalists, and the many addiction
researchers who frame their articles and grant proposals with the boilerplate (and incorrect
claim) that “addiction is a chronic, relapsing disease” have routinely failed to focus on or even
mention the findings emphasized in this chapter. Put another way, widely held ideas about
addiction are at odds with what research says about addiction, particularly research on the
manner in which addiction changes over time and the factors that influence these changes.

2  How to tell whether addiction is a disease

2.1  Expert opinion
When addiction specialists say that addiction is a disease, they mean that drug use has become
involuntary. In support of this perspective, the American Psychiatric Association’s handbook
of diagnostic criteria for mental disorders (DSM-​IV 1994) states that “drug dependence” (the
technical term for addiction) is “compulsive” drug use, which the dictionary would translate
as “an irresistible urge” to use drugs that “goes against one’s conscious will” to not use drugs.
Miller and Chappel, psychiatrists who specialize in addiction, write that the “sine qua non”
of the disease interpretation of addiction is “loss of control” over drug use (1991). In the
journal Science, Alan Leshner, a previous director of NIDA and now head of the American
Association for the Advancement of Science, states that addicts start off as “voluntary” drug
users, but then drugs change their brains, turning them into involuntary “compulsive” drug
seekers (1997). In the British medical journal, the Lancet, O’Brien and McLellan (1996), fre-
quently cited addiction experts, draw similar conclusions. They find that “At some point after
continued repetition of voluntary drug taking, the drug ‘user’ loses the voluntary ability to
control its use. At this point the drug ‘misuser’ becomes ‘drug addicted’ and there is a com-
pulsive, often overwhelming involuntary aspect to continued drug use” (p. 237).
According to the experts cited above, major medical organizations, and government
research agencies, the term “addiction” is equivalent to “involuntary drug use.” However,
these accounts fail to define “involuntary” in measurable terms that are independent of
drug use. What is needed are formulas for defining voluntary and compulsive and then
a check if the formulas apply to drug use in those who meet the criteria for addiction.

2.2  Popular opinion

Every winter, offices across America display signs that read, “Stop the spread of flu.
Wash your hands.” This becomes relevant to addiction when we imagine a possible
  387

Characteristics of voluntary behavior relevant to addiction 387

sign that is not posted:  “Stop the spread of flu. Don’t sneeze.” Elimination of sneez-
ing would likely reduce the spread of flu viruses as effectively as handwashing, but
no public notice makes this point. This difference reflects the widely shared intui-
tion that hand washing can be influenced by exhortations, bribes (e.g. from parents),
admonitions of others, and cultural values such as “cleanliness,” whereas sneezing
cannot. There is little or no record of “anti-​sneezing” policies, which in turn, reflects
the understanding that the tools of social policy will have little or no direct influence
on sneezing. Put more generally, widely shared intuitions about behavior distinguish
between activities that can be readily influenced by their consequences and those that
are relatively immune to their consequences.

2.3  Howresearchers distinguish between voluntary

and involuntary behavior
Informal public norms regarding voluntary behavior have a scientific analog. Researchers
interested in basic behavioral processes routinely distinguish between elicited behaviors
and learned instrumental behaviors. Reflexes and other automatic responses are triggered
by stimuli, as when the male stickleback attacks objects with “red bellies” in mating sea-
son. This is the subject matter of ethology. In contrast, learned responses take shape in
ever-​evolving, dynamic feedback loops that involve reward and/​or punishment. This is
the subject matter of psychology and economics.
Consequence-​driven behavior is voluntary in several senses. It emerges “spontaneously”
(without an obvious external prod) and then evolves. Planning plays a role, as when indi-
viduals enact self-​control measures such as removing temptations in order to achieve some
long-​term end (Ainslie 1975). It overlaps with the activities that in the courts are considered
matters of free will and individual responsibility. In contrast, reflexes have explicit proximal
causes, are not as easily avoided by self-​control measures, and are not under the jurisdic-
tion of the legal system. Thus, a fundamental feature of behavior is the degree to which it
is influenced by its consequences, which is to say, its “voluntariness.” For simple creatures,
such as the much-​studied drosophila fly, the list of possible consequences is relatively short
and likely does not go far beyond those things that it can sense. For humans, the number
of possible consequences is vast, ranging from the concrete (e.g. food) to the intangible
(e.g. ethical values). Humans routinely risk everything for what they believe others imagine
about themselves. We can, then, test whether drug use in addicts is voluntary by assessing
the degree to which it responds to its costs and benefits–​where costs and benefits may be
concrete or abstract. Do addicts use drugs regardless of the costs and benefits, as implied by
the claim that they are compulsive users who cannot say no? Or do addicts cut back on their
drug use, or maybe even quit, as the costs of drug use increase and the benefits decrease?

3  Characteristics of voluntary behavior relevant

to addiction
Implicit in this approach to testing whether addicts are involuntary drug users is (1) that
choice is not free will, (2) that voluntariness is a continuous dimension, (3) that choice is
388  

388 What addicts can teach us about addiction: a natural history approach

not necessarily optimal or rational, as is often assumed in economics and psychology, and
(4) that choice has a biological basis.

3.1  Not free will

The degree to which consequences influence behavior distinguishes voluntary and invol-
untary behavior. Thus, the issue is not free will but a distinction between causal relations.

3.2  The voluntary/​involuntary continuum

According to the dictionary, tics are brief, repetitive, spasmodic, involuntary movements
involving voluntary muscles. However, careful observation reveals that those who tic
chronically have some degree of voluntary control as a function of the situation. Oliver
Sacks (1995) introduces his readers to a physician who suffered from Tourette’s syndrome
yet reliably was tic free when doing surgery or flying his private plane. The documentary
Twitch and Shout (Chiten et al. 1994) features a chanteuse who does not tic when she is
on stage. Nevertheless, it would be unreasonable as well as inhumane to punish someone
with Tourette’s for ticing: one cannot always hold the stage or the surgery amphitheater.
A quite different but perhaps more familiar example makes the same point. When and
what we eat is voluntary. But this is true only in the short run. Well-​established records
show that as deprivation increases, eating becomes less voluntary. To be sure there is
an occasional successful hunger striker, but this is an unrealistic standard for voluntary
action. To be useful, we need feasible criteria for judging whether consequences signifi-
cantly influence behavior. While on stage many Touretters might not tic, and at the point
of a gun, we are confident that most, if not all, addicts would abstain. But these are neither
feasible nor reasonable examples for classifying drug use in addicts. Rather, we ask if
less than extraordinary pressures, consequences, repercussions and non-​drug alternatives
convince addicts to become controlled drug users or to quit using drugs altogether.

3.3  Choice is not necessarily rational

In economics and psychology, it is often assumed that individuals make choices so as to
maximize the consequences of their actions. For instance, textbook chapters on consumer
choice begin with the assumption that individuals choose the optimal market basket of
goods, taking into account price, budget, and diminishing marginal utility (e.g. Baumol and
Blinder 1994). In contrast, the idea that voluntary behavior refers to activities that vary sys-
tematically with their consequences presumes nothing about rationality. Actions primarily
controlled by the most immediate consequences (e.g. over the next few minutes) are likely
to seem irrational from a perspective that includes the future as well as the present and the
welfare of others. Moreover, as discussed elsewhere (Heyman 2009), the local frame of refer-
ence is usually more salient than the future so that, in experiments, subjects typically make
choices as predicted by the local frame of reference, thereby “irrationally” reducing their
overall benefits (Herrnstein et al. 1993). Consequently, we are not asking if the persistence
of drug use in addicts is rational, but whether it is susceptible to its consequences.
  389

The criteria for identifying addicts 389

3.4  Drug-​inducedbrain changes are not prima

facie evidence of disease
Drugs change the brain. Leshner and his successors at NIDA assume that this is sufficient
evidence that addiction is a disease state. Leshner (1997) writes: “drugs change the brain,
this is what makes addiction a brain disease.” Nora Volkow, the current Director of NIDA,
adds that because drugs can affect the orbito-​frontal cortex, addiction is a disease (Volkow
and Fowler 2000). However, drugs do not have privileged status when it comes to affect-
ing the brain. Everything that influences behavior does so by way of the brain. Similarly
our preferences have a genetic basis—​even political preferences (Waller et al. 1990). For
anyone who takes evolution seriously, how could it be otherwise? The proper question is
not whether drugs change the brain, but whether they change the brain so that drug use is
no longer voluntary. This is a behavioral question that can only be answered by studying
the natural history of drug use. To determine whether drug addicts are compulsive drug
users, we need to know what influences drug use in those who meet agreed-​upon criteria
for addiction. If the factors are similar to those that affect voluntary actions, then drug use
in addicts remains voluntary, albeit irrational and self-​destructive. Conversely, if the fac-
tors that affect voluntary action have little influence on drug use in addicts, as is the case
for diseases, then it may be reasonable to identify addiction as a disease.

4  The criteria for identifying addicts: the American

Psychiatric Association’s Diagnostic and Statistical
Manual of Mental Disorders
The American Psychiatric Association’s nosological handbook (1994), entitled the
Diagnostic and Statistical Manual of Mental Disorders (DSM), has become the gold stan­
dard for identifying psychiatric disorders for clinicians, researchers, and the courts in the
United States. The manual substitutes the term “substance dependence” for “addiction.” It
begins its description of Substance Dependence in the following words:
The essential feature of Substance Dependence is a cluster of cognitive, behavioral, and physio-
logical symptoms indicating that the individual continues use of the substance despite significant
substance-​related problems. There is a pattern of repeated self-​administration that usually results
in tolerance, withdrawal, and compulsive drug-​taking behavior.

(APA 1994, p. 176)

Following this passage is a list of seven observable, measurable signs related to drug use,
such as tolerance, withdrawal, using more drug than initially intended, or failing to stop
using after vowing to do so. If three or more of these symptoms are present in the previ-
ous 12  months then the drug user is considered drug-​dependent. These classification
rules have proven reliable and useful. Different raters agree as to who is an addict (e.g.
Helzer et al. 1987; Spitzer et al. 1979), and research based on these criteria yields orderly
distinctions. Drug users who meet the criteria for addiction reliably differ from drug
users who do not meet the criteria (Anthony and Helzer 1991). Thus, it is reasonable to
390  

390 What addicts can teach us about addiction: a natural history approach

use the APA criteria for distinguishing addicts from non-​addicts. Indeed there is no bet-
ter set of guidelines to follow. (There are several editions of the DSM. We cite the 1994
version, as its approach was used by most of the researchers cited in this chapter.)

4.1  Is addiction a chronic disorder?

The claim that addicts are involuntary drug users has two parts. First, that addiction is
a chronic disorder, requiring lifelong treatment. This is because there is no known cure.
Second, that addicts use drugs against their will. We examine chronicity first.

4.2  Remission in the general population

Addiction researchers have typically recruited their subjects from treatment centers.
However, a well-​known principle in medical research is that studies based on treatment
populations are susceptible to systematic biases. For a given illness, those who are in treat-
ment are more likely to suffer from additional illnesses than those who have the same
illness but are not in treatment (Berkson 1946). The additional illnesses are linked to
poorer outcomes so that treatment studies tend to yield a darker picture of the illness of
primary interest. For addiction, this is particularly important since the additional disor-
ders typically are ones that affect how long someone continues to use drugs. Moreover,
most addicts do not seek treatment (Anthony and Helzer 1991; Stinson et al. 2005). Thus,
to understand addiction’s natural history it is essential to ensure that subjects in treat-
ment are not overrepresented. American epidemiologists solved this problem by organiz-
ing large study samples that numbered in the thousands and mimicked the demographic
characteristics of the American public, using such criteria as gender, age, and ethnicity
(Robins and Regier 1991). The results have proved reliable and interesting.

4.3  Stationary
remission rates according to cumulative
remission functions
Figure 21.1 summarizes the remission rates for dependence on an illicit drug for three of
the four major, psychiatric epidemiological studies based on the DSM. On the X-​axis are
the studies and years of data collection (Conway et al. 2006; Kessler et al. 2005a, 2005b;
Stinson et al. 2005; Warner et al. 1995). On the Y-​axis is the percentage of those who met
the lifetime criteria for dependence on an illicit drug, but at the interview and the pre-
ceding 12-​month period did not. That is, they had been in remission for at least a year.
The graph shows that this was the case for most lifetime addicts. Between 76 percent
and 83 percent were in remission. According to the interviews and diagnostic criteria,
they once met the criteria for addiction but, for at least a year prior to the interview, they
no longer did so. The average age of the participants in these studies was about 41 years
old, and the average age of those who met the criteria for lifetime dependence on drugs
was about 37 years. Since dependence typically begins in early adulthood (Kessler et al.
2005a), the age trends suggest that more than half of those who were ever dependent on
an illicit drug remitted by age 30.
  391

The criteria for identifying addicts 391

Remission in Representative Samples

of Addicted Drug Users (N = approx 2700)

80%

% Cases Remitted
60%

40%

20%

0%
92

03

02
–

–
90

01

01
19

20

20
CS

CS

RC
N

A
ES
N
Figure 21.1  Percentage of those with a lifetime diagnosis of dependence on an illicit drug who
did not meet the criteria for dependence for at least a year prior to the diagnostic interview in
the three major US psychiatric surveys that distinguished between drug dependence and drug
abuse. The dates indicate the period in which the research was conducted.

4.4  Did remission vary as a function of type of drug?

Since the results in Figure 21.1 are at odds with current opinion, it is reasonable to sup-
pose that the findings are misleading. For instance, the national epidemiological studies
included marijuana users, so it is possible that the high remission rates reflect a drug that
many consider not addictive. Figure 21.2 tests this idea. It shows the results for the two
epidemiological studies that provided information on specific illicit drugs. On the X-​
axis is type of drug. On the Y-​axis is percentage in remission. For this analysis, it seemed
reasonable to include the first major national community survey, the Epidemiological
Catchment Area study. (We omitted the ECA survey results from the first graph because
it collapsed abuse and dependence into a single category, and claims regarding “compul-
sivity” are reserved for dependence. However, marijuana abuse and marijuana depen­
dence are difficult to differentiate, so it not unreasonable to include the ECA when the
focus is marijuana.)
Remission did not vary greatly as a function of drug type, and marijuana users were
not the most likely to remit. However, remission did vary by study. For every drug,
remission rates were higher in the NESARC survey. This may reflect how the two stu­
dies grouped their subjects. Remission rates are often lower for abuse than for depen­
dence (e.g. Vaillant 2003), despite the fact that abuse is, by definition, the less severe
conditions. Thus, remission rates for the ECA should be lower since they conflated
abuse and dependence.
392  

392 What addicts can teach us about addiction: a natural history approach

90%

75%

% Remission 60%

45%

30%

15%
ij

s
s

im
ar

od
M

St
pi
O

Type Drug

ECA Survey, 1980–1984

NESARC Survey, 2001–2002

Figure 21.2  Percentage of those in remission as a function of type of drug dependence. The

ECA survey did not distinguish between dependence and abuse.

5  Possible methodological problems in the study

of remission
5.1  Are the remission rates stable?
Marijuana users do not explain the high remission rates. Possibly, then, the results are
misleading because those who had remitted relapsed in the future, as expected according
to the phrase that “addiction is a chronic, relapsing disease.” We tested this possibility
graphically.
If remission in Figures 21.1 and 21.2 is temporary, then a graph of the cumulative
frequency of remission as a function of time will level off well below 50 percent and/​
or show some complex, fluctuating pattern. In contrast, if remission is stable, then
according to Figures 21.1 and 21.2, the cumulative frequency of remission should
increase as a function of time, settling in at an asymptote of 75 percent or higher. We
plotted these predictions in two ways:  first as a conventional cumulative frequency
plot, showing the overall percentage of those in remission as a function of time; second
as a linearized version of the same data but expressed in terms of those still addicted—​
a “survival plot.”
Figure 21.3 shows remission as a function of time since the onset of dependence. The
data are from the Excel data file compiled by NESARC researchers and provided by the
authors (Lopez-​Quintero et al. 2011). On the X-​axis is time since the onset of dependence.
  393

Possible methodological problems in the study of remission 393

Remission as Function of Time in Most

Recent Largest National Survey (NESARC)

1.0
0.9
Cumulative Probability 0.8
of Remission 0.7
0.6
0.5
0.4 Coc Rem = 0.981(1 – e–0.17yr)
0.3 Mj Rem = 0.941(1 – e–0.14yr)
–0.05yr
Alc Rem = 0.95(1 – e )
0.2
–0.015yr
0.1 Cig Rem = “1.38”(1 – e )
Fixed asymptote: 1(1 – e–0.024yr)
0.0
0 10 20 30 40 50 60 70
Years Since Dependence Onset
Figure 21.3  Cumulative probability of remission as a function of the time since the onset of
dependence in National Epidemiological Survey on Alcohol and Related Conditions. The graph is
based on two diagnostic interviews (see text).

On the Y-​axis is the percentage of those who have been in remission for a year or more.
This is the conventional, most straightforward way of presenting remission results. It
shows that remission was stable.
The fitted equations are negative exponentials. The multiplier that precedes the paren-
theses estimates the asymptotic rate of remission (so that it should not be greater than
1.0), and the exponent for the constant term e is the yearly rate of remission. For example,
the equation describing cocaine remission says that each year 17  percent of those still
dependent on cocaine remitted, whereas the equation for alcohol says that 5 percent of
those still dependent on alcohol remitted each year. The equations fit the data well, with
one exception. For cigarettes, the best fitting asymptote was greater than 1.0, which is
not sensible. However, it is apparent that this was because the rate of quitting markedly
increased among those who had smoked for fifty years.

5.2  Exponential cumulative remission functions: discussion

These results are orderly but surprising. First, the asymptotes are 90 percent or higher, mean-
ing that most of those who were dependent have remitted or will do so. Second, each drug
had its own characteristic rate of remission, a result that has no established theoretical basis.
For instance, there is no biological or behavioral theory that says the expected duration of
alcoholism is about 3.4 times longer than for cocaine addiction. Third, as illustrated by this
last example, remission rates were considerably higher for the two illegal drugs. If legal sta-
tus reflects the supposed addictiveness or dangerousness of a drug, these results are contrary
to what should occur. However, the results are in perfect accord with the premise that drug
availability plays a critical role in the persistence of addiction. Fourth, the equations fit the
394  

394 What addicts can teach us about addiction: a natural history approach

data almost perfectly even though they are based on the assumption that the likelihood of
remitting is constant and independent of the duration of dependence. These characteristics
hold for the decay of radioactive particles and assembly line errors, but surely the trajecto-
ries for remission from addiction must be more complex—​either increasing or decreasing
as a function of time since the onset of dependence.
The unlikely results prompted a reanalysis of the remission results. We sought guid-
ance from researchers with experience fitting “decay” data with exponential functions.
Radioactive isotopes have a characteristic decay rate that follows an exponential distri-
bution. Different isotopes have different decay rates, and a given sample of radioactive
material may contain more than one species of radioactive material. According to texts
on radioactive decay, physicists deal with the heterogeneity by plotting decay as a survival
function in logarithmic coordinates, rather than plotting the cumulated decay in linear
coordinates. If there is but one isotope that decays at a constant rate, the survival graph
will follow a linear trajectory. However, if there are two or more isotopes with different
decay rates or an isotope that spawns “daughter” isotopes with different decay rates, the
data points will deviate systematically from linearity. Put in terms of addiction, a single
linear function will describe the relationship between time since the onset of addiction
and those still addicted or, alternatively, more than one function will describe this rela-
tionship, and it may not be linear.

5.3  Linearizedcocaine and marijuana remission

rates reveal non-​stationarity
Figure 21.4 shows the percentage of individuals who are still dependent as a function of
years since the onset of dependence on either marijuana or cocaine. These are the same
data that appeared in Figure 21.3, except that the Y-​axis is logarithmic and the equation
is expressed as a survival function: those still addicted = ae-​λt (or in logarithms: log a –​λt,
so that a is the zero intercept and –​λ is the slope). In the top panel, the solid lines are the
best fitting linear models assuming a constant rate of remitting for up to 30 years since the
onset of dependence. This encompasses 92.5 percent of those dependent on marijuana and
98.5 percent of those dependent on cocaine. The dashed lines were fit to all the subjects,
including those who did not remit by 37 years. For both approaches, there are systematic
deviations from the assumption that the rate of remission remained constant for all addicts.
The bottom panel shows a piecewise fit (by eye), which allows for multiple remission
rates. Table 21.1 lists the rate parameters (λs) for a given span of years and the percentage
of addicts who remitted within that span of years. Remission rates decreased as years of
dependence increased. The changes are small in absolute terms but not trivial in relative
terms. For instance, according to the best fitting straight lines, the rate of remitting from
cocaine was 17 percent per year for those who remitted within 6 years of the onset of
dependence (60  percent of lifetime cocaine dependent informants) but 12  percent per
year for those who remitted after 14.5  years of dependence (9  percent of infor­mants).
In addition, 2.8  percent of those dependent on marijuana and 0.8  percent of those
dependent on cocaine did not remit within the 37-​year period for which the researchers
had records.
  395

Possible methodological problems in the study of remission 395

Decrease in Dependence

100.0
Cocaine
Marijuana
36.8
% Still Dependent

13.5

5.0

1.8

0 10 20 30 40

100.0
“Piecewise Fit,” Assuming
Changes in Rate of Remitting
36.8
% Still Dependent

13.5 7.5%

5.6%
5.0
2.8%

1.8
1.5%
0.8%

0 10 20 30 40
Years Since Onset of Dependence
Figure 21.4  Percentage of those who are still dependent on cocaine and marijuana as a function
of time since the onset of dependence. The y-axis is logarithmic, resulting in a linear relationship
between the variables. The data are the same as in Figure 21.3, but here are plotted as a survival
function.

5.4  Linearized cigarette and alcohol remission rates

Figure 21.5 shows the percentage of those still dependent on alcohol and cigarettes as a
function of time since the onset of dependence. The pattern is somewhat different than
for marijuana and cocaine. First, a higher percentage of those dependent on alcohol and
cigarettes did not remit, even though they had more than 50 years to do so. Second, for
alcoholics the rate of remission changed little. Third, as noted earlier and in contrast
to marijuana and cocaine, the remission rate for smoking increased after 50  years of
dependence on cigarettes, which was the case for 25 percent of dependent smokers. Logic
demands that the increase in remission rate after many years of smoking must be due to a
396  

396 What addicts can teach us about addiction: a natural history approach

within-​individual changes, rather than different rates of remission for different individu-
als. For instance, if the remission rate at the onset of smoking were as high as it was after
54 years of smoking then the percentage of those still smoking after 30.5 years would be
13 percent rather the observed 50 percent. Presumably, the health consequences of smok-
ing become more persuasive as smokers age.

5.5  Are the high remission rates due to missing addicts?

Figure 21.2 showed that the high remission rates did not reflect a bias for selecting sub-
jects who used marijuana, and Figures 21.3, 21.4, and 21.5, showed that remission was not
temporary. Nevertheless, the high remission rates could be misleading. If the researchers

Decrease in Dependence

100.0 Cigarettes
Alcohol
% Still Dependent

36.8

13.5

0 10 20 30 40 50 60 70

100.0 Cigarettes
Alcohol
% Still Dependent

25%
36.8

16%

13.5
9%

0 10 20 30 40 50 60 70
Years Since Onset of Dependence
Figure 21.5  Percentage of those who are still dependent on cigarettes and alcohol as a function
of time since the onset of dependence. See Figure 21.4 legend and text for details.
  397

Possible methodological problems in the study of remission 397

Table 21.1  Remission rate parameter for a given period of dependence and percentage of those
dependent

Cocaine Marijuana Alcohol Tobacco

Yrs Rem percent Yrs Rem percent Yrs Rem percent Yrs Rem percent
Rate Rate Rate Rate

0–​5.5 –​0.172 60 0 –​5.5 –​0.127 51 0–​54 –​0.042 91 0 –​49 –​0.022 67

5.5–​14.5 –​0.151 29 5.5–​14.5 –​0.102 29 —​ —​ —​ 49–​59 –​0.066 17
14.5 –​ 30 –​0.124 9 14.5–​ 28.0 –​0.076 13 —​ —​ —​ —​ —​ —​

failed to count a significant number of addicts, particularly those who did not remit, then
the findings would be biased in favor of remission. As it is reasonable to suppose that the
most severely addicted would be least likely to cooperate with researchers, this is a plausible
hypothesis—but missing addicts are hard to count. However, we can calculate how many
addicts would have to be missing to significantly change the results presented so far. To push
remission much below 50 percent in the US today, it is necessary to assume about 20 mil-
lion missing addicts. This seems unlikely. (See Heyman 2013 for details of this calculation.)

5.6  Doepidemiological surveys yield valid self-​reported drug

use histories?
The survey results, like much of research in psychiatry, are based on self-​report. This
raises the possibility that the participants either exaggerated earlier drug use and/​or
underestimated current drug use, thereby elevating remission rates. Stated more gener-
ally, do participants tell researchers the truth?
Researchers have dealt with this issue by comparing metabolic tests of drug use with
what drug users say about their recent drug use. The basic findings are that self-​reports
match the metabolic results when drug users have no apparent fear of negative conse-
quences, whereas the reports are palpably false when censure or worse is possible (Darke
1998; Land and Kushner 1990; Weatherby et al. 1994). The national surveys adopted meth-
ods that are similar to those that produce good fits between biological and verbal measures.
In support of this point, self-​reported drug use histories predict demographic character-
istics (Anthony and Helzer 1991; Kessler et al. 1994; Robins 1993) and even laboratory,
experimental measures, such as working memory and IQ. For example, in a recent study,
self-​reported drug histories predicted working memory scores in span tests even though
one was a personal history and the other was performance in an on-​the-​spot experimental
procedure (Heyman et al. 2014). It is impossible to concoct this correlation. Thus, it seems
unlikely that the high remission rates are an artifact of inaccurate drug histories.

5.7  Remissionrates are orderly, but when someone remits

is unpredictable
It seems reasonable to suppose that the transition from dependence to non-​dependence
is often complex, entailing many steps, not all of which proceed in the same direction.
398  

398 What addicts can teach us about addiction: a natural history approach

Yet simple equations captured the temporal features of these transitions, accounting for
more than 95 percent of the variance in the data. On the other hand, according to the
logic of the fitted exponential functions, exactly when a particular individual remits is
unpredictable, albeit on average constant. Thus, the results reflect both orderliness and
unpredictability.
The analysis, though, is incomplete. The equations and graphs apply to populations.
We have not analyzed the processes that lead individuals to change. How those pro-
cesses relate to the results presented here is not understood. Possibly, there are addicts
who become increasingly likely to stop using drugs and those who become increas-
ingly likely to keep using drugs, but the population graphs fail to show this because
the two tendencies cancel each other out. And, of course, some drug users may go
through periods in which they are more or less likely to continue using drugs. This
sort of complexity is highly plausible, but has not been studied. However, we need not
wait for a successful theory of individual change to identify the correlates of remission.
These are reasonably well documented, and they provide important information on
the nature of addiction.

6  What prompts remission from addiction?

It is customary to say that addiction is in remission, rather than resolved. However,
“in a state of resolution” may be a more accurate descriptor of the endpoint for most
addicts. The quantitative analyses say that the asymptotic “resolution” rate for those
addicted to illicit drugs is not less than 94 percent and for legal drugs the asymptotic
resolution rate is not less than 90 percent. But what are the circumstances surround-
ing remission? If drug use in addicts is impervious to the factors that affect choice,
then quitting drugs should usually involve extraordinary events, such as medical
interventions, forced detoxifications, and draconian judicial measures. Conversely,
if drug use in addicts remains voluntary then the circumstances surrounding remis-
sion may consist mostly of the economic, familial, and moral concerns that influ-
ence non-​drug, day-​to-​day challenges. A related question is whether the factors that
predict remission from addiction also predict remission from the chronic diseases
to which addiction is often compared, such as diabetes, heart disease, and cancer. To
answer these questions, we first outline the logical implications of the high addiction
remission rates and then summarize what the literature on addiction says about the
correlates of remission.

6.1  Logicalimplications of high remission rates in addicts who

are not in treatment
There is wide agreement that most addicts do not make use of clinical services (Anthony
and Helzer 1991; Stinson et al. 2005). Since most addicts remit, quitting drugs must typi-
cally take place without benefit of professional interventions. What brings about change?
  399

Natural experiments in the history of addiction 399

Absent interventions, the logical answer is that the pressures of everyday life reduce drug
use in addicts.
In-​depth interviews with addicts (e.g. Waldorf et al. 1991) and memoirs (e.g. Burroughs
1959; Rettig et al. 1977) yield the following correlates of why those who meet the APA
criteria for drug dependence stop or greatly reduce drug use as they age: financial pres-
sures, legal pressures, family pressures, hardships associated with pursuing illegal and/​
or stigmatized activities, drug tolerance, witnessing an overdose, wanting to be a bet-
ter parent, the desire to make parents proud rather than embarrassed, involvement in
a self-​help group, an awakened spirituality, a new romantic relationship, the breakup
of a romantic relationship, and so on. The material and emotional costs and benefits
of everyday life, including existential and value laden self-​reflections, are the correlates
of remission from addiction. When ex-​addicts talk about quitting, their comments are
replete with references to values as well as instrumental concerns: “I wanted my parents
to be proud of me,” “I didn’t want to embarrass my children,” “I was not raised to be
bad parent,” “I was not put on earth to be an addict,” “I could no longer afford my drug
habit,” “my life had gotten out of control,” “I was sick of the hassles.” In many cases,
addicts quit abruptly, a process known as “cold turkey.” There is no disease that has this
profile. The narratives associated with diabetes or cancer or even schizophrenia include
no references to the therapeutic efficacy of embarrassment or pride, and the time course
of recovery for these diseases is never all at once. No-​one has stopped being a diabetic
“cold turkey.”
Research that used more objective methods for gathering data supports the memoirs
and interviews and adds the important point that the correlates of addiction distinguish
it from other psychiatric disorders as well as from diseases.

7  Natural experiments in the history of addiction

The history of addiction includes several abrupt, society-​wide changes in policy and
information relevant to drugs. These events share several key features with experiments.
There is a marked change in conditions relevant to drug use (e.g. the legal status of
opiates), and a public record of changes in drug use. However, in contrast to an experi-
ment, we do not have to guess as to how the results might apply outside of the labo-
ratory. For addiction, this distinction is important because drug effects and drug use
vary as a function of context, particularly drug availability, expectations, and societal
trends. Several sources provide detailed and entertaining accounts of these events (e.g.
Courtwright 1982; Musto 1973). Below, we briefly list them. (See Heyman 2009, 2015 for
more detailed summaries.)

7.1 Opiates
In 1914 President Wilson signed off on legislation that criminalized the nonmedical
use of opiates and cocaine—​referred to as the “Harrison Narcotics Tax Act.” Heretofore,
400  

400 What addicts can teach us about addiction: a natural history approach

pharmacies openly sold “medications” containing opiates and cocaine, and those
without access to a nearby pharmacy could order addictive drugs by mail from Sears
Roebuck and Company. The legislation set the stage for drug prohibition, the emer-
gence of heroin as the opiate of choice, a shift from sniffing heroin to injecting it, and
the emergence of the urban, minority, street addict, along with a thriving street addict
culture and economy.
Kolb and DuMez ([1924] 1981, p. 1191), the first to chronicle the effects of the Harrison
Act, write:
Addiction is becoming more and more a vicious practice of unstable people who by their nature
have abnormal cravings which impel them to take much larger doses than those which were taken
by the average person who so often innocently fell victim to narcotics some years ago. Normal
people now do not become addicted or are, as a rule, quickly cured leaving as addicts an abnormal
type with a large appetite and little means to satisfy it.

7.2  Alcohol Prohibition 1920–​1933

In the United States, Prohibition is widely perceived as a failed, counter-​productive social
experiment. Nevertheless, Prohibition reduced drinking in alcoholics, particularly in its
early years. This is reflected best in the decrease in cirrhosis of the liver. Cirrhosis is a stop-
pable but not reversible disease that results primarily from heavy drinking. Prohibition
increased the costs of alcohol as measured by dollars, risk of arrest, and effort. These
increases were accompanied by a marked decrease in the incidence of cirrhosis of the
liver in Canada (Seeley 1960)  and the United States (Miron and Zwiebel 1991). Since
alcoholism is the most common pathway to cirrhosis, this means that increasing the costs
of alcohol decreased alcoholism.

7.3  Surgeon General’s 1964 report on smoking

In 1964 the Office of the US Surgeon General published a 357-​page report on the health
risks of smoking (USDHHS 1964) that summarized hundreds of studies on the medical
consequences of smoking and stopping smoking. The take-​home message was that smok-
ing increases the likelihood of diseases that shorten life, and quitting smoking can reverse
much of the damage. The science was overwhelming. Virtually all scientists and physicians
agreed that smoking was unhealthy. However, individuals differ in how seriously they
regard scientific opinion. Given the assumption that college increases the likelihood of val-
uing scientific research, the correlation between smoking and education should increase
following the 1964 report. As predicted, those with more education have been much more
likely to quit smoking. To our knowledge the frequencies and correlates of diseases are not
so tightly correlated with how their characteristics are presented to the public.

7.4  Whatnatural experiments teach us about addiction

and disease
The natural experiments are informative. They show that social practices, cultural trends,
economics, and even new information can persuade addicts not to take drugs. Of course,
  401

Years of school reduces years of heavy drug use 401

not every addict is persuaded at the same moment, but neither is every voter persuaded by
a given political message. In the context of whether it makes sense to classify addiction as a
disease, the issue is whether judicial actions and public opinion campaigns reduce disease
symptoms. We have no record of legislation altering the frequency of cancerous moles or
diabetic thirst, and it seems hard to imagine these effects in the future. On the other hand,
these are the very measures that have dramatic effects on the frequencies of addiction. But the
distinction is one of degree. Healthy living interventions have had some success persuading
those with Type 2 diabetes to exercise and lose weight which, in turn improves insulin sen-
sitivity, a key symptom of diabetes (Sigal et al. 2006). However, exercise-​related decreases in
diabetes symptoms have a more gradual time course than do documented historical changes
in addiction rates, and the magnitude of behaviorally induced changes in diabetes symp-
toms are considerably smaller than the policy-​related decreases in dependence on drugs. For
instance, we do not see 90 percent of diabetes cases going into remission as a function of age,
even though diabetes victims are more likely to be in treatment than drug addicts.

8  Years of school reduces years of heavy drug use

The degree to which the negative consequences of drug use curtail future drug use
depends, in part, on access to non-​drug alternatives. Recently, some addiction specialists
have begun to refer to non-​drug activities as “recovery capital” or “human capital,” where
recovery/​human capital is loosely defined as “acquired useful abilities that can produce
wealth.” Education is a key component of human capital. This suggests that if non-​drug
alternatives are central to quitting drugs, years of education should predict lower rates of
dependence and, for those who become dependent, higher remission rates. A recent study
tested these predictions (Heyman et al. 2014).
The subjects were opiate addicts in treatment at Boston methadone clinics and volun-
teers, many of whom had been or were heavy drug users. The methadone clinic patients
met the criteria for drug dependence. The non-​clinic subjects were not evaluated accord-
ing to DSM criteria, but many had drug histories that matched the clinic subjects. The
study measures included years of heavy illicit drug use (three times/​week or more), IQ,
impulsivity, working memory, psychiatric symptoms, and years of school. The study’s goal
was to identify the variables that predicted the transition from any drug use to heavy use,
and persistence of heavy use once in place.
IQ, impulsivity, and years of school were highly correlated with illicit drug use (in
the expected ways) when there was no control for other factors. However, in a double-​
hurdle multiple regression analysis that controlled for correlated variables only years of
school continued to predict the transition to dependence and years of illicit heavy use.
Impulsivity remained a significant predictor for heavy drinking and regular smoking. Put
another way, IQ and impulsivity predicted years of school, and years of school predicted
illicit drug use. Figure 21.6 illustrates these relations.
On the Y-​axis is the number of occasions of opiate and stimulant use for high and low
risk drug users, where high-​risk was defined as low IQ and high impulsivity. The graph
shows that high-​risk drug users who attained 14 years or more of schooling were less likely
402  

402 What addicts can teach us about addiction: a natural history approach

Years of school, Impulsiveness, IQ and Drug Use

20000 Most Impulsive 20000 Lowest IQ Subjects

Subjects (>59th
Stimulant + Opiate

(>41st percentile)
percentile)
15000 15000
Occassions

10000 10000

5000 5000

0 0
8 10 12 14 16 18 20 8 10 12 14 16 18 20

Least Impulsive
20000 20000 Highest IQ Subjects
Subjects (<41st
Stimulant + Opiate

percentile) (>59th percentile)

15000 15000
Occassions

10000 10000

5000 5000

0 0
8 10 12 14 16 18 20 8 10 12 14 16 18 20
Years of School Years of School

Community Subjects Only

2900 Avg 14+ Yrs School

2300
Stimulant + Opiate

2800 Avg <14 Yrs School

2200 2700
Occasions

2100 2600
600
400
400
200 200
0 0
Low Impulse High Impulse Hi IQ Low IQ
Figure 21.6  The relationship between the frequency of stimulant and opiate use and years of
school for the most and least at-risk subjects as measured by the Barratt Impulsiveness Scale and
IQ. The top four panels show the results for clinic and community subjects. The filled triangles
indicate the averages for subjects with fewer than 14 years of school (gray) and 14 or more years
of school (black). The bottom panel shows the correlation between school and drug use for just
the community subjects.

to continue to use stimulants and opiates than low-​risk individuals with fewer than 14 years
of school. Years of school (and/​or its correlates) protected high-​risk drug users from further
drug use. This is significant in regard to efforts to reduce the costs of addiction and for
the understanding of addiction. For instance, it seems highly unlikely that years of school
trump the influence of obesity in diabetes.
  403

SUMMARY AND DISCUSSION 403

9  Interventions based on the view that drug use

in addicts is voluntary
The historical correlations between changes in the consequences of drug use and changes
in drug use in addicts have laboratory parallels. In controlled experiments with animals
and humans, including addicts, drug consumption is sensitive to its price and to the avail-
ability of competing non-​drug activities. For example, Cohen and her colleagues (1971)
conducted an interesting study of binge drinking that pitted a priming dose of alcohol
against money for not drinking. The subjects were late-​stage gamma alcoholics, the most
serious that Cohen could find. For every alcoholic, there was a priming dose that initiated
a binge. However, for every binge-​inducing priming dose, there was a monetary incen-
tive that would bring the binge to a halt. Behaviorally-​oriented clinicians tested whether
these findings could be applied to addicts seeking treatment. The answer was not obvi-
ous. The experiments took place in environments that reduced the value of drug use (e.g.
drug-​using friends were absent), and the subjects knew that they could resume drug use
when the experiment was over. Nevertheless, interventions that include contingencies have
proven more effective than those that do not. In the clinic, vouchers for abstinence improve
outcomes for addicts seeking help to quit cocaine (Higgins et al. 2000). In the workplace,
drug-​testing programs that carry the threat of job loss yield an 80  percent to 100  per-
cent decrease in drug use (e.g. Coombs 1997; see Heyman 2009 for review of outcomes
for workplace drug-​testing programs). In the justice system, the Project HOPE contin-
gency program reduced drug use in criminal offenders who were on parole by 72 percent
(Hawken and Kleiman 2009).

10  Summary and discussion

10.1  Basic findings and puzzles
The empirical findings reviewed in this chapter fit neatly together. Individuals who meet
the APA criteria for addiction remain voluntary drug users. The large majority cut back
on drug use or quit altogether, and the factors that persuade them to do so are the deter-
minants of choice. Modern, research-​based choice theory predicts semi-​stable suboptimal
outcomes. Thus, it is possible to explain why addiction is self-​destructive, yet significantly
correlated with its consequences. Reframing addiction as the persistence of voluntary
drug use has important implications for interventions, for social policy, for social rela-
tions with addicts, and for how addicts think about themselves. However, before briefly
addressing some of these points, we need to mention possible misconceptions and puz-
zles that follow from the conclusion that addicts, the label notwithstanding, remain vol-
untary drug users.
1 Reframing addiction as a disorder of choice does not in itself reduce the immense and
often overwhelming personal, societal, and financial costs associated with addiction.
Rather, this perspective more properly restates the nature of the problem and in doing
so promises more effective solutions.
404  

404 What addicts can teach us about addiction: a natural history approach

2 The idea that addicts remain voluntary drug users yields an apparent paradox. By defi-
nition choice is behavior that is governed by its consequences, and addiction is self-​
destructive behavior whose costs outstrip its benefits on balance. If this were not the
case, addiction would not be considered a disorder. Thus, if drug use does not become
compulsive, addiction should not emerge or, if it does, it should persist but for a short
while. Yet, addiction typically persists for years and, for some individuals, persists for a
lifetime, as shown in Figures 21.4 and 21.5. This apparent contradiction been discussed
elsewhere (Heyman 2009). The resolution is that, on balance (globally), heavy drug use
exacts more costs than benefits but, on a moment-​to-​moment basis (locally), just the
opposite may be true when there are complex interdependencies between choices and
consequences (Herrnstein and Prelec 1992). Thus, whether drug use produces benefi-
cial or costly outcomes varies as a function of the manner in which choices are framed,
and in many cases the local frame of reference prevails (Heyman 2009).
3 Why has the disease interpretation persisted if it is so out of step with the data on
remission? This has been discussed elsewhere at some length (Heyman 2009). Here we
will briefly discuss the issue as it relates to the understanding of remission.
Prior to the major epidemiological studies summarized by Figures 21.1, researchers almost
always recruited subjects from clinics or hospitals, many of which doubled as treatment
centers and prisons. These subjects differed from the representative community popula-
tions that inform the remission studies summarized in this chapter. The treatment addicts
were typically unmarried males, with little education, poor work skills, poor social skills,
and a criminal record (e.g. Duvall et al. 1963; Hunt and Odoroff 1962; Vaillant 1966). Clinic
addicts were also more likely to suffer from additional psychiatric and medical ailments
than addicts not in treatment (e.g. Regier et al. 1990). The typical result following treat-
ment was relapse (e.g. Hunt et al. 1971). However, as shown in this chapter, when addicts
are recruited independently of their treatment history, remission—​not relapse—​is most
characteristic of addiction (and, as mentioned earlier, most addicts do not seek treatment).

11 What to do?
If addicts retain the capacity to make non-​drug choices, then interventions should focus
on ways to encourage non-​drug choices. Examples of such programs already exist, as noted
in section 9. However, it should be pointed out that these programs do not simply provide
tangible advantages for abstinence, they also teach their clients to take better advantage of
existing social relations, work, and community resources (Higgins et al. 2003). Put in terms
of choice theory, choice-​based treatment programs help individuals reframe their options
globally. Figure 21.6 supports this point as well. It shows that years of school was correlated
with lower levels of stimulant and opiate use, and that this correlation held for those who
were most likely to use stimulants and opiates. Importantly, “school” included all sorts
of training, including beautician school, computer programming, plumbing, and so on.
Training that led to a career is what mattered, not a particular subject matter. The emphasis
on choice does not rule out psychopharmacological approaches to addiction. For instance,
  405

What to do? 405

methadone reduces the reward value of heroin. However, this approach is less direct and
does not provide addicts with what they need most—​higher valued non-​drug choices.

11.1  Should
insurance plans distinguish between addiction
and diseases?
Addicts cause immense amounts of harm and, given the increasing interdependency that
has accompanied technological change, these harms spread widely. For instance, under
the Affordable Care Act and the Mental Health Parity and Addiction Equity Act, everyone
is asked to pay the treatment costs for heroin addicts, alcoholics, and smokers. Yet the
data presented in this chapter show that the natural history of addiction and diseases are
quite different. Even chronic diseases that have important behavioral components, such
as diabetes, differ markedly from addiction. As emphasized, no-​one has ever quit diabetes
“cold turkey.” Thus, it is highly questionable whether it makes sense for publicly financed
insurance plans to fail to distinguish between addiction and disease. The same line of
reasoning suggests that individuals would make wiser, healthier choices if they directly
bore both the negative and positive consequences of their behavior, rather than having the
negative consequences differentially shared by others. These are difficult matters that may
not have any good solutions. Indeed, it is easy to imagine that one of the reasons that the
disease interpretation of addiction has persisted for so long is that it provides a convenient
way to avoid difficult policy and moral questions. Finally, and related to this last point,
should addiction remain in the Diagnostic and Statistical Manual of Mental Disorders? Its
natural history is quite unlike that of the other psychiatric disorders. There is no other
psychiatric disorder in the DSM whose frequency has varied so much as a function of
social factors. For instance, the overall frequency of addiction varied about 15-​fold for
generations born before and after World War II (the advent of 1960’s youth culture, see
Figure 2.3, Heyman 2009), and no other disorder is as closely linked with interventions
that stress faith and other nonmedical elements, such as Alcoholics Anonymous.

11.2  Catching up with what science says about addiction

Darwin and Wallace’s compelling accounts of evolution by natural selection imply that
voluntary behavior, like involuntary behavior, has a genetic basis and is mediated by the
brain. Countless twentieth-​and twenty-​first-​century studies and experiments reflect these
important and irrefutable implications. The theory of natural selection and its empirical
support should have precluded the argument that addiction is a disease because it has a
biological basis. The epidemiological studies of the 1990s showed that addiction is not
a chronic disorder, and their findings implied that the correlates of remission were the
correlates of choice. Newer research continues to support the original epidemiological
results. It is time for the discussion of addiction to take into account long-​established
understandings of the biology of voluntary behavior and well-​established research on
addiction. What we need now are policies that focus on the best way to reduce destructive
yet voluntary drug use. Such policies may work best when the addict’s capacity for posi-
tive change is given the credit it deserves.
406  

406 What addicts can teach us about addiction: a natural history approach

References
Ainslie, G. (1975). Specious reward: a behavioral theory of impulsiveness and impulse control.
Psychological Bulletin, 82, 463–​96.
American Psychiatric Association (1994). Diagnostic and Statistical Manual of Mental Disorders: DSM-​
IV (4th ed.). Washington, D.C.: APA.
Anthony, J.C. and Helzer, J.E. (1991). Syndromes of drug abuse and dependence. In: L.N. Robins,
and D.A. Regier (eds), Psychiatric Disorders in America: The Epidemiologic Catchment Area Study.
New York: Free Press.
Baumol, W.J. and Blinder, A.S. (1994). Economics: Principles and Policy (6th ed.) Fort
Worth: Dryden Press.
Berkson, J. (1946). Limitations of the application of fourfold table analysis to hospital data. Biometrics
Bulletin, 2, 47–​53.
Burroughs, W.S. (1959). Naked Lunch. New York: Grove Weidenfeld.
Chiten, L., Medley, P., and Russel, B. (1994). Twitch and Shout: A Documentary about Tourette
Syndrome. S.l.: Twitch and Shout Productions.
Cohen, M., Liebson, I.A., Faillace, L.A., and Speers, W. (1971). Alcoholism: controlled drinking and
incentives for abstinence. Psychological Reports, 28, 575–​80.
Conway, K.P., Compton, W., Stinson, F.S., and Grant, B.F. (2006). Lifetime Comorbidity of DSM-​
IV Mood and Anxiety Disorders and Specific Drug Use Disorders: Results From the National
Epidemiologic Survey on Alcohol and Related Conditions. Journal of Clinical Psychiatry, 67, 247–​57.
Coombs, R.H. (1997). Drug-​impaired Professionals. Cambridge, MA: Harvard University Press.
Courtwright, D.T. (1982). Dark Paradise: Opiate Addiction in America before 1940. Cambridge,
MA: Harvard University Press.
Darke, S. (1998). Self-​report among injecting drug users: a review. Drug and Alcohol Dependence, 51,
253–​63.
Duvall, H. J., Locke, B.Z., and Brill, L. (1963). Followup study of narcotic drug addicts five years after
hospitalization. Public Health Reports, 78, 185–​94.
Hawken, A. and Kleiman, M. (2009). Managing Drug Involved Probationers with Swift and Certain
Sanctions: Evaluating Hawaii’s HOPE: Executive Summary. Report submitted to the National
Institute of Justice. https://www.ncjrs.gov/pdffiles1/nij/grants/229023.
Helzer, J.E., Spitznagel, E.L., and McEvoy, L. (1987). The predictive validity of lay Diagnostic Interview
Schedule diagnoses in the general population: a comparison with physician examiners. Archives of
General Psychiatry, 44, 1069–​77.
Herrnstein, R.J., Loewenstein, G.F., Prelec, D., and Vaughan, W. (1993). Utility maximization and
melioration: Internalities in individual choice. Journal of Behavioral Decision Making, 6, 149–​85.
Herrnstein, R.J. and Prelec, D. (1992). A theory of addiction. In: G. Loewenstein and J. Elster (eds),
Choice Over Time. New York: Russell Sage Foundation, pp. 31–​60.
Heyman, G.M. (2009). Addiction: A Disorder of Choice. Cambridge, MA: Harvard University Press.
Heyman, G.M. (2013). Quitting drugs: quantitative and qualitative features. Annual Review of Clinical
Psychology, 9, 29–​59.
Heyman, G.M. (2015). Opiate use and abuse, history of. In: J.D. Wright (ed), International Encyclopedia
of the Social & Behavioral Sciences (2nd ed.), vol 17. Oxford: Elsevier, pp, 236–​42.
Heyman, G.M., Dunn, B., and Mignone, J. (2014). Disentangling the correlates of drug
use: A regression analysis of the associations between frequency of drug use, years-​of-​school,
impulsivity, working memory, and psychiatric symptoms. Frontiers in Psychiatry, 5, 70.
Higgins, S.T., Wong, C.J., Badger, G.J., Ogden, D.E., and Dantona, R.L. (2000). Contingent
reinforcement increases cocaine abstinence during outpatient treatment and 1 year of follow-​up.
Journal of Consulting and Clinical Psychology, 68, 64–​72.
  407

What to do? 407

Higgins, S.T., Sigmon, S.C., Wong, C.J., Heil, S.H., Badger, G. J., et al. (2003). Community
reinforcement therapy for cocaine-​dependent outpatients. Archives of General Psychiatry, 60,
1043–​52.
Hunt, W.A., Barnett, L.W., and Branch, L.G. (1971). Relapse rates in addiction programs. Journal of
Clinical Psychology, 27, 455–​56.
Hunt, G.H. and Odoroff, M.E. (1962). Followup study of narcotic drug addicts after hospitalization.
Public Health Reports, 77, 41–​54.
Kessler, R.C., Berglund, P., Demler, O., Jin, R., Merikangas, K.R., and Walters, E.E. (2005a). Lifetime
prevalence and age-​of-​onset distributions of DSM-​IV disorders in the National Comorbidity Survey
Replication. Archives of General Psychiatry, 62, 593–​602.
Kessler, R.C., Chiu, W.T., Demler, O., Merikangas, K.R., and Walters, E.E. (2005b). Prevalence,
severity, and comorbidity of 12-​month DSM-​IV disorders in the National Comorbidity Survey
Replication. Archives of General Psychiatry, 62, 617–​27.
Kessler, R.C., McGonagle, K.A., Zhao, S., Nelson, C.B., Hughes, M., et al. (1994). Lifetime and
12-​month prevalence of DSM-​III-​R psychiatric disorders in the United States. Results from the
National Comorbidity Survey. Archives of General Psychiatry, 51, 8–​19.
Kolb, L. and Du Mez, A.G. (1981). Prevalence and trends of drug addiction in the United States and
factors influencing it (Public Health Reports, v, 39, no, 21). In G.N. Grob (ed.), Public Policy and the
Problem of Addiction: Four Studies, 1914–​1924. New York: Arno Press.
Land, D. and Kushner, R. (1990). Drug abuse during pregnancy in an inner-​city hospital: Prevalence
and patterns. Journal of the American Osteopathic Association, 90, 421–​26.
Leshner, A.I. (1997). Addiction is a brain disease, and it matters. Science, 278, 45.
Lopez-​Quintero, C., Hasin, D.S., de los Cobos, J.P., Pines, A., Wang, S., et al. (2011). Probability and
predictors of remission from life-​time nicotine, alcohol, cannabis or cocaine dependence: results from
the National Epidemiologic Survey on Alcohol and Related Conditions. Addiction, 106, 657–​69.
Miller, N.S. and Chappel, J.N. (1991). History of the disease concept. Psychiatric Annals, 21, 196–​205.
Miron, J.A. and Zwiebel, J. (1991). Alcohol consumption during prohibition. American Economic
Review, 81, 242–​47.
Musto, D.F. (1973). The American Disease: Origins of Narcotic Control. New York: Oxford
University Press.
National Institute on Drug Abuse. (2010). Testimony before the Subcommittee on Domestic Policy,
Committee on Oversight and Government Reform, United States House of Representatives.
https://www.drugabuse.gov/about-nida/legislative-activities/testimony-to-congress/2010/06/
treating-addiction-disease-promise-medication-assisted-recove
O’Brien, C.P. and McLellan, A.T. (1996). Myths about the treatment of addiction. Lancet, 347, 237–​40.
Regier, D.A., Farmer, M.E., Rae, D.S., Locke, B.Z., Keith, S.J., Judd, L.L., et al. (1990). Comorbidity of
mental disorders with alcohol and other drug abuse. Results from the epidemiologic catchment area
(ECA) study. JAMA, 264, 2511–​18.
Rettig, R.P. (1977). Manny: A Criminal-​Addict’s Story. Boston: Houghton Mifflin.
Robins, L.N. (1993). Vietnam veterans’ rapid recovery from heroin addiction: a fluke or normal
expectation? Addiction, 88, 1041–​54.
Robins, L.N. and Regier, D.A. (1991). Psychiatric Disorders in America: The Epidemiologic Catchment
Area Study. New York: Free Press.
Sacks, O. (1995). An Anthropologist on Mars: Seven Paradoxical Tales. New York: Alfred A. Knopf.
Seeley, J.R. (1960). Death by liver cirrhosis and the price of beverage alcohol. Canadian Medical
Association Journal, 83, 1361–​66.
Sigal, R.J., Kenny, G.P., Wasserman, D.H., Castaneda-​Sceppa, C., and White, R.D. (2006). Physical
activity/​exercise and Type 2 diabetes: a consensus statement from the American Diabetes
Association. Diabetes Care, 29, 1433–​38.
408  

408 What addicts can teach us about addiction: a natural history approach

Spitzer, R.L., Forman, J.B., and Nee, J. (1979). DSM-​III field trials: I. Initial interrater diagnostic
reliability. American Journal of Psychiatry, 136, 815–​17.
Stinson, F.S., Grant, B.F., Dawson, D., Ruan, W.J., Huang, B., and Saha, T. (2005). Comorbidity
between DSM-​IV alcohol and specific drug use disorders in the United States: results from the
National Epidemiological Survey on Alcohol and Related Conditions. Drug and Alcohol Dependence,
80, 105–​16.
United States Surgeon General’s Advisory Committee on Smoking and Health (1964). Smoking and
health: Report of the Advisory Committee to the Surgeon General of the Public Health Service; Surgeon
General’s Report on Smoking and Health. Washington, D.C.: US Department of Health, Education,
and Welfare, Public Health Service; Supt. of Docs., US Government Printing Office.
Vaillant, G.E. (1966). A twelve-​year follow-​up of New York narcotic addicts: IV. Some characteristics
and determinants of abstinence. American Journal of Psychiatry, 123, 573–​85.
Vaillant, G.E. (2003). A 60-​year follow-​up of alcoholic men. Addiction, 98, 1043–51.
Volkow, N.D. and Fowler, J.S. (2000). Addiction, a disease of compulsion and drive: involvement of the
orbitofrontal cortex. Cerebral Cortex, 10, 318–​25.
Waldorf, D., Reinarman, C., and Murphy, S. (1991). Cocaine Changes: The Experience of Using and
Quitting. Philadelphia: Temple University Press.
Waller, N.G., Kojetin, B.A., Bouchard, T.J., Jr, Lykken, D.T., and Tellegen, A. (1990). Genetics and
environmental influences on religious interests, attitudes, and values: a study of twins reared apart
and together. Psychological Science, 1, 138–​42.
Warner, L.A., Kessler, R.C., Hughes, M., Anthony, J.C., and Nelson, C.B. (1995). Prevalence and
correlates of drug use and dependence in the United States. Results from the National Comorbidity
Survey. Archives of General Psychiatry, 52, 219–​29.
Weatherby, N., Needle, R., Cesari, H., Booth, R., McCoy, C.B., and Watters, J.K. (1994). Validity of
self-​reported drug use among injection drug users and crack cocaine users recruited through street
outreach. Evaluation and Program Planning, 17, 347–​55.
  409

Chapter 22

How a stigmatic structure

enslaves addicts
Beth Burgess

Abstract
To tackle addiction effectively, society must confront stigma against addicts,
which pervades at both an individual and an administrative level, and which
reinforces addiction rather than resolves it. This chapter illustrates how
addiction may stem from a wider problem: a “vicious cycle of suffering” fed
by socioeconomic factors. The chapter argues that discriminatory attitudes
and punitive policies are short-sighted and paradoxical, since they make it
harder for addicted people to return to society as healed and contributing
individuals, and that current practices also fail to break the wider cycle and
prevent addiction from ever taking root. The author asserts that society should
be held, in part, accountable for an addict’s choices, and that responsibility
for recovery and prevention should be shared between the individual and the
system that has hitherto judged, rather than helped, people who have usually
been exceedingly vulnerable throughout their entire lives.

1  Introduction: misunderstanding the misery of addiction

“I have no sympathy; he chose his own path” is a common public response when an alco-
holic or drug addict meets a sticky end. People’s sympathy is frequently mediated by the
degree to which they believe the person suffering had a choice in the matter; goodwill
tends to wane when pain is perceived to be self-​induced. People often think of addictions
as self-​inflicted because they zero in on the choice to use substances, disregarding the rest
of the picture. Drug addiction is particularly harshly judged because people assume users
know the risks of taking addictive drugs. The social causes and context of the choice to
use substances are rarely given due significance, and the biological factors preceding and
perpetuating addiction are seldom acknowledged or understood.
When it comes to addiction, people who have never experienced it naturally suffer from
an empathy failure, because they find it difficult to imagine themselves in the same posi-
tion. For example, if someone can easily control their use of alcohol, they will find it diffi-
cult to conceive that another person cannot. They may, therefore, think of alcoholism only
410  

410 How a stigmatic structure enslaves addicts

in terms of choice, without understanding that an alcoholic may have a different physical
response to alcohol than they have. Furthermore, many people do not understand what
biological addiction feels like or how it works. The word “addiction” has become degraded
by its overuse and has wrongly become synonymous with indulgence and not wanting to
stop doing something, rather than feeling completely unable to.
Judgments about addicts are often based on assumptions and misinformation, resulting
in a lack of sympathy. If we are to present the misuse of substances as a choice, we must
balance this by showing what may have influenced that choice. Only in this way can peo-
ple gain a more comprehensive understanding of addiction.
When educated about mitigating circumstances, people sometimes feel more sympa-
thetic to bad decisions, even if they still think a person contributed to their own suffering
(Weiner et al. 1988). Knowing the context enables people to see the parts for which the
individual is not responsible, and gives them greater chances to understand—​and perhaps
sympathize with—​particular choices. While addicts may have chosen to take substances,
people rarely appreciate how often that choice has been underpinned by suffering, social
injustice, and under-​privilege, as well as being influenced by biological factors.
An addict may choose to administer drugs or lift a bottle to their lips, but there are social,
physical, and psychic factors that usually precede substance misuse which it is not within
an addict’s power to choose or refuse. In this chapter, I will show how addiction is often
part of a cycle of suffering and exclusion, usually beginning with childhood trauma. In
many cases, what precedes addiction is traumatic enough to cause biological changes that
make addiction more physically likely and make the brain less resilient to stress. Through
educating people about the context of addiction, I hope to elicit a more an enlightened
response to the problem. Current misconceptions are damaging in the fight against addic-
tion; the problem can only ever be solved by facing up to the truth of its wider vicious cycle.
Other components of the cycle may be both a cause and effect of more suffering, such
as poverty, missed educational opportunities, homelessness, and a slide into crime.
Currently, societal structures do little to help break the painful pattern, preferring instead
to shame and punish addicts for choices made in the midst of suffering. Prisons are full of
addicts with mental health issues; damaged people who may have suffered for their whole
lives, addiction and crime being symptoms of their exclusion and pain. Considering the
back-​story to substance use, it becomes clear that mere punishment is an ineffective solu-
tion to a much larger, more complex problem.
The political system, the justice system, and medical and social services all stigmatize
addicts to some extent. This ill-​treatment may be designed to deter people from using
substances; in reality, it deters addicts from making the choice to recover. Stigma makes it
more likely that, alienated and spurned by society, addicts will return to addiction, com-
pounding everyone’s misery and frustration.
If society wants addiction to stop, it must structure itself, ideologically and politically, in
a way that makes it easier for recovery to be a realistic and rewarding choice for the addict.
For those that refuse to put their condemnation of addicts aside, even when educated about
the mitigating factors, I  hope this chapter will present rational reasons for supporting
  411

Where the vicious cycle of suffering begins 411

recovery anyway. Stigmatization of addicts only leads to more addiction, not less. Recovery
is definitely a choice; I invite the reader to consider whether that commitment to rehabili-
tate is stronger when it’s only a personal choice or when it’s a societal one too.

2  Where the vicious cycle of suffering begins

When I talk about addiction, I am referring to the inability to stop a session of substance
taking after having had an initial “hit” of that substance. This pattern is generally progres-
sive and long term, and is marked by cravings: powerful mental and biological urges to
use a certain substance.
A person must choose to repeatedly use a substance for a pathological addiction to
occur (Kalivas and O’Brien 2008). However, those who overuse substances are usually
predisposed to do so, in part by biological factors. For example, it has been shown that
people who become addicts may start out with malfunctioning dopamine systems, char-
acterized by a lower than average number of D2 dopamine receptors in the brain (Dalley
et al. 2007). There are also likely to be other physiological precursors to addiction, which
are only just being uncovered (see Burgess, Chapter 17, this volume, for more on genetic
factors which may predispose individuals to addiction). So, the first step on the path to
addiction may not be a choice after all, but an intrinsic biology.
The second step on the road to addiction isn’t a choice either. The addict’s risk of addic-
tion can be greatly increased by a stressful or traumatic childhood. The Adverse Childhood
Experiences Study (Felitti et al. 1998) asked thousands of people to report whether they
had suffered harmful childhood experiences, such as enduring psychological, physical,
or sexual abuse, witnessing domestic violence, living with unstable people, or being sep-
arated from biological parents. Incrementally, the more types of adverse event a child
endured, the higher the risk of suffering from addiction in adulthood became. People
reporting the most types of childhood trauma were 12 times more likely to abuse sub-
stances than those who had experienced none.
Exposure to damaging experiences early in life can increase the risk for addiction on a
biological level; it’s not simply a case of wanting to drown one’s sorrows because of pain.
Childhood experiences shape the brain because they occur while it is still growing, and
mistreatment can cause structural alterations that reduce the brain’s capacity for deal-
ing with stress. For example, young adults who have previously suffered from violence,
neglect, or emotional abuse have a smaller ventral subiculum compared with those who
have not been mistreated (Teicher et al. 2011). The subiculum is a brain region respon-
sible for regulating the stress response of the dopaminergic reward system—​a system
already strongly implicated in addiction (Di Chiara 1999; Dalley et al. 2007). A statistic
starkly demonstrating this diminished ability to cope with childhood stress, and its link
with adult addiction, is that running away from home in childhood is six times as likely in
people who later become alcohol dependent and 12 times as likely in those who become
drug-​dependent (Meltzer et al. 2002).
While future emotional and behavioral difficulties can be traced back to the first few
years of life, the cycle of suffering may start even earlier than this. Individual biology
412  

412 How a stigmatic structure enslaves addicts

develops from genes, which have the ability to behave, or “express” themselves, in varying
ways—​and these expressions can be influenced by in utero experiences as well as events
after birth (Gottlieb et al. 2006). Humans develop in ways designed to help them, above
all else, to survive (Godfrey et al. 2007). A growing fetus receives chemical information
from its mother about what to expect from the world; its gene expression varies according
to the data it is given. These gene expressions enable a baby to adapt to the environment
it is about to be born into and can account for differences in everything from bodyweight
to the disposition of the child.
A pregnant woman experiencing physical or mental strain will transmit her own stress
chemicals to her growing baby. A child that has been exposed to excessive maternal stress
in the womb is likely to have psychological or behavioral problems later on (Seckl and
Holmes 2007), because it may biologically “presume” the world to be hostile, vexing, or
unsafe. Prenatal experiences can continue to have an impact throughout someone’s life.
In one study, children whose mothers experienced domestic violence while pregnant had
changes in gene expression influencing the glucocorticoid receptor—​which affects devel-
opment and binds with the stress chemical, cortisol—​that lasted at least into adolescence
as a direct result of their in utero trauma (Radtke et al. 2011).
These studies show that biological setups can precede problems like addiction, some
of which are impossible for the individual to avoid. They show how a cumulative cycle of
damage and a reduced ability to cope with future damage begins and can lead to mental
health complications and addiction. Whatever precisely the difficulties are that heighten
an individual’s addiction risk, be they prenatal or childhood stresses, they are not self-​
inflicted and are not down to personal choice.

3  The social and emotional impact of early damage

Through the lens of the care system, it is possible to look at the likely longer-​term outcomes
of adversity in early life. Sadly, kids in care give us a clear model for tracing the results of
early abandonment, abuse, or neglect. Statistics covering 13‒17-​year-​olds placed in resi-
dential care units reveal the devastating impact of an unstable or traumatic upbringing—​
96 percent of them had psychiatric disorders or problems with substances, compared with
15 percent of the general adolescent population (McCann et al. 1996). Other figures showed
that, compared to other adolescents, young care leavers were four times more likely to have
a mental health problem, five times less likely to achieve good academic grades, and three
times more likely to be in trouble with the law (UK Department for Education 2012).
The example of children in care is used as a clear case of young people who have likely
been neglected or mistreated by biological parents. There are plenty of other people, who
never went into care, who will also have also experienced childhood stress, trauma, or
hardship. Whether or not you can track them through institutions—​in care, psychiatric
hospitals, custody, and prison—​you can certainly find them in community mental health
centers, addiction services, and rehabs. In my experience, most mental health and sub-
stance misuse problems only develop after emotionally challenging events—​and the most
entrenched forms can usually be traced back to childhood difficulties.
  413

Continuing the cycle: crime and addiction 413

Furthermore, once the cycle of suffering has been set up, it makes later adversities even
more likely. Separation, divorce, and serious injury, illness, or assault are twice as likely to
have happened to individuals with mental health problems or addictions (Meltzer et al.
2002). People with such disorders are also much more likely to have been victimized, with
bullying, violence, homelessness, or sexual abuse far more commonly suffered by peo-
ple with substance misuse or psychiatric symptoms than without. Traumatic life experi-
ences are strongly linked with the development of neurotic and psychotic symptoms and
chemical dependency. Stigmatizers may protest that this is merely an effect; intoxicated or
unstable people are more likely to get injured or become involved in fights. If this is true,
it only confirms the sad accumulation of problems following childhood trauma or stress;
it doesn’t negate it. However, at this juncture, consideration of the first self-​directed step
in the process of addiction is due: don’t substance misusers willfully choose to use sub-
stances to make themselves feel better when faced with difficulties? It is possible, I hope,
to contend that addicts sometimes choose to “self-​medicate” with substances, while still
having sympathy for that choice. If people can’t have sympathy for adult addicts, perhaps
they can feel tenderness towards the damaged children that are still inside them, naively
grabbing out for the comfort of a bottle, pipe, or syringe.

4  Continuing the cycle: crime and addiction

One key reason why people stigmatize addicts, particularly drug users, is that addiction
is often perceived to be linked with crime. When an addict crosses over into criminal-
ity, people largely lose any sympathy they had. But if crime can also be contextualized
within the cycle of suffering, perhaps criminality, too can be better understood. It is only
by thoroughly appreciating what is behind someone’s choices that they can be helped to
make better ones.
It is said that hurt people hurt people. Accordingly, not only does having been mis-
treated as a child predispose people to addiction, it also predisposes them to committing
crimes against others (Widom and Maxfield 2001). I  would add that hurt people hurt
themselves too. Crime and addiction can be seen, not only as acts of hostility against the
world, but as the manifestation of a suffering individual’s disregard for their own wellbe-
ing and fate. People know that if you commit crime, you risk condemning yourself to a life
behind bars, just as most understand that if you try heroin, there is some risk of becoming
addicted. If someone chooses to do these things anyway, it shows a lack of respect and
care for not only society, but for themselves.
This self-​destructiveness can partially be explained by what has already happened to
people in this cycle. Negative events during childhood are not experienced using the cog-
nitive skills or balanced reflection of an adult. Instead of realizing that these are rotten
things happening to a good person, children often assume responsibility for these events
(Toner and Munro 1996) and believe these occurrences mean they are inherently bad or
blameworthy. If you grow up feeling that you are somehow bad, unlovable or worthless,
which many of my clients do report in therapy, how is that going to be played out in terms
of what you do with your life? What choices are you more likely to make? Productive ones
414  

414 How a stigmatic structure enslaves addicts

that serve you over the long-​term or destructive ones that lead you to dead ends? If you
think so little of yourself, underneath it all, why wouldn’t you try to soothe yourself in
ways that also destroy you?
Research shows that childhood abuse or neglect increases the risk of juvenile offend-
ing by 59 percent and that formerly abused children tend to be recidivists (Widom and
Maxfield 2001). The risk of offending as an adult rises by a third if a person was maltreated
in childhood, and the risk of their crime being violent also grows by a similar amount.
Crimes committed by people who have experienced past mistreatment can be seen as
a reflection of that treatment—​and this can be played out in a number of ways. Violent
crime can be an outlet for unexpressed anger while petty crime may reflect a feeling of
worthlessness, of not having or deserving a place in normal society. In individuals suffer-
ing from addiction, further crimes may be committed to fund their habit.
Investigations into prison populations reveal that many felons have suffered from social
and economic exclusion and deprivation their whole lives—​and have substance misuse
levels and mental health symptoms that match that. Roughly a third of prisoners were
permanently excluded from school, compared with less than 1  percent of the general
public (Williams et al. 2012, p. 14). Basic skills among prison populations are well below
national averages, with half of young people in custody possessing literacy and numeracy
levels beneath those of an 11-​year-​old (Ecotec 2001). Only a third of adult prisoners were
employed in the four weeks before arriving in custody, and among those who were waged,
their pay was roughly half the national average (Hopkins 2012).
Before condemning criminals as wretched monsters or pointing to a lack of effort or intelli-
gence, consider that more than a quarter of prisoners have been in care compared to 2 percent
of the general population (Dodd and Hunter 1992) and that, as children, a third of prisoners
experienced abuse and 41 percent witnessed violence at home (Williams et al. 2012, p. 7).
These early difficulties are the same things that cause changes in gene expression and are the
same harbingers of later problems with emotional regulation, education, and employment.
In these statistics, it is easy to see how cycles of suffering are perpetuated. Unsurprisingly,
the criminal justice system deals with more than its fair share of mental health disorders
and addictions. Some 63 percent of male prisoners serving sentences and 39 percent of
their female counterparts admitted to hazardous drinking levels in the year leading up to
their arrest (Singleton et al. 1998). More than half of all prisoners reported using at least
one drug during the same time period. The same report showed that fewer than one in ten
prisoners have no reported mental health conditions and that over 70 percent of prisoners
have multiple disorders. Now it is clear why mere punishment can never work.
Although crime is always a choice, it is one which is made primarily by people from
startlingly similar backgrounds. It is far less likely to be a choice that people with more
stable places in society or a better sense of self-​esteem might make. I am not saying that
all addicts are criminals or vice versa, although the two do cross over; I am making a point
about choices. This evidence makes it clear that destructive choices have a sad inevitabil-
ity to them and I  hope this realization will already start to replace stigmatic attitudes
with sympathy. Addicts, prisoners, and other “antisocial” people are sometimes seen as
  415

The pre-cycle: difficult backgrounds 415

mere reprobates, deserving of their fate, but I hope the dark history driving these people’s
choices is now apparent. If we stigmatize these individuals, we punish them for what is
essentially a response to continuous and biologically unbearable stress and trauma that
has likely gone on their entire lives.

5  The pre-​cycle: difficult backgrounds

The public are more likely to condemn the addict’s choices than to consider society’s
own role in this state of affairs; yet, if we go back to the beginning of the cycle, we find
early traumatization. Certain factors have been shown to increase the risk of childhood
mistreatment—​and many of them include social and economic circumstances. Figures
suggest that families on very low incomes could be over 20 times more likely to be affected
by child abuse and neglect than those with higher incomes (Sedlak and Broadhurst 1996).
Local levels of poverty have been associated with various types of child maltreatment,
especially neglect (Drake and Pandey 1996). Unemployment has also been proven to
increase the risk of child mistreatment (Jones 1990).
The reasons why low economic status is a powerful risk factor in childhood abuse and
neglect are wide-​ranging. Financial hardship can cause stress, family isolation, and hope-
lessness, and negative choices are more likely under stressful circumstances. Parents may
be absent and unable to afford safe, high-​quality childcare or may live in more dangerous
neighborhoods. Knowing what we do about the vicious cycle, we must also consider what
may have gone before in previous generations. Parents who become abusive may have
already been through the cycle of suffering. They may be the product of many genera-
tions of the cycle, with their parents, their grandparents, and their great-​grandparents all
caught up in decades of untreated trauma. Intergenerational trauma has been shown to
transmit not only through parental behavior, but even through changes in their children’s
gene expression (Kellermann 2013; Seckl and Holmes 2007; Godfrey et al. 2007).
The stress of poor socioeconomic status growing up can damage children’s brain devel-
opment (Luby et al. 2013) and increase the risk of lifelong difficulties, including mental
health and substance misuse problems, even if that child later becomes wealthier (Conroy
et al. 2010). Societal and financial conditions can clearly contribute to the cycle of suffer-
ing. Since this is an established fact, distributing opportunities in a way that creates more
equal societies should be a priority for governments all over the world—​especially since
doing otherwise is likely to encourage addiction and crime in the long term.
People may insist that they would never hurt their child, no matter what the circum-
stances. I invite the reader to consider a more subtle and insidious type of mistreatment—​
one that often goes unnoticed and can traumatize children without causing cuts or
bruises. Emotional cruelty or neglect, even if it is unintentional, can deeply erode a child’s
self-​esteem. Invalidation, coldness, dismissiveness, ridicule, pushiness, or preferential
treatment of siblings can all make a child feel traumatized and lead to later difficulties.
Heroin users, for example, often report having grown up with parents who are inconsist-
ent or uninterested in their children’s lives (Baer et al. 1974). Since children do not have
416  

416 How a stigmatic structure enslaves addicts

the same powers of comprehension as adults, and are prone to self-​blame, these parent-
ing styles can be very damaging; and these can go on regardless of social or economic
status. When I’ve worked in public addiction services with ex-​offenders, care leavers, and
the homeless, histories of violence, poverty, and severe abuse and neglect are the norm.
When addicts come into my private therapy room, the suffering is more often caused by
emotional ill-​treatment, parental narcissism, or accidental traumatization by caregivers.
So, the cycle of mistreatment, stress, mental health problems, and addiction can happen
to people from all walks of life, but it is more common—​or at least more pronounced—​in
people from underprivileged backgrounds. Knowing this, it might be wiser to promote
policies that minimize these risks than to continue to punish and stigmatize people from
difficult backgrounds.

6  Stigma precludes the choice to recover

At every stage of the cycle of addiction and pain, recovery is possible, but effective help
needs to be in place for that to happen. Criminals can turn their lives around, addicts
can recover, and suffering individuals can heal their broken self-​esteem. Tragically, the
lifelines that are most needed by vulnerable people seem to be the very things that are
withdrawn from them.
For example, finding stable housing and employment are things that offenders say would
most help keep them on the straight and narrow (Williams et al. 2012, p.22). Statistics show
that reconviction rates do drop, in some cases quite dramatically, among ex-​offenders with
jobs and housing awaiting them on release from prison (May et al. 2008). Yet employers
often discriminate against ex-​offending job-​seekers, with some automatically rejecting any
applicant with a criminal record even if they are qualified (Metcalfe et al. 2001). According
to crime reduction charity NACRO (2003, p. 6), employers are afraid to develop policies
for recruiting ex-​offenders openly because they anticipate criticism from the media and
the public. So ex-​offenders are often left with no job, no money to get safe accommodation,
and no self-​worth. Reconviction beckons with a sadly inviting finger.
It is easy to see how stigma can diminish ex-​offenders’ chances of becoming healthy, con-
tributing members of society—​and the story is the same when it comes to addicts. They
find it harder to secure employment, to access non-​judgmental healthcare, and to socialize
safely. This stigmatization continues even after the addiction has ceased, with more than a
quarter of people admitting they would be less likely to hire an otherwise qualified job appli-
cant on the sole grounds of their addiction—​even if they were in long-​term recovery (Peter
D. Hart Research Associates/​Coldwater Corporation 2004). Such stigma actively prevents
addicts from rebuilding their lives and moving even further away from their addiction.
One of the most common reasons for addicts relapsing is experiencing negative emo-
tions. A  treatment center survey showed that relapsers reported boredom, anxiety, and
stress as the most common reasons for it (Levy 2008). Men often cited anger as the key
ingredient in a relapse while women blamed depression and loneliness. The way that soci-
ety treats addicts can contribute to these negative feelings; these painful emotions would
  417

Why is there no change? 417

occur less if addicts were not stigmatized, with all the consequences that stigma brings.
Stigma against vulnerable people can also lead to self-​stigmatization, which causes feelings
of low self-​worth and makes people less likely to seek treatment (Corrigan 2004). Either
they believe that they will not be able to get better—​because they feel inherently flawed—​
or they avoid seeking help because they don’t want to be labelled, judged, or suffer social
exclusion. If addicts don't get treatment, they risk sliding further down the spiral of suffer-
ing, potentially ending up mentally ill, homeless, or involved in crime. These unfortunate
conditions only add more stigma. It’s a fatalistic nightmare that affects the most vulnerable.
If an addict has become a slave to substances, they become even further enslaved by stigma.
The stigmatization of addicts just doesn’t make any sense if society truly wants them to
get better and start contributing. Studies show that improving social status can actually
reduce the risk of relapse. Experiments on groups of previously cocaine-​dependent mon-
keys showed that they were less likely to use cocaine again if they enjoyed a boost in social
status while abstinent. The availability of D2 receptors increased among the dominant
abstinent monkeys (Czoty et al. 2010), making them demonstrate fewer risky behaviors
associated with drug abuse. Previous studies show that monkeys given the opportunity
to take cocaine will take less of it if they have a higher social rank (Morgan et al. 2002).
Their higher social status was, again, correlated with having more D2 receptors available
than subordinate moneys.
Research on humans also found that elevated levels of dopamine receptors were cor-
related with both high social status and generous support from others (Martinez et al.
2010). It makes sense to give more support to people who have a biologically compro-
mised ability to deal with stress, because social reinforcement helps people deal with psy-
chosocial stressors. One study found that having social support reduces the body’s release
of cortisol, meaning that feelings of stress and anxiety are diminished (Ditzen et al. 2008).
When participants in the study also had a secure attachment figure, mood and anxiety
levels were more stable in the face of stress, even if cortisol was high.
Stigma against addicts clearly contributes to the risk of relapse and can inhibit sub-
stance misusers from seeking help to recover. Society making an effort to include, rather
than exclude, people suffering from addictions could play a positive and pivotal role in
helping addicts recover and stay recovered.

7  Why is there no change?

People who misuse substances usually have backgrounds that render it more likely that
they will make negative choices. Whether they are plagued by difficult feelings, have a bio-
logical inability to cope with stress or regulate emotions, or are trying to achieve a greater
feeling of wellbeing, I hope this chapter has made the choice to take substances more read-
ily and sympathetically understood. This begs the question, then, why isn’t more done by
society to help people resolve their current addictions and prevent the risk of it happening
to anyone else? And why does the addict seem to only choose to recover when they’ve
already caused society negative—​criminal, financial, and emotional—​consequences?
418  

418 How a stigmatic structure enslaves addicts

I view the inadequate efforts made to end both the individual’s cycle of suffering, and
the larger social context that feeds it, in terms of a social dilemma (Dawes 1980) and a
social trap (Platt 1973). The problem is perpetuated by two opposite sides starting out
with short-​term, selfish goals in mind.
On the level of individual addiction, addicts, in the short-​term, get a high pay-​off for
their use of substances. At first, they don’t want to stop, because substances help them to
feel better. A lot of addicts start using substances early in life (Warner and White 2003),
certainly before they’ve had the opportunity to fully educate themselves about addiction
or for their brain to have developed sufficiently to make sound judgments. Over time, an
addict may realize that their use of substances is escalating or start noticing negative con-
sequences. This would be the perfect point at which to ask for help. But it is often not in
their interests to ask for help to recover because there may be several personal sanctions
for doing so. They may suffer stigma, have their children taken away, or risk imprison-
ment for drug offenses, on top of having to deal with facing the pain they have been self-​
medicating. Overall, it is easier and safer, in the short-​term, to stay in denial, block out
their guilt and shame with another hit, and pretend they can manage.
In the short-​term, it is not in the interests of the general public to understand addic-
tion or to invest in real solutions either. By ignoring the truth, the public can spend more
of their money, energy, and time on things that promote their own self-​interest. They are
afforded all the self-​esteem benefits and moral superiority of rejecting addicts as delin-
quent rather than suffering. People sometimes stigmatize others for this very reason,
falling prey to human weaknesses, cognitive biases and ego-​protecting fallacies. People
generally feel more satisfaction when someone else is treated less fairly than them, even if
they acknowledge it is wrong (Brickman 1975). People make themselves believe that bad
things only happen to bad people in order to feel reassured and safe (Lerner 1980). It is
much less threatening for the general public to class addicts as people who have caused
their own misfortune than it is to believe that addiction could befall nice people like them
or, heaven forbid, their children. And why should they help addicts, if they’re “bad” people?
Furthermore, there is a real lack of understanding among the general public of how dif-
ficult it is for an addict to quit and avoid relapse. A non-​addict might think: “I would be
able to stay quit in their place,” not considering that an addict’s brain has become sensitized
to substances (see Holton and Berridge, Chapter 9, and Segal, Chapter 20, this volume, for
more detail). Sensitization means that addicts can suffer overwhelming cravings or urges
to use when clean and sober, even if the last thing they consciously want to do is use sub-
stances (Robinson and Berridge 1993). Addicts may try to communicate this bewildering
and frustrating phenomenon to others not suffering from addiction, but they are rarely
believed. This dismissive attitude and denial of addicts’ experience is part and parcel of
stigma; addicts are, after all, “untrustworthy,” “make excuses,” and “lack willpower.” And
even if this horrible snare really did exist, well, they chose their own path, didn’t they?
Predictably, in the long term, individuals’ addictions lead to massive crises for both the
addict and society in terms of health, crime and personal wellbeing. Addiction affects
everybody somehow in the end, and neither addicts nor members of the public win by
  419

Why is there no change? 419

letting the cycle of addiction run its course. Whatever moral judgments people may make,
the reality is that everyone is better off with promotion of, and investment in, recovery.
And what of the larger problem which can precede and perpetuate addiction—​the pov-
erty, poor parenting, and the punitive measures against vulnerable people? While it is
very much in the interests of addicts to heal the larger contextual problem, they have not
historically had the power to do so. Already marred by the fallout from adverse childhood
experiences, many are disempowered and trapped in the system. The most badly affected
may even be locked up in institutions or dying on the streets. If impoverishment and
underprivilege can fuel the cycle of suffering, then the people caught in it won’t have the
wealth, contacts, or education to elevate them to influential public positions. They won’t
even have the confidence to speak up or believe they can change things.
When it comes to the general public, ego-​protecting mechanisms again stop people see-
ing and admitting the truth. If they have to own up to the fact that there is rot in the world,
and that innocent people often get hurt by it through no fault of their own, terror and
sadness can ensue. Even addicts themselves may unconsciously try to minimize their own
suffering in an effort to protect their own egos and those of their parents; they often don’t
allow themselves to see what damage their own family or childhood environment did.
Acknowledging these things is too much to bear and addicts often choose to self-​destruct
rather than address their earliest issues. It is tricky to broach the topic of poor parenting,
especially when parents may have been well-​meaning, and it is upsetting to acknowledge
how many children may have genuinely suffered from their upbringings. But these truths
must be faced in order to deal with the greater problem.
It is not only addicts and the general public who refuse to deal with the true facts of
the matter or who fail to act accordingly to end the cycles of suffering and addiction.
Governments also enter the social dilemma, precluding real solutions through incompatible
priorities. People in power have different agendas from the rest of society—​they inevitably
see things only in the short term, because they need votes in order to retain governmental
seats come election time. While some politicians may understand the cycle of abuse, addic-
tion, crime, and poverty, by political standards the benefits of working on, prioritizing, and
investing in these areas take too long to surface. Effective efforts to eradicate the root causes
of these issues will not return their most substantive rewards until future generations. How
can politicians persuade voters to re-​elect them without some quick, concrete results to
show them for the money and time that has been spent? Rather, it is more in governments’
interests to do superficial work to get impressive-​looking figures than to invest in the long-​
term strategies that will truly pay off later. Hence, it is generally not on politicians’ agendas
to educate the public about the true nature of the addictive cycle, even though it could
reduce stigma and improve rehabilitation rates. Besides, these are complex issues and won’t
fit neatly on the back of a campaign leaflet. It’s quicker and easier to sanction, stigmatize, and
punish addicts, criminals, and people with mental health problems than it is to understand
their background and biology and relay that to the public.
Clearly, not enough is done to prevent people from becoming vulnerable in the first
place either. Most politicians and social services are fairly vocal about their commitment
420  

420 How a stigmatic structure enslaves addicts

to child protection. But, as we have already seen, proper child protection includes not just
working with children already at risk, but reducing any factors that put children at risk
of neglect or abuse. This must necessarily involve reducing inequalities. We have already
noted that it is not in the interests of politicians to introduce the sort of long-​term policies
which would really help deal with addiction. The same goes for inequality—​it cannot be
effectively dealt with using short-​term thinking, therefore, not much is done to truly solve
it where it first takes root.
Governments, then, must take some responsibility here. While an individual makes
the ultimate choice in how they behave, the choices people in power make affect every
human activity. Citizens can only make the best of the information, legislation, policies,
services, and opportunities that governments decide to offer. Therefore an individual’s
choices must necessarily be influenced by the choices that those in power themselves
make. If governments were to take a longer-​term, more practical and effective approach
to ending the cycles of addiction and suffering, many social ills would dissolve over time.
It would take a strong, courageous government to break the mold and convince voters
to wait for the results of prudent long-​term investment in social change, but one whose
reputation would be sealed forever.

8  Choosing a better solution

It is important to note here that by no means all addicts are violently abused children
who later become hardened criminals. These are the most extreme examples of the cycle
of suffering playing out. But all addicts are somewhere on the spectrum of suffering, and
the evidence presented in this chapter was designed to help readers see the clear pathways
of cause and effect in vulnerable people’s lives. I hope it is now easier to understand the
choices made by suffering people.
Substance misuse is a choice. Crime is a choice. Going back into addiction or crime
because no-​one gives you a job, a chance, or the time of day, is also a choice. They are
unwise, antisocial, selfish choices. They are also choices made in response to very chal-
lenging feelings and circumstances. Discrimination is also a choice. Stigma is a choice.
They are choices made due to fear, assumption, misinformation, prejudice, and discom-
fort. They are also unwise, antisocial, selfish choices. These choices, while partially under-
standable, are ineffective in ending the reality that everyone is facing. Addiction is here;
it is a problem that both causes, and is caused by, so many other issues. Making a new
choice to help rather than stigmatize addicts would start to break the cycle of suffering
and addiction and help the whole of society to recover from its impact.
Thoroughly educating both politicians and the public about addictive cycles is neces-
sary to ensure that the recognition of substance use as a choice does not lead to moralis-
tic, punitive attitudes. Perhaps the way forward is framing the information in a way that
promotes responsibility but not blame. Addicts and non-​addicts both distance themselves
from the full truth of addiction, its causes and fallout. Maybe this could be avoided if
people were taught to engage with the facts in a less defensive way. Facing up to the reality
  421

THE BIGGER PICTURE 421

of addiction should not be about blaming, which leaves people resentful and unable to
move forward; it should be about understanding, which gives everyone new power. If
people are frightened by the truth, perhaps engaging their own sense of compassion could
counterbalance the discomfort and terror of facing up to the sad injustices and suffering
in the world. Empowerment comes from understanding; only by everyone acknowledge-
ing the truth and taking responsibility for change can we start applying real solutions to
end the cycle.

9  The “recovery movement”

There are already signs that addicts themselves are starting to mobilize in order to help
people make wiser choices regarding personal recovery and public stigma. The “recov-
ery movement” is a group of recovering addicts who choose to speak openly about their
addictions and professionals who advocate for recovery (White 2006). Addicts have
started sharing their stories in order to prompt others to embark on the journey of recov-
ery and to help lessen the distance between addicts and the general public. Alexander and
Link (2003, pp. 271–​89) showed that increasing contact with people who have tradition-
ally been stigmatized can improve perceptions and reduce ostracism. When people could
be shown that the onset of drug abuse had some mitigating factors, even if it was still seen
as a choice, they were less angry and more willing to help (Weiner et al. 1988). Over time,
I hope the recovery movement will be able to encourage more people to support public
policies that promote all forms of rehabilitation and recovery over punishment.
It is important that the public is not asked to pity addicts, but to understand what has
happened to them and why. Pity is not empowering and could lead to further self-​stigma.
Instead, the recovery movement should focus on educating both addicts and non-​addicts
about the nature of addiction, why self-​destructive choices are made, and on sharing
positive recovery role models with them. If stigma is to be totally eradicated, it must be
destroyed at all levels, including structurally. The recovery movement should aim to share
their educational message with healthcare providers, politicians, doctors, nurses, police
officers, legislators, judges, prison governors, and employers—​people whose decisions
about addicts have practical implications for their recovery.

10  The bigger picture

People with a compromised biology and difficult childhood experiences should not be pun-
ished even further by stigma. That’s not a kind or effective way of dealing with problems. If
someone has already gone down the road of addiction, when they finally make the choice
to recover, they should be given opportunities to break the cycle of pain and addiction—​not
just for themselves, but for the rest of society too. If addicts are supported to get better, and
to redirect their lives in more positive ways, society will benefit. It is also less likely that
addicts’ children and future generations will be afflicted by the same dire cycle.
The matter of whether addicts chose their path or not is much less relevant than the
matter of finding effective solutions to the current problem in wider society. Any other
422  

422 How a stigmatic structure enslaves addicts

attitude would be bloody-​minded and self-​defeating. Even those who don’t have sympathy
for addicts’ choices must accept that current structures and policies have failed to resolve
cycles of damaging behavior. The statistics in this chapter show how deviant behavior
escalates through ignoring the clear pathways of suffering that lead to it. Punishing people
does not work. Making it more difficult for recovering people to rejoin society does not
work. Policies that allow healing and contribution once an addict has committed to the
journey of recovery need to be introduced.
If public opinion still blames the addict for their choice, perhaps restitution that also
builds self-​esteem in the recovering person would be a productive, appeasing measure.
There might be room for policies that include housing for addicts and ex-​offenders who
commit to doing voluntary charity or community work in their spare time. Projects could
be developed that both house and employ recovering addicts, like UK charity Emmaus,
a furniture workshop where homeless people can live in exchange for their labor. There
needs to be a general global message that turning your life around is possible—​and that
governments support it.
Despite some people’s predisposition to addiction, no-​one needs to become an addict.
There are ways to prevent addiction; and it would be better to employ practical preventa-
tive measures than to sweep up the pieces after the damage has already been done. Making
forward-​thinking changes to society could break both the cycle of addiction and the cycle
of suffering. If governments were to invest in progressively eradicating poverty and mak-
ing appropriate resources available to citizens from all social and economic backgrounds,
it could have a powerful role in making better choices more likely for people who have the
potential to become addicts. It could have a powerful role in preventing and ameliorating
adverse childhood experiences. It could have a powerful role in ending many of society’s ills.
For example, decent addiction education could help children to grow up with the tools
to make wiser choices for themselves. Instead of being told horror stories about drugs or
alcohol, addiction science should be part of every school curriculum, so young people
learn how addiction may be caused and how it works. If teens are going to try risky things
anyway, they should be taught the experiential signs that they might be developing an
addiction, so they have the chance to turn back before it’s too late.
There also needs to be a shift in how people view and consume substances which are
dangerous to the human body—​this goes for both the addicted and non-​addicts. There
are questions to be asked before looking at solutions. Why are we so uncomfortable as a
society that poison seems the only way to make life bearable? Why do we feel the need to
“take the edge off ” our own existence, blurring it and making it more palatable? How can
we go about helping people to stop considering substances as preferable to real life and
genuine experience? A common strategy for preventing relapse into addiction is learning
to develop inner happiness and contentment. This way of life makes sense for someone
whose genes and life circumstances have combined to cause them great stress and other
negative emotions. Many addicts find it helpful to turn away from a consumption cul-
ture which will never fulfill them, turning instead to a more spiritual way of life where
they practice gratitude, compassion, mindfulness, and balance. Perhaps the solution to
  423

The bigger picture 423

addiction could be the solution for everyone. I believe everyone should be taught, as chil-
dren, to develop powerful internal resources.
The most effective solution to substance misuse is not physically or legally stopping
people from using drugs or sanctioning them for doing so; it’s stopping people from ever
wanting to use them in the first place. Anything else is a sticking plaster that will lead
to a festering wound. Chronic emotional pain and low self-​esteem, triggered by adverse
childhood events, can lead people to turn against themselves, slowly killing themselves
with substances. It is somewhat cruel and ironic that the society which stigmatizes the end
product of traumatic experiences has not managed to provide relief from the pain. Social
policies have failed to protect many children from risk factors associated with adverse
experiences, such as socioeconomic deprivation. The system which failed to prevent chil-
dren from becoming vulnerable ends up punishing them for their difficulties as adults.
A system which either provides greater support for impoverished communities, or elimi-
nates inequality as far as humanly possible, makes much more sense.
While we can’t necessarily always stop poor parenting or tragedy striking, we can help
children manage feelings of traumatization or worthlessness early on. Years of future
pain and antisocial behavior could be avoided by stopping the suffering where it starts.
Children should be raised in a way that makes them feel valued, loved, and safe. If they
aren’t getting this at home, then it needs to be provided in classrooms and care systems.
That may mean putting therapists and mentors in schools or training teachers to present
themselves as warm, reliable attachment figures. Schools should devise curricula which
support the growth of healthy self-​esteem, self-​respect, and self-​compassion. People who
start out vulnerable, such as children in care, should be given early mentors and role
models—​perhaps the perfect voluntary work for someone in recovery. Personal and social
education should be as great a priority as academic work. Everyone could benefit from
stress management and emotional regulation tools being taught in childhood. Perhaps,
then, no-​one would ever feel the need to rely on a substance and would never set foot on
the slippery slope to addiction.
To end all pernicious vicious cycles, we should aim to build a more balanced society
overall, where every child has equal access to tools that can help them build their self-​
esteem and resourcefulness, so that they can make wise, healthy choices about their lives.
With early intervention in negative cycles, it is possible to break them, so that they do
not haunt generations to come. Hurt people hurt people; people who feel at peace with
themselves lead peaceful existences—​inside and out.

References
Alexander, L.A. and Link, B.G. (2003). The impact of contact on stigmatizing attitudes toward people
with mental illness. Journal of Mental Health, 12, 271‒89.
Baer, D.J., James, J., and Corrado, S.J. (1974). Heroin addict relationships with parents during
childhood and early adolescent years. Journal of Genetic Psychology, 124, 99‒103.
Brickman, P. (1975). Adaptation level determinants of satisfaction with equal and unequal outcome
distributions in skill and chance situations. Journal of Personality and Social Psychology, 32, 191–​98.
424  

424 How a stigmatic structure enslaves addicts

Conroy, K., Sandel, M., and Zuckerman, B. (2010). Poverty grown up: how childhood socioeconomic
status impacts adult health. Journal of Developmental & Behavioral Pediatrics, 31, 154–​60.
Corrigan, P. (2004). How stigma interferes with mental health care. American Psychologist, 59, 614–​25.
Czoty, P.W., Gage, H.D.. and Nader, M.A. (2010). Differences in D2 dopamine receptor availability
and reaction to novelty in socially housed male monkeys during abstinence from cocaine.
Psychopharmacology, 208, 585–​92.
Dalley, J.W., Fryer, T.D., Brichard, L., et al. (2007). Nucleus accumbens D2/​3 receptors predict trait
impulsivity and cocaine reinforcement. Science, 315(1), 267‒70.
Dawes, R.M. (1980). Social dilemmas. Annual Review of Psychology, 31, 69–​93
Di Chiara, G. (1999). Drug addiction as dopamine-​dependent associative learning disorder. European
Journal of Pharmacology, 375, 13–​30.
Ditzen, B., Schmidt, S., Strauss, B., Nater, U.M., Ehlert, U., and Heinrichs, M. (2008). Adult
attachment and social support interact to reduce psychological but not cortisol responses to stress.
Journal of Psychosomatic Research, 64, 479–​86.
Dodd, T. and Hunter, P. (1992). The National Prison Survey 1991. London: HMSO.
Drake, B. and Pandey, S. (1996). Understanding the relationship between neighborhood poverty and
specific types of child maltreatment. Child Abuse & Neglect, 20(1), 3–​18.
Ecotec (2001). An Audit of Education Provision within the Juvenile Secure Estate. London: Youth Justice
Board Publications.
Felitti, V.J., Anda, R.F., Nordenberg, D., et al. (1998). Relationship of childhood abuse and household
dysfunction to many of the leading causes of death in adults—​the adverse childhood experiences
(ACE) study. American Journal of Preventive Medicine, 14, 245–​58.
Godfrey, K.M., Lillycrop, K.A., Burdge, G.C., Gluckman, P.D., and Hanson, M.A. (2007). Epigenetic
mechanisms and the mismatch concept of the developmental origins of health and disease. Pediatric
Research, 61, 5R‒10R.
Gottlieb, G., Wahlsten, D., and Lickliter, R. (2006). The significance of biology for human
development: a developmental psychobiological systems view. In: W. Damon and R.M. Lerner
(eds), Handbook of Child Psychology, Volume 1: Theoretical Models of Human Development (6th ed.).
New York: Wiley, pp. 210‒57.
Hopkins, K. (2012). The pre-​custody employment, training and education status of newly sentenced
prisoners: results from the Surveying Prisoner Crime Reduction (SPCR) longitudinal cohort study
of prisoners. Ministry of Justice Research Series 3/​12. London: Ministry of Justice.
Jones, L. (1990). Unemployment and child abuse. Families in Society, 71, 579–​88.
Kalivas, P.W. and O’Brien, C. (2008). Drug addiction as a pathology of staged neuroplasticity.
Neuropsychopharmacology, 33, 166–​80.
Kellermann, N.P.F. (2013). Epigenetic transmission of holocaust trauma: can nightmares be inherited?
The Israel Journal of Psychiatry and Related Sciences, 50, 33–​39.
Lerner, M.J. (1980). The Belief in a Just World: A Fundamental Delusion. New York: Plenum Press.
Levy, M.S. (2008). Listening to our clients: the prevention of relapse. Journal of Psychoactive Drugs, 40,
167–​72.
Luby, J., Belden, A., Botteron, K., et al. (2013). The effects of poverty on childhood brain
development: the mediating effect of caregiving and stressful life events. JAMA Pediatrics,
167,1135–​42.
Martinez, D., Orlowska, D., Narendran, R., et al. (2010). Dopamine type 2/​3 receptor availability in the
striatum and social status in human volunteers. Biological Psychiatry, 67, 275–​78.
May, C., Sharma, N., and Stewart, D. (2008). Factors linked to reoffending: a one-​year follow-​up
of prisoners who took part in the Resettlement Surveys 2001, 2003 and 2004. Ministry of Justice
Research Summary No. 5. London: Ministry of Justice, p. 6.
  425

The bigger picture 425

McCann, J.B., James, A., Wilson, S., and Dunn, G. (1996). Prevalence of psychiatric disorders in young
people in the care system. British Medical Journal, 313, 1529–​30.
Meltzer, H., Singleton, N., Lee, A., Bebbington, P., Brugha, T., and Jenkins, R. (2002). The Social and
Economic Circumstances of Adults with Mental Disorders. London: Stationary Office.
Metcalfe, H., Anderson, T., and Rolfe, H. (2001). Barriers to employment for offenders and ex-​
offenders. Department for Work and Pensions, Research Report No.155. London: Corporate
Document Services.
Morgan, D., Grant, K.A., Gage, H.D., et al. (2002). Social dominance in monkeys: dopamine D2
receptors and cocaine self-​administration. Nature Neuroscience, 5, 169–​74.
Nacro (2003). Good Practice Report: Recruiting ex-​offenders: the employers’ perspective. London: Nacro
Publications.
Department for Education (2012). Outcomes for Children Looked After by Local Authorities in
England: 31 March 2012. London: DfE.
Peter D. Hart Research Associates/​Coldwater Corporation (2004). Faces and Voices of Recovery public
survey. Washington D.C.: Faces and Voices of Recovery.
Platt, J. (1973). Social traps. American Psychologist, 28, 641–​51.
Radtke, K.M., Ruf, M., Gunter, H.M., Dohrmann, K., Schauer, M., Meyer, A. et al. (2011).
Transgenerational impact of intimate partner violence on methylation in the promoter of the
glucocorticoid receptor. Translational Psychiatry, 1, 2011‒21.
Robinson, T.E. and Berridge K.C. (1993). The neural basis of drug craving: an incentive-​sensitization
theory of addiction. Brain Research Reviews, 18, 247‒91.
Seckl, J.R. and Holmes, M.C. (2007). Mechanisms of disease: glucocorticoids, their placental
metabolism and fetal “programming” of adult pathophysiology. Nature Clinical Practice
Endocrinology & Metabolism, 3, 479–​88.
Sedlak, A.J. and Broadhurst, D.D. (1996). Executive summary of the Third National Incidence Study of
Child Abuse and Neglect (NIS-​3). Washington, D.C.: US Department of Health and Human Services.
Singleton, N., Meltzer, H., Gatward, R., Coid, J., and Deasy, D. (1998). Psychiatric morbidity among
prisoners: Summary report. Office for National Statistics. London: Stationery Office
Teicher, M.H., Anderson, C.M., and Polcaria, A. (2011). Childhood maltreatment is associated with
reduced volume in the hippocampal subfields CA3, dentate gyrus, and subiculum. Proceedings of the
National Academy of Sciences, 109(3), 209–​10.
Toner, M.A. and Munro, D. (1996). Peer-​social attributions and self-​efficacy of peer-​rejected
adolescents. Merrill-​Palmer Quarterly, 42, 339–​57.
Warner, L.A. and White, H.R. (2003). Longitudinal effects of age at onset and first drinking situations
on problem drinking. Substance Use & Misuse, 38, 1983‒2016.
Weiner, B., Perry, R.P., and Magnusson, J. (1988). An attributional analysis of reactions to stigmas.
Journal of Personality and Social Psychology, 55, 738–​48.
White, W. (2006). Let’s Go Make Some History: Chronicles of the New Addiction Recovery Advocacy
Movement. Washington, D.C.: Johnson Institute.
Widom, C.S. and Maxfield, M.G. (2001). An update on the “cycle of violence.” National Institute of
Justice—​Research in Brief. Washington, D.C.: US Department of Justice.
Williams, K., Papadopoulou, V., and Booth, N. (2012). Prisoners’ childhood and family
backgrounds: Results from the Surveying Prisoner Crime Reduction (SPCR) longitudinal cohort
study of prisoners. Ministry of Justice Research Series 4/​12. London: Ministry of Justice.
426  

Chapter 23

Addiction, choice, and criminal law

Stephen J. Morse

Abstract
Some claim that addiction is a chronic and relapsing brain disease; others
claim that it is a product of choice; still others think that addictions have
both disease and choice aspects. Which of these views holds sway in a
particular domain enormously influences how that domain treats addictions.
With limited exceptions, Anglo-American criminal law has implicitly
adopted the choice model and a corresponding approach to responsibility.
Addiction is irrelevant to the criteria for the prima facie case of crime, it is
not an excusing or mitigating condition per se, and it does not contribute
relevant evidence to existing excusing conditions, such as legal insanity. This
chapter evaluates the criminal law’s model of responsibility using scientific
and clinical evidence and dominant criminal law theories. It concludes that
although the law’s approach is generally justifiable, current doctrine and
practice are probably too unforgiving and harsh. Recommendations for
reform conclude the chapter.

1 Introduction
There is a debate among addiction specialists about the degree to which addicts can
exert choice about seeking and using substances and about other behaviors related to
addiction. All agree, as they must, that seeking and using and related actions are human
actions, but there the agreement largely ends. Some, especially those who believe that
addiction is a chronic and relapsing brain disease, think that seeking and using are solely
or almost solely signs of a disease and that addicts have little choice about whether to
seek and use. In contrast are those who believe that seeking and using are constrained
choices but considerably less constrained on average than the first group suggests. This
group is also more cautious about, but does not reject, characterizing addiction as a
disorder. There is evidence to support both positions. There is a third group who believe
that addiction is simply a consequence of moral weakness of will and that addicts sim-
ply need to and can pull themselves up by their bootstraps. The empirical evidence for
the moralizing third view seems weak, although such attitudes play a part in explaining
the limited role the criminal law accords to addiction. The Nobel-​prize winning econo-
mist, Gary Becker, famously argued that addiction can be rational (1996).
  427

Preliminary assumptions about addiction 427

This chapter demonstrates that, despite the debate and claims based on psychological,
genetic and neuroscientific research to expand the mitigating and excusing force of addiction
in evaluating criminal responsibility, existing Anglo-​American criminal law is most consist-
ent with the choice position. It also argues that this is a defensible approach that is consistent
with current science and with traditional justifications of criminal blame and punishment.
The chapter first discusses preliminary issues to avoid potential objections that the
discussion adopts an unrealistic view of addiction. It then provides a general explana-
tion of the responsibility criteria of the criminal law and addresses false or distracting
claims about lack of responsibility. Then it turns to analysis of the criminal law’s doctrines
about addiction to confirm that the criminal law primarily adopts a choice model and that
addiction per se plays almost no role in responsibility ascriptions. It concludes with a gen-
eral defense of present doctrine and practice, but suggests beneficial liberalizing reforms.

2  Preliminary assumptions about addiction

Virtually every factual or normative statement that can be made about addiction is con-
testable. This section tries to be neutral.
The primary criteria of addiction commonly employed at present are behavioral, namely,
persistent drug seeking and using, especially compulsively or with craving, in the face of
negative consequences (Morse 2009). The neural mechanisms of addiction are debatable,
but are being intensively investigated (Hyman 2007), and environmental variables play an
important role in explaining addictive behavior (Kalant 2010). It is unsurprising that per-
sistent use of brain altering substances changes both the brain and behavior. For example,
there are effects on the brain’s reward circuits, memory, perception and motivation, all of
which contribute to the maintenance of addictive behavior.
The most important terms for criminal law purposes are “compulsive” and “negative
consequences.” The concept of compulsion or something like it is crucial to the no-​choice
model because without it addiction is just a very bad habit that is difficult to break. Despite
the current biologizing within the medical approach and scientific advances (e.g. Kasanetz
et al. 2010), there is still no clear understanding of the biology of compulsively and per-
sistently seeking and using substances. Seeking and using are actions, not mechanisms.
There is no gold standard definition of or psychological or biological test for compulsiv-
ity, which must be demonstrated behaviorally. There are extremely suggestive laboratory
findings, especially with non-​human animals (e.g. Everitt and Robbins 2005), but none is
yet diagnostic for humans. We still lack an adequate definition of compulsion that applies
to actions rather than to mechanisms to explore compelled action’s biological basis.
The usual behavioral criteria for compulsion are both subjective and objective. Addicts
commonly report feelings of craving or that they have lost control or cannot help them-
selves. If the agent persists in seeking and using despite ruinous medical, social, and legal
consequences and despite an alleged desire to stop, we infer based on common sense
that the person must be acting under compulsion. It seems that there is no other way to
explain the behavior, but it is not based on rigorous tests of a well-​validated concept.
428  

428 Addiction, choice, and criminal law

Negative consequences, both internalities and externalities, are not necessarily part
of the definition of addiction because, depending on the circumstances, it is possible to
be a high-​functioning addict who does not suffer or impose substantial negative conse-
quences. Contingent social norms and expectations play a role in explaining how negative
the consequences are, but addiction often has severely negative consequences (e.g. over-
dose, cancer, psychosis) independent of social norms and expectations.
There are many findings about the biology and psychology of addicts that differenti-
ate this group from non-​addicts, but none of these findings is independently diagnostic.
Addiction must be demonstrated behaviorally. Although the characterization of addiction
as a “chronic and relapsing brain disease” is widely used, the characterization, “chronic
and relapsing,” is not justified by the data (Heyman 2009, 2013; see also Chapter 21, this
volume). Brain causation and brain differences do not per se make associated behaviors
the signs or symptoms of a disease. All behavior has brain causes and one would expect
brain differences between any two groups exhibiting different behaviors. Moreover, the
relapse data were not gathered on random samples of addicts. They have been largely
gathered from addicts in treatment and this population is disproportionately co-​morbid
with other psychiatric disorders (Heyman 2009). Characterizing a return to maladaptive
behavior as a “relapse” begs the question of whether the behavior is the sign or a symptom
of a disease. The latter must be established first in order properly to refer to the return as a
“relapse” (Fingarette and Hasse 1979). Whether addiction should be considered a disease
like any other is still an open question. Even if addicts have difficulty controlling their
behavior, they are not zombies or automatons; they act intentionally to satisfy their desire
to seek and to use drugs (Hyman 2007; Morse 2000, 2007a, 2009).
Most users of even the most allegedly addictive substances do not become addicts, but some
substances increase the risk. Whether one moves from casual recreational use or medical use
to addiction is influenced by the agent’s set (psychological expectations) and by the setting
(the environment and its cues) (Zinberg 1984). The substance itself does not account for all
the variance in explaining addiction. Some substances appear to be particularly addictive, but
it is extremely difficult empirically to disentangle the causal variables. It is nonetheless clear
that the psychoactive properties of the drug alone do not turn people into helpless puppets.
A debated question is whether addiction should be limited to substances. Large num-
bers of people engage persistently and apparently compulsively in various activities,
often with negative consequences. Gambling is an example. If there are some activities
or non-​drug substances that can produce the same “addictive behavior” as drugs, then
the criminal law response should perhaps respond similarly by analogy. I believe that the
concept of addiction should be expanded beyond drugs, but for this chapter will confine
the analysis to drug-​related addictions.

3  The concept of the person and responsibility

in criminal law
This section offers a “goodness of fit” interpretation of current Anglo-​American criminal
law. It does not suggest or imply that the law is optimal “as is,” but it provides a framework
  429

The concept of the person and responsibility in criminal law 429

for thinking about the role addiction does and should play in a fair system of criminal
justice.
Criminal law presupposes the “folk psychological” view of the person and behavior.
This psychological theory, which has many variants, causally explains behavior in part by
mental states such as desires, beliefs, intentions, willings, and plans (Ravenscroft 2010).
Biological, sociological, and other psychological variables also play a role, but folk psychol-
ogy considers mental states fundamental to a full explanation of human action. Lawyers,
philosophers, and scientists argue about the definitions of mental states and theories of
action, but that does not undermine the general claim that mental states are fundamental.
The arguments and evidence disputants use to convince others itself presupposes the folk
psychological view of the person. Brains don’t convince each other; people do.
For example, the folk psychological explanation for why you are reading this chapter is,
roughly, that you desire to understand the relation of addiction to agency and responsi-
bility in criminal law, you believe that reading the chapter will help fulfill that desire, and
thus you formed the intention to read it. This is a “practical” explanation rather than a
deductive syllogism.
Folk psychology does not presuppose the truth of free will, it is consistent with the truth
of determinism, it does not hold that we have minds that are independent of our bod-
ies (although it, and ordinary speech, sound that way), and it presupposes no particular
moral or political view. It does not claim that all mental states are conscious or that people
go through a conscious decision-​making process each time that they act. It allows for
“thoughtless,” automatic, and habitual actions and for non-​conscious intentions. It does
presuppose that human action will at least be rationalizable by mental state explanations
or that it will be responsive to reasons under the right conditions. The definition of folk
psychology being used does not depend on any particular bit of folk wisdom about how
people are motivated, feel, or act. Any of these bits, such as that people intend the natural
and probable consequences of their actions, may be wrong. The definition insists only that
human action is in part causally explained by mental states.
Responsibility concepts involve acting agents and not social structures, underlying psy-
chological variables, brains, or nervous systems. The latter types of variables may shed light
on whether the folk psychological responsibility criteria are met, but they must always be
translated into the law’s folk psychological criteria. For example, demonstrating that an
addict has a genetic vulnerability or a neurotransmitter defect tells the law nothing per se
about whether an addict is responsible. Such scientific evidence must be probative of the
law’s criteria and demonstrating this requires an argument about how it is probative.
The criminal law’s criteria for responsibility, like the criteria for addiction, are acts and
mental states. Thus, the criminal law is a folk-​psychological institution (Sifferd 2006).
First, the agent must perform a prohibited intentional act (or omission) in a state of rea-
sonably integrated consciousness (the so-​called “act” requirement, sometimes mislead-
ingly termed the “voluntary act”). Second, virtually all serious crimes require that the
person had a further mental state, the mens rea, regarding the prohibited harm. Lawyers
term these definitional criteria for prima facie culpability the “elements” of the crime.
They are the criteria that the prosecution must prove beyond a reasonable doubt. For
430  

430 Addiction, choice, and criminal law

example, one definition of murder is the intentional killing of another human being. To be
prima facie guilty of murder, the person must have intentionally performed some act that
kills, such as shooting or knifing, and it must have been his intent to kill when he shot or
knifed. If the agent does not act at all because his bodily movement is not intentional—​for
example, a reflex or spasmodic movement—​then there is no violation of the prohibition.
There is also no violation in cases in which the further mental state required by the defini-
tion is lacking. For example, if the defendant’s intentional killing action kills only because
the defendant was careless, then the defendant may be guilty of some homicide crime, but
not of intentional homicide.
Criminal responsibility is not necessarily complete if the defendant’s behavior satisfies
the definition of the crime. The criminal law provides for so-​called affirmative defenses
that negate responsibility even if the prima facie case has been proven. Affirmative
defenses are either justifications or excuses. The former obtain if behavior otherwise
unlawful is right or at least permissible under the specific circumstances. For example,
intentionally killing someone who is wrongfully trying to kill you, acting in self-​defense,
is certainly legally permissible and many think it is right. Excuses exist when the defend-
ant has done wrong but is not responsible for his behavior. Using generic descriptive lan-
guage, the excusing conditions are lack of reasonable capacity for rationality and lack of
reasonable capacity for self-​control (although the latter is more controversial than the
former). The so-​called cognitive and control tests for legal insanity are examples of these
excusing conditions. Note that these excusing conditions are expressed as capacities. If
an agent possessed a legally relevant capacity but simply did not exercise it at the time of
committing the crime or was responsible for undermining his capacity, no defense will be
allowed. Finally, the defendant will be excused if he was acting under duress, coercion or
compulsion. The degree of incapacity or coercion required for an excuse is a normative
question that can have different legal responses depending on a culture’s moral concep-
tions and material circumstances. Addiction is always considered the potential basis for
an excusing or mitigating condition.
It may appear that the capacity for self-​control and the absence of coercion are the
same, but for purposes of addressing the relation between addiction and responsibility,
it is helpful to distinguish them. The capacity for self-​control or “willpower,” is conceived
of as a relatively stable, enduring trait or congeries of abilities possessed by the individual
that can be influenced by external events (Holton 2009). This capacity is at issue in “one-​
party” cases, in which the agent claims that he could not help himself in the absence of
external threat. In some cases, the capacity for control is poor characterologically; in other
cases it may be undermined by variables that are not the defendant’s fault, such as mental
disorder. The meaning of this capacity is fraught. Many investigators around the world
are studying “self-​control,” but there is no conceptual or empirical consensus. Indeed,
such conceptual and operational problems motivated both the American Psychiatric
Association Insanity Defense Work Group (1983) and the American Bar Association
(1989) to reject control tests for legal insanity during the 1980s wave of insanity defense
reform in the United States. In all cases in which such issues are raised, the defendant
  431

False starts and dangerous distractions 431

does act to satisfy the allegedly overpowering desire. In contrast, compulsion exists if the
defendant was compelled to act by being placed in a “do-​it-​or-​else,” hard-​choice situation.
For example, suppose that a miscreant gunslinger threatens to kill me unless I kill another
entirely innocent agent. I have no right to kill the third person, but if I do it to save my
own life, I may be granted the excuse of duress. Note that in cases of external compulsion,
unlike cases of no action, the agent does act intentionally. In addition, note that there is no
characterological self-​control problem in these cases. The excuse is premised on external
threats, not on internal drives and deficient control mechanisms.
This account of criminal responsibility is most tightly linked to traditional retribu-
tive justifications of punishment, which hold that punishment is not justified unless
the offender morally deserves it because the offender was responsible. With exceptions
that need not detain us and prove the point, desert is at least a necessary precondition
for blame and punishment in Anglo-​American law. The account is also consistent with
traditional consequential justifications for punishment, such as general deterrence. No
offender should be punished unless he at least deserves such punishment. Even if good
consequences might be achieved by punishing non-​responsible addicts or by punishing
responsible addicts more than they deserve, such punishment would require very weighty
justification in a system that takes desert seriously.

4  False starts and dangerous distractions

This section considers four false and distracting claims that are sometimes made about
the responsibility of addicts (and others): (1) the truth of determinism undermines genu-
ine responsibility; (2) causation, and especially abnormal causation, of behavior entails
that the behavior must be excused; (3)  causation is the equivalent of compulsion, and
4) addicts are automatons.
The alleged incompatibility of determinism and responsibility is foundational.
Determinism is not a continuum concept that applies to various individuals in various
degrees. There is no partial or selective determinism. If the universe is deterministic or
something quite like it, responsibility is possible or it is not. If human beings are fully subject
to the causal laws of the universe, as a thoroughly physicalist, naturalist worldview holds,
then many philosophers claim that “ultimate” responsibility is impossible (e.g. Pereboom
2001; Strawson 1989). On the other hand, plausible “compatibilist” theories suggest that
responsibility is possible in a deterministic universe (Vihvelin 2013; Wallace 1994).
There seems no resolution to this debate in sight, but our moral and legal practices do
not treat everyone or no-​one as responsible. Determinism cannot be guiding our prac-
tices. If one wants to excuse addicts because they are genetically and neurally determined
or determined for any other reason to be addicts or to commit crimes related to their
addictions, one is committed to negating the possibility of responsibility for everyone.
Our criminal responsibility criteria and practices have nothing to do with determinism or
with the necessity of having so-​called “free will” (Morse 2007b). Free will, the metaphysical
libertarian capacity to cause one’s own behavior uncaused by anything other than oneself,
432  

432 Addiction, choice, and criminal law

is neither a criterion for any criminal law doctrine nor foundational for criminal responsi-
bility. Criminal responsibility involves evaluation of intentional, conscious, and potentially
rational human action. And few participants in the debate about determinism and free will
or responsibility argue that we are not conscious, intentional, potentially rational creatures
when we act. The truth of determinism does not entail that actions and non-​actions are
indistinguishable and that there is no distinction between rational and non-​rational actions
or compelled and uncompelled actions. Our current responsibility concepts and practices
use criteria consistent with and independent of the truth of determinism.
A related confusion is that, once a non-​intentional causal explanation has been identified
for action, the person must be excused. In other words, the claim is that causation per se is an
excusing condition. This is sometimes called the “causal theory of excuse.” Thus, if one identi-
fies genetic, neurophysiological, or other causes for behavior, then allegedly the person is not
responsible. In a thoroughly physical world, however, this claim is either identical to the deter-
minist critique of responsibility and furnishes a foundational challenge to all responsibility, or
it is simply an error. I term this the “fundamental psycholegal error” because it is erroneous
and incoherent as a description of our actual doctrines and practices (Morse 1994). Non-​
causation of behavior is not and could not be a criterion for responsibility because all behav-
iors, like all other phenomena, are caused. Causation, even by abnormal physical variables,
is not per se an excusing condition. Abnormal physical variables, such as neurotransmitter
deficiencies, may cause a genuine excusing condition, such as the lack of rational capacity, but
then the lack of rational capacity, not causation, is doing the excusing work. If causation were
an excuse, no-​one would be responsible for any action. Unless proponents of the causal theory
of excuse can furnish a convincing reason why causation per se excuses, we have no reason to
jettison the criminal law’s responsibility doctrines and practices.
Third, causation is not the equivalent of lack of self-​control capacity or compulsion. All
behavior is caused, but only some defendants lack control capacity or act under compul-
sion. If causation were the equivalent of lack of self-​control or compulsion, no-​one would
be responsible for any criminal behavior. This is clearly not the criminal law’s view.
A last confusion is that addicts are automatons whose behavioral signs are not human
actions. We have addressed this issue before, but it is worth re-​emphasizing that even if
compulsive seeking and using substances are the signs of a disease, they are nonetheless
human actions and thus distinguishable from purely mechanical signs and symptoms,
such as spasms. Moreover, actions can always be evaluated morally (Morse 2007a).
Now, with a description of addiction and responsibility criteria in place and with an
understanding of false starts, let us turn to the relation of addiction to criminal responsi-
bility, beginning with the law’s doctrines.

5  Criminal law doctrine and addiction: background

The introduction to this chapter suggested that the law’s approach to addiction is most
consistent with the choice model. The ancient criminal law treated the “habitual” or “com-
mon” drunkard as guilty of a status offense and drunkenness was considered wrong in
itself. The choice model is older than Blackstone, the great eighteenth-​century judge best
  433

Criminal law doctrine and addiction: background 433

known for his Commentaries, which tried to systematize English law. Although the legal
landscape has altered, the choice model is still dominant.
To provide background, this section discusses three illustrative, iconic cases concern-
ing addiction: Robinson v. California (1962), Powell v. Texas (1968), and United States
v. Moore (1973). Although these cases are older, their holdings and reasoning continue to
be robustly emblematic of the criminal law’s response to addiction and to a compulsion
defense based on addiction. Section 6 canvasses current doctrine.
Walter Lawrence Robinson was a needle-​injecting drug addict who was convicted of a
California statute that made it a crime to “be addicted to the use of narcotics” and he was
sentenced to 90 days in jail. The only evidence that he was an addict was needle marks.
Robinson appealed to the Supreme Court on the ground that punishing him for being an
addict was a violation of the 8th and 14th Amendment’s prohibition of cruel and unu-
sual punishment. There were many different opinions written in the case, but a majority
agreed that punishing for addiction was unconstitutional. (As a sad footnote, Robinson
died of an overdose before the case was decided.)
It is difficult to determine precisely what reasoning was the foundation for the Court’s
constitutional conclusion, but for our purposes three stand out: it is unconstitutional to
punish for status alone or because addiction is a disease or because addiction is “involun-
tary.” Herbert Fingarette and Anne Fingarette Hasse demonstrated conclusively decades
ago that the disease rationale collapses into either the status rationale or the involuntari-
ness rationale (1979), so let us examine what implications follow from each of the two.
The status rationale is far more modest and simply builds on the general criminal law
requirement that criminal liability generally requires action (or intentional omission in
appropriate cases). Robinson was not charged with possession or use, but simply with
the status of being an addict. In dissent, Justice White pointed out that if it was unfair to
punish an addict for his status, why would it not be equally unfair to punish him for the
actions that are signs of that status. It is a clever question, but ignores the view of addic-
tion as a chronic and relapsing disorder. On this view, one can be an addict even if one
is not using at the moment. Again, the status argument is modest because it betokens
no genuine widening of non-​responsibility conditions. Indeed, it is a narrowing holding
because the older common law permitted punishment for prohibited statuses.
The “involuntariness” claim more extensively suggests that punishing people for condi-
tions and their associated behaviors that they are helpless to prevent is also unconstitu-
tional. Adopting the involuntariness position would be an invitation to undermining the
choice model in light of some strains of thought about addiction.
Those who wanted to test the meaning of Robinson did not have long to wait. The
defendant-​appellant in Powell, Mr Leroy Powell, was a chronic alcoholic who spent all his
money on wine and who had been frequently arrested and convicted for public drunken-
ness. In the present case, his defense counsel argued that because Mr Powell was afflicted
with “the disease of chronic alcoholism … his appearance in public [while drunk] was
not of his own volition” (p.  517) and thus to punish Mr Powell for this symptomatic
behavior would be a violation of the Eighth Amendment prohibition of cruel and unusual
434  

434 Addiction, choice, and criminal law

punishment. Powell appealed his conviction to the Supreme Court. The Court was asked
to hold that it was unconstitutional to punish a person if a condition essential to the defi-
nition of the crime charged is “part of the pattern of his disease and is occasioned by a
compulsion symptomatic of the disease.” Note that this is an extremely sympathetic case
for a involuntariness excuse. The crime was not serious and the criminal behavior, public
intoxication, was a typical manifestation of his alcoholism.
The Supreme Court rejected Mr Powell’s claim for many reasons, including that it went too
far on the basis of too little knowledge and that it was unclear that providing a defense in such
cases would improve the condition of people suffering from alcoholism. But Justice Marshall’s
plurality opinion was also skeptical of the underlying involuntariness claim and in the course
of the opinion quoted the expert testimony extensively and part of Mr Powell’s testimony in
full. Mr Powell’s proposed defense was supported by the testimony of an expert psychiatrist,
Dr David Wade, who testified that “a ‘chronic alcoholic’ is an ‘involuntary drinker,’ who is
‘powerless not to drink,’ and who ‘loses his self-​control over his drinking’ ” (p. 518). Based on
his examination of Mr Powell, Dr Wade concluded that Powell was, “a ‘chronic alcoholic,’ who
‘by the time he has reached [the state of intoxication] … is not able to control his behavior,
and … has reached this point because he has an uncontrollable compulsion to drink” (p. 518).
Dr Wade also opined that Powell lacked “the willpower to resist the constant excessive
consumption of alcohol.” The doctor admitted that Powell’s first drink when sober was a
“voluntary exercise of will,” but qualified this answer by claiming that alcoholics have a
compulsion that is a “very strong influence, an exceedingly strong influence,” that clouds
their judgment. Finally, Dr Wade suggested that jailing Powell without treatment would
fail to discourage Powell’s consumption of alcohol and related problems. One could not
find a more clear expression of the medicalized, disease concept of addiction to ethanol.
Powell himself testified about his undisputed chronic alcoholism. He also testified that
he could not stop drinking. Powell’s cross-​examination concerning the events of the day
of his trial is worth quoting in full:
Q:  You took that one [drink] at eight o’clock [a.m.] because you wanted to drink?
A: Yes, sir.
Q:  And you knew that if you drank it, you could keep on drinking and get drunk?
A:  Well, I was supposed to be here on trial, and I didn’t take but that one drink.
Q: You knew you had to be here this afternoon, but this morning you took one drink and then
you knew that you couldn’t afford to drink anymore and come to court; is that right?
A:  Yes, sir, that’s right.
Q: Because you knew what you would do if you kept drinking, that you would finally pass out or
be picked up?
A: Yes, sir.
Q:  And you didn’t want that to happen to you today?
A: No, sir.
Q: Not today?
A: No, sir.
Q:  So you only had one drink today?
A: Yes, sir.
(Powell, pp. 519–​20)
  435

Criminal law doctrine and addiction: background 435

On redirect examination, Powell’s attorney elicited further explanation.

Q: Leroy, isn’t the real reason why you just had one drink today because you just had enough
money to buy one drink?
A:  Well, that was just give to me.
Q:  In other words, you didn’t have any money with which you could buy drinks yourself?
A:  No, sir,that was give to me.
Q:  And that’s really what controlled the amount you drank this morning, isn’t it?
A: Yes, sir.
Q:  Leroy, when you start drinking, do you have any control over how many drinks you can take?
A: No, sir.
(Powell, p. 520)
Powell wanted to drink and had that first drink, but despite that last answer his compul-
sion did not cause him to engage in the myriad lawful and unlawful means he might easily
have used to obtain more alcohol if his craving was desperately compulsive. Although
Powell was a core case of an addict, he could refrain from using if he had a good enough
reason to do so.
Although this was a sympathetic case, Justice Marshall for a plurality was simply unwill-
ing to abandon the choice model that guides legal policy and to impose a “one size fits
all” constitutionally required compulsion defense. The case interpreted Robinson as bar-
ring punishment for status and not as imposing a constitutional involuntariness defense.
Finally, note that if the Court had accepted Powell’s argument, it would not have created a
specific “addiction” defense. Rather, it would have adopted a general compulsion defense
in any case in which criminal behavior was a symptom allegedly compelled by a defend-
ant’s disease, whether the disease was addiction or any other.
Now let us turn to Moore. Raymond Moore was almost certainly a trafficking heroin
addict in Washington D.C. who was charged with possession of heroin. Moore’s expert
witness, Dr Kaufman, testified out of the hearing of the jury that Moore was a long-​term
addict, that Moore’s addiction was a disease, and that as a result, Moore was “helpless to
control his compulsion to obtain and use heroin” (p. 1143). Moore requested the judge to
charge the jury that this condition could be a basis for a defense to the possession charges.
Like Leroy Powell, Raymond Moore presents an apparently sympathetic case. Mere pos-
session of heroin is more serious than public intoxication, but it is not a very serious
crime—​at least not in my opinion. Possession is a necessary part of the diagnostic criteria
of the disorder because one cannot use a substance unless one possesses it and there was
uncontested evidence that Moore could not control his compulsion to obtain (possess)
and use the substance. Nevertheless, despite this testimony and in the absence of counter-
vailing evidence, the trial judge refused to instruct that jury that addiction might be the
basis for a compulsion defense, even for a non-​trafficking addict.
Moore was convicted and appealed to the influential United States Court of Appeals for
the District of Columbia Circuit, claiming that his conviction was improper because he
was a heroin addict with an overpowering need to use heroin and should not, therefore,
have been held criminally responsible for being in possession of the drug. According to
Moore, the case had one central issue: “Is the proffered evidence of … long and intensive
436  

436 Addiction, choice, and criminal law

dependence on (addiction to) injected heroin, resulting in substantial impairment of his

behavior controls and a loss of self-​control over the use of heroin, relevant to his criminal
responsibility for unlawful possession” (p. 1144).
Many judges wrote separately, but a majority voted to affirm the conviction, thus rejecting
Moore’s proposed defense. The judges who voted to affirm Moore’s conviction noted vari-
ously that: (1) there was controversy over whether addiction is a disease and whether we are
able to know an addict’s genuine capacity to refrain from using; (2) the defense would apply
to any defendant with impaired behavioral controls, even in the absence of an allegedly
objective cause such as a disease; (3) it would apply not only to possession, but also to any
other crimes committed to support the addiction; and (4) adopting such a defense would
undermine the strong public policy supporting the prohibition of sale and possession of
controlled substances. For these reasons, they rejected adopting Moore’s proposed defense.
There were two very strong dissents. In one, the judge wrote that the common law
should embrace a new principle according to which a drug addict who lacks substantial
capacity to conform his conduct to the requirements of the law as a result of drug use
should not be held criminally responsible for mere possession for his own use. The opin-
ion rejected as speculative the claim that deterrence would be undermined. The judge
recognized that the compulsion claim might be difficult to limit to mere possession, but
evaded the problem by arguing that Congress intended that the defense should not go this
far. In a second, partial dissent, the chief judge of the circuit, David Bazelon, argued that
the principle behind adopting the defense applied to crimes other than mere possession
and that juries should also hear evidence about compulsion arising from addiction when
other crimes were charged, including armed robbery or trafficking.
Taken together, these cases appear to adopt the choice perspective for two rea-
sons:  addicts have sufficient choice, and the public policy supporting criminalization
would be undermined by providing a defense, even if it could be shown that addicts have
little choice about mere possession and perhaps other crimes related to their addiction.
With these background cases in mind, let us now turn more generally to current doctrine
to explore the criminal law’s choice model.

6  Current doctrine and the choice model

Recall that crimes are defined by their elements and that affirmative defenses are avail-
able even if the prosecution is able to prove all the elements of the crime. This section will
first discuss the affirmative defenses, then it will address the use of intoxication to defend
against the elements of the crime charged, which is termed “negating” an element, and
will finally discuss the role of addiction in sentencing and diversion.
Given that there is still controversy about how much choice addicts have, it is per-
haps unsurprising that the conclusions in Powell and Moore are still regnant. The criminal
law has avoided expanding a defense based on addiction raised by the Moore dissenters.
Addiction is not an affirmative defense per se to any crime in the United States, England
or Canada. With one limited and somewhat unsettled exception in English homicide law
  437

Current doctrine and the choice model 437

(Ashworth and Horder 2013, pp. 271–​72; R. v Bunch 2013), it is also not the basis for any
other affirmative defense, such as legal insanity. Indeed, some United States jurisdictions
explicitly exclude addiction (or related terms) as the basis for an insanity defense despite
the inclusion of this class of disorder in the American Psychiatric Association’s Diagnostic
and Statistical Manual of Mental Disorders, Fifth Edition (DSM-​5) (2013). The claim that
an intoxicated addict might not have committed the crime if he had not been intoxicated
has no legal purchase, although some, such as the great English criminal law scholar,
Glanville Williams (1961, p.  564), disagree. Indeed, addiction does not even merit an
index entry in most Anglo-​American criminal law texts, except in the context of the use
of alcohol intoxication as a defense in some instances that will be explored below.
The only exception to the bar to using addiction as an affirmative defense or the basis of
one is what is known in the United States as “settled insanity.” If a defendant has become
permanently mentally disordered beyond addiction, say, suffers from delirium tremens
as a result of the prolonged use of intoxicants, the defense of legal insanity may be raised.
An enormous number of crimes are committed by people who are under the influ-
ence of intoxicating substances. In what follows I shall discuss the use of intoxication to
negate the elements of the crime charged, but readers should know that these doctrines
apply generally to addicts and non-​addicts alike. Of course, addicts are more likely to
be high than non-​addicts and thus these rules will disproportionately affect them, but
the application to addicts will be the same as to non-​addicts. Whether the criminal law
should distinguish addicts from non-​addicts for these purposes will be discussed in the
next section of the chapter.
Recall that most crimes require a mens rea, a culpable mental state that accompanies the
prohibited conduct. How evidence of intoxication might be used to negate the elements of
the crime charged is the question of logical relevance: does the evidence of intoxication in
fact tend to show that an element was not present? First, the defendant might be so drunk
that his consciousness is sufficiently dissociated to negate the act requirement. Second, the
defendant’s intoxication may be relevant to whether he formed the mental state, the mens
rea, required by the definition of the crime. For example, imagine a very drunk defendant in
the woods with a gun. In the drunken belief that he is shooting at a tree because his percep-
tions are so altered, he ends up killing a human being wearing camouflage gear. If he really
believed that he was shooting at a tree, he simply did not form the intent to kill required for
intentional homicide. To take another example, imagine that a very drunk patron at a bar
walks out without paying the bill. Suppose the bar owner claims that he has been defrauded
by the drunk patron, a form of theft. The patron claims that he was so drunk that he forgot
to pay the bill but formed no intent to steal. If this form of criminal behavior requires the
intent to steal and we believe the patron, he simply did not form that intent.
The logical relevance point is straightforward. If the defendant did not act or lacked the
mens rea for the crime charged, how can he be guilty of that crime (although he may be
guilty of some other offense for which he does have the mens rea)? Despite this logic, a
substantial minority of United States jurisdictions refuse to admit into evidence undeni-
ably factually relevant and probative voluntary intoxication evidence proffered to negate
438  

438 Addiction, choice, and criminal law

mens rea. The remaining United States jurisdictions and English law admit it only with
substantial restrictions.
The reasons for complete exclusion and for restriction of the admissibility of relevant
evidence of voluntary intoxication result, I believe, primarily from the choice model and
from fears for public safety. In the case of restricted testimony, the rules are highly techni-
cal, but typically evidence of intoxication is admitted to negate the mens reas for some
crimes but not for others, even if mens rea in the latter case might actually be negated.
The defendant will therefore be convicted of those crimes for which intoxication evi-
dence is not admissible even if the defendant lacked mens rea. The rules are a compromise
between culpability and public safety and the apparent unfairness of convicting a defend-
ant of a crime for which he lacked mens rea is in part justified by his own fault in becom-
ing intoxicated, a classic choice model rationale.
Leading precedents in the United States and England adopt choice reasoning explicitly.
In Montana v. Egelhoff (1996) the United States Supreme Court held that complete exclu-
sion of voluntary intoxication evidence proffered to negate mens rea was not unconstitu-
tional. Justice Scalia’s plurality opinion provided a number of reasons why a jurisdiction
might wish on policy grounds to exclude otherwise relevant, probative evidence. Among
these were public safety and juror confusion. But one is a perfect example of the choice
model. “And finally, the rule comports with and implements society’s moral perception
that one who has voluntarily impaired his own faculties should be responsible for the
consequences” (p. 50). This view is standard in both common law and continental crimi-
nal law (in which it is called actio libera in causa): a defendant should not benefit from a
defense that he has culpably created. The choice model is strongly at work.
In D.P.P. v. Majewski (1977), a unanimous House of Lords upheld one of the technical
distinctions alluded to above that permit defendants to introduce intoxication evidence
to negate the mens reas of only some crimes. Most of the Lords recognized that there was
some illogic in the rule, but all upheld it as either a justifiable compromise or as sound in
itself and it had long provenance. Most striking for our purpose, however, is one passage
from Lord Elwyn-​Jones’ opinion for the Court. He wrote:
If a man of his own volition takes a substance which causes him to cast off the restraints of reason
and conscience, no wrong is done to him by holding him answerable criminally for any injury he
may do while in that condition. His course of conduct in reducing himself by drugs and drink to
that condition in my view supplies the evidence of mens rea, of guilty mind certainly sufficient for
crimes of basic intent. It is a reckless course of conduct and recklessness is enough to constitute the
necessary mens rea in assault cases … The drunkenness is itself an intrinsic, an integral part of the
crime, the other part being the evidence of the unlawful use of force against the victim. Together
they add up to criminal recklessness.

(pp. 474–​75)

In other words, the culpability in getting drunk—​itself not a crime—​is the equivalent of
actually foreseeing the consequences of one’s actions even if the intoxicated defendant
did not foresee them. Such reasoning—​Majewski chose to get drunk, after all—​presaged
Justice Scalia’s argument in Egelhoff and is clearly based on the choice model.
  439

Addiction-related legal practices 439

Despite massive academic criticism of the Majewski rule and numerous Law
Commission reform proposals, it remains the rule and many think it works reasonably
well. Some Commonwealth countries, such as Australia, New Zealand and Canada, have
the more expansive logical relevance rule and it seems not to have opened the floodgates
of alcohol-​awash crime (Ashworth and Horder 2013). Apparently, however, juries in those
jurisdictions seldom fully acquit, suggesting that the culpability based on choice model is
implicitly guiding decision-​making even if the law is more lenient.Finally, even the Model
Penal Code in the United States, which has had major influence on law reform and which
strongly emphasizes subjective culpability and rejects strict liability of the sort Majewski
potentially imposes, adopted a similar rule in Section 2.08(2) of the Code (American Law
Institute 1962). If an intoxicated defendant was not aware of a risk he would have been
aware of if he was sober, then he will be held to have been aware of the risk. When sub-
stances are involved, the choice model seems recalcitrant to change.
In short, Anglo-​American rules concerning the effects of voluntary intoxication on
prima facie culpability strongly reflect the choice model with no or some qualifications.
The need for completeness compels me at this point to mention involuntary intoxica-
tion, that is, intoxication occasioned through no fault of the agent. Examples would be
mistakenly consuming an intoxicant, or being duped into or forced to consume one. The
law treats such cases more permissively than cases of voluntary intoxication by providing
a limited complete defense and the ability to negate all mens rea. But it does not apply
to intoxication associated with addiction because the law currently treats such states of
intoxication as the agent’s fault even though many addiction specialists would vehemently
disagree. The law’s view of involuntariness in this context could apply to addicts and non-​
addicts alike. Even addicts could be duped or coerced into becoming intoxicated on a
given occasion.

7  Addiction-​related legal practices

There are two United States contexts in which addiction has potential mitigating
force: sentencing, particularly capital sentencing, and diversion to specialized drug courts.
There are no studies that empirically examine the degree to which evidence of addiction
is sought to be used as a mitigating factor during non-​capital sentencing and it is never
listed as a statutorily specified mitigating factor. It is probably the case that the same con-
siderations about its impact would apply in both non-​capital and capital sentencing, so
I shall discuss only the latter.
Beginning in 1978, the United States Supreme Court has repeatedly held that capital
defendants can produce virtually any mitigating evidence (Lockett v. Ohio 1978) and the
bar for the admissibility for such evidence is low. Thus, even if addiction is not a statutory
mitigating factor, an addicted defendant convicted of capital murder may certainly intro-
duce evidence of his condition for the purpose of showing that addiction diminished his
capacity for rationality or self-​control or to support any other relevant mitigating theory.
Doing so also raises the danger that addiction will be thought to aggravate culpability
440  

440 Addiction, choice, and criminal law

based on the choice perspective—​especially the moralistic strain—​and it is possible that

it will make the defendant seem more dangerous, which is a statutory aggravating factor
in some jurisdictions. Addiction is a knife that could cut both ways in capital and non-​
capital sentencing.
Drug courts are an increasingly common phenomenon in the United States. The sub-
stantive and procedural details vary across jurisdictions, but these courts aim to divert
from criminal prosecution to the drug courts addicted criminal defendants charged with
non-​violent crimes whose addiction played a role in their criminal conduct. If diverted
defendants successfully complete the drug court imposed regimen of staying clean and
in treatment, they are discharged and the criminal charges are dropped. This approach
seems eminently sensible and these courts have fervent supporters, but they also have
critics on the grounds that they do not afford proper due process and genuinely solid evi-
dence for their cost-​benefit justified efficacy is lacking. Whatever the merits of the debate
may be, drug courts are now an entrenched feature of criminal justice in a majority of
United States jurisdictions and they do permit some number of addicts to avoid criminal
conviction and punishment.

8  A defense of current criminal law

Criminal law is generally unforgiving towards addicts specifically and those doctrines
that might sometimes favor addicts, such as the rules about negating mens rea, are not
specific to addicts but apply more generally. Given the profound effects of addiction,
can such unyielding rules be fair? Although many addicts are responsible for becom-
ing addicted, the following discussion will assume that an addict is not responsible for
becoming an addict, say, because he became addicted as a youth or because he was in
pathological denial about what was happening. I shall also assume that the rules apply to
adults and that juveniles require special treatment.
Let us begin with affirmative defense. Consider an addict who is broadly mentally
debilitated by chronic intoxication. Recall that the law is already forgiving in such cases,
permitting the addict suffering from “settled insanity” to raise the full excusing condition
of legal insanity. Most addicts are not so severely debilitated, however, so let us turn to the
more “typical” addict.
I believe that there are roughly two accounts for why addicts might not be responsi-
ble for addiction-​related crimes, including possession and other crimes committed to
obtain drugs (Morse 2011). The first is irrationality. As a result of various psychological
factors, including cue salience, craving, memory, and other variables, at times of peak
desire the addict simply cannot “think straight,” cannot bring to bear the good reasons
to refrain. This assumes that addicts do have good reasons to refrain, but this may not
always be true (Burroughs 2013, esp. pp.  144–​7). The irrationality theory is consistent
with the view that regards self-​control difficulties as resulting from an agent’s inability to
consider distant rather than immediate consequences. The other account uses a different
form of self-​control that analogizes the addict’s subjective state at times of peak craving as
akin to the legal excuse of duress. The addict is threatened by such dysphoria if he doesn’t
  441

A defense of current criminal law 441

use substances that he experiences the situation like a “do it or else” threat of a gun to
one’s head. Whether one finds these accounts or another convincing, there is surely some
plausible theory of excuse or mitigation that would apply to many addicts at the time of
criminal behavior. A very attractive case for a more forgiving legal response arises if one
believes that once an agent is addicted, he will inevitably be in an excusing state at the
time of his crimes on some and perhaps most occasions,
There are at least three difficulties with this position, one of which seems relatively deci-
sive. First, much is still not understood about the actual choice possibilities of “typical”
addicts. Maybe most can in fact think straight at the times of their crimes but choose not
to or they are not substantially threatened by dysphoria or, even if they are threatened with
severe dysphoria, they retain the capacity not to give in. The criminal law is justified in
adopting the more “conservative” approach under such conditions of uncertainty. Second
and relatedly, unforgiving criminal law doctrines enhance deterrence. The demand for
and use of drugs is price elastic for addicts. Addicts retain capacity for choice. The threat
of criminal sanctions might well deter addiction-​related criminal behavior on the margin.
The third and seemingly most decisive reason is the potential for diachronous respon-
sibility (Kennett 2001) for addicts who do not suffer from settled insanity. Even if they are
not responsible at the times of peak craving, as previously discussed, at earlier quiescent
times they are lucid. They know then from experience that they will again be in a psy-
chological state in which they will find it subjectively very difficult not to use drugs or to
engage in other criminal conduct to obtain drugs. In those moments, they are responsible
and know it is their other-​and self-​regarding duty not to permit themselves to be in a
situation in which they will find it supremely difficult to refrain from criminal behavior.
They then must take whatever steps are necessary to prevent themselves from allowing
that state to occur, especially if there is a serious risk of violent addiction-​related crimes
such as armed robbery or burglary. If they do not, they will be responsible for any crimes
they commit, although they might otherwise qualify for mitigation or an excuse.
An analogy from criminal law may be instructive. In a famous case, a person suffering
from epilepsy and subject to seizures had a seizure and blacked out while driving on the
public highway (People v. Decina, 1956). His automobile ran up on the curb and killed
four pedestrians. Because he was blacked out, the killing conduct was not his act and he
had no mens rea at the time of the killing. Nevertheless, he was held liable for negligent
homicide as a result of his careless previous act of driving while knowing he was subject to
seizures. Unless addicts are always non-​responsible, an assertion contradicted by the clin-
ical facts, diachronous responsibility is a sufficient ground to deny an excuse to addicts.
For similar reasons, the criminal law is justified in not providing addicts with enhanced
ability to negate mens rea. Recall that the law limits the use of intoxication evidence to
negate mens rea in part because it views most cases of intoxication as the user’s fault.
Many would claim, however, that the intoxication of addicts is a sign of their disorder
and not their fault. Thus, a crucial part of the rational for limitations on mens rea nega-
tion should not apply to addicts. Nevertheless, for the reasons addressed just above, when
addicts are not intoxicated and not in peak craving states, they know they will become
442  

442 Addiction, choice, and criminal law

intoxicated again unless they take steps to avoid future intoxication, which they are capa-
ble of doing when lucid. Consequently, the law need not be more relaxed about mens rea
negation for addicts than non-​addicts.
Two counter-​arguments to the above reasons to retain current law are denial and lack of
opportunity. As people slide into addiction—​and almost no-​one becomes an addict after
first use—​they may well deny to themselves and others that they are on such a perilous
path. This suggests that they may not be fully responsible or responsible at all for becom-
ing addicts. Genuine addicts, or at least most of them, know they are addicted or at least
understand that there is a “problem.” Assertions to the contrary are again inconsistent
with the clinical facts. Even if denial, anyway a vexed concept in psychiatry, prevents
addicts from understanding that they are addicted, if they get into trouble with the law as
a result of drug use, they know that they at least have a “problem” resulting from use. At
that point, they also know in their lucid moments that they have the duty to take the steps
necessary to avoid criminal behavior. Diachronous responsibility still obtains.
By lack of opportunity, I  mean the limited treatment resources available in many
places to addicts who wish to exercise their diachronous responsibility and to refrain
from further criminal behavior. We know from spontaneous remission rates that most
addicts can apparently quit using permanently without treatment, but typically they do
so after numerous failed attempts and only after they have recognized the good reasons
to do so, usually involving family obligations, self-​esteem or the like (Heyman 2009,
2013). Fear of criminal sanctions appears to be an insufficient reason for many. Thus,
especially when the typical addict is young, having trouble quitting, and at higher risk
for crimes other than possession, it may be too much to ask of such addicts to refrain
without outside help. If outside help is unavailable, diachronous responsibility would be
unfair. I think that there is much to this counter-​argument, although it certainly weakens
as the addiction-​related crimes become more serious, such as armed robbery or even
homicide.

9  Criticism of current legal regulation of addiction

Having offered a principled defense of current legal doctrines concerning addiction,
I should now like to suggest that on both rights and consequential grounds, the crimi-
nal law concerning controlled substances and addictions is misguided. Space limitations
prevent me from offering anything but the most superficial, sketchy gesture towards my
preferred regime, but here it is. There is a powerful case based on a liberal conception
of negative liberty that would grant citizens the right to consume whatever substances
they wish as long as they internalize foreseeable externalities through insurance or other
means. I fully recognize that decriminalization would be fraught and unpredictable and
that the dangers may be great (MacCoun and Reuter 2011), but the risks are worth tak-
ing in the name of liberty. Even if the law did not decriminalize drugs, no-​one should
be prosecuted for possession of small amounts of any drug for personal use. The moral
and political arguments for the right to consume what a competent adult chooses are too
powerful (e.g. Husak 1992, 2002).
  443

Conclusion 443

The second ground is consequential. The “war on drugs” in the United States is such an
abject failure that I am willing to take the risk of decriminalization to reduce the overall
harms to individuals and to society at large. I do not base this position on the success of
other places, such as British Columbia or Portugal, in moving towards decriminalization
without catastrophe striking. The United States is simply too different. Rather, my view
is based on the observation that the strongly moralistic view towards drug consumption
prevents our society from recognizing that the regime of criminalization produces vast
costs. Harm would be substantially reduced in a decriminalized system. Possession would
not be a crime and the cost of drugs would be sufficiently low so that addicts would not
have to commit crimes other than possession to support their habit. If they committed
crimes while intoxicated, the usual rules would apply with no unfairness. The vast sums
now spent on law enforcement could be used to support research and treatment. The
money would be far better spent in this way.
Finally, I believe that the substantive law of criminal responsibility is too harsh. In par-
ticular, there is no generic mitigating doctrine that would apply to all defendants who
might have substantial rationality or self-​control problems that do not warrant a complete
excuse. Taking such problems into account is largely limited to sentencing and is thus
discretionary. Assuming that the problem of diachronous responsibility could be finessed
generally or did not obtain in particular cases, many addicts might qualify for such miti-
gation. I have proposed such a doctrine (Morse 2003) and believe that the problem of
diachronous responsibility might not loom so large if defendants were simply seeking
mitigation and not a full excuse.
In short, the current criminal law response to drugs and addiction is defensible, but it
is far from optimum.

10 Conclusion
Current Anglo-​American law concerning addiction is most consistent with the choice
model of addictive behavior and the no-​choice model has made few inroads despite
the enormous advances in the psychological, genetic and neuroscientific understand-
ing of addiction. The law’s conservatism is defensible, even in the face of the chronic
and relapsing brain disease model of addiction, which often unjustifiably assumes that
addicts have essentially no choice about use and other crimes committed to support
use. Nevertheless, sound legal policy should move away from a primarily criminal law
response and should move towards a more liberal regime based on rights and good
overall consequences.

References
American Bar Association (1989). ABA Criminal Justice Mental Health Standards.Washington,
DC: American Bar Association.
American Law Institute (1962). Model Penal Code. Philadelphia: American Law Institute.
American Psychiatric Association (2013). DSM-​5: Diagnostic and Statistical Manual of Mental
Disorders (5th ed.). Arlington, VA: American Psychiatric Publishing.
444  

444 Addiction, choice, and criminal law

American Psychiatric Association Insanity Defense Work Group (1983). American Psychiatric
Association statement on the insanity defense. American Journal of Psychiatry, 140, 681–​88.
Ashworth, A. and Horder, J. (2013). Principles of Criminal Law (7th ed.). New York: Oxford
University Press.
Becker, G.S. (1996). A theory of rational addiction. In: G. Becker (ed.), Accounting for Tastes.
Cambridge, MA: Harvard University Press, pp. 50–​76.
Burroughs, A. (2013) This is How: Surviving What You Think You Can’t. New York: Picador.
D.P.P. v Majewski [1977] AC 443.
Everitt, B.J. and Robbins, T.W. (2005). Neural systems of reinforcement for drug addiction: from
actions to habits to compulsion. Nature Neuroscience, 8, 1481–​89.
Fingarette, H. and Hasse, A.F. (1979). Mental Disabilities and Criminal Responsibility.
Berkeley: University of California Press.
Heyman, G. (2009). Addiction: A Disorder of Choice. Cambridge, MA: Harvard University Press.
Heyman, G. (2013) Quitting drugs: quantitative and qualitative features. Annual Review of Clinical
Psychology, 9, 29–​59.
Holton, R. (2009). Willing, Wanting, Waiting. New York: Oxford University Press.
Husak, D. N. (1992). Drugs and Rights. Cambridge, UK: Cambridge University Press.
Husak, D. N. (2002). Legalize This! The Case for Decriminalizing Drugs. London: Verso.
Hyman, S. (2007). The neurobiology of addiction: implications for the control of voluntary behavior.
American Journal of Bioethics, 7, 8–​11.
Kalant, H. (2010). What neurobiology cannot tell us about addiction. Addiction, 105, 780–​89.
Kasanetz, F., Deroche-​Gamonet, V., Berson, N., et al. (2010). Transition to addiction is associated with
a persistent impairment in synaptic plasticity. Science, 32, 1709–​12.
Kennett, J. (2001). Agency and Responsibility: A Common-​Sense Moral Psychology. New York: Oxford
University Press.
Lockett v. Ohio, 438 U.S.586 (1978).
MacCoun, R.J. and Reuter, P. (2011). Assessing drug prohibition and its alternatives: a guide for
agnostics. Annual Review of Law and Social Science, 7, 61–​78.
Montana v. Egelhoff, 518 U.S. 37 (1996).
Morse, S. J. (1994). Culpability and control. University of Pennsylvania Law Review, 142, 1587–​660.
Morse, S.J. (2000). Hooked on hype: addiction and responsibility. Law and Philosophy, 19, 3–​49.
Morse, S. J. (2003). Diminished rationality, diminished responsibility. Ohio State Journal of Criminal
Law, 1, 289–​308.
Morse, S. J. (2007a). Voluntary control of behavior and responsibility. American Journal of Bioethics,
7, 12–​13.
Morse, S. J. (2007b). The non-​problem of free will in forensic psychiatry and psychology. Behavioral
Sciences & the Law, 25, 203–​20.
Morse, S.J. (2009). Addiction, science and criminal responsibility. In: N. Farahany (ed.). The Impact of
the Behavioral Sciences on Criminal Law. Oxford: Oxford University Press, pp. 241–​89.
Morse, S. J. (2011). Addiction and criminal responsibility. In: J. Poland and G. Graham (eds), Addiction
and Responsibility. Cambridge, MA: MIT Press, 159–​99.
People v. Decina, 2 NY 2d 133 (1956).
Pereboom, D. (2001). Living Without Free Will. Cambridge, UK: Cambridge University Press.
Powell v. Texas, 392 U.S. 514 (1968).
R v Bunch [2013] EWCA Crim 2498.
  445

Conclusion 445

Ravenscroft, I. (2010). Folk Psychology as a theory. [Online] Available at: http://​plato.stanford.edu/​

entries/​folkpsych-​theory/​. Accessed 05/​05/​2016.
Robinson v. California, 370 U.S. 660 (1962).
Sifferd, K. (2006). In defense of the use of commonsense psychology in the criminal law. Law and
Philosophy, 25, 571–​612.
Strawson, G. (1989). Consciousness, free will and the unimportance of determinism. 1nquiry, 32, 3–​27.
United States v. Moore, 486 F. 2d 1139 (D.C. Cir. 1973).
Vihvelin, K. (2013). Causes, Laws and Free Will: Why Determinism Doesn’t Matter. New York: Oxford
University Press.
Wallace, R. J. (1994). Responsibility and the Moral Sentiments. Cambridge, MA: Harvard
University Press.
Williams, G. (1961). Criminal Law: The General Part (2nd ed.). London: Stevens & Sons.
Zinberg, N. (1984). Drug Set and Setting: The Basis for Controlled Intoxicant Use. New Haven, CT: Yale
University Press.
446  
  447

Section VII

Conclusions
448  
  449

Chapter 24

Ambiguous terms and false

dichotomies
Gabriel Segal

Abstract
Different people mean different things by the key terms that feature in
debates about addiction. The terms in question include “addiction,”
“compulsion,” “disease,” “brain,” and “hijack.” Terminological differences
have been mistaken for disagreements over matters of substance. This
chapter analyses and discusses some of the different interpretations of
the key terms, and the logical structure of debates in which they feature.
Ultimately, however, the basic affliction of most actual addicts is a neurally
based malfunctioning of choice-making systems that has the potential to
cause great distress, and this can necessitate the same contingencies in
public spending and health care irrespective of whether this falls into a
generally agreed definition of disease.

1 Introduction
After Heather and I decided to bring a collection such as this together, we spent some
time deciding what to call it. We considered Addiction, Disease and Choice. We rejected
that idea and removed the word “disease” because we were agreed that addiction is a mat-
ter of impaired choice making, and wanted to advance the proposal that questions about
addiction were best approached with that lead:  what is the nature of the impairment?
How is it acquired? What are the consequences of the answers to those questions for ques-
tions about management, treatment, law and social and philosophical matters? The ques-
tion of whether addiction is a disease is best approached in the same way, as secondary to
the question of the nature of the impairment: given that the nature of the impairment is
thus and so, is addiction a disease?
When I began to research the question “Is addiction a disease?” I was struck by how
difficult it was to get a clear understanding of what the issue is supposed to be. While
the World Health Organization and most governmental and medical bodies officially
held the view that addiction is a disease, many physicians and psychiatrists expressed
doubts about the classification. After all, they don’t think they have pills that cure it.
Academic psychologists and philosophers argued back and forth. Some journalists,
450  

450 Ambiguous terms and false dichotomies

writers, therapists, addicts, their friends and families, and even people with no par-
ticular connection with addicts, seemed to have very strong views about it. Yet it
was entirely unclear what all the fuss was about and what the question even meant.
In spite of this lack of clarity, there was a tendency for people first to take a stand on
the disease question, and then to draw from that conclusions about the nature of the
impairment, its management and treatment, and matters social and philosophical.
I think this has been a mistake: at the end of the day “disease” is a folk term, not one of
science or even of medicine. It is just a label. And nothing of substance should hang on
whether the label correctly applies to addiction. Two individuals could agree on every-
thing else about addiction, but still disagree on whether to call it a “disease,” just because
they mean different things by “disease.” I believe that much of what have appeared to be
substantive debates about the nature of addiction, appearing under the guise of debates
about whether it is a disease, boil down to terminological differences. People mean differ-
ent things by “addiction,” “disease,” “brain,” and “compulsion.” And terminological differ-
ences have obscured substantive agreements (or at least absences of clear disagreement).
I will elaborate on this theme throughout this chapter.
Historically, discussions about how best to treat addicts and addiction have tended to be
framed in terms of a binary distinction: addiction is a disease versus addiction is simply bad (or
immoral or weak-​willed) choice making. For example, in 1886 Joseph Parrish MD, President
of the American Association for the Cure of Inebriates, wrote (Parrish 1886, pp. 21–​22):
Drunkenness … has proved itself to be beyond the reach of human laws. Law cannot penetrate
the occult forces of humanity and search out the intricate by-​ways of the morbid element which is
its source … Resolutions and edicts, pledges and covenants … intended to strengthen the moral
sense and restore the moral nature, must succumb at times to that which is inherent, vital and
overwhelming. When the crave comes, it demands satisfaction, like hunger.

The reference to a “morbid” element at the source of drunkenness is motivated by ref-

erences earlier in the pamphlet to an extensive collection of testimonials from doctors,
health managers, and others to the effect that inebriety is a disease. Parrish was con-
cerned to promote support for and public expenditure on homes for inebriates that would
include scientific and medical treatment and “aetiological and pathological researches”
(p. 23) for the benefit of the community.
Parrish and the learned people he quotes make two further claims about the nature of
the “disease” of inebriety that are now familiar from neuroscience and psychiatry: first,
that the process of becoming an inebriate (alcoholic) involves a transition from occa-
sional use to one in which “self-​control is suspended or annihilated … the victims of the
condition obey only an overwhelming craving for stimulants” (p. 5) and, second, that the
condition has a physiological root.
Parrish presents a package:  disease, lack of choice, physical basis, and public health
spending. There are four distinct concepts (or topics) here, each of which involves consid-
erable complexity. They and the logical relations among them require careful analysis and
consideration. My aim in what follows is to provide some, beginning with the big issue of
“disease” versus “choice.” I myself think that Parrish and the “disease” campers are more
right than wrong, and I shall not strive to be neutral in my presentation.
  451

“Disease” or “choice” 451

2  “Disease” or “choice”
Heyman and Mims (Chapter 21, this volume) say “[w]‌hen addiction specialists say that
addiction is a disease, they mean that drug use has become involuntary.” That seems to be
consistent with the quotes from Parrish above. And in a recent article that echoes many
of the themes from Parrish (1886), Alan Leshner, the then director of the (American)
National Institute on Drug Abuse (NIDA) wrote similarly (Leshner 2001, p. 3):
Over time the addict loses substantial control over his or her initially voluntary behavior, and it
becomes compulsive. For many people these behaviors are truly uncontrollable, just like the behav-
ioral expression of any other brain disease. Schizophrenics cannot control their hallucinations and
delusions. Parkinson’s patients cannot control their trembling.

So at least some addiction specialists talk as if addiction is a progressive disease and that
once it has reached a severe stage, an addict’s using ceases to be subject to any voluntary
control at all, and is as uncontrollable as the shaking of a Parkinson’s patient.
Heyman and Mims present a great deal of evidence that most people meeting standard
diagnostic criteria for addiction quit or moderate their use, that many among those do
so without treatment and that their doing so results from normal, understandable condi-
tions, such as financial, social or familial incentives. These data contradict that type of
“disease” view as applied to all who meet standard diagnostic criteria for addiction.
The remission data are, though, consistent with addictive use being completely involun-
tary, like Parkinson’s shaking, for some addicts, under some conditions. Parish, Alcoholics
Anonymous, Leshner and many others in the “disease” camp talk of cravings. Consistently
with all the data about high remission rates, it remains possible that when an addict expe-
riences a craving, perhaps set off by a priming dose or a high level of stress, their behavior
becomes entirely involuntary, like Parkinson’s shaking. However, as Heyman and Mims
(Chapter 21, this volume) say, the behavior of even severe addicts after a priming dose is
still subject to some voluntary control, since they will desist, at least for a time, given a
large enough incentive.
But many in the disease camp do not think of addictive behavior as quite so involuntary.
For example, in 1992 the American Medical Association offered a definition of alcohol-
ism as a disease “characterized by impaired control over drinking, preoccupation with the
drug alcohol, use of alcohol despite adverse consequences, and distortions in thinking,
most notably denial.” And more recently, the American Society of Addiction Medicine
defined addiction as a disease (ASAM 2011) “characterized by inability to consistently
abstain, impairment in behavioral control, craving, diminished recognition of significant
problems with one’s behaviors and interpersonal relationships, and a dysfunctional emo-
tional response.”1 Nora Volkow, Leshner’s successor as director of NIDA, writes (Volkow
2014): “The initial decision to take drugs is typically voluntary. However, with continued
use, a person’s ability to exert self-​control can become seriously impaired; this impair-
ment in self-​control is the hallmark of addiction.” And Holton and Berridge (Chapter 9,

1 That is their short form definition. They also offer a long form that makes interesting reading, beginning
“Addiction is a primary, chronic disease of brain reward, motivation, memory and related circuitry.”
452  

452 Ambiguous terms and false dichotomies

this volume) endorse a disease model that involves pathological desires: “consumption

of addictive drugs gives rise to pathologically intense desires or cravings, states that are
largely insulated from the subject’s beliefs and other desires” (p. 157).
“Addiction” is very often defined as “compulsive drug-​seeking and use.” But different
people mean very different things by the term “compulsive.” It can be characterized with
expressions such as “like the shaking of a Parkinson’s patient,” “annihilated” self-​control, a
(100 percent) “irresistible urge,” “beyond motivation,” done “regardless of the costs,” and
“can’t say no” (Heyman and Mims, Chapter 21, this volume, p. 387), an “urge to use” that
“is more than an urge … a command that cannot be resisted” Lewis (Chapter 10, this vol-
ume, p. 000). But “compulsive” does not always mean anything so extreme. Even people
with obsessive-​compulsive disorder can say “no” and resist their “compulsions” to some
extent (Abramowitz 2006).
In fact, no contemporary theoretical model of addiction offered by neuroscientists in
the “disease” camp treats addictive behavior as completely beyond voluntary control, like
Parkinson’s shaking (Hyman 2007). What, then, do theorists in the camp mean by “com-
pulsive?” Steven Hyman explains (Hyman 2007, p. 2):
The term compulsion is imprecise, but at a minimum implies diminished ability to control drug
use, even in the face of factors (e.g. illness, failure in life roles, loss of job, arrest) that should moti-
vate cessation of drug use in a rational agent willing and able to exert control over behavior.

And Barry Everitt says “it is indeed compulsive drug use that is widely seen as a core
aspect of addiction, and by that is meant repeated, persistent use, despite placing an indi-
vidual in danger, compromising their health, family and social lives” (Everitt 2014, p. 6).

3  “Compulsion” or “choice”
Consistently with Hyman and Everitt, Narcotics Anonymous (1993, pp. 6–​7) tells us that
“In our experience, compulsion is the irrational impulse to continue using drugs, no mat-
ter what happens as a result.” A  very familiar scenario is one in which the addict has
sworn off using, has a very strong conscious motive for abstaining and very much wants to
abstain, yet their action is in fact motivated by the more primitive impulse to use. While it
is true that most addicts—​or at least most people meeting standard diagnostic criteria for
substance dependence (e.g. under DSM-​IV-​TR’s definition)—​quit or control their using
(mostly without treatment), and do so for understandable reasons, there also exists the
subset of addicts who try repeatedly to control their use or quit, and repeatedly fail. These
more severely afflicted chronic addicts (meeting 5–​7 of the DSM-​IV-​TR criteria for sub-
stance dependence (American Psychiatric Association 2000)), the type who make up the
majority of NA (Willenbring et al. 2009), continue to use in spite of clear health, social,
financial and familial incentives to quit. Their use is not nearly as responsive to normal
incentives as that of the less ​severely afflicted individuals who meet the three DSM-​IV-​TR
criteria required to qualify as an addict under DSM-​IV-​TR’s particular definition.
In these more severe cases, the addicts’ using is “compulsive” in the sense of going
against their own wills in the three different ways discussed in Segal (Chapter 21, this
  453

“Compulsion” or “choice” 453

volume): acting on a desire that they would rather not have or act on; acting in a way
that they recognize is not in their own best interests; and breaking a previous resolve for
no good reason. (The first two of these amount to acting on a “compulsion” in the sense
articulated by Berridge and Robinson (2011).)
Further, as Henden (Chapter  7, this volume) argues, in active addiction the addict’s
will is affected by attentional bias towards cues. Attentional bias can exacerbate in-​the-​
moment delay discounting (whether or not the specific individual is already innately pre-
disposed to high delay discounting) and it can directly bias addicts’ beliefs, desires, and
reasoning toward addictive behaviors. It also disrupts the functioning of the will by con-
stantly drawing attention away from higher goals (reasons for abstention), thereby deplet-
ing its limited resources (Vonasch et al., Chapter 16, this volume). For the experienced
severe addict, external and internal cues are ubiquitous: for example, places, people, times
of day (evening), withdrawal, anxiety, fear, depression, boredom, relief, and joy. For these
individuals, the disruption to proper functioning of the will is constant and therefore
grave. The will’s resources get yet further depleted by the persistent and powerful urges
that the distracting cues generate.
Moreover, addictive behavior is repetitive and often “automatic” in the sense of being
driven by largely unconscious processes (Berridge 2009)  that are relatively stimulus-​
bound, insensitive to rational considerations and that lead quickly, directly and
thoughtlessly to action unless something stops them (Naqvi and Bechara, Chapter 12,
this volume; Holton and Berridge, Chapter 9, this volume). Addictive behavior is cer-
tainly compulsive in the sense of being driven by the forces of internal mechanisms over
which the addict’s rational thought processes have little control and which is often, in
their actual circumstances, itself beyond the control of their wills. No matter how hard
they try to control or refrain, they often fail. The urge to use often overpowers their very
best efforts of will.2
Added to all of this are failures of self-​awareness (Moeller and Goldstein, Chapter 11,
this volume) and denial (Pickard and Ahmed, Chapter 2, this volume; Segal, Chapter 20,
this volume). Addicts frequently underestimate the amounts they themselves have been
using. Many also continually fail to learn from their previous experiences:  believing
promises made to themselves (This time I will stop after just a few, Next Sunday I will defi-
nitely take a break), though they have repeatedly broken such promises before.
And, finally, there are the devastating “mental blank spots” (Crowther, Chapter 5, this vol-
ume; Segal, Chapter 20, this volume), during which the addict’s previous resolve to abstain
and the good reasons he had for it fail to surface in consciousness or deliberation or to act as

2 Ainslie (Chapter 13, this volume) says that it is best to reserve the term “compulsive” for behavior that
is strictly confined by personal rules, over-​controlled rather than under-​controlled. But the key idea of
compulsion, being driven by forces that make the behavior very difficult to control (the “pel” in “com-
pel” comes from the Latin “pellere,” meaning to drive), applies equally in both cases.
454  

454 Ambiguous terms and false dichotomies

any kind of defense against the desire to use. In these last cases, the will is not just overpow-
ered or malfunctioning, it is pushed off-​line altogether and fails to function at all.3
As Pickard and Ahmed (Chapter 2, this volume) say, denial has received surprisingly
little attention in the philosophical and scientific literature. As they point out, it involves a
cognitive impairment. A better understanding of this impairment needs to be developed
to improve prevention and help management. It is also crucial for all those who want a
good understanding of the addict’s predicament—​family, employers, counsellors, and so
on. It also matters to the law.

3.1  Responsible or not?

Morse (Chapter  23, this volume) argues that existing Anglo-​American criminal law is
most consistent with the choice position and that this is a defensible approach, consistent
with current science and with traditional justifications of criminal blame and punish-
ment. Given our current limited understanding of the nature of the impairments to the
functioning of the will, and the great difficulty in assigning responsibility (or lack of it)
in particular cases, this seems reasonable. But discussions need to be, and are, ongoing.
For example, under current England and Wales law, it is possible to enter a plea of
“diminished responsibility” for murder (which would reduce the charge from murder
to voluntary manslaughter), if one killed someone while under the influence of drugs or
alcohol, and the consumption was due to an “abnormality of mind,” such as might arise
from a recognized medical condition (e.g. an addictive disorder). Specifically:
(a) A person who kills or is a party to the killing of another is not to be convicted of mur-
der if they were suffering from an abnormality of mental functioning which:
• arose from a recognized medical condition
• substantially impaired their ability to do one or more of the following: understand
the nature of their conduct, form a rational judgement or exercise self control
• provides an explanation for [their] acts and omissions in doing or being a party to
the killing.
(b) An abnormality of mental functioning provides an explanation for the conduct if it causes,
or is a significant contributory factor in causing, the defendant to carry out that conduct.
(Coroners and Justice Act 2009)

3 Denial and the mental blank spots are not too hard to explain from Ainslie’s (2001) intertemporal
bargaining perspective. On that view, beliefs are not formed for accuracy per se, but for usefulness as
tools in obtaining reward. He writes: “Given temporary preference for present comfort, it isn’t hard to
picture a mechanism for repression. Many ways have been described whereby selective attention can
systematically distort the information you collect. Experiments have shown that we tend to label our
memories with their emotional meanings, and retrieve them by their labels. What comes to mind first
when I see someone walking toward me isn’t her name or where I saw her last, but a sense of whether I’d
like to see more of her or avoid her. The same is true for a book on the shelf, or a place I have a chance
to visit. If that first sense spells trouble, it’s easy enough to steer in another direction without ever going
into why I want to, or whether I’ve an obligation not to” (Ainslie 2001, p. 77). Memories of reasons for
abstention are not accessed when the emotional pressure to use is great enough.
  455

“Disease” or “disorder” 455

Previous interpretations of diminished responsibility (in relation to alcohol dependence,

in particular) were framed in terms of “disease,” “brain injury,” “involuntary” drink-
ing, and “irresistible” impulse (Galappathie and Jethwa 2010). (Galappathie and Jethwa
remark “The dichotomy of viewing control over drinking as present or absent [was]
myopic.” Cf. Heyman and Mims (Chapter 21, this volume): “voluntariness is a continuous
dimension.”) The 2009 act is free of such terms, and remains to be interpreted.
Suppose that a severe alcoholic realizes he has a problem and succeeds in staying absti-
nent for some months. Then he suffers a mental blank spot and has a drink believing it
will be just the one. Cravings result, the alcoholic ends up on a major binge, and then
commits a crime that he would have not have committed sober. An abnormality of mind
would be a significant contributory factor in causing the taking of the first drink, and then
the subsequent binge. While these actions were not entirely involuntary, the idea that the
alcoholic’s responsibility for them is diminished makes good sense. Just as actions may be
more or less voluntary, so also agents may be more or less responsible for what they do.

4  “Disease” or “disorder”
While it does not matter for legal purposes whether addiction is called a “disease,” it does
seem to matter that it be a recognized medical condition, a psychiatric disorder. Heyman
and Mims (Chapter 21, this volume) question whether addiction should be recognized
as such on the grounds (1) that its frequency varies considerably as a function of social
factors and (2) that it is closely linked with interventions that stress faith and other non-​
medical elements (e.g. Alcoholics Anonymous). It is not hard to explain why the inci-
dence of addiction might be highly culturally variable. For example, it is not unlikely that
it is much easier to get addicted if you are a teenager who binges regularly than if you are
not. That would explain why the incidence of addiction is much higher in cultures where
teenage bingeing is common than in those where it isn’t. That doesn’t show that addic-
tion is not a psychiatric disorder. The reason for (2) is explained by Segal (Chapter 20,
this volume). Twelve-​step programs are actually premised on the disease-​theoretic idea
that moderate use is, in practical terms, impossible for addicts, due to the chronic “physi-
cal allergy” (neurological abnormality). They are abstinence programs and they work by
managing emotions, so that addicts don’t stress out and relapse.
Addicts differ from non-​addicts in the way they react to cues, such as feelings of stress.
The cues cause a reaction in the mind-​brain—​an abnormally powerful motivation to
behave in a characteristic way (take a drug, place a bet). Whether the actual behavior
results depends on the circumstances. Whether it is correct to call addiction a “disease,” a
“behavioral” or “psychiatric” “disorder,” or something else depends in part on what sort of
“abnormality” is involved. A major part of what is involved in the idea of a disease is the
idea of an organ or system not functioning correctly.
For example, an autoimmune disease occurs when the immune system mistakenly
attacks and destroys healthy body tissue. In subjects with an autoimmune disease,
the immune system can’t tell the difference between healthy body tissue and antigens.
The response is a counterproductive hypersensitivity reaction similar to the response
in allergies. Analogously, an addiction occurs when a motive-​forming system in the
456  

456 Ambiguous terms and false dichotomies

mind-​brain—​almost certainly the mesocorticolimbic system—​mistakenly places a very

high value on harmful substances or behaviors. In addicts, the mesocorticolimbic sys-
tem can’t tell the difference between healthy and unhealthy substances or behaviors. It
mistakes the harmful for the beneficial, just as the diseased immune system mistakes the
benign for the harmful. So, I would say, addiction is a disease like an autoimmune disease.
Why is it a “mistake” to place a high value on drugs? Heyman and Mims (Chapter 21,
this volume) say:  “Darwin and Wallace’s compelling accounts of evolution by natural
selection imply that voluntary behavior, like involuntary behavior, has a genetic basis and
is mediated by the brain.” The choice-​making systems in the mind-​brain have evolved
to function in ways that promote fitness. Obviously, taking drugs does not promote fit-
ness. But that does not by itself show that systems are malfunctioning when they produce
addictive urges, since evolution does not design perfect systems. Perhaps even in addicts,
these systems are doing what they are designed to do. Are they?
Marc Lewis (Chapter 10, this volume) offers a helpful evolutionary perspective for this
question. Like many of the contributors to this volume, he focuses on delay discounting
and the tendency of addicts to place very high values on immediate rewards, discounting
superior rewards in the future. Lewis writes: “It seems likely that there have been distinct
evolutionary advantages to the motivational highlighting of immediate gains. Thus, crav-
ing, and its amplification in now appeal, may be the by-​product of a brain designed to be
maximally responsive to immediate rewards.” He continues:
Not surprisingly, dopamine activation of the striatum and its neighbours is to blame for now
appeal—​an unfortunate bias in the firmware that’s come to us over evolution. Like the slouching that
results in backaches—​a painful spin-​off of the achievement of upright locomotion—​now appeal is
an evolutionary side effect. Yet how could it be otherwise? The striatum evolved to get the animal to
go for low-​hanging fruit, available sexual partners, whatever is most accessible—​a habit it has kept to
this day. Dopamine rises with anticipation, rushing in to rev up the striatum, when rewards are just
around the corner. The release of dopamine by immediately available goodies distorts the perspec-
tive, the considered insight, we could have (otherwise) achieved using our more advanced cognitive
abilities (generally housed in the prefrontal cortex). We lunge for the immediate … Addicts are
excessively now-​oriented, more prone to delay discounting than the population average.

Lewis himself does not want to call addiction a disease.4 His reasons for this (I think)
boil down to the claim that the processes of addiction and recovery are those of normal
reward learning, development, and growth, and do not involve departure from and return
to a “normal” state.
I am not convinced. Consider the following analogy: suppose a group of programmers
design a generation of computers to learn to play, and to play chess. They build slightly
different biases into different learning programs. All the systems have a subsystem, BS,
that controls the bishops. It turns out that some of the systems—​call them “A-​systems”—​
are such that when they encounter certain types of sequences of game-​situations early

4 Lewis (Chapter 10, this volume) calls it an “ailment.” In personal communication, he did not object to
calling it a “psychiatric disorder.”
  457

“Disease” or “disorder” 457

in their learning histories, the BS gradually comes to assign higher and higher values to
taking opponents’ pieces. Over many thousands of games, the BS puts the whole program
out of kilter, taking up more and more computational resources to increase its efficacy in
taking pieces, and thereby reducing the efficacy of other subsystems. The result is that the
A-​systems end up playing dreadful chess, taking lots of opponents’ pieces with bishops,
irrespective of costs, and losing as a result.
We can imagine that for some A-​systems, the program will rectify itself given relatively usual
subsequent types of sequences of game situations. For others, programmers need to inter-
vene and adjust the programming (perhaps by subjecting the machine to controlled learning
regimes). With the most seriously afflicted machines, the BS remains permanently affected,
even after extensive reprogramming. These machines can adapt so that they function well in
many game situations. But not all. Certain sequences of game-​situations will always trigger
their old patterns and cause them to revert to playing dreadful chess, exactly as before.
The A-​systems are analogous to addicts, some of whom recover spontaneously, some
of whom can do so with help, and some of whom need to go abstinent if they are to
avoid relapse and have a decent chance of surviving. We can suppose that there is nothing
wrong with the BSs in their initial state, except that they are particularly prone to posi-
tive reinforcement from certain results of their moves. Nevertheless, the outcome of this
process, given certain types of learning environment, is a crippling bug.
Of course, one might adopt a liberal (anti-​psychiatric) view of chess-​playing machines
and say it is up to them set their own goals and ideals—​who are we to say that A-​systems,
or their BSs, are doing anything “wrong?” If they prefer to take opponents’ pieces with
bishops than to win games of chess, then that is their prerogative. But if we suppose that
the systems are designed to learn how to play, and then to play, good chess, and that BSs
are there to contribute to that goal, then these systems are not doing what they should.
Evolution designed our choice-​making systems to make choices in ways that promote
fitness and to do this by assigning motivational values that are roughly commensurate with
the actual values of motivated behaviors. The problem with addictive motivations is not
just that they are excessively now-​oriented. At least in the case of the severe addict, there
are at least two further dysfunctional aspects of the motivations to use. The first comes
from the disparity between “wanting” and “liking” (noted by Lewis himself): addicts con-
tinue to use even when they no longer enjoy the effects of the drugs. The motivational
system grossly miscalculates the value of the rewards, even in the moment. For the very
severely addicted, often the only genuine “reward” of a craving-​caused ingestion is very
temporary and only partial relief from the craving itself. That craving would not even have
existed had the brain not been addicted, and no other hedonic value accrues to the inges-
tion. Scratching a painful itch brings satisfaction in the moment. But the moment would
be more enjoyable if there were no painful itch at all.
The further dysfunctional aspect of addictive motivations is that using tends to increase
the motivation to use, rather than to decrease it. The scratched “itch” disappears or lessens
for a short while, but soon it returns with increased intensity. One of the key features of a
functional motivation (such as a desire or urge) is that acting upon it satisfies and relieves
458  

458 Ambiguous terms and false dichotomies

it, so it disappears. If you want something and you get it, it is not good if you go on want-
ing it! Priming effects have their evolutionary point: if there is good food available, then
it is important to eat well and stock up on energy. So it is a good idea for a taste to give a
boost to the appetite. But in addicts the priming effect is excessive and they tend to go on
using until there is none left or they pass out.
Lewis points out that the same sort of high-​dopamine processes that underlie addiction
also underlie important aspects of sexual attraction, love and mating. The implication is
that if addiction is a disease, then so is love (Lewis 2011). There are two points to be made
in response to that. Both can be made graphic by parallels with the case of the man who
mistook his wife for a hat, famously recounted by Oliver Sacks (Sacks 1985). The man,
who suffered from prosopagnosia, had great difficulties recognizing people. At one point
he confused the head of his wife with his hat, grabbed it, and tried to put it on his head.
Treating the head of one’s wife’s as a hat is a sign of mental dysfunction. And so would be
treating one’s wife the way an addict treats his drug, with extreme and obsessive neediness.
If it is indeed true that a man’s pursuit of sex, love, and mating could become dopamine-​
driven in the manner of a drug addict’s pursuit of their substance, then it would become
highly dysfunctional. For example, he might well: become immobilized or seriously dis-
tracted by romantic or sexual obsessions or fantasies; avoid responsibility for himself by
attaching himself to people who are emotionally unavailable; stay enslaved to emotional
dependency, romantic intrigue, or compulsive sexual activities; assign magical qualities
to others; confuse love with neediness; feel empty and incomplete when alone. The list
goes on and the rest of it can be found on the Sex and Love Addicts Anonymous website
in SLAA (1990). I  am not myself convinced that the dopaminergic processes underly-
ing the difficulties suffered by members of SLAA are quite the same as those underlying
substance addiction. Evidence of genuine dopamine sensitization in relation to romantic
stimuli is yet to be found. And nothing like a sex or love addiction is listed as a disorder
in DSM-​5 (American Psychiatric Association 2013). My point is that if they are the same,
then sex and love addiction is just as dysfunctional as drug addiction.
The second point is that it is also dysfunctional to treat a dangerous substance like
heroin or alcohol as one would treat a particularly beloved wife, as something to be prized
above all else. But that is what addicts do. Indeed, in many cases, they end up more dedi-
cated to their drug than they are to their loved ones, and it is the latter that lose out. Just
as addiction could be called “faux starvation,” so also it could be called “faux love.”
It is true that there is an arbitrary and subjective component involved in deciding
what counts as “excessively” powerful motivations to use or to do something. I have just
focused on the extreme cases. For much of the time many addicts use happily: they enjoy
the effects of using and they themselves might argue, perhaps even successfully, that the
immediate rewards are worthwhile. However, the fact that there is an arbitrary, subjec-
tive element to the judgment does not by itself mean that addiction is not a disease (cf.
Boorse 1977). Diseases can be like that. Diabetes, for example, is a disease in which the
body cannot regulate the amount of sugar in the blood. It can be caused by too little
insulin, resistance to insulin, or both. What counts as too high a level of blood sugar
for diagnostic purposes is arbitrary. At what point does the addict’s motivation to use
  459

“Psychiatric disorder” or “brain disease” 459

become pathological, rather than just great? Roughly: when it is so powerful and persis-
tent that, were the individual repeatedly to try not to act on it, they would repeatedly fail
(cf. Heather, Chapter 1, this volume). The fact there is an arbitrary and subjective element
to where we choose the line does not mean that it is a bad choice to make.
To conclude this section: addiction is like an autoimmune disease. It involves a (statisti-
cally) abnormal reaction of an organ (a neural system) to certain stimuli that is dysfunc-
tional from an evolutionary point of view.

5  “Psychiatric disorder” or “brain disease”

Parrish and the Anonymous fellowships say that there is a physiological element to alco-
holism. The claim the addiction is a brain disease is familiar in neuroscience and psy-
chiatry. If what I have been arguing is right, is it then also right to call addiction a “brain
disease?” There are two aspects to this question, one purely semantic, the other largely
empirical.
The semantic aspect can be clarified by the analogy of the chess-​playing computer. The
problem with the A-​systems is a software bug—​the problem is with the program. There
is nothing wrong with the machines’ hardware: every silicon chip, every flip-​flop, is func-
tioning perfectly well. Analogously, if what I have been arguing is right, addiction is a
disorder at the psychological level: one component of the choice-​making systems in the
mind assigns too high a value to certain activities. This leaves open the possibility that
there is nothing wrong with the addict’s brain at the physiological level, the level of indi-
vidual neurons.
If that possibility is actual, then one could with equal justification call addiction a “brain
disease” or not. It would be a brain disease in the sense that choice-​making systems in the
addict’s brain are malfunctioning. The brain (as opposed to, for example, the liver) is the
physiological realization of the mind, and as such it is the organ that is not functioning
right. (Cognitive scientists use the expression I used above and talk of the “mind-​brain”.)
On the other hand, one might prefer to reserve the term “brain disease” for a physiologi-
cal malfunction, apparent at the level of cells.5
It is worth noting that even if addiction is not a brain disease in the stricter sense,
it nevertheless seems to have crude physiological/​chemical components. Parkinson’s
patients are sometimes treated with dopamine-​receptor stimulating medications (D2/​D3
agonists). As a result, these individuals often develop compulsions to shop, gamble, seek
sex, binge eat, take drugs, and even sometimes to take their medication in an addict-​like
fashion (O’Sullivan et al. 2011).6

5 Indeed, some people reserve the term “disease” for disorders that include a physiological dysfunction,
so a purely psychological dysfunction (a software bug) would automatically not count as a “disease,”
and addiction would be a “disease” only if it were a “brain disease” in the stricter sense. The way I use
the terms, psychiatric disorders are a subset of diseases.
6 There is also considerable anecdotal and some scientific evidence that Baclofen, a GABA receptor
agonist, is an effective treatment for alcoholism and perhaps other addictions as well (Segal 2013;
Addolorato 2006).
460  

460 Ambiguous terms and false dichotomies

The second and largely empirical aspect of the question whether addiction should be
called a “brain disease” is, of course, the question “What is the actual neural basis of
addiction?” Or, indeed, since there is no guarantee that the neural basis of addiction is
the same in all individuals, “What are its actual neural bases?” It may be the case that
addiction always involves some kind of hardware problem. If so, then low D2 receptivity,
which might well be considered a dysfunction at the level of neurons—​and which might
be innate, or acquired, or both—​is probably the likeliest candidate (Burgess, Chapter 17,
this volume; Foddy, Chapter 3, this volume; Volkow et al. 2004).
It is not, though, established that low D2 receptivity has any causal role in core addictive
processes. It used to be thought that dopamine was a neural basis of pleasure, and that
low D2 receptivity would lead to a failure of normal stimuli to cause happiness. People
with low D2 receptivity would naturally be drawn to the dopamine-​stimulating effects of
drugs. However, there is very good evidence that dopaminergic activity associates with
“wanting” rather than with pleasure (Berridge 2007). So that theory is flawed. Certainly,
there is evidence that low D2 associates both with addiction and predisposition to it.
However, it is unclear what the causal relations are. Low D2 could be a causal factor in
addiction, or an effect of it, or a joint effect of a common cause. There are a variety of pos-
sible mechanisms that might underlie these relations. For example: low D2 can be a cause
of high dopamine release (when D2 activity on presynaptic dopamine neurons serves as
a brake on dopamine production) or a consequence of high dopamine release (when D2
receptors on postsynaptic neurons become downregulated as a result of continual high
dopamine stimulation). In the latter case, the high dopamine activity could itself be part
of a predisposition to addiction or the result of drug-​taking.
If what underlies all addiction is incentive sensitization of the mesolimbic system, then
there may well be nothing wrong with the brain at the cellular level. For example, one
sensitization process seems to involve elongation and branching of dendrites of cells in
the nucleus accumbens and prefrontal cortex. Each individual cell is perfectly healthy. But
as far as the design of the whole system is concerned, it is overpowered and does more
than it should.

6  “Hijacking” or “seduction”
The idea that addiction involves a failure of the mind-​brain to function well from an evo-
lutionary point of view is sometimes expressed in terms of “hijacking” (e.g. Hyman 2007).
Drugs “hijack” neural systems that were designed for the pursuit of natural rewards that
promote fitness. This is in line with the idea that addicts suffer feelings of “faux-​starvation”
or “faux-​love”: the dopamine system comes to treat drugs as if they were vital for survival
or reproduction. Drugs hi-​jack natural-​reward systems by giving the mesolimbic dopa-
mine system an artificially high boost, one that is independent of any actual value (includ-
ing pleasure) that they might provide.
However, not all addictive substances boost the dopamine system artificially. Sugar at
least does not. And there is evidence that some behaviors are addictive in the same way
  461

“Hijacking” or “seduction” 461

as drugs. There is evidence, for example, of sensitization in relation to gambling, the only
behavioral addiction to make it into DSM-​5 (American Psychiatric Association 2013;
Joutsa et al. 2012; Leyton and Vezina 2013).
Bennett Foddy (Chapter 3, this volume) argues that even video-​gaming is addictive.
But surely, says Foddy “a pattern of flashing light and sound cannot be enough to ‘hijack’
the brain’s reward system, as Hyman claims that drugs do.” Again, this is just a question of
semantics. “Hijack’ ” can mean “take over (something) and use it for a different purpose.”
Videogaming brings pleasure because its rewards symbolize natural ones: doing well in
the videogame is a proxy for doing well in the game of life, winning money or power
to attract mates and buy food. Fantasy stimulates our natural reward systems easily and
well. Think of sex. Getting a large reward as a quick payoff for an action generates rapid
reward-​learning, whether the reward is a sugary taste, an imagined victory or a flashing
light that has come to be associated with one.
If you prefer to reserve “hijacking” for the specific interventionist behavior of drugs,
then perhaps “seduce” would be a better choice. If videogames are genuinely addictive,
then they are so because they seduce the natural-​reward system in the mind-​brain.
Is addiction a brain disease characterized by compulsive reward-​seeking behaviors? It
all depends on what you mean by “addiction,” “brain disease,” and “compulsive.” If you
want to use the word “addiction” for a lifestyle (Flanagan, Chapter 4, this volume) and
would use the word “addict” for someone who lives the lifestyle of an addict, but who
could, if they wished, easily moderate their use or quit, then the answer is “no.” That’s all
fine. But whatever you call it, the basic affliction of most actual addicts is what it is: a neu-
rally based malfunctioning of choice-​making systems that has the potential to cause great
distress. That should justify time off work, insurance coverage, and public health spending
on research and rehabs in the same way as would a disease.

Acknowledgments
Thanks to George Ainslie, Kent Berridge, Tom Crowther, Nick Heather, Edmund Henden,
Louise King, and Hanna Pickard for helpful comments on an earlier draft.

References
Abramowitz, J. S. (2006) The psychological treatment of obsessive-​compulsive disorder. Canadian
Journal of Psychiatry, 51, 407–​16.
Addolorato, G., Leggio, L., Agabio, R., Colombo, G., and Gasbarrini, G. (2006). Baclofen: a new drug
for the treatment of alcohol dependence. International Journal of Clinical Practice, 60, 1003–​1008.
Ainslie, G. (2001). Breakdown of Will. New York: Cambridge University Press.
American Psychiatric Association (2000). Diagnostic and Statistical Manual of Mental Disorders (4th
ed., text rev.). Washington, DC: American Psychiatric Association.
American Psychiatric Association (2013). Diagnostic and Statistical Manual of Mental Disorders (5th
ed.). Washington, DC: American Psychiatric Association.
ASAM (2011). Definition of addiction. http://​www.asam.org/​for-​the-​public/​definition-​of-​addiction.
Accessed 05/​05/​2016.
462  

462 Ambiguous terms and false dichotomies

Berridge, K.C. (2007). The debate over dopamine’s role in reward: the case for incentive salience.
Psychopharmacology, 191, 391–​431.
Berridge, K.C. (2009). Wanting and liking: observations from the neuroscience and psychology
laboratory. Inquiry (Oslo). 52(4): 378.
Berridge, K., and Robinson, T. (2011). Drug addiction as incentive senstization. In: J. Poland, and G.
Graham (eds), Addiction and Responsibility. Cambridge MA: MIT Press, pp. 21–​54.
Boorse, C. (1977). Health as a theoretical concept. Philosophy of Science, 44, 542–​73.
Everitt B.J. (2014). Neural and psychological mechanisms underlying compulsive drug seeking
habits and drug memories—​indications for novel treatments of addiction. European Journal of
Neuroscience, 1–​20.
Galappathie, N. and Jethwa, K., (2010). Diminished responsibility and alcohol. Advances in Psychiatric
Treatment, 16, 193–​98.
Hyman, S. (2007). The neurobiology of addiction: implications for voluntary control of behavior.
American Journal of Bioethics, 7, 8–​11.
Joutsa, J., Johansson, J., Niemelä, S. et al. (2012). Mesolimbic dopamine release is linked to symptom
severity in pathological gambling. Neuroimage, 60, 1992–​99.
Leshner, A.I. (2001). Addiction is a brain disease. Issues in Science and Technology Online, 17(3). http://​
issues.org/​17-​3/​leshner/​. Accessed 23/​05/​2016.
Lewis, M. (2011). Dopamine and the neural “now”: essay and review of Addiction: A Disorder of Choice.
Perspectives on Psychological Science, 6, 150–​55.
Leyton, M. and Vezina, P. (2013). Striatal ups and downs: their roles in vulnerability to addictions in
humans. Neuroscience and Biobehavioral Reviews, 37(9 Pt A), 1999–​2014.
Narcotics Anonymous (1993). It Works How and Why: The Twelve Steps and Twelve Traditions of
Narcotics Anonymous. Chatsworth, CA: Narcotics Anonymous World Services.
O’Sullivan, S. S, Wu, K., Politis, M. et al. (2011). Cue-​induced striatal dopamine release in Parkinson’s
disease-​associated impulsive-​compulsive behaviors. Brain, 134, 969–​78
Parrish, J., (1886) Inebriety and homes got inebriates in England. Quarterly Journal of Inebriety.
Hertford, CT: Press of The Case, Lockwood & Brainard Company.
Sacks, O. (1985). The Man Who Mistook His Wife for a Hat and Other Clinical Tales. New York: Simon
and Schuster.
Segal, G. (2013). Alcoholism, disease and insanity. Philosophy, Psychology and Psychiatry, 20, 297–​315.
SLAA (1990). Characteristics of Sex and Love Addiction. Boston, MA: Augustine Fellowship, SLAA,
Fellowship-​Wide Services.
Volkow, N.D., Fowler J.S., Wang G.-​J., and Swanson, J.M. (2004). Dopamine in drug abuse and
addiction: results from imaging studies and treatment implications. Molecular Psychiatry, 9, 557–​69.
Volkow, N. (2014) Drugs, brains, and behavior: the science of addiction. Science of Drug Abuse and
Addiction. National Institute on Drug Abuse. http://​www.drugabuse.gov/​publications/​drugs-​brains-​
behavior-​science-​addiction/​preface. Accessed 05/​05/​2016.
Willenbring, M.L., Massey, S.H., and Gardner, M.B. (2009). Helping patients who drink too much: an
evidence-​based guide for primary care physicians. American Family Physician, 80, 44–​50.
  463

Chapter 25

Overview of addiction as a disorder

of choice and future prospects
Nick Heather

Abstract
This concluding chapter offers an overview of some issues raised in the
book. Future developments in theory and research include seeing addiction
as dynamic inconsistency, advancing the explanation of temptation and
self-control, clarifying the nature of “compulsion,” and moving toward an
inclusive dual-systems theory. Implications for treatment and brief intervention
include correcting an imbalance between motivational and self-control
approaches, and challenging the assumption that new pharmacotherapies
and other biological interventions are the only way forward. The main
implication for prevention is the need to reduce socioeconomic deprivation.
For public understanding of addiction, a disease view is not the only way to
combat stigma and punitive attitudes, though the disordered choice model
may face special risks of being misunderstood. The final conclusion is that the
role of volition in addiction must be clearly recognised and its implications for
how we respond to addiction honestly debated.

1 Introduction
In his chapter for this book, and in order to illustrate the range of assumptions underlying
different models of addiction, George Ainslie uses the ancient Indian parable of the blind
men seeking to understand the essential nature of an elephant (see Chapter  13). Each
feels a different part of the animal, such as the trunk, side, tail, or tusk, and all are then in
complete disagreement about what kind of thing an elephant is. The parable might also
be said to apply to the chapters in this volume. Although all authors ostensibly address
the same question of the relationship between addiction and choice, they come up with a
variety of answers depending on what aspects of addiction stand out to them as obvious
and important.
For this reason, if for no other, this concluding chapter will not attempt the task of
summarizing the contents of the book. Even if such an attempt were feasible, it would
be unlikely to be helpful. Rather, while bearing in mind as many of the original ideas
and insightful conclusions to be found in the book as possible, what will be offered is a
464  

464 Overview of addiction as a disorder of choice and future prospects

personal overview of how the science of addiction as a disorder of choice can be taken
forward. As well as some generalizations on future directions for theory and research,
I will pay attention to the implications of this new perspective1 for progress in the treat-
ment and prevention of harm resulting from addiction. Lastly, I will discuss what has
perhaps been insufficiently covered in the book—​the implications of the new perspective
for the public understanding of addiction and what special difficulties this may present.

2  Willing and unwilling addicts

A part of the addiction elephant that stands out for me is that addicts are unwilling par-
ticipants in their addictive behaviour—​that engagement in addiction implies behaving, or
acting, against what considers to be one’s best interests and, in some sense of the phrase,
against one’s will. Based on this premise, I arrived in Chapter 1 at a definition of addiction
as “a repeated and continuing failures to refrain from or radically reduce a specified behav-
ior despite prior resolutions to do so.” Given this definition, the idea of “willing addiction”
is an oxymoron; people may engage in behavior that we think is bad for them or of which
we disapprove but, if they do so willingly and make no attempt to change their behavior,
we have no grounds for attributing to them, from this point of view, anything other than a
strong and unusual preference. Later in Chapter 1 I accepted Skog’s (2003) counterfactual
supplement to my definition that a person can be considered addicted if there are good
reasons to believe that they would demonstrate continuing failures to refrain from or radi-
cally reduce the behavior under different personal or environmental circumstances.
This extension to the definition, however, still does not accommodate at least two
portrayals of addiction to be found in this book—​in the chapters by Bennett Foddy
(Chapter 3) and Owen Flanagan (Chapter 4). Foddy argues that it is possible to become
genuinely “addicted” to any behavior that produces operant reward; there is evidence
that a wide range of behaviors involving, for example, food, gambling, and videogames
all activate the D2 dopamine receptors to produce neural reward and reinforcement of
behavior in the same way that conventionally regarded addictive drugs are thought to do.
Addiction need not involve suffering or distress and whether it is engaged in willingly
or unwillingly is, from Foddy’s perspective, beside the point. In his chapter, Flanagan is
explicitly concerned with examples of the willing addict, which he defines as someone
“who reflectively endorses their addiction” (p. 66), who prefers addiction all things con-
sidered. He reminds us that a pattern of behavior that is often thought of as addiction can
be seen as a lifestyle choice, as valid and defensible in its own terms as any other.
Both these chapters are interesting and valuable contributions to the literature on the
relationship between addiction and choice and to addiction studies broadly considered.
As a clinical psychologist, however, my primary interest is not on behavior willingly
undertaken but on addiction as something that people suffer from, complain of, feel
enslaved by, and wish to liberate themselves from. I also venture to say that this applies to

1 New at least to the general public and to the majority of workers in the addiction field—​see Preface.
  465

Developments in theory and research 465

the majority of authors in this book. It is unwilling addiction, therefore, that will be the
focus of the remainder of this concluding chapter.

3  Developments in theory and research

A premise of the book is that choice is somehow involved in addiction and all authors would
subscribe to this basic, albeit vague, assumption. A  further assumption is that the truth
about addiction lies somewhere between the extremes of completely free choice and no
choice whatever, and most authors subscribe to this general position in one way or another,
some more explicitly than others (e.g. Holton and Berridge, Chapter 9). So too, most would
presumably accept that something is wrong with the choices addicts make—​that addiction
is in some sense a disorder of choice, though some may not wish to use that phrase as a sum-
mary of their position because of its association with views they eschew or for other reasons.
This much having been agreed, there remain a large number of possibilities as to what
kind of disorder of choice addiction represents. If addiction lies between total compulsion
and total free choice, is there a single dimension along which degrees of disordered choice
can be said to exist? If so, what is that dimension and how can it be measured? If one dimen-
sion is insufficient to describe the range of ways in which choice is disordered, what other
dimensions are needed? Though addiction shows neither complete presence nor absence of
free choice, there may nevertheless be a way(s) in which addiction as a disorder of choice,
under some definition of “addiction,” differs qualitatively from some lesser kind of disorder.
If so, how should this difference be characterized? Needless to say, theoretical speculations
along these lines need to result in falsifiable propositions that can be tested in research.

3.1  Addiction as dynamic inconsistency

One answer to the question of what kind of disorder of choice is involved in addiction is
to say that it is a disorder of choice over time or, in other words, a kind of failure to make
consistent choices over time. Such a view has been well articulated by Neil Levy (2006).
Although addicts respond to incentives and so are free to choose to use or not to use at
any one time, their autonomy is impaired when their pattern of choices is considered over
time. Thus, a person makes a strong resolution at time t1 to behave in a certain manner at
time t2 but fails to carry out that resolution; when that happens repeatedly and distress-
ingly, we can describe this pattern of behavior as addiction (Heather 1994). If we identify
the addict’s resolution at t1 with their will, on the ground that this reflects “that set of
considerations which he (the addict)—​in a cool and non-​deceptive moment—​articulates
as definitive of the good” (Watson 1982, p. 105; see also Kennett 2013), it is in this sense
that addictive behavior when it occurs is against the will. In short, in respect of their
addictive behavior, addicts cannot effectively extend their will over time. “It is because
addiction undermines extended agency, so that addicts are not able to integrate their lives
and pursue a single conception of the good, that it impairs autonomy” (Levy 2006, p. 427).
An attractive feature of this view of addiction is that it does not offer a sharp divide
between addicts and the rest of humankind; we are all, nearly every day, passing victims of
466  

466 Overview of addiction as a disorder of choice and future prospects

dynamic inconsistency when we fail to do what we had intended to do, though obviously
not to the extent and with the same devastating consequences that apply to addiction.
Though simply stated here, this view is consistent with a huge philosophical literature on
akrasia or weakness of will that can provide a rich and nuanced account of the dynamic
inconsistency in question and can hopefully lead to specific hypotheses that can be tested
by detailed observations of addictive behavior.

3.2  Explaining temptation and self-​control

Of course, to assert that addiction can profitably be seen as dynamic inconsistency of
choice does not itself explain very much; we need to know why this inconsistency occurs.
Two important leads here are the work of George Ainslie on hyperbolic discounting
(see Ainslie 1992, 2001, Chapter 13, this volume) and of Roy Baumeister and colleagues
on so-​called ego depletion (see Baumeister et al. 1998; Muraven and Baumeister 2000;
Chapter 16, this volume). In later work, Levy (2011a) outlines how these two theoretical
frameworks can be applied to explain the dynamic inconsistency and weakness of will
in addiction. More widely, this approach is a continuation of what is now an established
tradition of thought about addiction (e.g. Loewenstein and Elster 1992; Elster 1999; Elster
and Skog 1999; Levy 2013). The suggestion here is that both hyperbolic discounting and
ego-​depletion, either singly or in combination, should continue to be employed to formu-
late testable hypotheses as to why weakness of will in addiction occurs.
An unresolved issue is whether, at the time that the breakdown in resolution occurs, the
addict continues to be aware that succumbing to temptation is not in her best interests
or whether she has “changed her mind” and now believes that it is the addictive behavior
that is in her best interests. I  referred to these different accounts of the breakdown in
resolution in my chapter on addiction as a form of akrasia (see Chapter 8, this volume).
Recognizing that the dispute over “clear-​eyed akrasia” has a long history in philosophical
discussions of weakness of will (see, e.g., Levy 2011b; Kennett 2013), I stated my view
that the change-​of-​mind position seemed implausible and did not fit with an intuitive
understanding of addictive failures of self-​control (cf. Fitzpatrick 2008; Levy 2011b). But
the important question here is whether this is, or could be made to be, an empirical issue.
In more general terms, it seems clear that the crucial explanandum in the study of
addiction is the breakdown of the resolution at the point at which the addict succumbs to
temptation. What are the key determinants of this breakdown? Conversely, what deter-
mines avoidance of breakdown or, as Richard Holton puts it, “How do agents succeed
in persisting with their resolutions in the face of strong contrary inclinations?” (Holton
2003, p. 39). These are large and, indeed, ancient questions and much sweat and ink have
been expended on them already. The suggestion here is simply that they must continue to
be pursued and be given prominence in scientific enquiry into addiction.

3.3  Clarifying compulsion

As made clear in the Preface, in addition to rejecting the idea that addiction is a completely
free choice, a part of the logic supporting a view of addiction as a disorder of choice is a
rejection of the notion of complete compulsion. It should be unnecessary to point to the
  467

Developments in theory and research 467

crucial role played by the idea of compulsion in most disease theories of addiction in legiti-
mizing the assumption of a disease condition and in absolving the addict for responsibility
and blame for her addictive behavior (see Preface). Several critics have itemized the sources
of evidence and logical argument against compulsion in addiction (see, e.g., Pickard, in
press; Heyman and Mims, Chapter 21, this volume): (1) the extensive evidence showing
that addictive behavior is operant behavior and therefore under voluntary control; (2) the
success of contingency management programs in establishing abstinence in addicts in treat-
ment; (3) large-​scale epidemiological studies showing that in the natural environment the
majority of heroin and cocaine addicts mature out in their early 30s (Heyman and Mims,
Chapter 21, this volume); (4) evidence that these “recoveries” are due to ordinary incentives
and include many of the factors that influence everyday choices, such as economic pressures,
family pressures, the desire to be a better person, and the desire to lead a more meaningful
life; (5) the famous “Rat Park” (Alexander et al. 1978) and similar experiments, showing that
when apparently morphine-​addicted rats are taken out of their isolated cages and placed
in a naturalistic setting where they are able to co-​habit, nest and reproduce, they prefer
plain water to morphine-​laced water even when they experience withdrawal symptoms;
(6) “natural experiments” derived from historical accounts of the effects of legislation and
policies affecting drug availability (Harrison Act, 1914; National Prohibition in the USA,
1919–​1933; US Surgeon General’s Report on smoking 1964) and showing large changes in
rates of addiction inferred as a result of these policies (see Heyman and Mims, Chapter 21,
this volume); (7) evidence of the findings from the follow-​up of Vietnam Veterans show-
ing that, though 20% of service personnel were addicted to heroin in Vietnam, when they
returned to their home towns in the USA the great majority simply gave up using heroin
(Robins et al. 1975); (8) conceptual analysis of what addicts mean when they say that they
“can’t” resist their addictive behavior (see Pickard, in press).
This might be thought a formidable body of evidence against the view that addiction
represents compulsive behavior and in favor of the conclusion that a radical revision of
the popular conception of addiction is called for (see Hari 2015). However, it could be
claimed that this is evidence only against compulsion in a strong sense, a sense captured
perhaps by Aristotle’s image of someone being carried along by a wind or, at least, by an
internalized version of that image in which all control has been removed from a person’s
own motivational states (see Stephens and Graham 2009). Indeed, my co-​editor makes
a claim of this kind in his concluding chapter to the book (see Segal, Chapter  24, this
volume, p. 452). He points out, rightly, that different people mean different things by the
term “compulsive.” He continues:

… no contemporary theoretical model of addiction offered by neuroscientists in the “disease” camp
treats addictive behavior as completely beyond voluntary control, like Parkinson’s shaking (Hyman
2007). What, then, do theorists in the camp mean by “compulsive?” Steven Hyman explains
(Hyman 2007, p. 2): “The term compulsion is imprecise, but at a minimum implies diminished
ability to control drug use, even in the face of factors (e.g. illness, failure in life roles, loss of job,
arrest) that should motivate cessation of drug use in a rational agent willing and able to exert con-
trol over behavior.” And Barry Everitt says “it is indeed compulsive drug use that is widely seen as a
core aspect of addiction, and by that is meant repeated, persistent use, despite placing an individual
in danger, compromising their health, family, and social lives.” (Everitt 2014, p.6).
468  

468 Overview of addiction as a disorder of choice and future prospects

Neither of these quotations provides sufficient ground, in my view, to justify calling addiction
compulsive. They are simply restatements of the central puzzle of addiction, that of explaining
why a certain behavior is continued despite, sometimes severe, negative consequences or, more
crucially, despite the individual’s awareness of those negative consequences and a frequently-​
stated desire to eliminate or reduce them. It does not get us any further in explaining this
puzzle, or even in approaching such an explanation, merely to label the behavior compulsive.
In Section 3 of his chapter, Segal goes on to discuss the senses of the word in which he
sees it as legitimate to describe addiction as involving a compulsion (see Segal, Chapter 24,
this volume, pp.  452–5). Whatever one makes of Segal’s arguments, he is undoubtedly
right to call for a closer examination of the possible meanings of “compulsion” in addic-
tive behavior and whether those meanings justify the retention of the term, particularly
in view of its importance for the public understanding of addiction (see below). As Segal
notes, even those openly hostile to calling addiction compulsive concede that “voluntari-
ness is a continuous dimension” (Heyman and Mims, Chapter 21, this volume, p. 387).
In their chapter, Cox and colleagues say that the dichotomy that is sometimes drawn
between voluntary and automatic behavior is a false one (see Chapter 15, this volume).
Thus, what is recommended here is further review and analysis of the role of possible
meanings of compulsion in addiction, hopefully with empirical consequences.

3.4  Toward a dual-​process theory of addiction

All the foregoing suggestions can be subsumed under the more general proposal that addic-
tion should be explained within the framework of a dual-​process theory of human behavior
and experience. Such a proposal was adumbrated in my Chapter 1 (pp. 16–7) and Chapter 8
(pp. 145–6), is explicit in the title of Henden’s Chapter 7, is specifically referred to in Cox and
colleagues’ Chapter 15, and, while the term “dual-​process” may not be used, can be inferred
from several others (e.g. Papineau and Butlin, Chapter 6; Holton and Berridge, Chapter 9;
Lewis, Chapter  10; Naqvi and Bechara, Chapter  12; Rachlin, Chapter  14; Vonasch et  al.,
Chapter 16; Segal, Chapter 20). The proposal is by no means new and the idea that addiction
can profitably seen in dual-​process terms has been expressed in several publications in recent
years (e.g. Evans and Coventry 2006; Moss and Albery 2009; Redish et al. 2008; Redish 2010;
Vandermeeren and Hebbrecht 2012; Wiers and Stacy 2006; Wiers et al. 2014).
A difficulty is that there are many different versions of dual-​process theory and that
they have been applied to various aspects of human psychology—​attitudes, judgments,
thinking, reasoning, etc. (Evans 2009). Nevertheless, we can say here, in simple terms,
that dual-​process theory provides an account of how behavior can be the result of one
of two different kinds of information processing and of their interaction—​implicit, auto-
matic and mainly nonconscious processes and explicit, controlled, and mainly conscious
processes. A  distinction can be made between dual-​process theories and dual-​system
theories; the former assume that two distinct processes of thinking compete for control
of behavior, while the latter, stronger theories “ascribe the origin of these dual processes
to biologically distinct cognitive systems with sharply differing evolutionary histories”
(Evans and Coventry 2006, p. 30). For example, in an influential model, Strack and Deutsch
  469

Developments in theory and research 469

(2004) assume that behavior is a joint function of two parallel and interacting process-
ing systems that follow different operating principles: a reflective system that generates
behavioral decisions based on knowledge about facts and values, and an impulsive system
that elicits behavior through associative links and motivational orientations2. In neutral
language, Kahneman (2003) refers to System 1 and System 2 styles of cognitive process-
ing; the former is fast and automatic, usually with strong emotional bonds, whereas the
latter is slower and subject to conscious judgments and attitudes. System 1 processes are
very difficult to change or manipulate, while System 2 processes are much more volatile
and subject to influence and disturbance. Theorists agree that System 2 and its neural
substrates evolved much later and now work alongside, and often in competition with, the
evolutionarily older autonomous sub-​systems of System 1; we share System 1 with higher
animals but System 2 is uniquely and distinctively human (see, e.g., Stanovich 2004).
Ongoing behavior, then, is a result of a dynamic interaction and balance between
System 1 and System 2 processes. It is easy to see that, in addiction, this balance has
become disturbed. Automatic processes, like cue-​elicited craving, urges, attentional bias,
automatic approach tendencies, implicit memory associations and cognitions, promote
the maintenance of addictive behavior, while poor goal-​directed planning, limited self-​
control capacity, and distorted judgments and evaluations undermine attempts to resist it.
The “final common pathway” of overt behavior has become biased in favor of automatic,
System 1 processes. Another way to put this, more in tune with a neuroscientific per-
spective, is that addiction involves a failure of top-​down regulatory control of bottom-​up
automatic processing, a failure due to an unusually strong System 1, an unusually weak
System 2 or a combination of both (cf. Bechara et al. 2006).
The above is a short and simplified—​hopefully not over-​simplified—​account of the
foundations for a dual-​process theory of addiction. Several difficulties in the way of devel-
oping such a theory have not been touched on (e.g. ways in which reflective system pro-
cesses may themselves be influenced by impulsive processes and vice-​versa, the effects on
the interaction between the two systems of substance use and of substance deprivation,
etc.) but I trust enough has been said to indicate the possibilities of such a theory and
how it may be compatible with other leads on addiction described in this chapter and this
book—​with philosophical work on weakness of will and akrasia (cf. Strack and Deutsch
2004), with the theoretical implications for addiction of ego depletion and hyperbolic
discounting, with the idea of addiction as dynamic inconsistency, with research on the
psychology of temptation and self-​control, and with a clarification of the role of “com-
pulsion” in the explanation of addiction. Furthermore, although it remains to be con-
vincingly demonstrated, the suggestion is that a comprehensive dual-​process theory of
addiction would be roughly compatible with all the major neuroscientific accounts of
addiction to have appeared over the last two decades (i.e. Robinson and Berridge 2003;
Bechara 2005; Everitt and Robbins 2005; Hyman 2005; Kalivas and Volkow 2005; Koob
and Le Moal 2006; cf. Wiers et al. 2014).

2 For an application of this model to addiction, see Deutsch and Strack (2006).
470  

470 Overview of addiction as a disorder of choice and future prospects

4  Implications for treatment and brief intervention

This topic was addressed directly in Section V of this volume. From the viewpoint of a
frontline provider of counselling for those with addictive disorders and using the met-
aphor of the blindfold of addiction, in Chapter  17 Beth Burgess explains how addicts
are prevented from making better choices in relation to their addictive behavior. In
Chapter 20 Gabriel Segal argues that, in addiction, the power to choose has been lost but
can be regained through participation in 12-​step programs. In Chapter 18 Murphy and
colleagues describe the development and testing of a novel form of brief intervention,
based jointly on principles of behavioral economics and motivational interviewing, and
aimed at reducing impulsive choices. In addition to these, however, nearly every other
chapter in the book has some bearing on the implications of choice models of addiction
for the treatment of addictive disorders, while many include recommendations, vary-
ing in tentativeness and specificity, as to how treatment should be carried out (see, e.g.,
Chapters 2, 8, 11, 12, 14, 15, 16, 12, and 22).

4.1  Enhancing motivation or improving self-​control?

Arguably the main implication of the contents of this book for formal treatment of addic-
tive disorders is that clinical research and practical applications should renew a focus
on improving methods to assist people achieve greater control of their behavior. This
is certainly true if the promise of a dual-​process account of addiction is accepted (see
above). Self-​control training was an important component of the more general cognitive-​
behavioral approach to treating behavioral problems developed during the 1970s (e.g.
Thoresen and Mahoney 1974; Meichenbaum 1977); it formed a key part of Marlatt and
Gordon’s (1985) highly influential relapse prevention program for addictive behaviors
and was the basis for the most popular form of treatment program aimed at a moderation
drinking goal for problem drinkers (Hester and Miller 1989).
Following publication of the first edition of Motivational Interviewing (MI; Miller and
Rollnick 1991) and the enormous popularity of this and subsequent editions of the book
(Miller and Rollnick 2002, 2013) among workers in the addictions field, enthusiasm for
self-​control methods appeared to have waned; the achievements of previous work in the
1970s and 1980s remained available to therapists but research and development on more
effective techniques was retarded by the enormous volume of research and applications of
MI.3 This does not mean that research and implementation of MI should be abandoned or
even decreased; it must continue to play an essential part in the response to the harmful
use of substances and other harmful behaviors. However, the roles played by MI and self-​
control methods in treatment can be clearly distinguished and are equally important (cf.
Henden, Chapter 7, this volume). In terms used in this book, MI is principally concerned
with eliciting resolutions to change behavior but not so much with advising and training
people in how such change can be achieved and maintained; it assumes that in many

3 I do not suggest for one moment that this was the intention of the originators of MI or its enthusiasts.
  471

Implications for treatment and brief intervention 471

cases, once motivation has been enhanced, the client will be able to change behavior with-
out additional help. While it is true that natural recovery from addictions of various kinds
is a frequent occurrence in the population (see, e.g., Heyman and Mims, Chapter 21), it is
also true that many have difficulty changing behavior even after firmly resolving to do so.
Indeed, it is an important sense of “addiction” that it reflects the fact that people do have
trouble making desired changes to behavior, a sense of the word that has been recom-
mended in this book (see Heather, Chapter 1). For these people, teaching and practice in
how temptations can be resisted and self-​control over unwanted behavior sustained is the
necessary supplement to enhancing motivation to change. This applies most obviously to
formal treatment for those with moderate or severe degrees of addiction, but applies also
to brief interventions among those at lower levels of addiction for whom simple advice or
brief MI are insufficient to bring about long-​lasting change (see Heather 2010).

4.2  Recent
developments in self-​control
and “willpower” training
Despite the neglect of self-​control methods mentioned above, there are signs that they are
coming back into fashion, stimulated no doubt by just the kind of theoretical and research
advances described in this book. First, there are the various kinds of “cognitive bias modi-
fication” interventions briefly described by Cox and colleagues in Chapter 15. These tech-
niques are clearly in early stages of development and testing but show promise in their
ability to modify automatic impulsive processes subserving addictive behavior. Secondly,
an obvious contribution to the armory of the self-​control therapist would be a logical
extension of ego-​depletion theory to practical training in “willpower.” Given that willpower
apparently works like a muscle that gets tired in the short term but can be strengthened
in the long term (see Vonasch et al., Chapter 16, this volume), systematic training in and
regular practice of self-​control techniques is an obvious direction for the future treatment
of addictive disorders (see Muraven et al. 1999). Preliminary but promising findings of this
method applied to smoking cessation have been reported by Muraven (2010) but surely
much more can be expected from this direction. Thirdly, comparing an older version of
cognitive therapy for addiction by Beck and colleagues (1993) with a more recent version
by Ryan (2013a), it is clear that the latter has benefitted from the advances in cognitive sci-
ence and neuroscience that have occurred during the intervening period and, in particular,
from dual-​process formulations of addictive disorders. To these developments should be
added the potential of mindfulness training as a means of establishing greater cognitive
control over impulsive urges and preventing relapse (see Bowen et al. 2014).
Lastly, there has recently been a rash of popular self-​help books offering instruc-
tion on how to increase will​power and self-​control (e.g. Baumeister and Tierney 2011;
Martela 2013; McGonigal 2013; Ryan 2013b). All these books have been authoritatively
written by PhDs (and in Baumeister’s case by a world expert) with considerable knowl-
edge of the scientific foundations for their claims and all mention addictions as a prime
target for the application of improved self-​control, though obviously their effectiveness
in reducing the prevalence of addictive behaviors among the general public remains
472  

472 Overview of addiction as a disorder of choice and future prospects

to be demonstrated. They also illustrate that there is clearly a popular market for such
instruction and illustrate a way in which stigmatic attitudes to addiction and addicts
could be reduced (see below).

4.3  Treatments based on addiction as a brain disease

As this volume clearly demonstrates, neuroscience has made significant contributions to our
understanding of addiction and can be expected to continue to do so. It must be recognized,
however, that the leading neuroscientific theories of addiction differ in important ways—​
for example, in whether addiction should be seen as primarily a disorder of learning and
memory or of motivation (see Berridge and Robinson 2011). Although all major theorists
agree that dopamine is centrally involved in the development and maintenance of addiction,
the precise functions of this neurotransmitter are debated (Berridge 2007). Even the key role
of dopamine in substance addictions in general has recently been disputed (Nutt et al. 2015).
Leaving these theoretical differences aside, however, what are the consequences of the neuro-
science of addiction, now and in future, for the treatment of addictive disorders?
It is now nearly 20 years since the then Director of the National Institute on Drug
Addiction (NIDA) in the USA published what might be termed a manifesto for a con-
cept of addiction as a chronic, relapsing brain disease—​or the brain disease model of
addiction (BDMA) for short (Leshner 1997). Included in the manifesto were promises
that the BDMA would deliver more effective and accessible medical treatments based
on reversing or compensating for the fundamental changes in brain function believed to
be at the core of the disease. Recently, in addition to arguing that the BDMA as a whole
has not been supported by evidence, Wayne Hall and colleagues (2014) have claimed
that the better treatments promised by Leshner and NIDA have failed to materialize
(see also Kalant 2010). They point out that: (1) few new drugs have been approved for
the treatment of addiction over the past two decades; (2) the most widely used drugs in
addiction (e.g. methadone and nicotine replacement therapy) preceded the BDMA by
more than 30 years; (3) the few drugs derived from neurobiological research into addic-
tion (e.g. naltrexone and varenicline) are only modestly superior to older drugs like
disulfiram and various forms of nicotine replacement therapy; and (4) NIDA invest-
ment in research on vaccines against nicotine and cocaine dependence has produced
disappointing results.4 Meanwhile, owing to the influence of the BDMA on policy and
on research funding5 and the resulting pursuit of expensive, high technology cures,

4 Disappointing results apply not only to vaccines and conventional medications of various types aimed
at preventing relapse but include implantable neurotransmitter agonists and antagonists, DNA tests
aimed at patient-​treatment matching, drugs to reverse epigenetic changes in chronic drug use, ablative
neurosurgery, and deep brain stimulation.
5 Leshner himself testified in 1998 that NIDA supported more than 85% of the world’s research on drug
abuse and addiction and this percentage may well be even greater now. The majority of this funding is
devoted to basic and clinical neuroscience research and development of pharmacotherapies (see Hall
et al. 2014).
  473

Implications for treatment and brief intervention 473

there has been a neglect of public health policies aimed at reducing harmful drug use in
the general population and of ways to increase access to cheap and effective treatments
that already exist.
In their reply, Nora Volkow, the current Director of NIDA, and George Koob, the
recently appointed Director of NIDA’s sister organization, the National Institute on
Alcohol Abuse and Alcoholism (NIAAA), reject all the claims made by Hall and col-
leagues (see Volkow and Koob 2015). They wonder why, given all the evidence they cite
in its favor, the BDMA is controversial. In their reply to this reply, Hall et al. (2015) say
that Volkow and Koob simply repeat the promise of future treatment advances and put
a highly favorable spin on the modest achievements in BDMA-​inspired treatments that
have appeared since Leshner’s (1997) article. They restate their belief that “a narrow
focus on the neurobiological causes of severe cases of addiction has over-​allocated scarce
research and health resources towards neurobiological solutions that have benefitted and
will benefit few people with addictions” (p. 867).6
Of course, the BDMA is not identical with neuroscience-​based theories of addiction;
the latter are wider in scope than the BDMA and do not necessarily lead to the conclu-
sion that addiction is a chronic, relapsing brain disease (see, e.g., Lewis, Chapter 10, this
volume). At the same time, it is a serious mistake to imagine that improvements to phar-
macotherapies and other biological treatments follow inevitably from advances in the
neuroscience of addiction or that such improvements, assuming they were possible, are
the only or best way forward. It too often seems to be assumed in scientific and govern-
ment policy circles that our ability to improve the treatment of addiction relies on the
invention of new medications and other biological interventions,7 an assumption that can
and should be constantly challenged.
From the viewpoint of the framework for explaining addiction shown in Chapter  1
(Figure 1.1, p. 12), medications presumably work at the lowest level by altering the actions
of relevant neurotransmitters and brain systems, and hence at the middle level by reduc-
ing abnormal desires, conscious or nonconscious, for the effects of drugs or addictive
behaviors (or, from a wider dual-​process perspective, automatic System 1 responses to
drugs and drug-​related cues). Because they thereby eliminate or substantially reduce the
experience of temptations leading to the breakdown of resolutions to quit or cut down, the

6 Readers should, of course, make up their own minds on these issues and I encourage them to con-
sult the original texts. It may be noted, however, that addiction scientists and practitioners around
the world are by no means unanimous in their support for the BDMA, as shown when 94 addiction
researchers and clinicians signed a letter to the editors of the journal Nature to protest about their
assumption that the BDMA represented the consensus view in the addictions field (see Heim 2014).
7 In 2005 I was invited to attend a Foresight project run by the UK Government’s Office of Science and
Innovation and entitled, Drug Futures 2025. The project was set up to consider how to manage the use
of psychoactive substances in the future to the best advantage of the individual, the community and
society (see Nutt et al. 2007). I attended one meeting but did not return because it was clear to me it
had already been decided that the development of new medications for the brain disease of addiction
was the only possible way forward.
474  

474 Overview of addiction as a disorder of choice and future prospects

problem of self-​control over behavior is avoided and the probability of relapse decreased.
But, at the highest level of akrasia, this can only be a temporary solution to the continu-
ing problem of self-​control (unless it is envisaged that medications should be taken for a
lifetime, an undesirable outcome for many individual and societal reasons). Thus medica-
tions can only ever be adjunctive in the treatment of addictive behaviors. Apart from in
difficult detoxification, they can play a vital part by cutting through the vicious cycle of
mounting problems and increasing, chaotic consumption, leading to further problems
and so forth, and by making possible a period of stability in which other life problems
in relationships, accommodation, livelihood etc., may be addressed. But this essential
period of stability leaves the more lasting problem of temptation and self-​control largely
untouched; any permanent or long-​lasting solution to this problem can only be achieved
by learning ways to strengthen self-​control over addictive behavior, whether by cognitive-​
behavioral therapy, mindfulness training, membership of a mutual aid group, or any other
relevant method.

5  Implications for the prevention of addiction

Implications for prevention were included in Section V and were addressed by Jalie Tucker
and her colleagues in Chapter 19. From a behavioral economics framework, they describe
how human choices are vulnerable to delay discounting and other common biases, and
then discuss the implications of these biases for public health strategies to prevent and
reduce addictive behaviors. Because one of the main aims of brief intervention is the
secondary prevention of more serious problems, the chapter by Murphy and colleagues
(Chapter 18) is relevant here too.
The most obvious way to reduce the prevalence of addiction in society is by population-​
based measures to reduce the consumption, or hazardous/​harmful consumption, of sub-
stances to which they are formed (Babor et al. 2010a, 2010b). This applies also to activities,
like gambling, that can become the basis for addictive behaviors (Gallagher 2013). These
observations on prevention are entirely consistent with choice models of addiction but
obviously not unique to them. They will therefore not be considered in detail here.

5.1  Socioeconomic deprivation and addiction

There is one major issue that has not been covered in this book as much as it should
have been but was mentioned by Pickard and Ahmed in Chapter 2. This is the existence
of external as well as internal limitations on choice and how this affects impoverished
and culturally deprived members of society (see Pickard in press; Orford 2013; Burgess,
Chapter 22, this volume). As Pickard and Ahmed point out, there is good evidence that
addiction, together with mental health problems, is associated with low socioeconomic
status (e.g. Compton et al. 2007). On the other hand, because addiction entails continuing
use despite negative consequences, wealth and other forms of privilege provide protec-
tion against those negative consequences and thus against addiction. Lacking any such
protection, as Pickard and Ahmed say, “Some addicts may choose to continue to use, not
  475

Implications for the public understanding of addiction 475

only despite harrowing consequences, but because of them” (p.  33, italics original)—​in
other words, because the benefits seem to exceed the costs on a realistic evaluation of life
circumstances and the options available to them. While evidence clearly shows that, in
the natural environment, the majority of those addicted to illicit drugs mature out in their
early 30s (see Heyman and Mims, Chapter 21, this volume), more chronic addiction is
shown mainly by marginalized people from underprivileged backgrounds who also suffer
from co-​morbid psychopathology—​those who lack personal, social and economic capital
and whom Storbjoerk and Room (2008) termed “the spare people” in society (cf. Burgess,
Chapter 22, this volume).
As Pickard and Ahmed make clear, externally restricted choice applies as much to the
possibilities for recovery from addiction as to the development of addiction in the first
place. This does not mean, to my mind, that addiction for the culturally and economically
underprivileged is a rational choice; that, at least, is a debatable point. The message for the
prevention of addiction, however, is simple and unavoidable: to reduce the incidence and
chronicity of addiction we must strive to eliminate poverty and deprivation.

6  Implications for the public understanding of addiction

This book has proposed that seeing addiction as a disorder of choice is necessary on
scientific grounds for a better understanding of the nature of addiction and hence for
progress in the treatment and prevention of addictive behavior. But what are the implica-
tions of this revised account of addiction for public understanding—​for the views of the
general public on the extent to which public funds should be allocated to research on and
treatment of addiction, for the sense family members make of their loved one’s otherwise
incomprehensible behavior and, importantly, for the sense addicts themselves make of
their own troubles?
Traditionally, the disease view of addiction, based on the idea that addictive behavior
is compulsive and therefore outside voluntary control, has been used to try to persuade
the general public (and policy-​makers) that addicts are not responsible for their behavior
and should not therefore be blamed and punished for it but should instead be offered
treatment. It is invariably suggested in these pronouncements that this is the only way to
persuade the public to withhold blame from addicts and to support a caring and compas-
sionate response to them. Seeing addiction as a disease is the only way of combatting the
stigma addicts will otherwise be subjected to (see, e.g., Leshner 1997, but many other
examples could have been chosen.)
As a general principle, the idea that calling addiction a disease is the only way that the
public can be persuaded to avoid blaming and punishing addicts is simply untrue. It is
obviously possible to present a behavioral, a psychosocial or a biopsychosocial model of
addiction without implying either that it is a disease or that addicts should be blamed
and punished. It should go without saying that blame and punishment have no place in
science and that one can’t derive an “ought” from an “is” (Hume [1739] 2004). The true
alternative to a moralistic view of addiction is a scientific account, not merely a disease
476  

476 Overview of addiction as a disorder of choice and future prospects

account. It might be claimed, nevertheless, that the attribution of disease is the only way
the general public can understand the argument that addicts should not be blamed and
punished (i.e. “they can’t help it because they’ve got a disease”). Patronizing though this
seems to be, is it true? If so, does this mean that a scientific account of addiction must
necessarily be a disease account? If not, how can the public be educated to accept non-​
disease explanations of addiction without applying blame? To these questions I now turn.

6.1 Stigma
Though the disease concept has not been successful in persuading Anglo-​Saxon law that
addiction should be an excusing or mitigating condition for criminal acts (see Morse,
Chapter 23, this volume), it is undoubtedly true that, over the years, it has been respon-
sible for diverting many users of illicit drugs and addicts from prison to treatment and
this is no mean achievement. A wider claim, however, is that, because addicts are not held
to be responsible for their behavior, it can remove the stigma that would otherwise be
attached to addiction in the public mind.
The interplay between addiction and stigma is more complicated than this suggests.
First, stigma against heavy drinking or other substance use is an important informal
social control that inhibits the progression to heavier use and more serious harm (Room
2005). Secondly, the straightforward notion that seeing a behavior as the consequence of
disease necessarily leads to benign attitudes to it can be empirically examined and is, in
fact, found wanting. Some years ago, a colleague and I conducted a survey of the general
public in Scotland regarding their attitudes to the disease concept of alcoholism and the
treatment of alcoholics (Crawford and Heather 1987). The results showed that whether
or not people believed in the disease concept of alcoholism had little directly to do with
whether or not they had a sympathetic attitude to the treatment of alcoholics and believed
that public money should be devoted to treatment, etc.; the latter was better predicted by
more general, non-​condemnatory and humanitarian attitudes to socially deviant groups
as a whole. Other surveys have also produced findings inconsistent with the equation of
disease beliefs and sympathetic attitudes to alcohol addicts (see Heather and Robertson
1997, pp. 100–​101).
More recently, Kvaale and colleagues (2013) carried out the first meta-​analytic review
of studies looking at the effects on stigma of biogenetic explanations of mental disorders,
including substance use disorders. The main finding was that such explanations reduced
blame but induced pessimism over the prognoses for those suffering from these disor-
ders. They also found that biogenetic explanations increase endorsement of the stereotype
that people with psychological problems are dangerous, an understandable reaction to the
idea that addiction is the result of permanent changes to brain mechanisms over which
the sufferer has no control. Moreover, Bell and colleagues (2014) found very mixed sup-
port for the BDMA among clinicians involved in the treatment of addictions in Australia,
echoing earlier findings of ambiguous and ambivalent attitudes to the disease concept of
alcoholism among clinicians in the UK (see Heather and Robertson 1997, p. 101). This
appears to be a growing area of research and more findings of interest can be expected.
  477

Implications for the public understanding of addiction 477

What can be concluded is that the claim that the disease concept of addiction, and the
BDMA in particular, removes stigma among the general public and is wholly of benefit in
the treatment of addiction is currently unsupported by evidence.

6.2  The risk of being misunderstood

The above considerations apply to the merits and drawbacks of disease models of addiction
versus any non-​disease model. However, there might reasonably be thought to be special
problems attaching to disordered choice models of addiction. The very concepts and terms
used in some of these models—​willpower, weakness of will, failure of self-​control—​might
suggest to the layperson that there is a reversion to the stance that addicts are weak or bad
people, choosing to lead immoral lives and unwilling to control their impulses for imme-
diate gratification. No matter how often it is repeated that it is a disordered choice that is
being proposed, not a free choice, it is likely that this distinction will be ignored when
choice in relation to addiction is merely mentioned. If a theory of addiction along these
lines describes agents actively making choices rather than passively reacting to internal
and external stimuli, the same kinds of distortion are likely to arise. Though it is quite pos-
sible to have a model of addiction involving “responsibility without blame” (i.e. that, for
explanatory purposes, addicts should be regarded as agents responsible for their behavior
but without any implication that they should be blamed for it; Pickard 2012), it is unlikely
that such a distinction can acquire much traction outside academic circles.
Public misunderstanding of choice models will be inevitably be fuelled by over-​
simplified, distorted and sensationalist portrayals in the media, not to mention by scien-
tists and clinicians with vested interests in the BDMA. There is also the danger that choice
models play into the hands of those who profit from the sale of products with addiction
potential, especially those in the smoking, alcohol and gambling industries who appeal
to the consumer’s “freedom to choose” and their “personal responsibility” for controlling
their consumption. Such discourses are also endorsed, of course, by governments of a
neoliberal complexion. It would be naïve and short-​sighted not to recognize these societal
and political risks of proposing that addiction is a disorder of choice.
The irony is that it is perfectly true that disordered choice models, or at least those
that invoke the construct of akrasia, regard addicts as weak-​willed. But this is only
because we are all weak-​willed. Addicts struggle with extreme variants of a difficulty
in controlling behavior that affects all members of the human race on a daily basis and
there is no good evidence that, outside the area of their addictions, addicts are any more
weak-​willed than other people. Hoffman et al. (2012) carried out a large-​scale sampling
study of adult citizens of Germany and found that people reported fighting against a
desire for approximately one-​quarter of their waking hours, giving in on roughly half of
occasions. In the USA, when people were asked about reasons for failing to meet their
goals for healthy living, they named lack of willpower as the most important contributor
(American Psychological Association 2010). This suggests that, if addiction were pre-
sented to the public in the way advocated here—​as an extreme version of a problem with
which we are all familiar—​and buttressed perhaps by slogans along the lines of, “There
478  

478 Overview of addiction as a disorder of choice and future prospects

but for the grace of God go I,” understanding and compassion might be increased and
stigma avoided or, at least, reduced. The recent flurry of self-​help books on willpower
(see p.  471 above) suggests that the public might be receptive to such messages. This
kind of education would have the very opposite effect to telling people that addicts have
a brain disease.
If it is thought that the general public is unable to accept the idea that addiction is a
disorder of choice without reverting to moralistic and punitive attitudes, then the only
solution would be to countenance an increasing bifurcation between the scientific and
public understandings of addiction. This bifurcation is the case in many areas of science
where progress in science far outstrips the layperson’s view. It would be unfortunate if this
happened for addiction because the decisions of policy-​makers are based more on what
they imagine the public believes than on scientific evidence.

6.3  A risk worth taking?

Despite all the problems just described, the reader will not be surprised to learn that my
answer to this question is in the affirmative. The highest priority of all is to strive for as
true an understanding of addiction as it is possible to reach and this can only be done by
beginning with the assumption that addiction is a disorder of choice. This priority trumps
any risks that may apply.
At the very least, there should be a serious discussion of these risks. Ten years ago,
Volkow and Li (2005) wrote:
despite these advances in understanding the neuroplastic changes to drugs and alcohol, addicted
individuals continue to be stigmatized by the pernicious yet enduring popular belief that their
affliction stems from voluntary behavior. The loss of behavioral control in the addicted individual
should spur a renewed discussion of what constitutes volition, challenge us to identify the neuro-
biological substrates that go haywire, and influence our evolving strategies to direct our efforts to
prevent and treat substance abuse and addiction more effectively.

(p. 1430)

There are things to agree with and disagree with in this short passage. First, while addic-
tion does not “stem” from voluntary behavior, it consists, as the contents of this book have
amply demonstrated, of both voluntary and involuntary behavior. As I argued in section
6.2, the task is to explain this to the public in a way that does not lead to stigmatization.
It is certainly true, however, that there should be discussion between scientists and aca-
demics, assisted by empirical research, on what constitutes volition in addiction and how
precisely voluntary and involuntary processes interact.
As Heyman and Mims recommend in Chapter 21, there should also be a conversation
with the general public in which they are urged to accept that breaking free of addiction
is possible and told what we know about how this can be accomplished. The palpable fact
that addicts, despite very strong temptations over which they have little control, neverthe-
less engage in the final resort in intentional, voluntary behavior can no longer be ignored
and swept under the carpet but must be clearly recognized, and its implications for how
we respond to addiction honestly debated.
  479

Implications for the public understanding of addiction 479

Acknowledgments
Thanks to Matt Field and Gabriel Segal for helpful comments on an earlier draft of this
chapter.

References
Ainslie, G. (1992). Picoeconomics: The Strategic Interaction of Successive Motivational States Within the
Person. Cambridge, UK: Cambridge University Press.
Ainslie, G. (2001). Breakdown of Will. Cambridge, UK: Cambridge University Press.
Alexander, B.K., Coambs, R.B., and Hadaway, P. (1978). The effect of housing and gender on morphine
self-​administration in rats. Psychopharmacology, 58, 175–​79.
American Psychological Association (2010). Americans report willpower and stress as key obstacles
to meeting health-​related resolutions. http://​www.apa.org/​news/​press/​releases/​2010/​03/​lifestyle-​
changes.aspx. Accessed 05.05.2016.
Babor, T., Caetano, R., Casswell, S., et al. (2010a). Alcohol: No Ordinary Commodity—​Research and
Public Policy (2nd ed.). Oxford: Oxford University Press.
Babor, T.F., Caulkins, J.P., Edwards, G., et al. (2010b). Drug Policy and the Public Good. Oxford: Oxford
University Press.
Baumeister, R.F., Bratlavsky, E., Muraven, M., and Tice, D.M. (1998). Ego depletion: is the active self a
limited resource? Journal of Personality and Social Psychology, 74, 1252–​65.
Baumeister, R.F. and Tierney, J. (2011). Willpower: Why Self-​Control is the Secret to Success.
London: Penguin.
Bechara, A. (2005). Decision making, impulse control and loss of willpower to resist drugs: a
neurocognitive perspective. Nature Neuroscience, 8, 1458–​63.
Bechara, A., Noel, X., and Crone, E.A. (2006). Loss of willpower: abnormal neural mechanisms of
impulse control and decision making in addiction. In: R.W. Wiers and A.W. Stacy (eds), Handbook
of Implicit Cognition and Addiction. Thousand Oaks, CA: Sage, pp. 215–​32.
Beck, A.T., Wright, A.T., Newman, C.F., and Liese, B.S. (1993). Cognitive Therapy of Substance Abuse.
New York: Guilford.
Bell, S., Carter, A., Mathews, R., Gartner, C.E., Lucke, J., and Hall, W. (2014). Views of addiction
neuroscientists and clinicians on the clinical impact of a “brain disease model of addiction.”
Neuroethics, 7, 19–​27.
Berridge, K.C. (2007). The debate over dopamine’s role in reward: the case f​ or incentive salience.
Psychopharmacology, 191, 391–​431.
Berridge, K.C., and Robinson, T.E. (2011). Drug addiction as incentive sensitization. In: J. Poland and
G. Graham (eds.), Addiction and Responsibility. Cambridge MA: MIT Press, pp. 21–​54.
Bowen, S., Witkiewitz, K., Clifasefi, S.L., et al. (2014). Relative efficacy of mindfulness-​based relapse
prevention, standard relapse prevention, and treatment as usual for substance use disorders: a
randomized clinical trial. JAMA Psychiatry, 71, 547–​56.
Compton, W.M., Thomas, Y.F., Stinson, F.S., and Grant, B.F. (2007). Prevalence, correlates,
disability, comorbidity of DSM-​IV drug abuse and dependence in the United States: results
from the national epidemiologic survey on alcohol and related conditions. Archives of General
Psychiatry, 64, 566–​76.
Crawford, J. and Heather, N. (1987). Public attitudes to the disease concept of alcoholism. International
Journal of the Addictions, 22, 1129–​38.
Deutsch, R. and Strack, F. (2006). Reflective and impulsive determinants of addictive behavior. In: R.W.
Wiers and A.W. Stacy (eds), Handbook of Implicit Cognition and Addiction. Thousand Oaks,
CA: Sage, pp. 45–​58.
480  

480 Overview of addiction as a disorder of choice and future prospects

Elster, J. (ed.). (1999). Addiction: Entries and Exits. New York: Russell Sage Foundation.

Elster, J. and Skog, O.-​J. (eds) (1999). Getting Hooked: Rationality and Addiction. Cambridge,
UK: Cambridge University Press.
Evans, J.St.B.T. (2009). How many dual process theories do we need? One, two or many? In: J.St.B.T.
Evans and K. Frankish (eds), In Two Minds: Dual Processes and Beyond. Oxford: Oxford University
Press, pp. 33–​54.
Evans, J.St.B.T. and Coventry, K. (2006). A dual-​process approach to behavioral addiction: the case
of gambling. In: R.W. Wiers and A.W. Stacy (eds), Handbook of Implicit Cognition and Addiction.
Thousand Oaks, CA: Sage, pp. 29–​44.
Everitt, B.J. and Robbins, T.W. (2005). Neural systems of reinforcement for drug addiction: from
actions to habits to compulsion. Nature Neuroscience, 8, 1481–​89.
Fitzpatrick, W.J. (2008). Moral responsibility and normative ignorance. Ethics, 118, 518–​613.
Gallagher, P. (2003). Addiction soars as online gambling hits £2bn mark. Independent, Sunday, February
3. http://​www.independent.co.uk/​news/​uk/​home-​news/​addiction-​soars-​as-​online-​gambling-​hits-​
2bn-​mark-​8468376.html. Accessed 05/​05/​2016.
Hall, W., Carter, A., and Forlini, C. (2014). The brain disease model of addiction: is it supported by the
evidence and has it delivered on its promises? Lancet Psychiatry, 2, 105–​10.
Hall, W., Carter, A., and Forlini, C. (2015). Brain disease model of addiction: misplaced priorities?
Lancet Psychiatry, 2, 867.
Hari, J. (2015). Chasing the Scream: The First and Last Days of the War on Drugs. London: Bloomsbury
Circus.
Heather, N. (1994). Weakness of will: a suitable topic for scientific study? (Editorial). Addiction
Research, 2, 135–​39.
Heather, N. (2010). Breaking new ground in the study and practice of alcohol brief interventions. Drug
and Alcohol Review, 29, 584–​88.
Heather, N. and Robertson, I. (1997). Problem Drinking (3rd ed.). Oxford: Oxford University Press.
Heim, D. (2014). Addiction: not just brain malfunction (Letter). Nature, 507, 40.
Hester, R.K. and Miller, W.R. (1989). Self-​control training. In: W.R. Miller and R.K. Hester (eds), Handbook
of Alcoholism Treatment Approaches: Effective Alternatives. New York: Pergamon, pp. 141–​49.
Hofman, W., Baumeister, R.F., Forster, G., and Vohs, K.D. (2012). Everyday temptations: an experience
sampling study of desire, conflict and self-​control. Journal of Personality and Social Psychology, 102,
1318–​35.
Holton, R. (2003). How is strength of will possible? In: C. Tappolet and S. Stroud (eds), Weakness of Will
and Practical Irrationality. Oxford, UK: Oxford University Press, pp. 39–​67.
Hume, David ([1739] 2004). A Treatise of Human Nature. London: Dover.
Hyman, S.E. (2005). Addiction: a disease of learning and memory. American Journal of Psychiatry, 162,
1414–​22.
Kahneman, D. (2003). A perspective on judgement and choice: mapping bounded rationality. American
Psychologist, 58, 697–​720.
Kalant, H. (2010). What neurobiology cannot tell us about addiction. Addiction, 105, 780–​89.
Kalivas, P.W. and Volkow, N.D. (2005). The neural basis of addiction: a pathology of motivation and
choice. American Journal of Psychiatry, 162, 1403–​13.
Kennett, J. (2013). Just say no? Addiction and elements of self-​control. In: N. Levy (ed.), Addiction and
Self-​Control. Oxford, UK: Oxford University Press, pp. 144–​64.
Koob, G.F. and Le Moal, M. (2006). The Neurobiology of Addiction. New York: Academic Press.
Kvaale, E.P., Haslam, N., and Gottdiern W.H. (2013). The “side effects” of medicalization: a meta-​
analytic review of how biogenetic explanations affect stigma. Clinical Psychology Review, 33, 782–​94.
  481

Implications for the public understanding of addiction 481

Leshner, A.I. (1997). Addiction is a brain disease, and it matters. Science, 278, 45–​47.
Levy, N. (2006). Addiction and autonomy. Canadian Journal of Philosophy, 36, 427–​47.
Levy, N. (2011a). Addiction, responsibility, and ego depletion. In: J. Poland and G. Graham (eds),
Addiction and Responsibility. Cambridge, MA: MIT Press, pp. 89–​112.
Levy, N. (2011b). Resisting “weakness of will.” Philosophy and Phenomenological Research, 82, 134–​55.
Levy, N (Ed.). (2013). Addiction and Self-​Control: Perspectives from Philosophy, Psychology and
Neuroscience. Oxford, UK: Oxford University Press.
Loewenstein, G. and Elster, J. (eds) (1992). Choice Over Time. New York: Russell Sage Foundation.
Marlatt, G.A. and Gordon, J.R. (eds). (1985). Relapse Prevention: Maintenance Strategies in the
Treatment of Addictive Behaviors. New York: Guilford Press.
Martela, F. (2013). Willpower: The Owner’s Manual—​12 Tools for Doing the Right Thing. Helsinki
Finland: Filosofian Akatemia.
McGonigal, K. (2013). The Willpower Instinct. New York: Avery.
Meichenbaum, D.H. (1977). Cognitive-​behavioral Modification. New York: Plenum.
Miller, W.R. and Rollnick, S. (1991). Motivational Interviewing: Preparing People to Change Addictive
Behavior. New York: Guilford.
Miller, W.R. and Rollnick, S. (2002). Motivational Interviewing: Preparing People for Change (2nd ed.).
New York: Guilford.
Miller, W.R, and Rollnick, S. (2013). Motivational Interviewing: Helping People Change (3rd ed.).
New York: Guilford Press.
Moss, A.C. and Albery, I.P. (2009). A dual-​process model of the alcohol-​behavior link for social
drinking. Psychological Bulletin, 135, 516–​30.
Muraven, M. (2010). Practicing self-​control lowers the risk of smoking lapse. Psychology of Addictive
Behaviors, 24, 446–​52.
Muraven, M. and Baumeister, R.F. (2000). Self-​regulation and depletion of limited resources: does self-​
control resemble a muscle? Psychological Bulletin, 126, 247–​59.
Muraven, M., Baumeister, R.F., and Tice, D.M. (1999). Longitudinal improvement of self-​regulation
through practice: building self-​control strength through repeated exercise. Journal of Social
Psychology, 139, 146–​57.
Nutt, D., Lingford-​Hughes, A., Erritzoe, D., and Stokes, P.R.A. (2015). The dopamine theory of
addiction: 40 years of highs and lows. Nature Reviews—​Neuroscience, 16, 305–​12.
Nutt, D., Robbins, T.W., Stimson, G.V., Ince, M., and Jackson, A. (eds) (2007). Drugs and the
Future: Brain Science, Addiction and Society. London: Elsevier.
Orford, J. (2013). Power, Powerlessness and Addiction. Cambridge, UK: Cambridge University Press.
Pickard, H. (2012). Responsibility without blame. In: B. Fulford, R. Gipps, and J. Sadler (eds), Oxford
Handbook of Philosophy of Psychiatry. Oxford: Oxford University Press.
Pickard, H. (in press). Addiction. In: N. Levy, M. Griffith, and K. Timpe (eds), Routledge Handbook of
Free Will. London: Routledge.
Redish, A.D. (2010). Addiction as a breakdown in the machinery of decision making. In: D. Ross, H.
Kincaid, D. Spurrett, and P. Collins (eds), What is Addiction? Cambridge, MA: MIT Press, pp. 99–​130.
Redish, A.D., Jensen, S., and Johnson, A. (2008). A unified framework for addiction: vulnerabilities in
the decision process. Behavioral and Brain Sciences, 31, 415–​87.
Robins, L.N., Helzer, J.E., and Davis, D.H. (1975). Narcotic use in Southeast Asia and afterward: an
interview study of 898 Vietnam returnees. Archives of General Psychiatry, 32, 955–​61.
Robinson, T.E. and Berridge, K.C. (2003). Addiction. Annual Review of Psychology, 54, 25–​53.
Room, R. (2005). Stigma, social inequality and alcohol and drug use. Drug and Alcohol Review, 24,
143–​55.
482  

482 Overview of addiction as a disorder of choice and future prospects

Ryan, F. (2013a). Cognitive Therapy for Addiction: Motivation and Change. Chichester,

UK: Wiley-​Blackwell.
Ryan, F. (2013b). Willpower for Dummies. Chichester: John Wiley.
Skog, O.-​J. (2003). Addiction: definitions and mechanisms. In: R.E. Vuchinich and N. Heather (eds),
Choice, Behavioural Economics and Addiction. Kidlington: Elsevier, pp. 157–​75.
Stanovich, K.E. (2004). The Robot’s Rebellion: Finding Meaning in the Age of Darwin.
Chicago: University of Chicago Press.
Stephens, G.L. and Graham, G. (2009). An addictive lesson: a case study in psychiatry as
cognitive neuroscience. In: M.R. Broome and L. Bortolotti (eds), Psychiatry as Cognitive
Neuroscience: Philosophical Perspectives. Oxford: Oxford University Press, pp. 203–​20.
Storbjoerk, J. and Room, R. (2008). The two worlds of alcohol problems: who is in treatment and who
is not? Addiction Research and Theory, 16, 67–​84.
Strack, F. and Deutsch, R. (2004). Reflective and impulsive determinants of social behavior Personality
and Social Psychology Bulletin, 8, 220–​47.
Thoresen, C.E. and Mahoney, M.J. (1974). Behavioral Self-​Control. New York: Holt, Rinehart and
Winston.
Vandermeeren, R. and Hebbrecht, M. (2012). The dual process model of addiction: towards an
integrated model? Tidshrift voor Psychiatrie, 54, 731–​40.
Volkow, N., and Koob, G. (2015). Brain disease model of addiction: why is it so controversial? Lancet
Psychiatry, 2, 677–​9.
Volkow, N. and Li, T.-​K. (2005). The neuroscience of addiction. Nature Neuroscience, 8, 1429–​30.
Watson, G. (1982). Free Will. New York: Oxford University Press.
Wiers, R.W. and Stacy, A.W. (eds) (2006). Handbook of Implicit Cognition and Addiction. Thousand
Oaks, CA: Sage.
Wiers, R.W., Field, M., and Stacy, A.W. (2014). Passion’s slave? Conscious and unconscious processes
in alcohol and drug abuse. In: K.J. Sher (ed.), Oxford Handbook of Substance Use Disorders.
Oxford: Oxford University Press; doi: 10.1093/oxfordhb/9780199381678.013.009.
 Index

abstinence violation effect  294 addiction as non-​akratic weakness of will  121–​2

acetylcholine  179–​80 addiction as social choice  245–​7
addiction and harm  33–​34 developing self-​control  250–​2
addiction as a brain disorder  77–​9 discount function  252–​4
addiction as a disease extended self  247–​8
distinguishing between voluntary and involuntary figure and ground  254–​6
behavior 387 social interaction as substitute and
expert opinion  386 displacement  256–​8
popular opinion  386–​7 teleological behaviorism  248–​50
addiction as a dual process  116–​17, 130–​1 addiction
compulsion  124–​5 beliefs about  292–​3
attention, bias, and will  125–​9 components 229
freedom of will  117–​18 counterfeit reward  232–​3
irresistible desires and rational capacity  118–​19 extra reward  229
weakness of will  119–​24 habit  231–​2
illness and choice  129–​30 hot thinking  230–​1
addiction as akrasia  133, 147–​8 hyperbolic delay discounting  233
links with current theory and research  145 weakness of will  230
behavioral economic theories  145 withdrawal avoidance  229–​30
dual-​process theories  145–​6 consequences of  29–​30
ego depletion  146 continuation 314
neuroscience theories  146–​7 controlled processes  290–​1
Mele’s definition of akrasia  135–​6 decision to become addicted  311–​12
ordinary akrasia  133–​4 definition  308–​9
philosophical accounts of akrasia  134–​5 emotional choosing  313–​14
requirements 137 endogenous reward is basis of addictive
against one’s better judgment  137–​8 habits  238–​40
consciously recognized as against one’s better extent of  167–​8
judgment  138–​41 genuine addictions  49–​50
freely chosen  141–​5 definition  50–​1
intentional action  137 identifying specific types  315–​16
three differences between ordinary akrasia and initiation 32
addiction  136–​7 intertemporal bargaining  240–​1
addiction as akratic weakness of will  122–​4 misery of  409–​11
addiction as blindfold  307–​8 nature of  79–​80
addiction as disorder of choice  463–​4 rationality  167–​8
developments in theory and research  465 trying substances  310–​11
addiction as dynamic inconsistency  465–​6 addiction potential  326
compulsion  466–​8 addiction Stroop test  271, 272
temptation and self-​control  466 addictive drugs  51–​4
towards a dual-​process theory of addiction  468–​9 addictive personality type  310
implications for prevention of addiction  474 addict’s power of choice  365–​7
socioeconomic deprivation and addiction  474–​5 nature of impairment
implications for public understanding of disease theory  367–​70
addiction  475–​6 incentive sensitization  370–​1
risks of being misunderstood  477–​8 mental obsession concept  368–​70
stigma  476–​7 physical allergy concept  367–​8
implications for treatment and brief stress  372–​3
intervention 470 wantings, willpower and denial  371–​2
enhancing motivation of improving twelve-​step programs
self-​control  470–​1 efficacy  373–​4
treatments for addiction as brain disease  472–​4 rationale  377–​81
willpower training  471–​2 spiritual solution to spiritual
willing and unwilling addicts  464–​5 malady  374–​7
484 Index

addicts akratic weakness of will  122–​4

attitude to addiction  32 alcohol
bigger picture  421–​3 demand and expenditure  328, 329
crime and addiction  413–​15 metabolism 267
difficult backgrounds  415–​16 reduction in self-​awareness  288
emotional impact of early damage  412–​13 remission rates  395–​6
‘insanity’ 39 Alcohol Attention Control Training Program
knowledge of negative consequences of drugs  36–​9 (AACTP) 275
lack of social change  417–​20 alcohol attentional bias  270
‘maturing out’  34 alcohol consumption as a goal  265–​6
misery of addiction  409–​11 cognitive bias modification (CBM)  275–​6
recovery movement  421 cognitive biases  271–​2
stigma precludes choice to recover  416–​17 approach-​avoidance bias  273–​4
vicious cycle of suffering  411–​12 interpretation bias  273
willing addicts value bias  273
alcoholics 74 cognitive mediators  270
borderline alcoholism  72–​3 current factors  268
drinkers before alcoholism  71–​2 incentive comparison effects  268–​70
drugstore cowboy  70–​1 distal determinants  267–​8
Keith Richards  69–​70 enhancing self-​control  274
morphine Mary  68–​9 importance of substance-​related cognitive
willing, unwilling, and resigned  66–​8 biases  270–​1
addicts teaching about addiction  385–​6 past drinking experiences  267–​8
interventions based on voluntary drug use in alcohol flush reaction  267
addiction 403 alcohol prohibition  400
issues  404–​5 alcohol use disorder  51
awareness of current scientific research  405 DSM-​5 diagnostic criteria  9–​10
health insurance plans  405 Alcoholics Anonymous (AA)  82, 83–​4, 256, 286,
summary  403–​4 365, 366
Ainslie, George  178–​9 defects of the wakeful state
akrasia  12–​14, 16, 82, 133–​4 emotional difficulties  92–​3
alcoholism  87–​8 lack of self-​knowledge  92
clear-​eyed akrasia  138 mental state  89–​91
definition 133 reasons for relapse  91–​2
definition of addiction  19–​20 definition of alcoholism  85
in addiction  83 bodily abnormality  85
Mele’s definition  135–​6 mental abnormality  86
philosophical accounts  134–​5 spiritual malady  86–​8
weakness of will  120 functions  96–​7
akrasia and addiction  133, 147–​8 questions about definition  88–​9
features 137 Steps 374
against one’s better judgment  137–​8 alcoholism  82–​4, 95–​7
consciously recognized as against one’s better akrasia  87–​8
judgment  138–​41 as a brain disorder  78
freely chosen  141–​5 chronic  93–​5
intentional action  137 defects of the wakeful state
links with current theory and research  145 additional defect  93–​5
behavioral economic theories  145 emotional difficulties  92–​3
dual-​process theories  145–​6 lack of self-​knowledge  92
ego depletion  146 mental state  89–​91
neuroscience theories  146–​7 reasons for relapse  91–​2
Mele’s definition of akrasia  135–​6 definition according to AA  85
ordinary akrasia  133–​4 bodily abnormality  85
philosophical accounts of akrasia  134–​5 mental abnormality  86
three differences between ordinary akrasia and spiritual malady  86–​8
addiction  136–​7 initial approach  84–​5
akratic action  135 questions about AA account  88–​9
against one’s better judgment  137–​8 Alcohol-​Savings Discretionary Expenditure Index
as intentional action  137 (ASDE)  330–​1
consciously recognized as against one’s better Alexander, Bruce  30–​1
judgment  138–​41 ambiguity in terms  449–​50
freely chosen  141–​5 responsibility in addiction  454–​5
 Index 485

amphetamine 198 brain disease model of addiction (BDMA) 

amygdala 180 459–​60, 472–​4
anabolic steroids, addictive potential  4 brief motivational interview (BMI)  333
anticipation of reward  58 outcomes  338–​40
appetite motivational system  208–​9
approach-​avoidance bias  273–​4 caffeine, addictive potential  4
approval 5 carrots, addictive potential  4
arousal of appetite  228, 230–​1 causal knowledge  36–​7
artists and drug use  75–​6 causal theory of excuse  432
assessing drug choice  186–​7, 198–​9 childhood experiences  313–​14, 411–​12
drug-​cue reactivity tasks  189–​92 abuse 415
study limitations  192 negative experiences  413–​14
future directions  196–​8 social and emotional impact of early
self-​administration paradigms  187–​9 damage  412–​13
study limitations  189 chocolate, addictive potential  4
simulated drug choice  192–​5 choice architecture  346–​7
self-​awareness  195–​6 choice biases  346–​7, 351–​2
attention bias  190–​1 choice in addiction  171–​4, 182–​3
attention deficit disorder (ADD)  178 brain changes over time  179–​82
attentional bias  271–​2 craving 176
in addiction  125–​6 cues, conditioning, and dopamine  174
automatic behaviors  264–​5 choice model of addictive behavior  116–​17, 130,
automatic drug seeking  212–​13 152–​6, 168
addiction as malfunctioning of wanting
behavioral addictions  4–​6 system  163–​4
behavioral choice theories of addiction  145 extent and rationality of addiction  167–​8
behavioral economic theories of addiction  145 modeling wanting system  162–​3
behavioral economics and choice  346–​7 pure choice models  156–​7
future directions for drug-​control strategies  357–​9 self-​control  164–​6
improving intervention strategies  353–​7 role of belief  166–​7
established strategies  353–​5 role of desires  166
promoting health decisions and outcomes  356–​7 strengths 166
remediating choice biases  355–​6 wanting and liking, distinguishing between  160
behavioral economics for brief motivational causal relationship  160–​1
interventions  325–​6 relation of wanting to expected liking  161–​2
alcohol and drug demand  329 choice to use drugs, explaining  31–​3
Alcohol-​Savings Discretionary Expenditure Index addiction and harm  33–​34
(ASDE)  330–​1 addictiveness  51–​4
assessment strategy  327 humans compared with rats  43–​4
clinical utility of assessment indices  331 ignorance of negative consequences in rats  39–​43
future directions knowledge of negative consequences  36–​9
assessment approaches  340–​1 psycho-​socio-​economic status  34
intervention approaches  341–​2 temporal discounting  34–​5
outcomes  338–​40 cigarette smoking  see smoking
proportionate substance-​related classification of addictions  6
reinforcement  329–​30 clear-​eyed akrasia  138
reinforcement pathology model of clinical addict  15, 16
addiction  326–​7 cocaine
Substance-​Free Activity Session (SFAS)  331–​2 addictive potential  4
delayed reward discounting (DRD)  332–​3 remission rates  394–​5
substance-​free reinforcement  332 selection by rats  40–​2
Substance-​Free Activity Session (SFAS) treatment cognitive behavioral change  320–​1
elements cognitive behavioral therapy (CBT)  217
feedback elements  334–​8 cognitive bias modification (CBM)  264, 275–​6
goal setting  338 cognitive biases  271–​2
initial investigations  333–​4 approach-​avoidance bias  273–​4
belief, role of  166–​7 countering 320
benzodiazepines, addictive potential  4 interpretation bias  273
binge flying  5 value bias  273
binge-​eating disorder (BED)  50, 55–​6 cognitive craving model  145
Borland, Ron  18–​19 cognitive desires  112
brain changes over time  179–​82 cognitive incentives  112
486 Index

cognitive mediators of alcohol consumption  270 dependence

cold turkey  30 definition 4
comfort food  314 physical 309
compulsion and addiction  116–​17, 124–​5, 145, Descartes, René  247
152–​6, 167–​8 desire 12
attention, bias, and will  125–​6 desire system  99, 103
malfunctioning of the will  127–​9 desires
freedom of will  117–​18 irresistible  118–​19
irresistible desires and rational self-​control  166
capacity  118–​19 desires, normal and addictive  99–​101, 113–​14
weakness of will  119–​24 behavioral control systems
conditioned craving  237–​8 conflict between systems  106–​7
consonant addict  15 habit system  101–​2
Contingency Management (CM)  118–​19, 217, 331 planning system  105–​6
controlled drug seeking  212–​13 standing and occurrent desires  103–​4
controlled processes  286–​7, 290–​1 three systems  101
corticotropin-​releasing factor (CRF)  372 theories of drug addiction
counterfeit reward  232–​3 comparison with existing theories  110–​11
craving  12, 113–​14, 164, 175–​6, 176–​9 dopamine  107–​8
dopamine 100 incentive salience theory  111–​13
sudden craving as positive feedback reward learning theory  113
phenomenon  237–​8 standing desire theory  109–​10
crime and addiction  413–​15 determinism 431
criminal law and addiction  426–​7, 443 detour behavior  265
background  432–​6 Diagnostic and Statistical Manual of Mental
concept of person and responsibility  428–​31 Disorders (DSM)  7–​10
current situation  436–​9 alcohol use disorder diagnostic criteria  9–​10
criticism of  442–​3 dielactical behavioral therapy (DBT)  321
defence of  440–​2 discount function  252–​4
legal practices  439–​40 discount rate  121, 351
potential hurdles  431–​2 disease models of addiction  157
preliminary assumptions about addiction  427–​8 habit accounts  157–​8
cue-​triggered wanting  174 learning accounts  159–​60
culpability in crime  437 pleasure accounts  158–​9
cultural aspects  see social and cultural permissions dispositional desires  164
for drug use distress  50–​1
current concerns  260 domain effect bias  352
nonconscious nature  261–​2 dopamine hijacking  20, 30, 32, 52, 174, 179
addictive substances  100
decision-​making modes  127–​9 craving 100
definitions of addiction  3–​4 food addictions  54–​5
as akrasia  19–​21 gambling  57–​8
as excessive appetite  17–​18 incentive salience theory  111, 167
as hard-​to-​maintain (HTM) behavior change  18–​19 levels of dopamine  163–​4
behavioral addictions  4–​6 neuroscience theories of addiction  146
common usage  6–​7 reward learning theory  113
DSM and ICD  7–​10 role of dopamine  162
etymology 7 theory of addiction  107–​8
liberal account  18 drug-​cue reactivity tasks  189–​92
proposed definition  10–​11 study limitations  192
as a relapsing condition  11–​12 drug demand  329
counterfactual objection  14–​16 reduction strategies  354
denial and self-​deception  16–​17 drug Stroop task  190
theoretical definition  12–​14 drug supply reduction strategies  354
type of definition needed  6 dual-​process theories of addiction  145–​6
delay discounting  177–​8, 245–​6, 253, 312, 350 dual-​task paradigm  289
delay duration bias  352 dynamic inconsistency effect  352
delay of gratification  143
delayed reward discounting (DRD)  332–​3 economic status and addiction  415
denial in addiction  16–​17, 38 economics demand for drugs  326
addict’s power of choice  371–​2 education about addiction  420–​1
 Index 487

ego depletion  146, 289, 296 hard-​to-​sustain (HTS) behaviors  19

Elliot, Carl  30 harm in addiction  11, 12
endogenous reward  238–​40 harm reduction strategies  354
event-​related potentials (ERPs)  196 Harry Potter books  5
excessive appetite definition of addiction  17–​18 health insurance and addiction  405
executed movement  273–​4 hedonic impact of activities  176
executive control  125 hedonic importance  239
Executive System (ES)  19 hippocampus 180
exercise addiction  5, 51 Holton, Richard  138–​9
expectancy × value theory  262 hospices, morphine use in  68–​9
alcohol consumption  266 hot thinking  228, 230–​1
expected utility theory (EUT)  228, 241 hyperbolic delay discounting  233
extended self  247–​8 features  233–​5
extent and rationality of addiction  167–​8 recursive self-​prediction  235
eye-​tracking  272 endogenous reward is basis of addictive
habits  238–​40
failures of rationality and self-​knowledge  82–​4, 95–​7 intertemporal bargaining creates willpower and
alcoholism according to AA  85 also failures  236–​7
bodily abnormality  85 sudden craving as positive feedback
mental abnormality  86 phenomenon  237–​8
spiritual malady  86–​8 hyperbolic discounting  121–​2, 227–​9
defects of the wakeful state hypothalamus 180
additional defect  93–​5
emotional difficulties  92–​3 Impulse Control Disorders Not Elsewhere
lack of self-​knowledge  92 Specified 4
mental state  89–​91 incentive comparison  268–​70
reasons for relapse  91–​2 incentive disengagement cycle  263
initial approach  84–​5 incentive sensitization  58, 111–​13, 174
questions about AA account  88–​9 addict’s power of choice  370–​1
‘fall off the wagon’  294, 298–​9, 318 incentives 262
false dichotomies in describing addiction  449–​50 indicated prevention of substance use
compulsion or choice  452–​4 disorders 349
disease or choice  451–​2 indoor tanning  5
disease or disorder  455–​9 inhalants, addictive potential  4
hijacking or seduction  460–​1 ‘insanity’ of addicts  39
psychiatric disorder or brain disease  459–​60 insufficiency 123
fear in alcoholism  87 insula, role in drug seeking and choice  205–​6
flicker change blindness paradigm  272 appetite motivational system  208–​9
flow, state of  57 automatic versus goal-​directed drug
folk addictions  50 seeking  212–​13
food addictions  5, 54–​6 clinical implications  217
free will  431–​2 decision system in addiction  211–​12
freedom of will  117–​18 effects of insula damage on cigarette
smoking  206–​8
gambling, pathological  4, 56–​8 integrative model  213–​16
Gestalt psychologists  254 interoceptive representation  209–​11
goal theory for current concerns  259–​60 location and anatomical subdivisions  210
determinants of commitments  262–​3 understanding addiction as choice  217–​18
nature of current concerns  261–​2 International Classification of Diseases (ICD)  7–​10
nature of goals  260–​1 Internet gaming disorder  4, 50, 58–​9
goal-​directed actions  125, 212–​13 interpretation bias  273
grade-​point average (GPA)  335 intertemporal bargaining  236, 240–​1
gratification, delayed  143 intertemporal choice  172
grey matter volume  181 involuntary actions  387
irresistible desires  118–​19
habit system  99, 101–​2, 157–​8, 231–​2
habitual actions  212–​13, 315–​16 James, William  152, 228
habituation, definition  4 joy-​riding  5
happy addict  15
hard-​to-​maintain (HTM) behavior change  19 Kant, Immanuel  254
hard-​to-​reduce (HTR) behaviors  19 kleptomania 5
488 Index
larger, later (LL) rewards  233–​4, 236, 245–​7
late positive potential (LPP)  196
learning models of addiction  159–​60
Levels of Emotional Awareness (LEAS) tasks  196
liberal account of addiction  18
life circumstances and addiction  32
liking  122–​4, 128, 160, 232
causal relationship with wanting  160–​1
relation of wanting to expected liking  161–​2
loss of control  117, 128
love addiction  5
magnitude effect bias  351
malfunctioning of the will  127–​9, 130
marijuana/​cannabis
addictive potential  4
remission rates  394–​5
Maté, Gabor  20
‘maturing out’ of addicts  34
medical model of addictive behavior  116–​17, 130
Mele, Alfred R.  135, 142–​4
definition of akrasia  135–​6
mental health difficulties and addiction  ignorance of in rats  39–​43
313–​14, 321–​2
mental health problems and addiction  32
mesolimbic dopamine  154, 158
midbrain dopamine neurons  108
Miller, Jonanthan Wolfe, Sir  249, 251
Mischel, Walter  143
misery of addiction  409–​11
money as a token reinforce  57
moral issues  254
morphine use in hospices  68–​9
motivated behaviors  264–​5
motivational interviewing (MI)  293, 331
motivations to quit addictions  291–​2, 294, 315
muscle dysmorphia  5
naïve addict  14–​15
Narcotics Anonymous  377–​8
natural history approach to addiction  385–​6
addict identification criteria  389–​90
chronic disorder  390
remission according to drug type  391
remission in general population  390
statutory remission rates  390–​1
characteristics of voluntary behavior  387–​8
choice not necessarily rational  388
drug-​induced brain changes  389
not free will  388
voluntary/​involuntary continuum  388
experiments 399
alcohol prohibition  400
lessons learnt  400–​1
opiates  399–​400
smoking report  400
interventions based on voluntary drug use in
addiction 403
issues  404–​5
awareness of current scientific research  405
health insurance plans  405
methodological problems in remission studies
epidemiological surveys  397
exponential cumulative remission
functions  393–​4
individual remission timing  397–​8
linearized remission rates for cocaine and
marijuana  394–​5
linearized remission rates for smoking and
alcohol  395–​6
missing addicts improving rates  396–​7
stability of remission rates  392–​3
reason for remission from addiction  398
logical implication of non-​treatment high
remission rates  398–​9
summary  403–​4
years of school and heavy drug use  401–​2
negative consequence of drugs, knowledge of  29–​31
addiction  36–​9
explaining choice to use drugs  31–​3
addiction and harm  33–​34
psycho-​socio-​economic status  34
temporal discounting  34–​5
humans compared with rats  43–​4
nested systems model  106–​7
neuroscience theories of addiction  146–​7, 174–​5
nicotine
addictive potential  4
bolus hypothesis  53
non-​addictive drug use  31
non-​akratic weakness of will  121–​2
nonconscious motivational influences on cognitive
processes 276
cognitive bias modification (CBM)  275–​6
cognitive biases  271–​2
approach-​avoidance bias  273–​4
interpretation bias  273
value bias  273
determinants of commitments  262–​3
disengagement from failed goals  263–​4
addiction as motivated or automatic  264–​5
enhancing self-​control  274
goal of drinking alcohol  265–​6
cognitive mediators  270
current factors  268–​70
distal determinants  267–​8
importance of substance-​related cognitive
biases  270–​1
past drinking experiences  267–​8
motivational goal theory for current
concerns  259–​60
nature of current concerns  261–​2
nature of goals  260–​1
normative biases in decision-​making  350–​3
occurrent desires  100, 103–​4
initiated by standing desires  109
operant behavior  49–​50, 141
any reward can lead to addiction  60–​1
food addiction  54–​6
gambling  56–​8
pleasures and perils  61–​2
videogames  58–​60
 Index 489

operant conditioning  53–​4 Relative Addiction model  326–​7

Operational System (OS)  19 relative reinforcing efficacy  327
Opiates remission
criminalization  399–​400 methodological problems in studies
model of addiction  4 epidemiological surveys  397
Orford, Jim  17 exponential cumulative remission functions  393–​4
outcome devaluation  103 individual remission timing  397–​8
OxyContin 67 linearized remission rates for cocaine and
marijuana  394–​5
parallel systems model  106 linearized remission rates for smoking and
Parfit, Derek  253–​4 alcohol  395–​6
partner addiction  180 missing addicts improving rates  396–​7
Pascual-​Leone, Alvaro  181 stability of remission rates  392–​3
pathological gambling  4, 50 reason for remission from addiction  398
Pavlovian conditioning  174 logical implication of non-​treatment high
Pavlovian values  113 remission rates  398–​9
personal rules  236 resigned addicts  15–​16, 67–​9
phasic dopamine  108, 113 revealed preferences  253
phenomenology 254 reward
physical dependency  309 counterfeit  232–​3
planning system  99, 105–​6, 110 nature of  238–​9
plasticity of the brain  181 reward as driver of addiction
pleasure models of addiction  158–​9 any reward  60–​1
population perspective on substance misuse  348 food addictions  54–​6
positive feedback  237–​8 gambling  56–​8
prefrontal cortex (PFC)  177–​8 videogames  59–​60
premenstrual syndrome (PMS)  289 reward learning theory  111, 113
prevention paradox  349 reward prediction error signal  108
probe-​detection paradigm  272 reward value  327
prodynorphin 198 Rush, Benjamin  152
psycho-​education  317–​19
psycho-​socio-​economic status and addiction  32, 34 satiety system  61
public understanding of addiction  420–​1, 475–​6 school years and heavy drug use  401–​2
risks of being misunderstood  477–​8 selective prevention of substance use disorders  348–​9
risks of being misunderstood worthwhile  478 self-​absorption  375
stigma  476–​7 self-​administration paradigms of drug choice  187–​9
study limitations  189
quitting, motivations for  291–​2, 294, 315 self-​awareness  195–​6
reduction by alcohol  288
Rainbow Group  76–​7 self-​centred fear  87
rational capacity  118–​19 self-​centredness  94–​5, 375, 376–​7
rational commitment to action  138 self-​control  164–​6, 296
rationality of addiction  167–​8 developing  250–​2
rationality of choice  388 enhancing 274
rats failure of  14
addiction, nature of  43–​4 role of belief  166–​7
ignorance of negative consequences of role of desires  166
drugs  39–​43 strengths 166
studies of addiction  53 self-​deception in addiction  16–​17
recursive self-​prediction  235 self-​efficacy  295–​6
endogenous reward is basis of addictive self-​esteem, building  319–​20
habits  238–​40 selfishness 375
intertemporal bargaining creates willpower and self-​medication hypothesis  34
also failures  236–​7 reasons for  313
sudden craving as positive feedback self-​monitoring  251
phenomenon  237–​8 self-​prediction  235
reductionism 253 endogenous reward is basis of addictive
reinforcement pathology model of addiction  326–​7 habits  238–​40
reinforcing efficacy (RE)  326, 327 intertemporal bargaining creates willpower and
relapse in addiction  11–​12 also failures  236–​7
self-​regulation  297–​8 sudden craving as positive feedback
avoiding temptation  298–​9 phenomenon  237–​8
490 Index

self-​regulation  286–​7 subjective utility theory  262

fighting against addiction  293 Suboxone 67
motivations for quitting  294 substance dependence  50
self-​efficacy  295–​6 substance misuse, identifying specific types  315–​16
standards  293–​4 Substance-​Free Activity Session (SFAS)  326, 331–​2,
willpower  296–​7 331–​2
fighting against relapse  297–​8 delayed reward discounting (DRD)  332–​3
avoiding temptation  298–​9 future directions  341–​2
theory  287–​90 outcomes  338–​40
self-​seeking  375 substance-​free reinforcement  332
self-​will  375–​6 Substance-​Free Activity Session (SFAS) treatment
sensitization 367 elements
sequence effect bias  352 feedback elements  334
sexual addictions  5 coping with stress and depression  337
shock  90–​1 graduating from college  334–​5
shopping addiction  5 intended career requirements  335
sign effect bias  351 personalized career-​related activities  335
Simon, Julian  253 substance-​free recreational activities  337–​8
simulated drug choice  192–​5 time allocation  335–​6
self-​awareness  195–​6 weighing immediate and future benefits  336–​7
situational altruism  256 goal setting  338
Skog, Ole-​Jørgen  14, 16, 21 initial investigations  333–​4
sleep 90 substance-​related conitive biases  270–​1
smaller, sooner (SS) rewards  233–​4, 236, 240, 245–​7 sugar, addictive potential  4
smartphone addiction  5 suicide option  35
smoking 400 systematic motivational counselling  274, 275–​6
effects of insula damage  206–​8
remission rates  395–​6 teleological behaviorism  248–​50, 255
social and cultural permissions for drug use  75 temporal discounting  34–​5
artistic permission  75–​6 temptation, avoiding  298–​9
Rainbow Group  76–​7 theoretical definition of addiction  6, 12–​14
social discount function  252, 253 tolerance to drugs  32, 51
social interaction as substitute and displacement for treatment for addiction  286–​7, 299, 322
addiction  256–​8 behavioral economics  325–​6
social status as reward  60 alcohol and drug demand  329
social support  257 Alcohol-​Savings Discretionary Expenditure
socioeconomic deprivation and addiction  32, 34, Index (ASDE)  330–​1
415–​16, 474–​5 assessment strategy  327
spiritual solutions to addiction  374–​7 clinical utility of assessment indices  331
spirituality 377 future directions  340–​2
standing desire theory  99–​100, 103–​4, 109–​10 outcomes  338–​40
stigmatic structure enslaves addicts proportionate substance-​related
better solutions  420–​1 reinforcement  329–​30
bigger picture  421–​3 reinforcement pathology model of
crime and addiction  413–​15 addiction  326–​7
difficult backgrounds  415–​16 Substance-​Free Activity Session (SFAS)  331–​3
lack of change  417–​20 Substance-​Free Activity Session (SFAS)
misery of addiction  409–​11 treatment elements  333–​8
public understanding  476–​7 behavioral economics and choice  346–​7
recovery movement  421 cognitive behavioral change  320–​1
stigma precludes choice to recover  416–​17 fight against addiction  291
vicious cycle of suffering  411–​12 beliefs about addiction  292–​3
street life/​gang activity  5 motivations to quit  291–​2, 294
stress  372–​3 future directions for drug-​control strategies  357–​9
stress management  320–​1 improving intervention strategies  353–​7
striatum  177–​8 established strategies  353–​5
Stroop interference  126 promoting health decisions and
Stroop tasks  190 outcomes  356–​7
subclinical addict  15 remediating choice biases  355–​6
subiculum 411 providing 317
subjective rationality  138 public health foundations  347–​50
 Index 491

self-​esteem, building  319–​20 causal relationship with liking  160–​1

self-​regulation  293 incentive salience theory  112
motivations for quitting  294 modeling  162–​3
self-​efficacy  295–​6 relation of wanting to expected liking  161–​2
standards  293–​4 willpower and denial  371–​2
willpower  296–​7 water, addictive potential  4
self-​regulation fighting against relapse  297–​8 weakness of the will and addictions  116–​17, 119–​24
avoiding temptation  298–​9 compulsion  124–​5
self-​regulation theory  287–​90 compulsion: attention, bias, and will  125–​9
stress management  320–​1 freedom of will  117–​18
suggestions 299 irresistible desires and rational capacity  118–​19
efficacy  373–​4 weakness of will  230
type-​1 decision-​making processes  127–​9 Weber–​Fechner law  234
type-​2 decision-​making processes  127–​9 ‘what the hell’ effect  294
Wikler, Abraham  51–​2
unhappiness and addiction  313–​14 willing addicts  66–​7
universal prevention of substance use disorders  348 examples
unwilling addicts  66–​8 alcoholics 74
utility theory  178 borderline alcoholism  72–​3
drinkers before alcoholism  71–​2
value bias  273 drugstore cowboy  70–​1
value of activities  255 Keith Richards  69–​70
videogame addiction  58–​60 morphine Mary  68–​9
visceral biases, countering  320 willpower  289–​90, 296–​7
visual-​neurological reward  58 addict’s power of choice  371–​2
voluntary actions  212–​13, 387 training  471–​2
characteristics  387–​8 Wilson, Bill  82, 83
choice not necessarily rational  388 withdrawal  32, 51–​2
drug-​induced brain changes  389 dopamine levels  159
not free will  388 work addiction  5, 51
voluntary/​involuntary continuum  388 Wyvell, Cindy  161, 163

wanting  122–​4, 128, 160, 232 yoked control  53

addiction as malfunctioning  163–​4 Young Adult Alcohol Consequences Questionnaire
addict’s power of choice  370–​1 (YAACQ) 338