You are on page 1of 38

Schizophrenia and the Narrative of Enlightened Geneticization

Author(s): Adam Hedgecoe

Source: Social Studies of Science, Vol. 31, No. 6 (Dec., 2001), pp. 875-911
Published by: Sage Publications, Ltd.
Stable URL:
Accessed: 11/06/2010 08:30

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact

Sage Publications, Ltd. is collaborating with JSTOR to digitize, preserve and extend access to Social Studies of
ABSTRACT In this paper I explore the way in which scientists attempt to construct
schizophrenia as a genetic disease using various discursivestrategies. These strategies
involve: referring back to earlier twin and adoption studies on schizophrenia
inheritance;constructing genetic models to allow some r6le for single gene versions
of the condition; a cautious and responsible limit on the strength of current research;
and the construction of a range of bordering conditions which contribute to the
genetic riskof schizophrenia.These various rhetorics produce a 'narrative'about
schizophreniawhich subtly prioritizesgenetic explanations, while appearing to allow
a r6le for non-genetic factors; I term this the 'narrativeof enlightened
geneticization'. This research has implicationsfor scientific controversies,specifically
the way in which criticsattack schizophreniagenetics, since many of their points are
already incorporated into the narrative.It also contributes to the ongoing debate
over the r6le of social science research in the study of geneticization.

Keywords complex disease, controversy,discourse analysis,explanation, genomics

Schizophrenia and the Narrative of

Enlightened Geneticization
Adam Hedgecoe

Much work in S&TS revolves around the nature of scientific contro-

versies.1This study contributes to this literature by analysing how scientists
involved in controversial research defend themselves against opponents'
complaints, enabling them to continue their research while presenting a
sophisticated and inclusive appearance of their work to others. My argu-
ment in this paper is that the predominant narrative in schizophrenia
genetics is one of 'enlightened geneticization'. The central theme of this
narrative is the presentation of current genetic thinking as reasonable, non-
extremist, and accepting a role for non-genetic factors in schizophrenia
causation. There is far more to the narrative of enlightened geneticization
than simply calling schizophrenia a 'genetic disease', and proposing a 'gene
for' schizophrenia, with little role for environmental influence. My aim is
to demonstrate how the discourse surrounding the genetics of schizo-
phrenia is constructed to prioritize genetic explanations, and subtly to
undermine non-genetic factors, while at the same time accepting that they
have a role in its aetiology.
With the completion of the analysis of the human genome, commenta-
tors have raised hopes that the genetic basis to a number of complex
diseases will soon be revealed.2 However, human molecular genetics has

Social Studies of Science 31/6(December 2001) 875-911

C SSS and SAGE Publications (London, Thousand Oaks CA, New Delhi)
[0306-3127(200112)31:6;875-91 1;022049]
876 Social Studies of Science 31/6

been producing results for years now, so there is already a great deal of
genetic information available to medical researchers. How is this molecular
genetic information used to describe diseases? How do researchers con-
vince their peers that the most productive research into a specific condition
(and thus that which requires the most funding) explores genes rather than
environmental, social or even other biological causes? This paper examines
the rhetorical techniques which researchers use to encourage genetic
research into a condition (in this case schizophrenia) and to reshape it on
genetic grounds. There is a great deal of sociological interest in the new
genetics;3 much of this has focused on the attitudes of professionals
(scientists/geneticists) towards the new genetics, and this paper relates to
work focusing on how professional geneticists frame their research.4 For
example, as part of a larger study, Anne Kerr, Sarah Cunningham-Burley
and Amanda Amos interviewed a number of senior UK geneticists to
examine how they saw social and ethical issues, and their responsibilities.
They show how these scientists construct objective, disinterested positions
with regard to the political use of their research results, but at the same
time feed into the policy debates surrounding, for example, the use of
genetic information by the insurance industry.5 In a later paper, they also
show how these scientists draw distinct boundaries between their own
research and 'eugenics', by defining eugenics as bad science, and by
emphasizing both the need for personal choice and the complex relation-
ship between nature and nurture. It is this last point that forms the starting
position for my own research. As these authors note:
[R]hetoricof a balanced approachto nature and nurture disguisesthe
privilegingof nature over nurture which genetics research necessarily
entails.Nature,this time in the form of a combinationof genes leadingto
a genetic predisposition,is still viewed as a fundamentalinfluence on
behavior(or disease).6

My research shows how geneticists construct a narrative around schizo-

phrenia that subtly privileges genetic explanations without succumbing to
hard-line genetic determinism.
As well as work in the sociology of science, my research relates to
broader debates within the literature around the term 'geneticization', and
is part of a research project investigating that term.7 Coined by Abby
Lippman in the early 1990s,8 and since adopted by a number of other
authors,9 'geneticization' describes:

The evergrowingtendencyto distinguishpeople one fromanotheron the

basis of genetics;to define most disorders,behaviors,and physiological
variationsas whollyor in part genetic in origin.'?

For my own research, I have adopted a 'stripped down' definition which

states simply that in medicine, geneticization takes place when a condition
is linked to a specific stretch of DNA." There are various reasons for
choosing this narrow definition over broader ones such as Lippman's, or
that used by Sue Sherwin and Christy Simpson: 'Geneticization is the
Hedgecoe: Schizophrenia and the Narrative of Enlightened Geneticization 877

attitude that the differences among people can be reduced to differences in

their genetic makeup; it assumes that most disorders are largely attribut-
able to genetics'.12 An underlying assumption among many who use the
concept of geneticization is that such a process is negative and undesirable.
If 'geneticization is a process of colonization with genetic technologies', it
is very hard, in these post-colonial times, to say anything good about it.13
Thus any analysis of the ethics of geneticization is inherently circular. My
'stripped down' definition removes the concept of geneticization from its
current ethically loaded context and places it within the more neutral,
historical context of 'molecularization'. This means that the use of genetic
explanations is seen as simply the latest in a long line of attempts to analyse
the body in terms of molecules, rather than just an opportunistic tactic
employed by doctors to gain power over patients.14The danger with this
approach is that it prioritizes the role of molecular genetics over older
investigations of inheritance. The failure of geneticization to acknowledge
the hereditary nature of much family information has been noted by
Celeste Condit and Melanie Williams, who criticize the term geneticization
as 'an overly general description that obscures the fact that contemporary
genetics discourse is a permutation of long-standing hereditarian dis-
courses that reflect age-old recognitions by human beings that heredity
plays a role in the characteristics of living things'.15 While this is a fair
point, it does not undermine the fact that molecular genetics is qual-
itatively different from population genetics, both in terms of the methods
used and the results reached. With this in mind, my molecular definition of
geneticization also has the advantage of pinpointing a specific time period
as central to this process. In addition to these ethical and historical aspects
to geneticization, there has also been some methodological debate over
whether the term can be used for research in the social sciences or whether
it is best seen as a philosophical concept which should not be assessed in
empirical terms.'6 This paper aims to refute this position, and to show how
the concept of geneticization can contribute to empirical research in the
sociology of science.

Method and Materials

In their work on geneticists' discourse, Kerr, Cunningham-Burley and
Amos based their analysis on the techniques developed by Nigel Gilbert
and Michael Mulkay in their classic analysis of scientific discourse, Open-
ing Pandora's Box.17 In this research, a central theme is the comparison
between scientists' use of two separate discourses (or 'repertoires'). The
'empirical repertoire' is used in published articles (although it also occurs
in interviews and conversations), and follows certain genre rules concern-
ing disputes with other scientists and prioritization of one's own results.
The 'contingent repertoire' only occurs in informal settings (such as in an
interview), and allows one to provide social explanations for others' failings
and empirical support for one's own position. Interesting though geneti-
cists' verbal discourse is, it is not the side of their work that most people,
878 Social Studies of Science 31/6

even most scientists, see. Rather, it is in the published article that scientists
present their 'approved' views of, for example, the causes of schizophrenia.
Therefore my work develops Gilbert and Mulkay's approach by focusing
exclusively on the empiricist repertoire, and thus on published scientific
articles, specifically review articles.
There are two main reasons for using review articles in my research.
First, their length allows authors greater freedom and detail in description
and explanation; they are a 'rich' site for discourse analysis. Their length
also means that authors have more space to address controversies and
social and ethical issues, discussions of obvious interest to me. Second,
review articles have already proved a valuable genre for analysis, as shown
by the work of Greg Myers, and his use of the concept of 'narratives'.18A
central theme of Myers' research is that review articles provide a textual
space within which knowledge is constructed, allowing certain experi-
mental reports to be seen as key papers. Even the idea of a specific event as
the discovery of a particular fact depends upon review articles to organize
the claims and techniques in a particular direction.19 Jan Golinski terms
this type of research 'narratological hermeneutics', where attention is paid
to those elements that construct a story about a particular topic.20 Much of
Myers' work has focused on the alternative narratives employed by differ-
ent parties in a scientific controversy,21 but he has also shown how
narratives operate in the reporting of scientific discoveries.22 Since, in
narratives, the focus is on the parts that go towards making up the story,
rather than the motivations that lie behind them, it is easier to maintain a
neutral position with regard to the ethical rights or wrongs of a process like
geneticization. This is important, since one of the overall aims of my
research is to distance the concept of geneticization from the idea that it is
necessarily an unwelcome (or welcome) process. Focusing on the con-
struction of narratives helps achieve both sociological symmetry and an
element of ethical neutrality with respect to geneticization. Within the
narrative I present, I claim that various different strategies are used to
construct and persuade the reader, where a strategy is 'simply a theme or
theory enveloping and reflecting patterns in the formation of objects,
subjects and concepts within a discourse'.23 In methodological terms, my
work extends Myers' research into new areas, while staying close to his
constructivist position and empirical rigour.
The articles selected for analysis in this case study are all drawn from
the journal SchizophreniaBulletin, which is published by the US National
Institute of Mental Health. Its mission is to 'facilitate the dissemination
and exchange of information about schizophrenia',24 and is therefore
aimed at a wide audience of clinicians and researchers. While the journal
makes clear that the views it expresses are not those of the NIMH, this link
does imply that its articles are representative of particular positions within
the research community. In 1976, SchizophreniaBulletin published its first
wide-ranging review of schizophrenia genetics, written by two pioneers of
twin studies in schizophrenia, Irving Gottesman and James Shields.25The
three articles analysed in my work are all 'successors' to this one, in that
Hedgecoe: Schizophrenia and the Narrative of Enlightened Geneticization 879

they are all review articles on schizophrenia genetics from the same
journal, are heavily cited in their own right,26 focus on the molecular
aspects of schizophrenia genetics, and represent the dominant themes in
schizophrenia genetics as a whole. They cover the period from 1989 to
1997; there are more recent review articles available (although not in
Schizophrenia Bulletin) in the literature, but the ones chosen for this
research cover an important time-span in terms of genetics, from pre-
genomic research to a stage when molecular genetic investigation of
complex disease was more common-place. The three articles I selected

* McGue & Gottesman (1989): 'Genetic Linkage in Schizophrenia: Per-

spectives from Genetic Epidemiology';27
* Kendler & Diehl (1993): 'The Genetics of Schizophrenia: A Current
Genetic-Epidemiologic Perspective';28
* Moldin & Gottesman (1997): 'Genes, Experience and Chance in Schizo-
phrenia-Positioning for the 21st Century'.29

What these reviews construct is the narrative that occurs in a disease where
genetic explanations are proposed, but where critics fiercely resist the use
of such explanations.30

Schizophrenia, Genetics and Controversy

This paper explores the way in which geneticization is taking place in the
professional discourse surrounding schizophrenia. There are two reasons
for choosing this condition for my research. The first is that schizophrenia
presents a challenge to those scientists who feel that genetic factors
underpin most aspects of human health and illness, and that the future of
medicine lies in extensive searches for the genes that predispose us to
different conditions. The complex nature of schizophrenia (both as a
medical condition and in terms of its genetics) means that research into the
molecular genetics of schizophrenia is slow, and characterized by many
false hopes and unreplicated results.31There have been attempts to find
genes (or their markers) which might play a role in schizophrenia aetiology
but, so far, none have stood up to the extensive scrutiny that such claims
undergo. For example in 1988, Robin Sherrington and his colleagues
announced the discovery of a locus on chromosome 5 that they claimed
indicated susceptibility to schizophrenia.32The resulting high-profile media
discussion became embarrassing when other investigators failed to repli-
cate these results; even more so when the same research team also failed to
repeat their earlier findings. It is still not clear what went wrong, and how
to explain this incident [Kendler & Diehl (1993)]. However, research into
the molecular genetics of schizophrenia has not stopped. A recent study
has announced the discovery of two genes, labelled 'Disrupted-in-
Schizophrenia (DISC) 1 and 2', linked with schizophrenia in a specific
Scottish family.33Whether this result will stand the tests of time, analysis
880 Social Studies of Science 31/6

and repeatability that have undermined other putative 'schizophrenia

genes' remains to be seen.
The complex nature of schizophrenia relates to my second reason for
choosing it as a topic of research: schizophrenia genetics is controversial.34
In many ways, schizophrenia is a touchstone topic in the debate between
scientists who support a broadly genetic view of human disease and
behaviour, and those critics and activists who resist such genetic explana-
tions as 'reductionistic' and 'deterministic'. Schizophrenia has been on the
front-line between geneticists and critics for several decades, covering not
just modern molecular genetics, but also earlier debates over the use of
twin and adoption studies. Almost since its designation in the 19th
century, researchers have seen schizophrenia as a genetic disease. Mary
Boyle, who is extremely critical of genetic explanations of schizophrenia,
claims that...

before any empiricalliteratureexisted, the belief that 'schizophrenia'...

had a genetic basis was apparently strong.... [B]efore any attempt at
systematicdata collectionwas ever made, the two most prominentusers
of the concepts of dementiapraecoxand schizophreniawere disseminat-
ing the view that whateverphenomenathey includedunder these terms
were largelyinherited.35

This argument that, in schizophrenia, genetic causation is an axiom rather

than a hypothesis, implies that there is something rather circular about
investigating the possibility of a genetic cause for a disease which has to be
seen as genetic to exist as a disease entity at all. As Richard Marshall has

attemptsto establishschizophreniaas geneticallydeterminedhave been

pursuedwith greateror lesser convictionsince disorderedconduct was
first construed as some form of illness... this medicalizationprocess
requiredas a cornerstonethe resortto biologicalexplanation,and this, in
turn, could only gain scientificrespectabilityif the source of the disorder
could be locatedin genetic transmission.

Thus the critics' claim is that the concept of schizophrenia as a disease

entity and the idea that it is genetically caused are 'mutually reinforcing',
each reifying the other and helping to turn it into a valid object of
Some readers' knowledge of schizophrenia genetics may have been
informed by the controversies surrounding the use of population methods
(that is, familial, twin and adoption studies). Comparing identical (that is,
monozygotic) twins is one of the oldest methodologies in genetics research,
and Luxenburger performed the first twin studies for schizophrenia in
1928 and 1934 - not to determine whether there was a genetic component
to schizophrenia (which was the general assumption of that time), but to
discover whether or not it was transmitted in a Mendelian fashion.37 Franz
Kallman, Eliot Slater and James Shields carried out the largest of the early
twin studies in the 1940s and 1950s: these seemed to confirm the existence
Hedgecoe: Schizophrenia and the Narrative of Enlightened Geneticization 881

of a strong genetic component to schizophrenia, with pairs of monozygotic

twins (MZ) having a much higher concordance rate than non-identical,
dizygotic twins (DZ).38 As has become the trend in this area of research, a
series of articles were then published, criticizing the assumptions, method-
ologies and results of this early twin study work. Those who supported this
research tacitly accepted many of the criticisms, changed their research
methodologies and continued their research.
Criticism of twin studies focuses on a number of different aspects,
particularly those of methodology, underlying concepts, and the statistical
analyses used. The main debates focus on different ways of calculating
concordance rates and the use of the 'schizophrenia spectrum' (see below),
although other possible disputes include: sample selection (drawing
samples from hospitalized twins biases the concordance rate);39 zygosity
determination (accurate blood typing is not extended to whole samples to
confirm MZ/DZ status);40 and the diagnostic criteria used in different
studies to determine schizophrenia.
In the case of different concordance rates, the dispute revolves around
the fact that there are a number of different ways of calculating the
concordance from the same data - pair-wise, proband-wise and case-wise -
each of which produces a different figure for the concordance.41 While all
three methods are appropriate in certain circumstances, disputes focus on
whether the required circumstances pertain for the data offered by the twin
and adoption researchers.
The concept of the 'schizophrenia spectrum' relates to both twin and
adoption studies, and involves using broad definitions of schizophrenia to
include extra cases as research subjects. Researchers claim that the genetic
basis of schizophrenia may underpin other, borderline conditions, and thus
that they are justified in counting relations who suffer from these other
conditions in their studies. Critics suggest that these researchers are simply
trying to 'boost' their results, and that, without these borderline cases,
many twin and adoption studies do not produce statistically significant
results.42 In 1984, Steven Rose, Richard Lewontin and Leon Kamin
mounted the most high-profile attack on research using twin and adoption
studies in their book, Not in Our Genes.43Using a mixture of their own
original research and the work of others well known in the field,44 these
authors criticized the methodology, assumptions and ideology of those
genetic researchers, comparing the rates of schizophrenia between sepa-
rated identical twins and the children of people with schizophrenia adopted
by other families. Their work has been followed up by a number of authors
who attack various aspects of the genetic researchers' statistics, diagnoses
and motivations.45 Rose, Lewontin and Kamin's book revealed some
dubious practices within these genetic researchers' studies, the pinnacle of
which is probably the 'pseudo-interview' based on hospital records, with
the psychiatrist imagining a (dead) relative's responses to questions.46
Yet the critics of these studies fail to take into account the fact that the
approach and methodologies of these genetic researchers are continually
changing, partly in response to the critics' complaints. This is clear when
882 Social Studies of Science 31/6

one looks at the 'time-lag' which seems to affect critical reviews of twin and
adoption studies in schizophrenia: critics tend to focus on older twin
and adoption studies, while proponents of this research discuss the most
recent work. It stands out that Gottesman and Shields' classic sympathetic
review of genetic research in schizophrenia, written in 1976, gives compre-
hensive synopses of adoption study research published as late as 1975.47
The latest studies cited in a similarly supportive 1987 review are from the
mid-1980s.48 Meanwhile, critics writing in 1981 or 1983,49 in 1984,50 or in
1990,51 focus only on early reports of twin/adoption research from 1968 or
1971. Just as the proponents of twin studies in the 1960s and 1970s looked
back at previous research and admitted its methodological flaws, claiming
that newer studies were improved,52so the authors of more recent reviews
can defuse these detailed critiques, by focusing on re-analyses that have
been performed since those critiques were written. This discrepancy is clear
in Petter Portin andY.O. Alanen's (1997) critical review of the population
studies, in which the sources of their criticism of adoption studies are from
1976, 1983 and 1984, while the latest actual adoption study they cite is
from 1994.53 I suggest that this 'time-lag' is one reason why the critics'
attacks on schizophrenia genetics have been less effective than one might
expect; geneticists can always claim that they have taken on board criticism
of earlier twin and adoption studies, and that the latest research is thus
'insulated' from repeat attacks.54
Since the heyday of the twin and adoption studies debates, schizo-
phrenia has still been a topic of dispute between genetic researchers and
their critics, for the simple reason that, as one critic puts it:

If the biological basis of schizophreniais so uncertain,then it is quite

reasonableto questionthe validityof geneticizingotherhuman disorders
aboutwhich less is known.55

A number of authors who use the concept of geneticization in their work

are critical of schizophrenia genetics,56 and professional debate over the
r6le of genes in psychiatric disease is wide-ranging and intense.57 The
geneticization of schizophrenia is therefore a rich site to study the use of
genetic explanation for disease, and in the next section I sketch out one
way in which scientists write up research in such a controversial

Features of Enlightened Geneticization

The aim of the following analysis is to explore the narrative that predom-
inates in schizophrenia genetics - what I have termed the narrative of
'enlightened geneticization'. This section introduces some of the themes
and ideas present in the narrative, which distinguish it from crude genetic
determinism; a central theme to the narrative is the presentation of current
genetic thinking as reasonable, non-extremist, and accepting a role for
non-genetic factors in schizophrenia causation. What this paper reveals are
those elements and themes ('strategies') in the chosen review articles,
Hedgecoe: Schizophrenia and the Narrative of Enlightened Geneticization 883

which together form a coherent single narrative about the genetics of

schizophrenia.59The term 'enlightened' has been adopted from an earlier
(1987) review article by Irving Gottesman, Peter McGuffin and Anne
Farmer, who...
reviewa sampleof the highlightsrelevantto enlightenedgeneticthinking,
i.e., a broad... frameworkwith multifactoralcausationassumedand with
provisionfor the epigenetic interactionof psychosocialas well as neu-

The idea of such enlightened thinking is interesting in the light of current

criticisms raised against genetic research into schizophrenia, which some-
times describe it as 'simplistic'.61 As I will show in this paper, researchers
present this narrative as a sophisticated and politically enlightened ex-
planation of schizophrenia genetics. The narrative is called 'enlightened'
not just because it acknowledges (at the surface level at least) the role of
non-genetic factors, but also because it presents a genuinely enlightened
(that is to say, liberal) political position. In their 1997 review, Steven
Moldin and Irving Gottesman present a number of 'Common Misconcep-
tions and Current Realities' surrounding schizophrenia genetics. They
answer the misconception that 'Individuals with schizophreniashould not
reproduce;prevention of reproductionwill lead to the eradication of schizo-
phrenia' with the statement:
Given that there is an imperfect relationshipbetween genotype and
phenotype... there is unlikelyto be a way to predictwith great certainty
who will or will not become affected with schizophrenia.[Moldin &
Gottesman(1997): 555, emphasisin original]

This clearly has echoes of the observation by Anne Kerr and her colleagues
that the geneticists they interviewed dismissed eugenics as bad science
(hence objectively flawed), rather than as ethically unsound.62Yet the use
of uncertainty of prediction in this statement undermines, or even contra-
dicts, comments made earlier in the same article, about the 'enhancement
of diagnosis' and even gene therapy that may result from advances in
knowledge of schizophrenia genetics [Moldin & Gottesman (1997): 552].
This contradiction is unavoidable because, however ironic 'enlightenment'
might seem in this narrative, it does not sit comfortably with the authori-
tarian position expressed in the italicized misconception quoted above.
The concepts of 'improving diagnosis', 'aiding genetic counselling' and
'providing gene therapy' are all part of what Philip Kitcher has called
'Utopian Eugenics'.63This is the idea that individuals and couples, through
exercising their autonomy, will achieve eugenic results without the co-
ercion and paternalism of earlier, state-sponsored eugenic programmes. Yet
there is an uneasy balance to keep; awareness of the dark side of genetic
determinism is perhaps one reason why this narrative ostensibly admits the
r6le of environmental factors in schizophrenia aetiology. Letting the envi-
ronment have a role keeps the old eugenics at bay, yet the narrative is
constructed in such a way as to deny non-genetic factors decisive control.
884 Social Studies of Science 31/6

I suggest that there is far more to the narrative of enlightened geneti-

cization than simply calling schizophrenia a 'genetic disease' and proposing
a 'gene for' schizophrenia with little role for environmental influence,
although sophisticated construction of schizophrenia as a 'genetic disease'
does take place (see below). The difference between such crude genetic
determinism and enlightened geneticization is illustrated by the following
example. In the early 1980s, in a letter replying to criticism of adoption
studies, a supporter of genetic explanations claims that:
Each of these studies, using differentapproaches,has producedresults
that have to date demonstratedonly the action of genetic factors in the
etiologyof the schizophrenias.None of these studies has produceddata
consonantwith the hypothesisof psychosocialcontributionsto etiologyof
the schizophrenias.64

It is not uncharitable to view this as claiming that only genetic factors

contribute to the aetiology of schizophrenia.65This view is not compatible
with the more sophisticated position of enlightened geneticization, repre-
sented by Gottesman and Shields' claim, in their pioneer review article
from the same era, that 'the burden of proof has shifted from showing that
genes are important to showing that environment is important even though
adoption studies cannot prove that environment is unimportant'.66 Such a
challenge does not deny the relevance of environmental effects; it merely
requires evidence of such factors, in the same way as these authors claim
that adoption and twin studies provide evidence for inherited factors.
The themes of moderation and complexity of explanation are con-
sistent across the three review articles analysed in my research study. For
example, in the final section of their 1989 article, Matthew McGue and
Irving Gottesman provide a statement which can be seen as characteristic
of the narrative of enlightened geneticization. They emphasize the fact
Nobody has ever been able to demonstrate statistically that a single major
gene accounts for a large share of the overall risk for schizophrenia...
Approaches premised upon the hypothesis that schizophrenia is a unitary
single gene disorder run counter to a vast amount of empirical research
that suggests otherwise. [McGue & Gottesman (1989): 460]

And they conclude that:

Environmental influences play an essential role in the etiology of schizo-

phrenia. In the rush to molecular biology, it would appear shortsighted for
psychopathologists no longer to consider why it is that among two
individuals, both of whom inherit a genetic diathesis, one will go on to
develop the disorder while the other will not. [ibid.: 461]

Similar sentiments can be found in the 1997 review, where Steven Moldin
and Irving Gottesman reject the idea that environment does not matter in
schizophrenia aetiology, stating:
Environmental factors are clearly of importance in the etiology of schizo-
phrenia. Unfortunately, no specific factor has been found to trigger
and the Narrativeof EnlightenedGeneticization
Hedgecoe:Schizophrenia 885

schizophreniain many or most cases. Our best leadingcandidates... are

nonspecificand idiosyncratic.These factors are more likely to be impli-
cated in the course of illness, ratherthan in its distaletiology.[Moldin&
Gottesman(1997): 554]

Both of these papers emphasize the important role for environmental

effects in schizophrenia causation. The 1989 paper rejects a major role for
a single 'schizophrenia gene' (but note that it allows it some potential role),
but stresses that we need to view such environmental effects in the context
of a genetic baseline, or 'diathesis'. The 1997 paper goes further by
emphasizing the nonspecific, ungeneralizable nature of the environmental
factors, and the additional fact that these environmental factors are not
part of schizophrenia's real (that is, distal) causes. Thus, although the
narrative accepts some role for non-genetic causation, it is presented as a
minor factor in comparison with the regular, specific nature of genetic
This final point links with Gottesman and Shields' 1976 claim that
their model of schizophrenia causation involving a genetic background (or
diathesis) and an external initiator (or stressor) implies that schizophrenia
is a genetic disease.67 This 'diathesis-stressor' or 'threshold' model has
remained constant in schizophrenia genetics since the mid-1970s, and is
represented by the idea of an environmental 'trigger', which is also present
in geneticists' verbal discourse about schizophrenia.68 In a reply to critics
of their 1976 review, Gottesman and Shields state that:

Our emphasis... is on a large and rather specific genetic 'something'

interactingwith nonspecific,ratheruniversal,environmentalfactors.

They suggest that, on these grounds, schizophrenia should be called a

genetic disorder because of the 'relative prevalence of the genetic predis-
position to developing schizophrenia compared with the relative prevalence
of alleged environmental causes'.69 Their proposed 'threshold' model of
schizophrenia has the effect of introducing genetic causation as the basic
explanation of the condition. However valued the environmental contribu-
tion is, it is much harder to define rigorously, and thus the main causal
factor should be seen as genetic.
This outline of the narrative of enlightened geneticization presents a
complex, multifactoral vision of schizophrenia with a role for environmen-
tal influence. It also subtly: ensures that schizophrenia can only be con-
sidered in terms of some necessary genetic aetiology (even if it is not
sufficient); suggests that the variety of different possible environmental
factors should be seen against a 'genetic baseline' which is the only single
necessary condition for causation; and finally that non-genetic factors are,
in contrast to genetic ones, non-specific, unpredictable and thus less
Having sketched the main features of the narrative of enlightened
geneticization, I will now present in detail the different strategies used to
construct this narrative. These strategies are:
886 Social Studies of Science 31/6

* the use of history (reference to twin and adoption studies to support

molecular genetic research);
* genetic modelling (constructing explanatory models which fit empirical
data but which also prioritize genetic explanations);
* responsibilityand caution (accepting past mistakes and moderating claims
about the power of molecular approaches);
* the schizophreniaspectrum(constructing the spectrum to include border-
ing conditions with a known genetic component).

The Use of History

This strategy involves looking back to twin and adoption studies to provide
'proof' of the genetic component of schizophrenia aetiology, and hence
justification for molecular research. One problem for those concerned with
molecular genetic approaches to schizophrenia is that the complexity of
this disorder in phenotypic terms is mirrored by the complexity of the
genetic factors involved. From what geneticists have revealed about schizo-
phrenia, no single gene is involved; rather, a number of different genetic
factors are implicated, in addition to environmental ones. Thus the
challenge for the proponents of research into the molecular genetics of
schizophrenia is to justify research into such a complex and difficult topic.
The answer lies in the fact that for these scientists, twin and adoption
studies prove that there is a genetic component to schizophrenia; by
referring back to these old-fashioned studies (the use of history), authors
can justify support for continued genetic research. Whatever the defi-
ciencies of the twin and adoption studies, according to this narrative they
do provide hard evidence that genetic factors play a role in the aetiology of
As an example of the role historical perspective plays in constructing
the narrative of enlightened geneticization, I provide an analysis of the
1989 review which presents and summarizes epidemiologic data from
research drawing on twin and family studies, and includes a table (re-
produced as Table 1) which...

gives estimatedlife time risks... for developingschizophreniaamong the

relativesof schizophrenicprobandspooled fromsystematicstudiesunder-
taken in Western Europe since 1920. [McGue & Gottesman (1989):

There is no attempt to explain how a large number of different studies with

varying methodologies were combined to produce this table. Nor is there
any explanation of the fact that the early twin and family studies are
normally treated with extreme caution, since even the strongest exponents
of twin studies now admit that the lack of controls, and the non-blind
status of the investigators, render many of the early results doubtful.
Readers are merely referred to two other papers where the compilation of
the data is explained. There is no mention of the debates surrounding twin
and adoption studies; as presented here, any controversy over the r6le of
Hedgecoe: Schizophrenia and the Narrative of Enlightened Geneticization 887

Rates of Definite Schizophrenia among the Relatives
of Schizophrenia Cases

Familial relationship % Affected

Offspring of 2 schizophrenic parents 36.6

Monozygotic twins 44.3
Dizygotic twins 12.1
Siblings 7.3
Offspring of 1 schizophrenic parent 9.4
Half siblings 2.9
Nieces or nephews 2.7
Grandchildren 2.8
First cousins 1.6
Spouses 1.0

Source:McGue & Gottesman (1989): 454, Table 1.

genetics in the aetiology of schizophrenia has reached closure.70 As men-

tioned earlier, the critics of twin and adoption studies are largely erased in
all three of these reviews. The authors present twin and adoption studies as
unproblematic (although there may be internal debates over variations in
concordance rates).
Some problems with early twin and adoption studies are mentioned in
the 1993 article, where the authors again review family, twin and adoption
studies. The discussion introduces early family studies that suggest that
'schizophrenia consistently and substantially aggregates in families'
[Kendler & Diehl (1993): 262]. However, Kenneth Kendler and Scott
Diehl do suggest that these studies were deemed to be methodologically
flawed; in this case, non-blind, without controls and without organized
diagnostic criteria. Schizophrenia researchers then carried out subsequent
studies which repeated the earlier research under more rigorous condi-
tions. They produced largely the same results, suggesting that...

the earlier first-generation family studies were substantially correct in

concluding that schizophrenia strongly aggregates in families. Their
findings were not systematically biased by nonblind or unstandardized
diagnostic practices, the absence of control groups, or an overly broad
classification of the disease. [ibid.: 264]

This enlists even the oldest (and least reliable) studies in the cause of
schizophrenia genetics. If the original studies are 'substantially correct',
then modern research into schizophrenia genetics is part of an honourable
tradition. This appeal to history is also adopted by the authors of the 1997
review, who clearly state: 'Family, twin and adoption studies conducted
over the past 30 years have provided very strong evidence that both genes
and environment play a role in its [schizophrenia's] etiology'; and they note
that 'data from more than 40 family studies spanning seven decades of
research consistently show that risks to different relatives of affected
888 Social Studies of Science 31/6

individuals are in fact significantly greater than the population risk'

[Moldin & Gottesman (1997): 547].
However, the figures from the twin and adoption studies also show
that non-genetic factors appear to play a significant role (over 50%
concordance for monozygotic twins) in schizophrenia aetiology. McGue
and Gottesman admit that'the MZ twin concordance rate is substantially
less than 100 percent and thus... environmental factors... play a sig-
nificant role', but they present this role as almost that of background noise,
since the environmental influence 'obscures the mechanism of genetic
transmission, resulting in continued debate as to whether schizophrenia is
a single gene disorder' [McGue & Gottesman (1989): 454]. Citing an
average MZ concordance rate of 46%, Moldin and Gottesman note that
since this is 'significantly less than 100 percent... [it]... implicates the role
of nongenetic factors' [Moldin & Gottesman (1997): 547]. They then
discuss heritability scores, which is another method for measuring the
inherited component of a phenotype: they note that the high heritability
scores for schizophrenia gained from twin studies, 89% and 74%, are
'without contribution from the common environment', and that the non-
genetic contributions are likely to be the result of 'idiosyncratic environ-
mental events' [ibid.: 548].71 Such a discussion fits neatly into the pattern
maintained by the narrative in these reviews - namely, the acceptance of
environmental factors only if they are not regular and predictable.
Both the 1993 and 1997 reviews describe the results of research
carried out in Finland which challenges the theme that environmental
factors in schizophrenia aetiology are variable and unpredictable.72
Kendler and Diehl accept that Pekka Tienari's 1991 adoption study...
has also examined the role of familial-environmentalfactors in the
etiologyof schizophreniaand found a substantialcorrelationbetweenthe
functioningof the adoptivefamily and the psychiatricoutcome in the
adoptee. [Kendler& Diehl (1993): 266]

Such a finding resists the narrative of enlightened geneticization because it

presents an identifiable, repeating environmental factor, and contradicts
data from twin studies covered earlier in the same review which claimed
that 'familial environment makes little or no contribution to the liability for
schizophrenia' [ibid.: 265]. The most obvious way to defend against such
findings is simply to reverse the cause-effect implied in the correlation:
However,most of the familieswereevaluatedwhen the adopteeswerewell
into adult life. Thus this observed correlation... could also be due to

By making disturbed family life the resultof schizophrenic behaviour rather

than its cause, the authors prevent these results from undermining the
narrative.This kind of dismissal is reminiscent of many of the criticisms of
the twin and adoption studies, and shows how both sides in a scientific
controversy adopt similar strategies to make their case.73When addressing
the results achieved by Pekka Tienari and his Finnish colleagues in their
Hedgecoe: Schizophrenia and the Narrative of Enlightened Geneticization 889

adoption studies, Moldin and Gottesman, like the authors of the 1993
review, admit that this research 'found evidence for gene-environment
interaction' [Moldin & Gottesman (1997): 548]. The type of environmen-
tal factor which was found to be important was 'Communication deviance
in adoptive parents based on projective test results', but the authors
dismiss this as 'a nonspecific environmental factor commonly found in
many families'; their summary of the Finnish results is phrased as 'genes
controlling sensitivity to the environment', reinforcing the idea of a genetic
'baseline' in schizophrenia, with the environmental factors, nonspecific
and vague, merely layered on top [ibid.: 548].
By constructing the debate in this way, the narrative gains the benefits
of historical precedence, without having to accept the faults and weak-
nesses of the early research (as discussed by the critics of the twin and
adoption studies). This gives the narrative strong historical 'roots' in
population genetics, but also clears the early researchers of the kinds of
accusations of bias and ideology that critics often make. By extension, it
insulates modern researchers from any claims of bias. It may seem strange
that an important part of supporting molecular genetic studies of schizo-
phrenia involves detailed reference to twin and adoption studies - old ways
of investigating inheritance. However, as the next section of this paper
shows, the genetic factors in schizophrenia are thought to be extremely
complex. In this case, older, simpler studies which confirm a genetic
component to schizophrenia become extremely useful in supporting
further research.

This section shows how genetic models are used in these articles to
undermine non-genetic causes and allow strong, single-gene models some
r6le in the narrative, despite empirical evidence to the contrary. The
models are based on empirical evidence of how schizophrenia is inherited,
and are attempts to explain how different genetic and environmental
factors combine to produce the patterns of inheritance seen.
In their 1989 article, McGue and Gottesman review the single gene
concept (the so-called 'generalized single locus' model, or 'GSL'), describ-
ing it as the oldest and most prominent hypothesis in the genetics of
schizophrenia [McGue & Gottesman (1989): 454]. The attraction of the
GSL lies in its economy, and the fact that its simplicity would make
molecular searches for the gene involved much easier. Unfortunately, as
they also point out, 'the GSL model has repeatedly failed to account for
the observed pattern of familial risk in schizophrenia' [ibid.: 455]. For
them, this divergence from the empirical data makes it extremely unlikely
that the GSL is an accurate explanation of the genetic aetiology of
schizophrenia. In addition, the low penetrance (that is, power) of the gene
required by the GSL model is at odds with the 'relatively large MZ twin
concordance rate', which is, of course, still 'substantially less than 100
percent' [ibid.: 456].74 In their 1993 article, Kendler and Diehl also
890 Social Studies of Science 31/6

dismiss the likelihood of schizophrenia being caused by a single gene,

concluding that 'schizophrenia differs from the classic Mendelian disorders
in at least four crucial ways'. These differences are: the lack of full
penetrance; the existence of phenocopies;75 schizophrenia's relative fre-
quency (and heterogeneity); and the lack of distinct diagnostic boundaries
between affected and unaffected individuals [Kendler & Diehl (1993):
271]. This distancing of schizophrenia from Mendelianism avoids accusa-
tions of simplification and ignorance of empirical reality. To subtly con-
struct schizophrenia in terms of geneticization, one has to distinguish it
from less sophisticated versions of genetic disease.
When McGue and Gottesman turn to 'Multiple Gene Models', they
focus on their preferred candidate, the 'multifactoral threshold' (MFT)
model. With the MFT, 'genetic factors are assumed to be polygenic',
meaning that 'a large number of genes, each of small and equal effect,
combine additively with the effects of other genes and environmental
factors to influence schizophrenia liability'; under this model, schizo-
phrenia happens when 'an individual's combined liability exceeds some
threshold value along the unobserved liability continuum' [McGue &
Gottesman (1989): 456]. They claim that the MFT model successfully
accounts for patterns of familial risk, producing the characteristic ex-
ponential risk function seen in schizophrenia; but they admit that 'there is
a general reluctance to accept the MFT model as an explanation for the
transmission of schizophrenia' [ibid.: 451]. The reasons for this are not
intrinsic to the model itself, but are contingent factors - namely, the
failure, at the time of writing, to identify specific genes and environmental
factors which might contribute to the liability; and the fact that 'strict
polygenic inheritance... would likely preclude, for the near future, at-
tempts at identifying single gene effects on schizophrenia risk through
molecular genetic approaches' [ibid.: 456]. Essentially, they suggest that
the MFT is less popular because it is harder to investigate, and would be
harder to prove correct. One advantage of the MFT is that it allows for co-
existence with...

a single majorgene whose effect upon schizophreniarisk is largerelative

to the effects of other (poly)genes... raresingle gene disordersthat give
rise to schizophrenia... [and]... a tractableMFT model of a limited
number of polygenes (3, 4, or 5) each with a 'subcomponenteffect' on
schizophrenia. [ibid.: 456-57]

The strength of this strategy is that it allows one to remain committed to a

complex, polygenic model, which fits the empirical results but is hard to
investigate. At the same time, it justifies molecular research into schizo-
phrenia since it presupposes a number of other co-existing monogenic
varieties of schizophrenia which are easier to explore.
Kendler and Diehl suggest that although the results are 'rather incon-
clusive', most evidence from segregation analysis supports some form of
multifactoral model, since 'it appears unlikely that the distribution of
schizophrenia is due to the effect of a SML' (a 'Single Major Locus',
Hedgecoe: Schizophrenia and the Narrative of Enlightened Geneticization 891

synonymous with McGue and Gottesman's GSL). Yet, however strongly

schizophrenia is constructed as a multifactoral (or mixed) genetic disorder,
a back door is always left open to allow single major genes some role in its
aetiology: 'lack of evidence for schizophrenia should not be interpreted as
strong evidence against the possibility that such "major" genes exist in a
subset of families'. This strategy acknowledges the empirical reality and
complexity of schizophrenia genetics and the limitations of the methods of
molecular analysis used to investigate them. It also permits continued
investigation using the same techniques. The authors suggest that schizo-
phrenia is not inherited in a Mendelian fashion, yet they admit that the
main method used to investigate the molecular genetics of schizophrenia is
linkage analysis, which 'was developed originally for simple Mendelian
traits' [Kendler & Diehl (1993): 271].76 This allows some form of genetic
association with schizophrenia, permitting the use of linkage analysis but
allowing for the existence of a rare, Mendelian version of the disorder.
Such a strategy enlists the results of molecular genetic research without
them having to refer to all instances of schizophrenia.
In genetic modelling, as in the use of history, the authors have to find
a way to incorporate non-genetic causal factors into their descriptions,
without allowing them to gain any real explanatory power. One example of
this is the way in which McGue and Gottesman discuss the popular
division between familial versions of psychopathology and sporadic forms
(where there is assumed to be no inherited component, with causation due
to environmental factors). Using MZ twin rates from a study by Fischer,
they suggest that:

Relativelylow ratesof schizophreniaare expectedamongthe offspringof

discordant MZ twins if discordance among the genetically identical
parentsis due largelyto environmentallyinduced phenocopies. [McGue
& Gottesman(1989): 459-60]

The assumption is that the difference between the identical twins is

because one has a wholly environmentally induced type of schizophrenia
(hence phenocopy), while the other, not exposed to the same environmen-
tal factor, does not develop it. And neither do the children of either twin,
since the cause of the schizophrenia is environmental rather than genetic.
McGue and Gottesman then present a table (reproduced as Table 2).
They admit that the small size of the sample limits the conclusions one can
draw, but suggest that the data in this table support the concept of a
widespread genetic component to schizophrenia, even when it is not
expressed. The offspring of both the affected and unaffected MZ twins
have a similar rate of schizophrenia (represented in the MR column), and
both rates are comparable to the risk to the offspring of a parent with
schizophrenia (9.4%). The rates for the offspring of the DZ twin are
significantly different, with the children of the affected twin having a higher
risk than those of the unaffected. The authors draw the conclusion that this
is because 'the expression of schizophrenia depended upon both an in-
herited genetic diathesis, which was transmitted regardless of whether the
892 Social Studies of Science 31/6

Schizophrenia and Schizophrenia-like Psychosis in Offspring of
Discordant Twins

Parent Status Number Affected MR%

Monozygotic sample
Affected twin 14 1 10.0 + 9.0
Unaffected twin 24 4 17.4 + 7.7

Dizygotic sample
Affected twin 13 1 8.3 + 7.6
Unaffected twin 52 1 2.1 + 2.1

Note: 'Affected' gives the total number of offspring affected with

either schizophrenia or a schizophrenic-like psychosis; 'MR' gives the
estimated lifetime morbid risk of being diagnosed as schizophrenic.
Source:from McGue & Gottesman (1989): 460, Table 6.

diathesis was phenotypically expressed, and exposure to environmental

stressors, to which the twins are differently exposed' [McGue & Gottes-
man (1989): 460].77 In many ways this is a restatement of Gottesman and
Shield's (1976) point, that the construction of the 'genetic baseline'
presents the genetic liability to schizophrenia as the main requirement for
schizophrenia aetiology.78 Environmental factors are usually classed in
terms of being 'non-specific' and hard to identify; unless one or two
specific environmental influences can be identified, then the role of envi-
ronmental factors, while accepted, cannot be given priority. Although the
environmental stressors are needed for schizophrenia to develop, it is the
genetic diathesis which is ever-present, just needing the correct trigger.
Alternative phrasing, with the environmental stressors occurring in cases
where there is not the requisite genetic 'background' but having their
effect when it is present, is not suggested. The irony is that the narrative
of enlightened geneticization is not so strict about the identifiability of
genes. A single-gene model is not a credible explanation for schizophrenia
aetiology, yet the lack of a single geneticcause is not counted as a weakness
for genetic explanations overall.
This emphasis on the unpredictability of environmental factors is
repeated in Moldin and Gottesman's 1997 review, where the section on
'Gene-Environment Interaction' suggests that the environmental factors
that might increase liability to schizophrenia 'include nonspecific stressors,
obstetrical complications and illicit drug use'. But they moderate this by
claiming that 'the predictive power of these factors for a schizophrenia
phenotype... is low'. The authors debate possible models relating to
genetic and environmental risk factors. First, the 'additive model' supposes
that increases in risk from environmental stressors affect those with high
and low-risk genotypes alike. In short, environmental and genetic risk
factors operate separately. The alternative model, 'genetic sensitivity',
specifies genetic control of the sensitivity to environmental risk factors;
those with higher genetic risk are more sensitive to environmental factors.
Hedgecoe: Schizophrenia and the Narrative of Enlightened Geneticization 893

This second model offers a reduced role for environmental factors in

comparison to the 'additive' model, which allows a degree of independence
for non-genetic factors; there is always the possibility that a multitude of
environmental stressors could cause schizophrenia in a genetically low-risk
individual. But the 'genetic sensitivity' model ensures that environmental
stressors are only of relevance when high-genetic risk is also a factor.
Moldin and Gottesman cite research into depression which follows this
model, but admit that there is no evidence for either of these models in the
case of schizophrenia, since 'a careful assessment of putative environmental
risk factors... has not been done'. Despite this lack of evidence, the
authors prefer the second, genetic sensitivity model, because: 'Such a
model will likely permit a more accurate understanding of the etiology of
schizophrenia, in which genetic and environmental risk factors interact in a
complex way' [Moldin & Gottesman (1997): 548]. Such a model also
limits the role of these environmental factors in a way the first, additive
model does not; the environment only counts in the presence of the correct
genetic background. Support for it is on the basis of how useful it would be
if it were true, rather than the strength of particular empirical evidence.
This strategy constructs genetic models which strengthen the narrative by
allowing for the existence of single gene versions of schizophrenia, and the
undermining of non-genetic causes as non-specific and unpredictable.

Responsibilityand Caution
A strategy of caution deflects criticism that researchers are over-enthu-
siastic and guilty of genetic hype, and allows readers to see authors as
responsible and objective in their assessments. One of the most obvious
ways of displaying caution is to show how others have gone wrong in the
past. For example, Kendler and Diehl give a detailed analysis of the alleged
'chromosome 5' linkage which followed the observation that a young man
and his uncle, both diagnosed with schizophrenia, also had distorted facial
features. Analysis showed that they each had a chromosome 5ql 1.12-
q13.3 trisomy. As a result:

In 1988, Sherringtonet al. reportedlinkagein seven high-densitypedi-

grees... between schizophrenia-relatedphenotypesand two DNA mark-
ers in the regionof the 5q trisomy.[Kendler& Diehl (1993): 273]

This study was striking for a number of reasons:

* the pedigrees were very large and of extremely high density for
* there was very strong evidence in favour of the linkage between the DNA
markers and the phenotype;
* the evidence was strongest when a broad diagnosis was used (including
phobia, depression and alcoholism);
* the strong linkage was found in all of the pedigrees, implying more
homogeneity than expected. [ibid.: 273-74]
894 Social Studies of Science 31/6

Kendler and Diehl then list 45 other studies subsequently carried out, 11
of which looked at the same area of the genome; 'Attempts to replicate this
finding have been uniformly negative', including new studies carried out by
the same research team that reported the initial results [ibid.: 274-75]. The
authors ask: 'How can we explain the initial strong positive evidence of
linkage, followed first by uniformly negative results in all independent
studies and then by an inability of the original investigators to replicate
their own work?', and they are forced to conclude that 'we remain in the
unsatisfactory position of having no adequate explanation for this puzzling
and discouraging series of events' [ibid.: 276]. Moldin and Gottesman also
discuss the 'false step in the path of discovery' that was the 1988 chromo-
some 5 result, stating that it was 'welcomed with considerable optimism
and insufficient scientific criticism'. One of the reasons they give for this
acceptance is 'because localization of a locus was "confirmation" that
schizophrenia was in fact a "biological" and not a "psychosocial" disorder',
and thus undermines the stigma associated with having schizophrenia
[Moldin & Gottesman (1997): 554].79 The reason for the quotation marks
in this statement is not clear. Perhaps the most likely explanation is to cast
doubt on the idea that there is any useful distinction to be made between
genes and social causes. This is in keeping with the enlightened position
which is firmly in favour of multifactoral explanations, and for which old-
fashioned nature/nurture distinctions are too crude. It also casts doubt on
the idea that linkage analysis provides confirmation of anything. This may
be a cautious warning about the strength of such studies in the case of
complex disorders, but it can also be seen as a reinforcement of the
original twin and adoption studies. For many of the researchers in this
field, the earlier, non-molecular work has already confirmed the biological
nature of schizophrenia causation. They do not need molecular results to
confirm what, for them, is already a fact. In the narrative, the chromosome
5 story is a useful cautionary tale, covering as it does an over-optimistic
research result, enthusiastic media response and embarrassing withdrawal.
As far as these authors are concerned, this is exactly the wrong way to do
schizophrenia genetics.
The 1993 review article continues the theme of caution when it offers a
detailed and persuasive account of the difficulties faced by those looking for
a putative genetic marker for schizophrenia. The authors 'again emphasize
the special challenges'; schizophrenia cannot be seen as a 'simple disorder'
[Kendler & Diehl (1993): 271], making linkage studies 'qualitatively and
quantitatively more difficult' [ibid.: 273]. The qualitative changes are
required because the method has to be altered to be effective, and quantita-
tively because of the sheer size of the samples required in such a complex
case. Kendler and Diehl present a table (reproduced as Table 3) which
suggests the numbers of subjects needed to produce reliable results.
This table shows the intuitively obvious trend that the higher the
penetrance (that is, the stronger the gene's effects), the fewer the number
of families required to detect linkage. The figure 'a' is inversely pro-
portional to the degree of heterogeneity present in schizophrenia. The
Hedgecoe: Schizophrenia and the Narrative of Enlightened Geneticization 895

Number of Multiplex Families Needed to Detect Link-
age to Schizophrenia with 80% Power, Given Varying
Levels of Genetic Heterogeneity

a 0.5 0.6 0.7 0.8

1.0 47 43 40 34
0.9 61 56 51 44
0.8 76 71 64 56
0.7 99 93 84 73
0.6 137 126 114 99
0.5 196 180 160 140
0.4 300 277 246 213
0.3 522 483 434 370
0.2 1106 1045 922 794
0.1 4359 3824 3550 2994

Note: 'a' is the proportion of families in which the schizo-

phrenia-predisposing gene is linked to the marker.
Source:based on Kendler & Diehl (1993): 272, Table 5.

optimistic view that schizophrenia is homogeneous (that is, single gene,

a = 1), with 60% penetrance, means that only 43 families (301 in-
dividuals, assuming a family of seven) would be required to ascertain
linkage. But, assuming the same penetrance:
More likely, the single most common gene responsible for schizophrenia is
present in only 50 or 30 percent of high-density families, so that required
samples would amount to 180 families (1,260 individuals) or 483 families
(3,381 individuals). [ibid.: 272]
Should the heterogeneity of schizophrenia be even higher, with the most
common gene accounting for only 10% or 20% of these families, then
'sample sizes required to detect linkage by current methods may be
unobtainable with realistic resources' [ibid.: 272-73]. Similar, unmanage-
able sample sizes are required if one assumes that each high-density family
has a small number of genes (four or five) instead of just one [ibid.: 273].
A different way of displaying caution is offered by Moldin and Gottes-
man when they discuss linkage studies in their 1997 paper. They spend
over one page explaining the specifics of the level of proof required to
suggest linkage. Such proof is expressed as a logarithm of the odds (lod)
ratio of that linkage occurring, with the criterion for Mendelian disorders
set at a lod of 3. This score means that the observed data are 1,000 times
more likely to occur if there is linkage, than if there were none [Moldin &
Gottesman (1997): 549]. But schizophrenia is not a Mendelian disorder,
and the authors point out the apparent contradiction upon which linkage
studies for polygenic disorders are based:
Despite the inability of the single locus model to explain the familial
aggregation of schizophrenia, investigators have conducted traditional
linkage studies under single locus model assumptions. [ibid.: 548-49]
896 Social Studies of Science 31/6

The contradiction is heightened if we consider the 'use of history' strategy

that suggests that we know schizophrenia is genetic because of the family,
twin and adoption studies. Yet it is these same studies which produce the
models proving that schizophrenia is not a single locus model, and thus
cannot adequately be investigated using linkage analysis.
Avoiding this contradiction requires a higher standard of proof from
the linkage studies of polygenic disorders than one would need with
Mendelian ones. Thus Moldin and Gottesman adopt a scoring system
(devised by Eric Lander and Leonid Kruglyak) which describes a lod score
of 1.86 as 'Suggestive', 3.30 as 'Significant' and 5.10 as 'Highly Sig-
nificant'. They then assess actual linkage studies by these criteria, and find
them lacking. This review's Abstract states that 'the strongest evidence
implicates chromosomes 6 and 8, but these linkages are not yet confirmed'
[ibid.: 547]. This is an apparent understatement since, in the article itself,
Moldin and Gottesman agree that the chromosome 6 has 'the strongest
evidence thus far for linkage in schizophrenia', but also that correction of
these results for different models had reduced their strength. In addition,
'Nonreplications of chromosome 6 linkage have been reported', and 'the
markers inplicated [sic] by the studies reporting suggestive evidence... lie
within a very large chromosomal region that contains many genes'. As for
chromosome 8, which was also mentioned in the Abstract, two studies gave
evidence classed as 'suggestive' for linkage, but the results from one of
these would be down-graded to 'less than suggestive' if corrected for
'multiple transmission and disease models'. There has also been at least
one nonreplication study for this region. Summarizing the results for these
two regions, the authors conclude that 'the magnitude of the statistical
evidence and the existence of nonreplication demonstrate that these are
clearly not confirmed, convincing findings' [ibid.: 550]. As for the rest of
the genome, 'Reported linkages to other chromosomes (3, 5, 9, 20, 22) are
less compelling' [ibid.: 551].
Such admissions about mistakes in the past might seem odd, since the
non-replication and withdrawal of molecular results are often used by
critics to suggest how weak the research into schizophrenia genetics is.80It
might make more sense to skip over the embarrassments of the past, in
much the same way that problems and objections to the early twin and
adoption studies are glossed over in these reviews. The answer lies in the
way in which the narrative incorporates these problems, so that the
presentation of such a large amount of negative evidence may be construed
as less damaging. Genetic aetiology in schizophrenia is insulated from
doubt by the 'fall-back' position of the twin and adoption studies which
'prove' that there is a strong genetic component. Since the debate on these
studies has, for these authors at least, reached closure, they never have to
question this 'fact'. This means that the authors can use previous failings of
molecular genetic research as a springboard for future research. For
example, Moldin and Gottesman move from what might be seen as
depressing linkage results to outlining the US National Institute of Mental
Health's 'Genetic Initiative', a sampling project launched in 1989 to collect
Hedgecoe: Schizophrenia and the Narrative of Enlightened Geneticization 897

DNA samples and diagnostic data to use for large-scale genomic searches.
They also outline 'a new revolution in molecular genetic techniques' that
will help the large-scale analysis of the genome using parallel sequencing
with DNA arrays or genechips [Moldin & Gottesman (1997): 551]. They
propose 'New robust statistical techniques that rapidly extract inheritance
information provided by many genetic markers and permit estimation of
disease gene location'; and they even cite 'the successful cloning of a lamb
from the cells of an adult sheep', explaining that:

Largenumbersof identicalcopies of animalsfor genetic studies may be

generated more efficiently and used to develop epigenetic and neu-
rodevelopmentalmodels of processes aberrantin schizophrenia.[ibid.:

The speculative enlistment of these cutting-edge technologies advances the

narrative by implying that molecular results for schizophrenia genetics have
almost been achieved despite the limitations of traditional molecular
techniques like linkage and association studies. The story being told here is
of past failures and the bright future that awaits researchers thanks to
developments in technology. These technologies do not involve reassessing
the concept of a genetic basis to schizophrenia aetiology, but simply
provide ways of testing more genetic material at any one time than was
previously possible.
What purpose do these discussions serve in the narrative of en-
lightened geneticization? The strategy of caution works at two levels. First,
as I have already claimed, it denotes responsibility and reasonableness on
the part of the authors. In tune with claims to be outlining a 'third way'
between the extremes of the molecular pessimists and evangelists, such a
responsible strategy accepts how difficult the molecular investigation of a
complex disorder such as schizophrenia is. It deflects claims that the
authors are 'hyping' the prospects of linkage results too much, and not
accepting the mistakes that have been made in the past. By highlighting the
problems with the 5q13 trisomy, the narrative comes across as sober, solid
and objective. This reflects themes noted by Sarah Cunningham-Burley
and Anne Kerr in their own review of geneticists' writings, where responsi-
bility and education of both the public, press and clinicians about the
'facts' of genetics are called for by scientists.81
The second level at which these strategies work is more prospective. By
setting the hurdles high, no one should be surprised if linkage results are
not forthcoming for a long time. With the problems of such studies openly
acknowledged, this strategy combats disillusionment among researchers
and prevents possible critics of this research from claiming that 'no results'
means that there are no stretches of DNA that can be associated with
schizophrenia. For example, in the final section of their review, Kendler
and Diehl outline several of the elements that make up this narrative: the
straightforward use of twin and adoption studies which 'continue to
confirm... the major etiologic role played by genetic factors'; the call for
898 Social Studies of Science 31/6

moderation which characterizes this article ('we caution against pre-

maturely abandoning the more tried and true methods of psychiatric
genetic research'); and the need for realistic expectations with respect to
the outcomes of molecular research:
It is criticalthat we avoidprematuredisillusionmentwith linkagestudies
of schizophrenia...If there are manygenes, none of which has any more
thana smalleffecton liability,then currentmethodsand projectedsample
sizes are almost certainly inadequate and will yield negative or un-
The need for patience on the part of researchers is echoed again in the last
paragraph of this review which, while discussing the potential role of the
large numbers of genetic markers being made available, states: 'this
approach, if allowed sufficient time to mature, could yield truly unprece-
dented insights into the etiology of this disorder' [Kendler & Diehl (1993):
278, 279]. These authors manage to combine both lack of results and
optimism when they note that 'no replicated positive findings have yet
emerged from efforts to locate individual genetic loci that influence the
liability to schizophrenia. This is not surprising, nor should it be too
discouraging' [ibid.: 277]. This strategy uses the failures and disappoint-
ments of the past as a way of enlisting new technologies, and protecting
schizophrenia genetics from overly high expectations of what it may
achieve in the short-to-medium term.

Return of the SchizophreniaSpectrum

As I have explained earlier in this paper, the 'schizophrenia spectrum' played
a major r6le in the early, non-molecular research into schizophrenia genet-
ics. One continual point of criticism of twin studies was that their practi-
tioners did not limit themselves to including those twins clearly diagnosed
with schizophrenia, but also included in their figures 'borderline' diagnosis
of 'schizophreniform', or 'schizophrenia-like' conditions. This use of broad
criteria increases the possible number of participants, and thus adds to the
strength of the association. The use of broad criteria serves (and is served
by) the idea of schizophrenia existing as part of a wider spectrum of
disorders, which itself has a significant genetic aetiology.
The role of the schizophrenia spectrum in molecular research is as
important as that in twin and adoption studies, and is most explicit in
Kendler and Diehl's (1993) article. Following their review of the family,
twin and adoption studies, the authors turn to the section on 'Boundaries
of the "Schizophrenia Spectrum"', where they test four different hypo-
theses on the transmission of schizophrenia:
* only liability to typical schizophrenia is transmitted;
* liability to schizophrenia and schizophrenia-like personality disorders is
* liability to all nonaffective psychosis (e.g. schizoaffective disorder, delu-
sional disorder and atypical psychosis) is transmitted; and
* liability to all psychiatric illness is transmitted.
Hedgecoe: Schizophrenia and the Narrative of Enlightened Geneticization 899

Reviewing work in this area, the authors state that levels of schizotypal
personality disorders and paranoid personality disorders (SPD/PPD) are
higher in relatives of people with schizophrenia than in controls. This
proves the second hypothesis - namely, that 'family liability to schizo-
phrenia can, at least in part, express itself as a predisposition to a certain
set of schizophrenia-like personality traits' [Kendler & Diehl (1993): 267].
There are some doubts over the third hypothesis (liability to nonaffective
disorders), but: 'The preponderance of the evidence... supports the valid-
ity of the third hypothesis that the family vulnerability to schizophrenia
may manifest itself... as a predisposition to more broadly defined psycho-
sis'. The authors do not support the final hypothesis, that 'the risk for all
major forms of psychopathology ought to be increased in relatives of
schizophrenia' [ibid.: 269]. The 'familial liability to schizophrenia appears
to be neither extremely narrow nor extremely broad', thus reifying a
moderate version of the schizophrenia spectrum [ibid.: 270]. By presenting
the concept of the schizophrenia spectrum as neither the narrowest nor the
broadest possible hypothesis, the authors present the spectrum as a bal-
anced, reasonable proposition (in much the same way that their view of the
use of molecular methodology was presented as between two extremes).
The construction of the spectrum is carefully worded, and throughout this
section, the authors use the term 'familial liability' rather than 'genetic
liability'. The two are clearly distinct since, earlier in the review, the
authors state that:

Familystudies suggestthat schizophreniastronglyaggregatesin families.

Twin and adoptionstudies continueto suggestthat geneticfactorsaccount
for most of thisfamilial aggregation.[ibid.: 266, emphasis added]

Therefore, genetic and family effects are not one and the same thing.82
Clearly, this discussion of the varying degrees of the schizophrenia spec-
trum does not present evidence that supports a genetic basis to the whole
range of variations, but it may encourage this interpretation; the authors
state that: 'the familial liability to schizophrenia is expressed, at least in
part, as a predispositionto a set of schizophrenia-like or schizotypal person-
ality traits' [Kendler & Diehl (1993): 266, emphasis added]. The use of the
word 'predisposition', a word often preceded by the word 'genetic', could
easily imply an inherited factor linking these disorders to an extent that
might not be supported by the family and adoption studies. At the
beginning of this section, the authors state that 'An accurate definition of
the affected phenotype is crucial in any genetic investigation', and then
continue that it is helpful 'to use genetically informative designs to deter-
mine' associated disorders. In addition, 'Such an endeavour is also of
interest because it can provide insights into the nature of that familial
liability and hence into the etiology of schizophrenia itself' [ibid.: 266].
This section thus blurs the meaning between the words 'family liability'
and 'genetic predisposition' so that the former stands as a shorthand for
the latter, reinforcing the concept of the schizophrenia spectrum as a
genetic entity.
900 Social Studies of Science 31/6

The next section of the 1993 review addresses the alternative molec-
ular methodology of 'association studies'. While linkage focuses on the
cosegregation of genetic markers and disease phenotype within a family,
association studies examine large numbers of unrelated disease sufferers
and compare the frequency of individual genes [ibid.: 277]. One obvious
advantage of an association study is that it is far easier to find large
numbers of individuals with schizophrenia than it is to find high-density
schizophrenia families. Since association studies are looking for specific
genes (or alleles) rather than the larger areas of genome covered by linkage
analysis, they rely upon the close location of the target alleles on the
genome and the tendency of markers close to target alleles to be inherited
together ('linkage disequilibrium'). Thus association studies are limited to
'candidate genes' which are suspected of playing some form of aetiologic
role in schizophrenia. For example, studies have found association with the
porphobilinogen deaminase (PBGD) gene that causes a form of porphyria,
a mental disorder with a number of similarities to schizophrenia.83 But
linkage analysis for this gene has shown that between 60% and 80% of
high-density families do not carry it, so the evidence is equivocal [ibid.:
Association studies provide another means to investigate the genetics
of schizophrenia, but they also support enlightened geneticization, in that
they neatly complement linkage analysis by offering a form of investigation
which does not involve multiplex families, and which relates the genetic
research to actual diseases. By definition, the use of association studies
requires some acceptance of the schizophrenia spectrum. It is far easier to
see association studies contributing to schizophrenia research if the whole
concept of schizophrenia genetics is built on the idea that 'bordering'
diseases share a genetic aetiology (although porphyria is not normally
included in the schizophrenia spectrum). The introduction of association
studies means that diseases such as porphyria, which were previously
described as distinct from schizophrenia, are now enlisted, becoming an
important research resource.84

Discussion and Conclusion

The subtle, but powerful, nature of the narrative of enlightened geneticiza-
tion becomes clear if we consider criticism of molecular genetic research,
and how it often seems to miss the point about the recent discourse of
schizophrenia genetics. In their critique of schizophrenia genetics, Petter
Portin and Y.O. Alanen review association studies between schizophrenia
and specific, identified markers (especially Dopamine genes), and they
suggest that, as yet, there is no strong evidence for association with any
specific marker.This is perfectly in keeping with the narrative.Their review
of 'Studies involving the whole genome or large parts of it' cites many of
the same studies as Moldin & Gottesman (1997), and leads to the
conclusion that 'it appears that the hypothesis of a single major gene as a
risk factor is unlikely'. As Portin and Alanen note, this is 'in agreement
Hedgecoe: Schizophrenia and the Narrative of Enlightened Geneticization 901

with the viewpoints expressed previously by many psychiatric geneticists'.85

Here they cite four papers, two by McGue and Gottesman, one of which is
the 1989 review article. Portin and Alanen suggest that Lander and
Kruglyak's requirement of a lod score of equal to or greater than 3.6 is a
prerequisite to suggest significance in linkage; in their review, Moldin and
Gottesman make it the requirement for serious consideration.86 They
suggest that schizophrenia phenotype diagnosis is ambiguous, that the
disorder may in fact be 'a heterogeneous group of diseases rather than a
single disease', and that 'The use of over-simplified hypotheses of genetic
aetiology of schizophrenia... have led to reductionistic research strategies
and goals'.87 All these are uncontroversial for the narrative of enlightened
Portin and Alanen conclude by stating that they 'favour an integrative
approach to the study of schizophrenia', and that 'there is a need for
holistic perspectives' in schizophrenia research. But it is the last sentence
of the article that displays the weakness of their critique: 'However,
reductionistic views still prevail in the field'.89These authors have failed to
make clear what the 'reductionistic views' are. The claim that their conclu-
sions regarding the paucity of molecular results are in agreement with
'many psychiatric geneticists' is uncontroversial. Their descriptions of the
limitations of molecular linkage analysis are, if anything, less powerful than
Moldin & Gottesman's or Kendler & Diehl's, if only because they have less
space to devote to it. Even their preference for a multifactoral model with
'many non-specific minor genes which act together' over an oligogenic one
(where 'two or three genes together lead to a predisposition to schizo-
phrenia') is perfectly in tune with the narrative of enlightened geneticiza-
tion.90 Although the tone of Portin and Alanen's article suggests that they
are aiming to challenge a dominant, reductionistic position in psychiatric
genetics, the information they use and the assumptions they accept un-
questioningly suggest that they are also using this narrative. Part of the
strength of the narrative of enlightened geneticization then lies in this co-
option of opponents; it appeals to those who ostensibly oppose its core
theme, the dominant role of genetic factors in schizophrenia causation.
While the methodology used in this paper relies on a small sample size
rather than an extensive coverage of the literature, the strategies and
narrative presented can be found elsewhere in that literature, and are not
restricted to the three reviews analysed. I suggest that picking any review of
schizophrenia genetics will display some or all of these features. For
example, the 1999 review by Ming Tsuang, William Stone and Stephen
Faraone presents: the results of twin and adoption studies ('The majority
of studies show that schizophrenia runs in families'); the schizophrenia
spectrum ('genetic studies provide evidence for a spectrum of disorders
that are similar to schizophrenia and are caused by the same genes'); the
use of genetic (including single-gene) models ('Mixed models containing
coexisting SML and MFP components... have not yet determined mode
of transmission but may have value'); caution and responsibility over the
results of association and linkage studies, and a positive future direction
902 Social Studies of Science 31/6

('although much more work will be needed in the laboratory, progress in

linkage studies brings us closer to the goal of identifying and cloning
disease genes for the ultimate purpose of developing more-effective and
more-specific treatments for schizophrenia').91A more recent review, from
2001, employs the strategies of 'use of history', 'genetic models' and
'responsibility and caution'.92
I have claimed that the narrative of enlightened geneticization is the
dominant story told about genes and schizophrenia by geneticists. It is not
clear that it is dominant within schizophrenia discourse as a whole. Any
reading of the clinical literature on this disease suggests that there are large
gaps which are inadequately addressed in these review articles. The pre-
sentation of schizophrenia as a developmentaldisorder is largely ignored, as
is research into the medical imaging of the brains of people with schizo-
phrenia,93 and the idea of epigenesis in schizophrenia.94 Perhaps this lack
of a 'bigger picture' is understandable; these reviews are explicitly focused
on the role of genetics in schizophrenia aetiology. Yet the failure to put
genetic explanations within a broader framework is perfectly in keeping
with the narrative of enlightened geneticization, which avoids detailed
discussion of specific non-genetic causes, if at all possible.
In terms of my methodological aims, this paper shows what a rich
resource for discourse analysis the empiricist repertoire is, and how fruitful
examination of scientific controversy can be carried out, even if one
focuses on only one side of a debate. More specifically, my research adapts
and extends Greg Myers' approach to scientific texts, and supports his
position regarding the central role review articles play in the construction
of scientific facts and the shaping of disciplines. My work also extends the
analysis by Anne Kerr and her colleagues of geneticists' discourse, showing
how even when the nature/nurture distinction is rejected, genetic explana-
tions are prioritized.
At a broader level, a balanced approach to the nature/nurture relation-
ship is claimed, with implications for those who criticize molecular genetic
explanations of schizophrenia, just as Rose, Lewontin and Kamin opposed
twin and adoption studies. The enlistment of Portin and Alanen shows the
subtlety and strength of the narrative of enlightened geneticization. Simply
to criticize modern molecular genetic research in schizophrenia as 'deter-
ministic' is to ignore the rl1e it claims to offer to non-genetic factors. To
suggest that there is no single gene 'for' schizophrenia in the belief that this
undermines molecular approaches is to miss the point that genetic re-
searchers are the first to admit this.
Beyond this is the role my research plays in the debates surrounding
the concept of 'geneticization'. This paper shows how a particular defini-
tion of the concept of geneticization can be used to guide and inform
empirical research in the social sciences. My definition focuses attention on
attempts to link schizophrenia to specific stretches of DNA, rather than
larger debates about inheritance or broader topics such as evolutionary
psychology. It makes us think about the way schizophrenia is constructed
in genetic terms and hence geneticized, instead of raising issues about the
Hedgecoe: Schizophrenia and the Narrative of Enlightened Geneticization 903

broader social effect of this geneticization. There is the obvious question of

how enlightened is the narrative of enlightened geneticization. The authors
of these reviews are aware of the controversies and the previous mistakes
surrounding schizophrenia genetics. Yet the undermining of non-genetic
explanations suggests that enlightened geneticization still has a lot in
common with Abby Lippman's original definition in terms of a core
determinism. Hopefully the term enlightenedgeneticization highlights the
sophisticated way in which this narrative is constructed, the apparent
acceptance of environmental causation, and the careful avoidance of
This paper contributes, in a small way, to discussions concerning the
role of the social sciences in medical ethics and bioethics, and the need for
philosophical discussions of the ethics of new technologies to incorporate
empirical research.95 This becomes clearer if we return to Lippman's
original conception of geneticization as ethically loaded and implicitly
critical of genetic explanations in medicine; any discussion needs to note
that geneticists have begun to engage directly with the topic of geneticiza-
tion. In their piece in the Special Issue of the journal Science devoted to
'The Human Genome', Peter McGuffin, Brien Riley and Robert Plomin
note that:
[I]t has sometimesbeen suggestedthat geneticizationis likelyto increase
the stigma of mental disorders.To the contrary,far from increasingthe
stigma,advancesin geneticshavethe oppositeeffect.As a case in point, it
is now perfectlyacceptablefor an ex-presidentof the United States and
his family to acknowledgethat he has Alzheimer'sdisease... We predict
that this is the startof a trend...96
Kendler & Diehl (1993) also mention this 'anti-stigma' role for geneticiza-
tion in schizophrenia in their explanation for why molecular results are so
often hyped by the media. And, in the case of schizophrenia, there is some
evidence to support the idea that genetic explanations may reduce stigma-
tization of sufferers. As Jon Turney and Jill Turner show, many people with
schizophrenia (and their families) actually find some comfort in the idea of
a genetic cause for the condition.97 As well as this kind of empirical testing
of conventional wisdom, the social sciences can challenge the basic as-
sumptions made by bioethicists when they try to assess the ethical and
social impact of genetic technologies. Although I am critical of Abby
Lippman's position, I agree with her when she states that:
Regrettably... most of the critiquesof the 'new' genetics have concen-
tratedon the potentiallyharmfulconsequencesof using the information
that genetic projects... will provide... Though necessary,these critiques
are not sufficient;they are certainlynot primary.They appearto take
furtherdevelopmentand use of genetictechnology,and of geneticization,
as faits accomplis... Thus, even when the hyperbolicclaims made for
genetics are shown to be unsubstantiated,the tendency to see 'genes as
the answer'persists.98

My work contributes to the approach, labelled 'critical bioethics',99 which

seeks to refocus ethical attention on genetic science itself, and not only on
904 Social Studies of Science 31/6

its applications. By highlighting the way in which genetic explanations are

constructed, sociologists of science can expand the focus away from simply
debating the merits of particular technologies, and increase awareness of
how ethicists need to consider the process of scientific investigation, rather
than just the applications of science and technology. Although I have
approached these review articles in a symmetrical way, this does not mean
that my conclusions cannot be used in such ethical assessments. When
Greg Myers discusses the politics of discourse analysis, he argues that:

The texts I study are not obviously political, but they are central to the
processes of constructing facts, methods and authority... The political
strategy of discourse analysis is based on the assumption that this skepti-
cism about scientific authority is a good thing.100

In the same way, I argue that a symmetrical approach to geneticization

challenges the authority of scientific statements about the role of genetic
explanation, without Lippman's assumption that such explanations must
be inherently wrong.

I would like to thank Jon Turney, Brian Balmer and John Waller for commenting on early
versions of this paper. It has also benefited from the comments of SSS referees. This
research was supported by the UK Economic and Social Research Council, Postgraduate
Award No. R00429834453.

1. Dorothy Nelkin (ed.), Controversy:Politicsof TechnicalDecisions (London & Newbury

Park, CA: Sage, 3rd edn, 1992); H. Tristram Engelhardt, Jr, and Arthur L. Caplan
(eds), Scientific Controversies:Case Studies in the Resolutionand Closureof Disputes in
Science and Technology(Cambridge: Cambridge University Press, 1987).
2. John Bell, 'The New Genetics and Clinical Practice', British MedicalJournal, Vol. 316
(21 February 1998), 618-20.
3. For reviews of this literature, see Peter Conrad and Jonathon Gabe, 'Sociological
Perspectives on the New Genetics', Sociologyof Health and IllnessVol. 21, No. 5
(September 1999), 505-16; Anne Kerr and Sarah Cunningham-Burley, 'On
Ambivalence and Risk: Reflexive Modernity and the New Human Genetics', Sociology,
Vol. 34, No. 2 (May 2000), 283-304.
4. Joan H. Fujimura, CraftingScience:A Sociohistoryof the Questfor the Geneticsof Cancer
(Cambridge, MA: Harvard University Press, 1996); Sarah Cunningham-Burley and
Anne Kerr, 'Defining the "Social": Towards an Understanding of Scientific and Medical
Discourses on the Social Aspects of the New Human Genetics', Sociologyof Health and
Illness,Vol. 21, No. 5 (September 1999), 647-68; Elizabeth Ettorre, 'Experts as
"Storytellers" in Reproductive Genetics: Exploring Key Issues', ibid., 539-59; Jon
Turney and Jill Turner, 'Predictive Medicine, Genetics, and Schizophrenia', New Genetics
and Society,Vol. 19, No. 1 (2000), 5-22.
5. Anne Kerr, Sarah Cunningham-Burley and Amanda Amos, 'The New Genetics:
Professionals' Discursive Boundaries', The SociologicalReview,Vol. 35, No. 2 (May
1997), 279-303.
6. Anne Kerr, Sarah Cunningham-Burley and Amanda Amos, 'Eugenics and the New
Genetics in Britain: Examining Contemporary Professionals' Accounts', Science,
Technology,& Human Values,Vol. 23, No. 2 (Spring 1998), 175-98, at 188-89.
7. Adam Hedgecoe, Narratives of Geneticization:Cystic Fibrosis,Diabetes and Schizophrenia
(unpublished PhD thesis, University of London, 2000).
Hedgecoe: Schizophrenia and the Narrative of Enlightened Geneticization 905

8. Abby Lippman, 'Prenatal Genetic Testing and Screening: Constructing Needs and
Reinforcing Inequalities', AmericanJournal of Law and Medicine,Vol. 17, Nos 1 & 2
(1991), 15-50, esp. 16-17.
9. Ruth Hubbard and Elijah Wald, Explodingthe Gene Myth (Boston, MA: Beacon Press,
1992); Melanie Rock, 'Genetic Norms, Eugenic Logic and UNESCO's International
Bioethics Committee', EubiosJournal of Asian and InternationalBioethics,Vol. 7, No. 4
(1997), 108-10; Sally MacIntyre, 'Social and Psychological Issues Associated with the
New Genetics', PhilosophicalTransactionsof the Royal Society of London: Biological
Sciences,Vol. 352, Issue 1357 (29 August 1997), 1095-101; J. Fitzgerald, 'Geneticizing
Disability: The Human Genome Project and the Commodification of Self', Issues in
Law and Medicine,Vol. 14, No. 2 (Fall 1998), 147-63; Rogeer Hoedemaekers and
HenkTen Have, 'Geneticization: The Cyprus Paradigm', Journal of Medicine and
Philosophy,Vol. 23, No. 2 June 1998), 274-87; Jose Van Dijk, Imagenation:Popular
Images of Genetics(NewYork: NewYork University Press, 1998); Alan H. Goodman,
'Epistemology of Complex Traits in an Age of Geneticization', CurrentPsychologyof
Cognition,Vol. 18, No. 2 (April 1999), 176-81; Sue Sherwin and Christy Simpson,
'Ethical Questions in the Pursuit of Genetic Information: Geneticization and BRCA1',
in Alison Thompson and Ruth Chadwick (eds), GeneticInformation:Acquisition,Access
and Control(NewYork: Kluwer Academic/Plenum Publishing, 1999), 121-28; John M.
Opitz, 'Afterword:The Geneticization of Western Civilization: Blessing or Bane?', in
Phillip R. Sloan (ed.), ControllingOur Destinies:Historical, Philosophical,Ethical, and
TheologicalPerspectiveson the Human GenomeProject(Notre Dame, IN: University of
Notre Dame Press, 2000), 429-50. Other authors who address similar concerns but
who use different terminology include: Troy Duster, Backdoorto Eugenics (NewYork &
London: Routledge, 1990); Dorothy Nelkin and M. Susan Lindee, The DNA Mystique:
The Gene as a CulturalIcon (NewYork: Freeman, 1995); Steven Rose, Life Lines:
Biology, Freedom,Determinism(London: AlenLane, 1997); Barbara K. Rothman,
GeneticMaps and Human Imaginations:The Limits of Science in UnderstandingWhoWeAre
(NewYork & London: Norton 1998).
10. Abby Lippman, 'The Politics of Health: Geneticization versus Health Promotion', in
Sue Sherwin (ed.), The Politics of Women'sHealth: ExploringAgency and Autonomy
(Philadelphia, PA: Temple University Press, 1998), 64-82, at 64.
11. Obviously, the idea of a genetic 'link' is rather vague, but it covers a range of different
possible scientific claims that could be made, ranging from a simple correlation
between certain people with schizophrenia and a particular piece of DNA, through the
cloning of a gene, to the idea of gene therapy. This broadness of the idea of a 'genetic
link' is part of its appeal as an analytic term.
12. Sherwin & Simpson, op. cit. note 9, 123.
13. Abby Lippman, 'The Genetic Construction of Prenatal Testing: Choice, Consent or
Conformity for Women?', in Karen H. Rothenberg and Elizabeth J. Thomson (eds),
Womenand PrenatalTesting:Facing the Challengesof GeneticTesting(Columbus: Ohio
State University Press, 1994), 9-34. For critiques of current uses of 'geneticization',
see: Celeste Condit and Melanie Williams, 'Audience Responses to the Discourse of
Medical Genetics: Evidence Against the Critique of Medicalization', Health
Communication,Vol. 9, No. 3 (1997), 219-35; C.M. Condit, Nneka Ofulue and
Kristine M. Sheedy, 'Determinism and Mass-Media Portrayals of Genetics', American
Journal of Human Genetics,Vol. 62, No. 4 (April 1998), 979-84; Adam Hedgecoe,
'Geneticization, Medicalization and Polemics', Medicine, Healthcareand Philosophy:A
EuropeanJournal, Vol. 1, No. 3 (1998), 235-43; C.M. Condit, The Meanings of the
Gene: Public Debates about Human Heredity (Madison: University of Wisconsin Press,
1999); C.M. Condit, 'How the Public Understands Genetics: Non-Deterministic and
Non-Discriminatory Interpretations of the "Blueprint" Metaphor', Public Understanding
of Science,Vol. 8, No. 3 July 1999), 169-80; A. Hedgecoe, 'Reconstructing
Geneticization: A Research Manifesto', Health Law Journal, Vol. 7 (1999), 5-18; Carlos
Novas and Nikolas Rose, 'Genetic Risk and the Birth of the Somatic Individual',
Economyand Society,Vol. 29, No. 4 (November 2000), 485-513.
906 Social Studies of Science 31/6

14. For a historical introduction to the concept of molecularization, see Soraya De

Chadarevian and Harmke Kamminga (eds), MolecularizingBiology and Medicine: New
PracticesandAlliances, 1910s-1970s (Amsterdam: Harwood, 1998).
15. Condit & Williams, op. cit. note 13, 220.
16. Henk ten Have, 'Genetics and Culture: The Geneticization Thesis', Medicine, Health
Care and Philosophy,Vol. 4, No. 3 (2000), in press; for a defence of the sociological use
of 'geneticization' (in the same issue), see A. Hedgecoe, 'Ethical Boundary Work:
Geneticization, Philosophy and the Social Sciences', ibid., in press.
17. G. Nigel Gilbert and Michael Mulkay, OpeningPandora'sBox: A SociologicalAnalysis of
Scientists'Discourse(Cambridge: Cambridge University Press, 1984). For other work on
the discourse of science, see: John Law and R.J. Williams, 'Putting Facts Together: A
Study of Scientific Persuasion', Social Studies of Science,Vol. 12, No. 4 (November
1982), 535-58; Ann M. Blackeslee, 'The Rhetorical Construction of Novelty:
Presenting Claims in a Letters Forum', Science, Technology,& Human Values,Vol. 19,
No. 1 (Winter 1994), 88-100; Donald A. Dewsbury, 'Rhetorical Strategies in the
Presentation of Ethology and Comparative Psychology in Magazines after WorldWar
II', Science in Context,Vol. 10, No. 2 (Summer 1997), 367-86; Benoit Godin, 'The
Rhetoric of a Health Technology: The Microprocessor Patient Card', Social Studies of
Science,Vol. 27, No. 6 (December 1997), 865-902.
18. Greg Myers, WritingBiology:Textsin the Social Constructionof ScientificKnowledge
(Madison: University of Wisconsin Press, 1990); G. Myers, 'Stories and Styles in Two
Molecular Biology Review Articles', in Charles Bazerman and James Paradis (eds),
TextualDynamics of the Professions(Madison: University of Wisconsin Press, 1991),
45-75; Christiane Sinding, 'Literary Genres and the Construction of Knowledge in
Biology: Semantic Shifts and Scientific Change', Social Studies of Science,Vol. 26, No. 1
(February 1996), 43-70.
19. Greg Myers, 'Making a Discovery: Narratives of Split Genes', in Christopher Nash
(ed.), Narrative in Culture:The Uses of Storytellingin the Sciences,Philosophy,and
Literature(London: Routledge, 1990), 102-25.
20. Jan Golinski, Making Natural Knowledge:Constructivismand the History of Science
(Cambridge: Cambridge University Press, 1998), 130.
21. Myers (1990), op. cit. note 18, 101-40, 193-246.
22. Myers, op. cit. note 19.
23. Cynthia R. Haller, 'Foucault's Archaeological Method and the Discourse of Science:
Plotting Enunciative Fields', in John T. Battalio (ed.), Essays in the Study of Scientific
Discourse:Methods, Practiceand Pedagogy (Stamford, CT & London: Ablex Publishing,
1998), 53-69, at 64.
24. From 'Editorial Mission' of SchizophreniaBulletin on the inside cover of each issue, and
on its web page: < http://www.medscape.comlgovmt/NIMHSchizophreniaBulletin/public
about.SB. html>.
25. Irving I. Gottesman and James Shields, 'A Critical Review of Recent Adoption, Twin,
and Family Studies of Schizophrenia: Behavioral Genetics Perspectives', Schizophrenia
Bulletin,Vol. 2, No. 3 (1976), 360-401. This article is heavily cited (89 citations in the
Science Citation Index?), with the latest mention in April 2001.
26. Cited 72, 196 and 34 times respectively on the Science Citation IndexTMdatabase. These
figures do not 'prove' that these articles are influential, of course, but they do suggest a
good size of readership, as well as a degree of current importance; all three have been
cited since mid-2000.
27. Matthew McGue and Irving I. Gottesman, 'Genetic Linkage in Schizophrenia:
Perspectives from Genetic Epidemiology', SchizophreniaBulletin,Vol. 15, No. 3 (1989),
28. Kenneth S. Kendler and Scott R. Diehl, 'The Genetics of Schizophrenia: A Current
Genetic-Epidemiologic Perspective', SchizophreniaBulletin,Vol. 19, No. 2 (1993),
Hedgecoe: Schizophrenia and the Narrative of Enlightened Geneticization 907

29. Steven O. Moldin and Irving I. Gottesman, 'Genes, Experience and Chance in
Schizophrenia-Positioning for the 21st Century', SchizophreniaBulletin,Vol. 23, No. 4
(1997), 547-61.
30. It is important to reiterate that this kind of discourse analysis does not give one access
to the reasons why the narrative is being constructed in this way, only how: 'the focus is
on the process of construction by which elements of language are put together into
meaningful stories... The artificiality of the narratives thereby constructed is exposed
by this analysis, but the role (if any) that social interests play in the process remains
obscure' (Golinski, op. cit. note 20, 132). This may be unsatisfactory for some readers
who wish to determine the power structure or interests that underlie such discourse,
and one could easily suggest areas that might be worth examining. The r6le of the
pharmaceutical industry in encouraging genetic explanations is one possible topic.
Another is the issue of the professional interests involved, and the disciplinary
dominance of biological psychology. Finally, one might judge the motivations of people
whose entire careers are based on the existence of genetic aetiology in schizophrenia.
But, interesting as these topics are, they are not the subject of this paper.
31. Thus, in terms of my stripped-down definition, geneticization has not yet occurred.
32. Robin Sherrington, Jon Brynjolfsson, Hannes Petursson, Mark Potter, Keith
Dudleston, Brian Barraclough, John Wasmuth, Mark Dobbs and Hugh Gurling,
'Localization of a Susceptibility Locus for Schizophrenia on Chromosome 5', Nature,
Vol. 336 (10 November 1988), 164-67.
33. J. Kirsty Millar, Julie C. Wilson-Annan, Susan Anderson, Sheila Christie, Martin S.
Taylor, Colin A.M. Semple, Rebecca S. Devon, David M. St Clair, Walter J. Muir,
Douglas H.R. Blackwood and David J. Porteous, 'Disruption of Two Novel Genes by a
Translocation Cosegregating with Schizophrenia', Human Molecular Genetics,Vol. 9,
No. 9 (May 2000), 1415-23.
34. Controversy over schizophrenia begins with disputes over the idea of whether there
should even be a diagnostic category called 'Schizophrenia'. However, the aim of this
paper is not to analyse the social construction of 'schizophrenia' but of 'schizophrenia
genetics': see Mary Boyle, Schizophrenia:A Scientific Delusion? (London & NewYork:
Routledge, 1990).
35. Ibid., 118.
36. J. Richard Marshall, 'The Genetics of Schizophrenia: Axiom or Hypothesis', in Richard
P. Bentall (ed.), ReconstructingSchizophrenia(London & New York: Routledge, 1990),
89-117, at 89 & 91.
37. Gottesman & Shields, op. cit. note 25, 361.
38. These studies are discussed by Gottesman and Shields, ibid. Concordance is a measure
of the proportion of index subjects whose wife/husband/brother/twin shares a particular
characteristic with them - for example, the concordance between blue-eyed men who
have blue-eyed wives. MZ (monozygotic) twin concordance rate is the rate of incidence
of a condition in identical twins. It is normally expressed as a percentage, and is felt to
indicate the degree of genetic 'control' of a condition. For example, if, out of 100 twin
pairs, where one twin in each pair was schizophrenic, only 28 of the other twins were
also diagnosed as schizophrenic, then the concordance rate would be 28%.
39. This problem is avoided in studies such as those carried out in Scandinavia, which
access twin registers and then track down twins who are diagnosed with schizophrenia.
40. 'Fischer used these tests in only 39 percent of pairs, Gottesman and Shields in 50
percent and Kringlen in 71... The problem of establishing zygosity is confounded by
the fact that subjects included in a research sample may actually be dead at the time
the research is carried out': Boyle, op. cit. note 34, 123.
41. For the geneticists' position, see Gottesman & Shields, op. cit. note 25; Irving I.
Gottesman and James Shields, 'Rejoinder: Toward Optimal Arousal and Away from
Original Din', SchizophreniaBulletin,Vol. 2, No. 3 (1976), 447-53; Paul H. Wender,
'The Danish American Studies: Critiquing the Critique', AmericanJournal of Psychiatry,
Vol. 138, No. 10 (October 1981), 1392-93. For the critics' position, see J. Richard
Marshall and A.N. Pettitt, 'Discordant Concordant Rates', Bulletin of the British
908 Social Studies of Science 31/6

PsychologicalSociety,Vol. 38 January 1985), 6-9; Marshall, op. cit. note 36; Boyle, op.
cit. note 34.
42. Theodore Lidz, Sidney Blatt and Barry Cook, 'Critique of the Danish-American
Studies of the Adopted-Away Offspring of Schizophrenic Parents', AmericanJournal of
Psychiatry,Vol. 138, No. 8 (August 1981), 1063-68; Boyle op. cit. note 34.
43. Steven Rose, Richard C. Lewontin and Leon J. Kamin, Not in Our Genes:Biology,
Ideologyand Human Nature (London: Penguin Books, 1984).
44. Including, for example, Lidz, Blatt & Cook, op. cit. note 42.
45. Marshall & Pettitt, op. cit. note 41; Marshall, op. cit. note 36; Boyle, op. cit. note 34,
46. Rose, Lewontin & Kamin, op. cit. note 43, 224.
47. Gottesman & Shields, op. cit. note 25.
48. Irving I. Gottesman, Peter McGuffn and Anne E. Farmer, 'Clinical Genetics as Clues
to the "Real" Genetics of Schizophrenia: (A Decade of Modest Gains While Playing for
Time)', SchizophreniaBulletin,Vol. 13, No. 1 (1987), 23-47.
49. Lidz, Blatt & Cook, op. cit. note 42; Theodore Lidz and Sidney Blatt, 'Critique of the
Danish-American Studies of the Biological and Adoptive Relatives of Adoptees who
became Schizophrenic', American Journal of Psychiatry,Vol. 140, No. 4 (April 1983),
50. Rose, Lewontin & Kamin, op. cit. note 43.
51. Boyle, op. cit. note 34; Marshall, op. cit. note 36.
52. Gottesman & Shields, op. cit. note 25, 361.
53. Petter Portin andY.O. Alanen, 'A Critical Review of Genetic Studies of Schizophrenia.
I. Epidemiological and Brain Studies', Acta PsychiatricaScandinavica, Vol. 95, No. 1
January 1997), 1-5.
54. For example, the SchizophreniaBulletin is still publishing twin studies in 2001: Angus
W. MacDonald III, Michael F. Pogue-Geile, Thomas T. Debski and Stephen Manck,
'Genetic and Environmental Influences on Schizotypy: A Community Based Twin
Study', SchizophreniaBulletin,Vol. 27, No. 1 (2001), 47-58. Readers should note that
the critiques of twin and adoption studies (for example: Rose et al.; Lidz et al.; or
Marshall) are not cited in the review articles I analyse in this paper. There is no
suggestion that there is, or has been, sustained and informed criticism of recent
research into schizophrenia. But of course, ignoring your opponents' research is one of
the strongest tactics to adopt in a controversy: see Bruno Latour, Science in Action: How
to Follow Scientistsand EngineersThroughSociety (Cambridge, MA: Harvard University
Press, 1987), 40-41.
55. Gordon J. Edlin, 'Inappropriate Use of Genetic Terminology in Medical Research: A
Public Health Issue', Perspectivesin Biology and Medicine,Vol. 31, No. 1 (Autumn
1987), 47-56, at 49.
56. Hubbard & Wald, op. cit. note 9; Abby Lippman, 'Led (Astray) by Genetic Maps: The
Cartography of the Human Genome and Health Care', Social Science and Medicine,Vol.
35, No. 12 (1997), 1469-76. See also associated authors such as Duster, op. cit. note
57. See: Thomas A. Widiger and Lizabeth M. Sankis, 'Adult Psychopathology: Issues and
Controversies', Annual Review of Psychology,Vol. 51 (2000), 377-404; March 2001
issue of The Psychologist(Vol. 14, No. 3), which contains peer commentary on the need
for psychologists to adapt to a genetic future; or the issue-length peer commentary on
Rose's book Lifelines(op. cit. note 9) in Behavioral and Brain Sciences,Vol. 22, No. 5
(1999), 871-907. Rose has maintained a steady critique of schizophrenia genetics: see
Steven Rose, 'Neurogenetic Determinism and the New Euphenics', British Medical
Journal, Vol.318 (29 May 1998), 1707-08; S. Rose, 'Moving On from Old
Dichotomies: Beyond Nature-Nurture Towards a Lifeline Perspective', British Journal of
Psychiatry,Vol. 178, Suppl. 40 (2001), S3-S7.
58. On a reflexive note, it is only fair to acknowledge the rhetorical, discursive nature of
this paper and my own construction (through symmetrical analysis) of a position mid-
way between genetic 'hype' and radical criticism. Just as the authors of the review
Hedgecoe: Schizophrenia and the Narrative of Enlightened Geneticization 909

articles adopt a strategy of caution and responsibility to show how trustworthy and
believable their claims are, my moderate position with regard to the extent to which
geneticization is happening, and my refusal to assume that geneticization is a negative
process, strengthen my empirical claims and position as an 'objective' commentator.
On a more basic level, this paper is, of course, structured to persuade. For example,
rather than a repetitious working through of the three review articles in a chronological
order, I present each of the strategies used and show how they occur. This thematic
approach allows the reader to see how each strategy occurs consistently over time,
making my argument more persuasive. It also cuts down on the risk of boring the
reader. I also leave most of the quotes I use as they appear in the original text, avoiding
capitalization to add an air of authenticity.
That said, the kind of discourse analysis developed by Greg Myers keeps reflexivity
at arm's length, and this is a stance which I adopt in this paper. While I could provide a
running commentary on the rhetorical nature of my own text, I am not sure how this
advances my argument about the geneticization of schizophrenia. Since the main aim of
this paper is not a methodological/theoretical one (with regard to discourse analysis),
but is an empirical one (with regard to geneticization), reflexive comments will be kept
to a minimum.
59. This is, of course, only one interpretation of these texts, but it is one that arises out of
the analysis of the articles rather than being a prior assumption, for which selective
evidence was chosen. There are two obvious, related, objections to the idea of the
narrative of enlightened geneticization. The first is that the reader is not convinced by
the textual evidence that the narrative is as I claim. The second is that there are other,
more convincing interpretations of these articles, for example that geneticists adopt
different strategies in a flexible, varying manner as and when they need to. This latter
objection loses much of its force if we consider Stanley Cavell's point that by definition
there must be alternative interpretations, if my interpretation is to exist at all: Stanley
Cavell, Pursuit of Happiness:The HollywoodComedyof Remarriage(Cambridge, MA &
London: Harvard University Press, 1981), 36-37. The onus is on the critic to present
alternative interpretations in as much detail as I present mine. And it is this detail that
I hope neutralizes the first objection; I hope that the excerpts from the articles that I
present will convince that reader that I have given a fair reading to the authors
concerned, and have provided a comprehensive description of the narrative of
enlightened geneticization.
60. Gottesman, McGuffin & Farmer, op. cit. note 48, 23. By borrowing the term
'enlightened', I am not making any assumptions about the nature of the narrative,
merely adopting one of the scientists' own terms. That said, as I suggest below, the
narrative is framed in politically liberal (or, at least, non-authoritarian) terms, and can
thus be seen as 'enlightened'.
61. Portin & Alanen, op. cit. note 53; Petter Portin andY.O. Alanen, 'A Critical Review of
Genetic Studies of Schizophrenia. II. Molecular Genetic Studies', Acta Psychiatrica
Scandinavica, Vol. 95, No. 2 (February 1997), 73-80.
62. Kerr, Cunningham-Burley & Amos (1998), op. cit. note 4, 182.
63. Philip Kitcher, The Lives to Come: The GeneticRevolutionand Human Possibilities
(London: AllenLane, 1996), 205-19.
64. Wender, op. cit. note 41, 1393.
65. The first half of this statement might be saying that the adoption studies were only
designed to assess genetic factors, but the claim that none of the data from these
studies is at all in accordance with the environment making contributions to the
aetiology of schizophrenia would seem to belie that reading.
66. Gottesman & Shields, op. cit. note 25, 367.
67. Gottesman & Shields, op. cit. note 41.
68. For further discussion of this point, see Tumey & Turner, op. cit. note 4.
69. Gottesman & Shields, op. cit. note 41, 447.
70. This is an example of the way in which review articles reconstruct the history of a
discipline, to present a smooth, unruffled image: see Myers (1991), op. cit. note 18.
910 Social Studies of Science 31/6

71. It is possible to have high heritability figures from modest concordance rates. This is
another bone of contention between genetic researchers and their critics: see
Gottesman & Shields, op. cit. note 25, 377, and Seymour Kessler, 'Progress and
Regress in the Research on the Genetics of Schizophrenia', SchizophreniaBulletin,Vol.
2, No. 3 (1976), 434-39.
72. In particular, a study conducted by Pekka Tienari and published in 1991: P. Tienari,
'Interaction between Genetic Vulnerability and Family Environment: The Finnish
Adoptive Family Study of Schizophrenia', Acta PsychiatricaScandinavica,Vol. 84, No. 5
(November 1991), 460-65.
73. On the use of irony in the controversy surrounding Edward Wilson's book, Sociobiology,
see Myers (1990), op. cit. note 18, 214-38.
74. In the GSL model, the gene cannot be highly penetrant (powerful), otherwise far more
people would be diagnosed with schizophrenia. But a not very penetrant single gene
cannot explain the high MZ concordance rates.
75. Phenocopies are conditions similar to schizophrenia but which arise from a different
cause, for example drug misuse.
76. Segregation analysis uses large pedigrees or families of affected individuals, but looks at
the disease phenotype, rather than genetic markers involved.
77. One should also note that such results, while they might serve to explain discordance in
MZ twins, do not account for 'spontaneous' schizophrenia, unless it is implied that it is
merely a less common version of the familial genetic factor. The authors seem to
assume that what holds for schizophrenia in twins also holds for schizophrenia in single
persons - one of the underlying assumptions of twins studies as a whole.
78. Gottesman & Shields, op. cit. note 41, 447.
79. No evidence is offered for the destigmatizing effect of schizophrenia genetics.
80. See, for example, Rose (2001), op. cit. note 57; R. Grant Steen, DNA and Destiny:
Nature and Nurture in Human Behavior (New York & London: Plenum Press, 1996).
81. Cunningham-Burley & Kerr (1999), op. cit. note 4, 656-57.
82. This makes sense, especially in the light of the results of the Tienari study, which
suggested that family environment might play a part in schizophrenia aetiology. Of the
seven studies used to examine the second hypothesis ('liability to schizophrenia and
schizophrenia-like personality disorders'), only two are adoption studies while the
remaining five are family studies; and family studies do not provide the sort of
information with which to separate genetic from environmental effects.
83. A.R. Sanders, J.D. Hamilton, W.E. Fann and P.I. Patel, 'Association of Genetic
Variation at the Porphobilinogen Deaminase Gene with Schizophrenia' [Abstract No.
2011], AmericanJournal of Human GeneticsSupplement, Vol. 49, No. 4 (October 1991),
84. Peter Propping, 'Genetic Disorders Presenting as "Schizophrenia": Karl Bonhoeffer's
Early View of the Psychoses in the Light of Medical Genetics', Human Genetics,Vol. 65
(1983), 1-10.
85. Portin & Alanen (February 1997), op. cit. note 61, 75-76.
86. Although both Portin & Alanen and Moldin & Gottesman cite the same article, the
latter use the lod score of 3.30 to denote significance, which is more accurate. The lod
score of 3.6 is mentioned by Lander and Kruglyak, but refers to siblings and half-
siblings only, rather than full family studies: see Eric Lander and Leonid Kruglyak,
'Genetic Dissection of Complex Traits: Guidelines for Interpreting and Reporting
Linkage Analysis', Nature Genetics,Vol. 11, No. 3 (November 1995), 241-47.
87. Portin & Alanen (February 1997), op. cit. note 61, 77.
88. It is ironic that, to support this point, Portin and Alanen cite a paper co-authored by
Peter McGuffin who, in a 1994 article, suggests that schizophrenia is solely caused by
genetic factors, and that any apparently environmental input is actually due to random
variations in gene function and structure: Peter McGuffin, Philip Asherson, Michael
Owen and Anne Farmer, 'The Strength of the Genetic Effect: Is There Room for an
Environmental Influence in the Aetiology of Schizophrenia?', British Journal of
Psychiatry,Vol. 164, No. 5 (May 1994), 593-99.
Hedgecoe: Schizophrenia and the Narrative of Enlightened Geneticization 911

89. Portin & Alanen (February 1997), op. cit. note 61, 77.
90. Ibid., 76.
91. Ming T. Tsuang, William S. Stone and Stephen V. Faraone, 'Schizophrenia: A Review
of Genetic Studies', Harvard Review of Psychiatry,Vol. 7, No. 4 (November/December
1999), 185-207, quotes at (in turn) 186, 188, 192 & 196-97.
92. Bryan J. Mowry and Derek J. Nancarrow, 'Molecular Genetics of Schizophrenia',
Clinical and ExperimentalPharmacologyand Physiology,Vol. 28, No. 1-2 (January
2001), 66-69.
93. Susan K. Schultz and Nancy C. Andreasen, 'Schizophrenia', The Lancet,Vol. 353 (24
April 1999), 1425-30.
94. Arturas Petronis, Andrew D. Paterson and James L. Kennedy, 'Schizophrenia: An
Epigenic Puzzle', SchizophreniaBulletin,Vol. 25, No. 4 (1999), 639-55.
95. See: Barry Hoffmaster (ed.), Bioethics in Social Context (Philadelphia, PA: Temple
University Press, 2000); Raymond DeVries and Peter Conrad, 'Why Bioethics Needs
Sociology', in R. DeVries and Janardan Subedi (eds), Bioethics and Society:
Constructingthe Ethical Enterprise(Upper Saddle River, NJ: Prentice Hall, 1998),
233-57; Barry Hoffmaster, 'Can Ethnography Save the Life of Medical Ethics?',
Social Science and Medicine,Vol. 35, No. 12 (1992), 1421-31; Arthur Kleinman,
'Moral Experience and Ethical Reflection: Can Ethnography Reconcile Them? A
Quandary for "The New Bioethics"', Daedalus, Vol. 128, No. 4 (1999), 69-97.
96. Peter McGuffin, Brien Riley and Robert Plomin, 'Future Directions: Genomics and
Behavior: Toward Behavioral Genomics', Science,Vol. 291 (16 February 2001), 1232,
1249, at 1249.
97. Turney & Turner, op. cit. note 4.
98. Abby Lippman, 'Geneticization and the Canadian Biotechnology Strategy:The
Marketing of Women's Health', in Fiona Miller, Lorna Weir, Roxanne Mykitiuk,
Patricia Lee, Susan Sherwin and Sari Turner (eds), The Genderof GeneticFutures:The
Canadian BiotechnologyStrategy,Womenand Health, Proceedings of a National
Strategic Workshop,York University, 11-12 February 2000 (Toronto: York University,
2000), 32-40, at 36.
99. Lisa S. Parker, 'Breast Cancer Screening and Critical Bioethics' Gaze', Journal of
Medicine and Philosophy,Vol. 20 June 1995), 313-37.
100. Myers (1990), op. cit. note 18, 250-51.

Adam Hedgecoe is a Wellcome Trust Postdoctoral Research Fellow at the

Department of Science and Technology Studies, University College London,
where he also studied for his PhD. Originally trained in bioethics, he has
worked for the technology assessment units of both the European and UK
Parliaments, and is interested in the social construction of disease, the
representation and regulation of genetic technologies, and the relationship
between bioethics and the social sciences. His current research investigates
the role of pharmacogenetics in the genetic reclassification of common
Address: Department of Science and Technology Studies, University College
London, Gower Street, London WC1E 6BT, UK; fax: +44 20 7916 2425; email: