You are on page 1of 473

The Continuum

Companion to Hume

9780826443595_Prelims_Final_txt_print.indd i 2/9/2012 9:38:07 PM

9780826443595_Prelims_Final_txt_print.indd ii 2/9/2012 9:38:07 PM


Alan Bailey
Dan O’Brien

9780826443595_Prelims_Final_txt_print.indd iii 2/9/2012 9:38:07 PM

First published in 2012 by

Continuum International Publishing Group

The Tower Building 80 Maiden Lane

11 York Road Suite 704
London SE1 7NX New York, NY 10038

Companion to Hume
EISBN 978 1 44111 461 7

© Continuum, 2012

British Library Cataloguing-in-Publication Data

A CIP record of this title is available from the British Library

All rights reserved. No part of this publication may be

reproduced, stored in a retrieval system, or
transmitted in any way or by any means, electronic,
mechanical, photocopying, recording or otherwise,
without the written permission of the copyright holder.

Typeset in Sabon by
Newgen Publishing and Data Services
Printed and bound in Great Britain

9780826443595_Prelims_Final_txt_print.indd iv 2/9/2012 9:38:07 PM


List of Contributors vii

Abbreviations for Works Written by Hume ix
Acknowledgements xi
David Hume – A Timeline xiii

Emilio Mazza
Tom Seppalainen and Angela Coventry
Peter Millican
Lorne Falkenstein
Helen Beebee
Alan Bailey
Harold Noonan
Galen Strawson
Constantine Sandis

9780826443595_Prelims_Final_txt_print.indd v 2/9/2012 9:38:07 PM


10. FREE WILL 214

James A. Harris
Duncan Pritchard and Alasdair Richmond
Andrew Pyle
David O’Connor
Julia Driver
Dan O’Brien
Russell Hardin
Lívia Guimarães
Margaret Schabas
M. W. Rowe
Timothy M. Costelloe
Paul Russell

Bibliography 396
Index of Names 437
Index of Topics 441


9780826443595_Prelims_Final_txt_print.indd vi 2/9/2012 9:38:07 PM


Alan Bailey Lorne Falkenstein

Visiting Lecturer in Philosophy Professor of Philosophy
School of Law, Social Sciences and Department of Philosophy
Communications University of Western Ontario
University of Wolverhampton Canada
Lívia Guimarães
Helen Beebee Lecturer in Philosophy
Professor of Philosophy Department of Philosophy
Department of Philosophy Universidade Federal de Minas Gerais
University of Birmingham Brazil
Russell Hardin
Timothy Costelloe Professor of Politics
Associate Professor of Philosophy Department of Politics
Department of Philosophy New York University
College of William and Mary USA
James Harris
Angela Coventry Senior Lecturer in Philosophy
Associate Professor of Philosophy Department of Philosophy
Department of Philosophy St Andrews University
Portland State University UK
Emilio Mazza
Julia Driver Associate Professor
Professor of Philosophy Institute of Human Sciences, Language and
Department of Philosophy Environment
Washington University in St Louis Università IULM, Milan
USA Italy


9780826443595_Prelims_Final_txt_print.indd vii 2/9/2012 9:38:08 PM


Peter Millican Department of Philosophy

Gilbert Ryle Fellow and Professor of University of Edinburgh
Philosophy UK
Hertford College
University of Oxford M.W. Rowe
UK Senior Lecturer in Philosophy
Department of Philosophy
Harold Noonan University of East Anglia
Professor of Philosophy UK
Department of Philosophy
University of Nottingham Paul Russell
UK Professor of Philosophy
Department of Philosophy
Dan O’Brien University of British Columbia
Senior Lecturer in Philosophy Canada
Department of History, Philosophy and
Religion Constantine Sandis
Oxford Brookes University Reader in Philosophy
UK Department of History, Philosophy and
David O’Connor Oxford Brookes University
Professor of Philosophy UK
Department of Philosophy
Seton Hall University Margaret Schabas
USA Professor of Philosophy
Department of Philosophy
Duncan Pritchard University of British Columbia
Professor of Philosophy Canada
Department of Philosophy
University of Edinburgh Tom Seppalainen
UK Associate Professor of Philosophy
Department of Philosophy
Andrew Pyle Portland State University
Reader in Philosophy USA
Department of Philosophy
University of Bristol Galen Strawson
UK Professor of Philosophy
Department of Philosophy
Alasdair Richmond University of Reading
Senior Lecturer in Philosophy UK


9780826443595_Prelims_Final_txt_print.indd viii 2/9/2012 9:38:08 PM


Abs An Abstract of a Book lately Published; Entituled, A Treatise of Human Nature

Cited by paragraph and page number from the texts included in the editions of THN listed

DIS A Dissertation on the Passions

Cited by section and paragraph number from A Dissertation on the Passions; The Natural
History of Religion, The Clarendon Edition of the Works of David Hume, ed. T.L. Beauchamp
(Oxford: Clarendon Press, 2007). Hence DIS 1.2 = Section 1, paragraph 2.

DNR Dialogues concerning Natural Religion, ed. N. Kemp Smith, 2nd edn with supple-
ment (London and Edinburgh: Thomas Nelson, 1947).
Cited by part and page number (e.g. DNR 1.135 = Part 1, page 135).

E Essays, Moral, Political, and Literary, ed. E.F. Miller (Indianapolis: Liberty Classics,
Cited by page number.

EHU An Enquiry concerning Human Understanding, ed. T.L. Beauchamp (Oxford:

Oxford University Press, 1999).
Cited by section and paragraph number, and supplemented by the corresponding page num-
ber in Enquiries concerning Human Understanding and concerning the Principles of Morals,
ed. L.A. Selby-Bigge, 3rd edn with text revised and notes by P.H. Nidditch (Oxford: Clarendon
Press, 1975). Hence EHU 5.1 / 40 = Section 5, paragraph 1 in the Beauchamp edition and
page 40 in the Selby-Bigge edition.

EPM An Enquiry concerning the Principles of Morals, ed. T.L. Beauchamp (Oxford:
Oxford University Press, 1998).
Cited by section and paragraph number, and supplemented by the corresponding page num-
ber in Enquiries concerning Human Understanding and concerning the Principles of Morals,
ed. L.A. Selby-Bigge, 3rd edn with text revised and notes by P.H. Nidditch (Oxford: Clarendon


9780826443595_Prelims_Final_txt_print.indd ix 2/9/2012 9:38:08 PM


Press, 1975). Hence EPM 2.2 / 176–7 = Section 2, paragraph 2 in the Beauchamp edition and
pages 176–7 in the Selby-Bigge edition.

H The History of England, from the Invasion of Julius Caesar to the Revolution in
1688, ed. W.B. Todd, 6 vols (Indianapolis: Liberty Classics, 1983).
Cited by volume and page number (e.g. H1.155 = Volume 1, page 155).

LDH Letters of David Hume, ed. J.Y.T. Greig, 2 vols (Oxford: Clarendon Press, 1932).
Cited by volume, page number, and letter number (e.g. LDH 1.11–12, 5 = Volume 1, pages
11–12, letter 5).

LFG A Letter from a Gentleman to his Friend in Edinburgh, ed. E.C. Mossner and
J.V. Price (Edinburgh: Edinburgh University Press, 1967).
Cited by page number.

MOL My Own Life.

Usually cited by page number from Volume 1 of LDH above.

NHR The Natural History of Religion.

Cited by page or, where applicable, part and page number from Dialogues and Natural
History of Religion, Oxford World’s Classics, ed. J.C.A. Gaskin (Oxford: Oxford University
Press, 1993). Hence NHR 134 = page 134; NHR 1.135 = Part 1, page 135.

NLH New Letters of David Hume, eds. R. Klibansky and E.C. Mossner (Oxford:
Clarendon Press, 1954).
Cited by page and letter number (e.g. NLH 5–6, 3 = pages 5–6, letter 3).

THN A Treatise of Human Nature, eds. D.F. Norton and M.J. Norton (Oxford: Oxford
University Press, 2000).
Cited by book, part, section and paragraph number, and supplemented by the corresponding
page number in A Treatise of Human Nature, ed. L.A. Selby-Bigge, 2nd edn with text revised
and notes by P.H. Nidditch (Oxford: Clarendon Press, 1978). Hence THN / 263 =
Book 1, part 4, section 7, paragraph 1 in the Norton edition and page 263 in the Selby-Bigge

References to the Appendix of the Treatise make use of the abbreviation App. and are then
given by paragraph and page number (eg. THN App. 1 / 623 = paragraph 1 of the Appendix
in the Norton edition and page 623 in the Selby-Bigge edition).
References to the editorial material of the Oxford Philosophical Texts or Clarendon Critical
Edition versions of the Norton edition are by page number and the abbreviations THN-P and
THN-C respectively.

9780826443595_Prelims_Final_txt_print.indd x 2/9/2012 9:38:08 PM


Special thanks to our copy editor, Merilyn Holme, for coaxing and prodding the book to
completion; a far from easy task particularly given, in the final months, the looming festive
season. We are immensely grateful to her and to Sarah Douglas at Continuum for commis-
sioning the project. Thanks are due as usual to the Edwardian Tea Room in the Birmingham
Museum and Art Gallery for refreshments and scholarly ambience, and Birmingham Central
Library provided from its stacks a range of books and journals that would have done credit to
a major university library. We would also like to take this opportunity to express our gratitude
to the following colleagues at Oxford Brookes University, the University of Wolverhampton
and Keele University: Stephen Boulter, Mark Cain, Beverley Clack, Geraldine Coggins, Meena
Dhanda, Giuseppina D’Oro, Cécile Hatier, William Pawlett and Constantine Sandis.
A. B. and D. O’B.
Much of my work on this book has been timetabled around the ongoing DIY house project
that Lucy and I are undertaking. This is something that I am sure Hume would appreciate.
If not hands-on, he was certainly no slouch when it came to the upkeep of the home. When
the ‘[p]laister broke down in the kitchen’ in his house in James’s Court, Edinburgh, he tells us

[his repairman] having thus got into the house, went about teizing Lady Wallace [Hume’s
tenant], and telling her, that this and the other thing was wrong, and ought to be mended.
She told him, as she informs me, that everything was perfectly right, and she wou’d trou-
ble the Landlord for nothing. Yet the Fellow had the Impudence to come to me, and tell
me that he was sent by Lady Wallace to desire that some Stone Pavement under the Coal
Bunker shoud be repair’d. I, having a perfect Confidence in Lady Wallace’s Discretion,
directed him to repair it, as she desir’d. Having got this Authority, which cannot be good
as it was obtain’d by a Lye, he not only pav’d the Bunker anew, but rais’d a great deal of
the other Pavement of the Kitchen and laid it anew, nay, from his own head, took on him
to white-wash the Kitchen: For all which, he brought me in an account of 30 Shillings.
(NLH 115, 206)

The contributors to this book may have explored Hume’s contributions to metaphysics,
morality, religion and epistemology, but I have sympathetic appreciation of his knowledge of


9780826443595_Prelims_Final_txt_print.indd xi 2/9/2012 9:38:08 PM


that great human pursuit of house-building. And my deepest gratitude goes to Lucy my fellow
plasterer, drywaller, spark and plumber and to Dylan who is still ‘patiently’ waiting for the
kitchen to be finished. The skirting boards will be attached next week . . .
D. O’B.
Particular thanks on my behalf go to Ross Singleton for his longstanding friendship and our
many lengthy conversations about religion, philosophy and international politics. Linda Dai
has patiently coped with my tendency to introduce comments about Hume into a quite exces-
sive number of contexts, and her backing and encouragement have played a crucial role in
allowing me to complete my contribution to this volume. I also wish to thank my mother
Dorothy Bailey for her support during the writing and editing process.


9780826443595_Prelims_Final_txt_print.indd xii 2/9/2012 9:38:08 PM


1702 Death of William III and the accession of

Queen Anne
1704 Battle of Blenheim; Isaac Newton, Opticks
1707 Union of England and Scotland
1711 Hume born in Edinburgh
1712 Thomas Newcomen’s steam engine
1713 George Berkeley, Three Dialogues between Hume’s father dies
Hylas and Philonous; Anthony Collins,
Discourse of Free-Thinking
1714 Death of Queen Anne and the accession of
George I
Bernard Mandeville, Fable of the Bees
1715 Jacobite rebellion in Scotland; death of
Louis XIV; Louis XV becomes King of
France at the age of five
1718 Inoculation for smallpox introduced in
1720 Collapse of the South Sea Bubble; Edmond
Halley becomes Astronomer Royal at
1722 Robert Walpole becomes the equivalent of a Hume and his brother John enrol
modern British Prime Minister as students at the University of
1725 Work starts on Grosvenor Square, London;
Francis Hutcheson, The Original of Our
Ideas of Beauty and Virtue
1726 Jonathan Swift, Gulliver’s Travels Hume begins a legal education


9780826443595_Prelims_Final_txt_print.indd xiii 2/9/2012 9:38:08 PM


1727 Death of George I and the accession of

George II; Isaac Newton dies
Robert Greene, Principles of the Philosophy
of the Expansive and Contractive Forces
1728 John Gay, Beggar’s Opera
1729 Hume abandons the idea of
becoming a lawyer; his ‘new Scene
of Thought’
1733 Alexander Pope, Essay on Man
1734 Voltaire, Lettres philosophiques (Lettres Hume travels to London; takes
anglaises) up employment in Bristol as a
merchant’s clerk; resigns or is
dismissed; relocates to France and
works on the Treatise
1735 John Harrison’s chronometer; William Hume moves from Rheims to
Hogarth, Rake’s Progress La Flèche
1736 Joseph Butler, Analogy of Religion
1737 Returns from France to London
1739 Publication of Books One and Two
of A Treatise of Human Nature
1740 Start of the War of Austrian Succession; Book Three of the Treatise
Frederick II (Frederick the Great) becomes
King of Prussia
George Turnbull, Principles of Moral
1741 Samuel Richardson, Pamela Volume I of Essays, Moral and
Political published in Edinburgh
1742 Walpole falls from power; Handel’s
Messiah premieres in Dublin
1745 Second Jacobite rebellion; von Kleist Hume fails to secure the Chair of
discovers the ability of the Leyden jar to Ethics and Pneumatical Philosophy at
store electrical charge Edinburgh; becomes tutor in England
to the Marquess of Annandale
Hume’s mother dies
1746 Jacobite army is decisively defeated at Hume’s appointment as a tutor
Culloden comes to an acrimonious end
Hume becomes Secretary to
Lieutenant-General James St Clair;
accompanies St Clair on a military
expedition attacking the coast of


9780826443595_Prelims_Final_txt_print.indd xiv 2/9/2012 9:38:08 PM


1748 War of Austrian Succession concludes; Travels with St Clair on a

Treaty of Aix-la-Chapelle diplomatic mission to Vienna and
Excavations begin at Pompeii; Turin
Montesquieu, L’Esprit des lois; La Mettrie, Publication of Philosophical Essays
L’Homme machine concerning Human Understanding
1749 Buffon, first volumes of Histoire naturelle; Returns to Scotland and resides
David Hartley, Observations on Man; with his brother and sister at the
Henry Fielding, Tom Jones family home in Ninewells
1751 Diderot and d’Alembert, Volume I of Hume moves to Edinburgh and is
L’Encyclopédie later joined by his sister Katherine
Adam Smith becomes Professor of Logic at Publication of An Enquiry
the University of Glasgow concerning the Principles of Morals
1752 Adoption in Britain of the Gregorian Hume is unsuccessful in his
calendar candidacy for a chair at the
University of Glasgow
Elected Librarian to the Faculty of
Advocates in Edinburgh
1753 British Museum founded
1754 Publication of Volume I of The
History of Great Britain
1755 Lisbon earthquake; Samuel Johnson,
Dictionary of the English Language
1756 Seven Years’ War
1757 Robert Clive and the East India Company Hume resigns from his post as
are victorious at the Battle of Plassey Librarian
Edmund Burke, A Philosophical Enquiry
1758 Philosophical Essays published
under the new name of An Enquiry
concerning Human Understanding
1759 Laurence Sterne, Tristram Shandy;
William Wilberforce is born
1761 The Bridgewater Canal opens from Worsley Madame de Boufflers’s initial letter
to Manchester to Hume
1762 Catherine II becomes Empress of Russia;
Sarah Scott’s novel of a female utopian
community, A Description of Millenium Hall
1763 Seven Years’ War concludes; Peace of Paris Hume accompanies Lord Hertford
Catherine Macaulay, first volume of her to Paris and takes on the duties of
History of England Secretary to the British Embassy
Hume and Madame de Boufflers
meet for the first time


9780826443595_Prelims_Final_txt_print.indd xv 2/9/2012 9:38:09 PM


1764 Thomas Reid, Inquiry into the Human

Mind; Horace Walpole, The Castle of
Mozart’s Symphony No. 1 composed in
London during the family’s European tour
1765 Matthew Boulton finishes building the Soho Hume is officially confirmed as
Manufactory in Birmingham Secretary shortly before his post
comes to an end
1766 Immanuel Kant, Dreams of a Spirit-Seer Hume returns to London,
Elucidated by Dreams of Metaphysics accompanied by Jean-Jacques
Rousseau; Rousseau accuses Hume
of being part of a conspiracy
against him
A Concise and Genuine Account of
the Dispute between Mr. Hume and
Mr. Rousseau
Hume spends the final months
of the year in Ninewells and
1767 James Craig’s plan for New Town, Travels to London to take up the
Edinburgh is adopted; Royal Crescent, office of Under-Secretary of State in
Bath, started the Northern Department
1768 James Cook’s first voyage to the Pacific Hume retires from public office
1769 Josiah Wedgwood and Thomas Bentley open Hume returns to Edinburgh
their Etruria factory near Stoke-on-Trent
James Watt’s steam engine, James
Hargreaves’ spinning jenny, and Richard
Arkwright’s water frame are patented
Arthur Wellesley, later Duke of Wellington,
is born in Dublin
1770 The future Louis XVI marries Marie Hume has a house built for himself
Antoinette; Baron d’Holbach, Système de la in Edinburgh’s New Town; rumours
nature reach Paris that Hume might be
about to marry Nancy Orde
1771 Tobias Smollet, Humphry Clinker Hume and his sister move into the
new house – St Andrew Square, off
St David Street
1773 Boston Tea Party
Hester Chapone, Letters on the
Improvement of the Mind; Oliver
Goldsmith, She Stoops to Conquer


9780826443595_Prelims_Final_txt_print.indd xvi 2/9/2012 9:38:09 PM


1774 Death of Louis XV; Louis XVI becomes

King of France
Joseph Priestley discovers ‘dephlogisticated
air’ (oxygen)
1775 John Wilkinson’s cannon-boring machine; The ‘Advertisement’ repudiating the
Jane Austen born Treatise
1776 Declaration of American Independence Hume dies in Edinburgh
Edward Gibbon, Decline and Fall of the
Roman Empire; Adam Smith, The Wealth
of Nations; Thomas Paine, Common Sense;
Jeremy Bentham, Fragment on Government
1777 Publication of Life of David Hume,
written by Himself
1779 World’s first iron bridge is completed across Publication of the Dialogues
the River Severn at Coalbrookdale concerning Natural Religion
1782 Atheism openly avowed in print in Britain
for the first time – Matthew Turner, Answer
to Dr Priestley’s Letters to a Philosophical
1783 Peace of Versailles establishes independence
of American colonies
1785 Edmund Cartwright patents his power loom
1787 Founding in Britain of the Committee for
the Abolition of the African Slave Trade
1789 French Revolution


9780826443595_Prelims_Final_txt_print.indd xvii 2/9/2012 9:38:09 PM

9780826443595_Prelims_Final_txt_print.indd xviii 2/9/2012 9:38:09 PM

In 2005 the British Broadcasting Corporation metaphysics and epistemology. There is, in
ran a poll asking Radio Four listeners to say fact, an overwhelming case for saying that
whom they regarded as the greatest philoso- no other eighteenth-century writer’s account
pher of all time. Such familiar philosophical of English history came close to matching the
luminaries as Plato, Aristotle, Descartes and intellectual quality and non-partisan nature
Kant all featured prominently in the subse- of Hume’s own narrative, and in this particu-
quent voting, and Marx’s immense influence lar case those genuine merits were, for once,
within the political arena saw him, rather rewarded by the approbation of substantial
predictably, taking first place. However, the sections of the public.
pre-eminent British philosopher, and the Hume’s current reputation is, therefore,
philosopher with the second highest overall something that stands in need of explana-
number of votes, was David Hume. tion. How has an eighteenth-century Scottish
In his own lifetime Hume certainly pos- intellectual and writer who enjoyed his great-
sessed a substantial reputation as a public est success amongst his contemporaries as a
intellectual. In some respects, though, it would historian, economist and writer of polite
be more appropriate to talk in terms of his essays arrived at the status and, in the eyes
notoriety rather than his reputation. His sup- of the editors of this volume, the wholly
posedly sceptical epistemological views and deserved status of being viewed as the great-
the manner in which his writings seemed to est British philosopher?
develop a series of pointed criticisms of reli- In many respects the answer lies in the
gious belief attracted vituperative criticism very features of his philosophical writings
from many of his contemporaries. It is also that saw them subjected to so much criticism
a striking fact that much of his fame sprang when Hume was alive. Epistemological scep-
from his ostensibly non-philosophical writ- ticism, even of a radical variety, is no longer
ings. Until his death in 1776 Hume enjoyed a seen as constituting any kind of threat to
great deal of influence as a writer on matters morality and social order; so it is now pos-
of economics. Moreover, sales of his History sible to respond to the sceptical arguments
of England made him independently wealthy deployed within Hume’s writings as provid-
and brought him to the attention of far more ing us with a series of fascinating puzzles that
readers than were interested in works of may succeed in pointing us towards important

9780826443595_Intro_Final_txt_print.indd 1 2/9/2012 9:37:57 PM


truths about the nature of philosophy or the these staples of theistic apologetics. Just as
incoherence of certain aspects of our self- significantly, however, Hume’s evident will-
conception as inquirers and agents without ingness to question religious dogma at a
those arguments constituting instruments of time when the social and cultural pressure
intellectual self-annihilation. Moreover, once towards internalizing such beliefs was so
this fear of sceptical conclusions has been strong marks him out as a person who was
dissipated, it becomes psychologically easier prepared to be guided by argument and the
to acknowledge the inadequacy of so many available evidence instead of suppressing
of the standard supposed refutations of scep- his critical faculties at the behest of super-
tical arguments. Hume’s own recognition of stition and the power-structures of religious
the power of these arguments accordingly authority. In this respect, Hume amply meets
comes to be seen as compelling evidence of the essential requirement that a true philoso-
his own intellectual integrity and powers of pher, a philosopher of genuine integrity, must
analysis. answer only to the autonomous demands of
This issue of intellectual integrity also has the reflective intellect.
a bearing on present-day reactions to Hume’s It also seems to be the case that once
criticisms of religion. Britain in the eighteenth Hume’s epistemological and irreligious views
century was an overwhelmingly Christian are no longer predominantly seen as views
country, where overt expressions of disbelief that need to be repudiated as aggressively
could still attract substantial prison sentences as possible, other valuable aspects of his
and books regarded as attacking Christianity philosophical outlook become increasingly
were frequently subjected to determined easy to recognize. Given the disappointing
campaigns of suppression. Today, in contrast, results of attempts at a priori metaphysics,
there is substantial evidence that between 30 Hume’s denunciations of the application
and 40 per cent of the British population do of a priori reasoning outside the sphere of
not believe in God or any Higher Power anal- issues of ‘quantity and number’ (EHU 12.27
ogous to a person. And although the United / 163) seem amply vindicated by the histori-
States signally lags behind almost all Western cal record. Thus philosophical inquiry needs
European states in this regard, agnosticism an alternative methodology if it is not simply
and atheism are making some inroads even to repeat past errors in ever more complex
in that hitherto hostile environment. There forms. And Hume’s ‘experimental’ method,
is, accordingly, a far more receptive audi- with its commitment to being guided by
ence in the current climate for arguments experience, seems to meet this need.
challenging the metaphysical underpinnings There might perhaps be some worries that
of a religious world-view and the compla- this approach actually amounts to a simple
cent supposition that religious convictions repudiation of philosophy in favour of the
constitute crucial support for moral behav- investigative methods of the sciences. In
iour and an appreciation of the value of life. Hume’s hands, however, it constitutes not an
Hume’s writings provide such arguments in abandonment of philosophy but a confirma-
abundance, and his critiques of the argument tion that at least some philosophical conun-
to design and the credibility of testimony to drums can be satisfactorily dissolved by
alleged miracles still constitute some of the paying due attention to the empirical facts.
most trenchant attacks ever launched on Confronted, for example, by the question

9780826443595_Intro_Final_txt_print.indd 2 2/9/2012 9:37:57 PM


‘What obligation do we have to obey this or whist, to avoid challenging him to any card
indeed any government?’, we might initially games involving large sums of money. In a
be at a loss to know how to proceed. Science, sense, of course, Hume’s personal virtues do
we have been repeatedly told, cannot answer not add to the importance of his intellec-
normative questions. Conceptual analysis, tual achievements. But they do confirm one
it is tempting to suppose, can at best clarify important thing, namely that the philosophi-
the sense of the question, and a priori rea- cal outlook embraced by Hume is one that is
soning of a non-mathematical kind cannot entirely compatible with a flourishing human
be relied upon to yield anything more sub- life that combines generous concern for the
stantial than vacuous tautologies. Hume’s well-being of other people with ample enjoy-
account of human nature, in contrast, allows ment of a full range of social and intellectual
us to see this question as an idle one. There satisfactions.
may indeed be scope for choosing which This combination of the power of Hume’s
government to follow. But our psychologi- thought and the engaging quality of his per-
cal properties mean that some institutions sonality has undoubtedly helped in bringing
of government will inevitably arise in all cir- together the contributors to this volume. As
cumstances that are ever likely to persist for editors, we were repeatedly pleasantly sur-
a significant length of time. Moreover, once prised by the enthusiasm expressed for this
these institutions have arisen, their success in project by potential contributors, and we
securing high levels of obedience is equally hope that the finished anthology succeeds
inevitable irrespective of our normative both in illuminating Hume’s own achieve-
speculations. ments and in suitably showcasing the com-
It would be remiss, however, of any mitment to Humean scholarship manifested
account of Hume’s well-merited appeal to by all the authors represented in the follow-
present-day philosophers and anyone inter- ing pages.
ested in understanding the place of human Emilio Mazza opens the volume by draw-
beings in the world to ignore the question ing us into Hume’s world, one far from the
of Hume’s personal character. Although this ivory tower – a world of business, military
has frequently been traduced by defend- expeditions, international diplomacy and
ers of religion and people who mistakenly Parisian ladies. But Le Bon David always
suppose that seriousness of purpose must sought refuge from this heady world in
be evidenced by tortuous writing, pompous work, in friendships and in his pursuit of
pretentiousness and a complete absence of literary fame, his ‘ruling passion’. His aca-
humour, it is clear from the record of Hume’s demic legacy and fame, however, are perhaps
life that he was a benevolent man of ami- rather fortunate given that he would have
able temperament, a good and loyal friend, been happy to stay at home in the borders
and a master of comic self-deprecation and of Scotland, if his brother had not married,
subtle word-play. If one were planning a fan- or to join the army if he had discovered its
tasy dinner party, it is difficult to imagine pleasures and camaraderie at a younger age.
any philosopher in history who would make Mazza’s evocative biography illuminates
a more winning and entertaining guest or a a life of travel and friendships with a por-
more congenial host, though it would prob- trait of a cheery, avuncular man toddling
ably be advisable, given his reported skill at around Edinburgh, being dragged out of a

9780826443595_Intro_Final_txt_print.indd 3 2/9/2012 9:37:57 PM


bog, saving a man from execution after that is the source of the vivacity of our experi-
man’s unsuccessful attempt to commit sui- ence, of its intentional content, and of the
cide in Paris, and having deep and sometimes believability of our ideas. This account is
stormy relationships with the literati of his contrasted with Descartes’s theory of ideas
day including, amongst many others, Adam and with interpreters of Hume who see him
Smith, Rousseau, d’Alembert and Lawrence as a proto-logical positivist.
Sterne. We come away with the impression Peter Millican turns to Hume’s account
that Hume had a good life – one with much of inductive reasoning and his ‘famous argu-
friendship, fame and fortune – and if one can ment’ to the conclusion that we have no
ever say this, Hume also had a good death. warrant for our beliefs about unobserved
To the end he was in good spirits, reading his matters of fact. Millican spells out the steps
beloved classics, and revising his Dialogues of Hume’s argument as articulated in the
concerning Natural Religion. Treatise, the Abstract and the first Enquiry.
Tom Seppalainen and Angela Coventry All beliefs concerning matters of fact are
take a ‘fresh look’ at Hume’s theory of ideas grounded in causal reasoning but, Hume
and impressions. The notion of liveliness or argues, knowledge of causal relations can-
vivacity that distinguishes mere ideas from not be acquired a priori, nor can it be gained
beliefs – beliefs being vivid ideas – is usu- via inductive reasoning. In place of such sup-
ally taken to be a phenomenological one port Hume provides an account of belief
and various interpretations of the nature of grounded in custom or habit. However con-
vivacity are considered. It has been charac- vincing Hume’s arguments may be, there is
terized in terms of qualitative feel although undoubted tension between his seemingly
Seppalainen and Coventry argue that think- sceptical conclusions and his embrace of
ing of perceptions in this way ignores their inductive science, his ‘experimental’ approach
intentional content and the way perceptions to the study of human nature, and his empiri-
seem to be of the world. An improvement, cal approach to history. Some interpreters of
then, is to read the phenomenology of per- Hume take him just to be concerned with a
ception not in terms akin to those describ- psychological description of thinkers and not
ing the intensity of colour in a picture, but with issues concerning justification and war-
in terms of ‘presentedness’ (or, according rant. Millican, however, argues that Hume
to another intentional reading of Hume, in is interested in normative questions; it is
terms of verisimilitude and the feeling of important to be clear, though, on the target
reality). Experience presents the world to of Hume’s scepticism – and that is Locke’s
one. Seppalainen and Coventry applaud conception of reason, what Millican calls his
such intentional readings, but they argue perceptual model. Such scepticism, however,
that Hume does not use vivacity to refer to does not engender what has come to be called
the phenomenological qualities of individual ‘The problem of induction’. The purpose of
perceptions, but rather to sequences of ideas Hume’s form of ‘mitigated’ scepticism is not,
and impressions. Only patterns of change as with Descartes, to prompt us to discover
can be vivid in the requisite sense. Our very a sure path to certain knowledge, but rather
notion of the existence of the external world to instil in us a suitable level of modesty and
depends on the constant and coherent flux caution concerning our epistemic practices.
of our perceptions and, they argue, such flux Furthermore, Hume’s naturalistic account

9780826443595_Intro_Final_txt_print.indd 4 2/9/2012 9:37:57 PM


provides an explanation of how human the usually apparent regularity of nature.

beings can – and actually do – reason induc- And Hume’s account of the vivacity of belief
tively and this, given the impossibility of any can show how the strength of our beliefs
rational foundations for such reasoning, pro- depends on the uniformity of our experience.
vides us with all the support we require for This account does not depend on a mathem-
our causal inferences. atical calculation of probabilities, but rather
Lorne Falkenstein further explores Hume’s on associationist psychological processes of
account of causal reasoning and the ‘system’ vivacity transfer. Falkenstein concludes by
of the Treatise. It is constant conjunctions showing how Hume takes his scepticism as
in experience that impel the imagination to supporting his ‘system’ and his account of
form beliefs concerning causes and effects. empirical reasoning.
Habits of thought guide our reasoning and Helen Beebee heads into stormy water –
not rational argument or judgement. As dis- into Hume’s account of causation, a hotly
cussed by Coventry and Seppalainen, beliefs debated topic: the so-called New Hume
are seen as vivacious ideas, and Falkenstein Debate focusing on the question of whether
stresses that Hume does not think of vivacity Hume takes there to be real causal powers
in terms of the intensity of an image. Beliefs, in nature – oomphs pushing billiard balls
rather, amount to dispositions of the mind around tables – or whether he thinks that
(for example, to incline one to act in certain there are really only constant conjunctions
ways and to focus one’s attention). Beliefs, and regular brute patterns in the world.
then, are the product of the principle of asso- Beebee investigates this question, concentrat-
ciation of causation, although Falkenstein ing on three claims to which Hume seems to
also suggests how Hume might have included be committed. First, Hume suggests that we
the relations of contiguity and resemblance project causal connections onto the world,
in his account of reasoning. connections that are not really there. Second,
If, however, belief formation is just a mat- the concept of causation includes the notion
ter of habit, how can it be so that some ways of necessity. We think of causal connections as
of forming beliefs are seen as better than oth- those that are necessarily connected together:
ers? Hume suggests that we follow certain the red ball must move off in that direction
general rules, those that are learnt from past given that it was struck in that way by the
experience, such as, like objects in like cir- white ball. And third, causal talk is objective
cumstances will have like effects. Particular in that we can talk correctly of causes and
inferences we make can be assessed against we can sometimes make false causal claims
such general rules. This is Hume’s logic of about the world. These commitments seem
causal inference. to be inconsistent since it is not obvious how
Such rules can also be extended to cover causal talk can be objective when Hume does
probable reasoning. It is not the case that we not think that there are really causal connec-
have uniform experience, but this does not tions in nature and that they are projections
lead to us rejecting such general rules. Instead, of our cognitive processes. Beebee discusses
in cases where like causes do not seem to lead various ways to resolve this (perhaps only
to like effects we come to believe – again via apparent) inconsistency. The traditional
habit grounded in past experience – that interpretation of Hume claims that causation
there is some hidden cause that is disrupting is just constant conjunction. The sceptical

9780826443595_Intro_Final_txt_print.indd 5 2/9/2012 9:37:57 PM


realist, in contrast, sees Hume as accepting radical position than mere fallibilism. In
that there are causal powers in nature; it Bailey’s judgement, this scepticism is better
is just that we cannot come to have know- interpreted as a stance that does not endorse
ledge of them. The projectivist interpretation any beliefs as possessing a positive degree
adopts a non-cognitivist stance: our claims of epistemic justification except for beliefs
concerning causal relations are subject to about very simple necessary truths that can
norms, but these norms are constituted not be grasped without going through any proc-
by features of the world independent of our ess of inference and beliefs about the content
judgements concerning its causal structure – of our present ideas and impressions.
by real causal powers in nature – but by It is clear, however, that if such scepticism
certain ‘rules’ which we have come to appre- is an integral component of Hume’s thought,
ciate concerning how we judge of causes and then it co-exists with Hume’s assent to a
effects, rules that enable us to override errant detailed and carefully constructed account
judgements in particular cases. of human nature that is supposed to be both
A clue to the correct interpretation can be true and useful. Even a moment’s reflection
found in Beebee’s claim that Hume is driven on the Treatise’s subtitle, which is Being
by his opposition to the Cartesian Image of an Attempt to Introduce the Experimental
God Hypothesis. There are two aspects to this Method of Reasoning into Moral Subjects,
hypothesis: we are, as the Bible says, made indicates that it would be a disastrous mis-
in God’s image, and we have epistemic abil- reading of Hume’s views to construe him as
ities in line with such an origin. The nature simply a destructive sceptic.
of reality is accessible to human reason – we Bailey’s solution to this interpretative con-
can come, through a priori reasoning, to undrum is to claim that Hume views radic-
have knowledge of the world and specifically ally epistemological scepticism and properly
of its causal structure. Beebee notes certain conducted empirical inquiries as mutually
aspects of this picture in the sceptical realist supportive. Hume thinks that sceptical
approach and thus argues that this cannot be arguments are indeed successful in placing
the correct interpretation of Hume. Of the us in a position where only our acceptance
remaining options, Beebee favours projectiv- of the view that scarcely any of our beliefs
ism over the traditional interpretation. are rationally justified can allow us to deny,
Alan Bailey then undertakes an examina- without being guilty of bad faith, that such
tion of the equally vexed issue of the pros- sceptical arguments provide rationally com-
pects for providing a unified account of pelling grounds for that assessment of our
Hume’s philosophical outlook that satisfac- beliefs. However, Bailey argues that Hume
torily accommodates both his ambitions to does not see this as posing any threat to the
construct a science of human nature and the ability of our belief-forming mechanisms
sceptical elements of his thought. If Hume’s to generate and sustain in existence all the
putative scepticism actually amounted to beliefs we need to guide our actions. Nor
nothing more than a modest epistemologi- indeed does Hume view it as undermining
cal fallibilism, as some recent commentators our capacity to assent to relatively sophisti-
have supposed, there would be no real ten- cated scientific theories. Where such theor-
sion to overcome here. Bailey argues, how- ies are constructed using systematized and
ever, that Hume’s scepticism is a much more reflective versions of common sense methods

9780826443595_Intro_Final_txt_print.indd 6 2/9/2012 9:37:58 PM


of inquiry and are accordingly supported by that is, that the soul or mind is a substance,
experience and experiment, sceptical discov- be it physical or non-physical.
eries are incapable of preventing us from All we can do is provide an account of
giving our assent. And where those theories what causes us to have the mistaken belief
are not supported by experience and experi- that there are enduring selves. Such an
ment, Hume can, as an empiricist, rejoice account includes certain identity-ascribing
in their destruction. Thus Bailey holds that mechanisms of the imagination – those
Hume sees scepticism and the proper experi- grounded in the principles of association of
mental method of inquiry working in tan- resemblance and causation – mechanisms
dem. Sceptical arguments curb the power of that generate belief in the self as well as in
the imagination to generate beliefs that are the continued existence of the external world
not the products of the observed correla- and of bodies.
tions that give rise to causal inferences. And Hume, however, is dissatisfied with his
the experience-based beliefs towards which conclusion. He thinks that he is committed
we accordingly gravitate generate a plaus- to two inconsistent principles: that distinct
ible picture of the workings of the human perceptions are distinct existences, and that
mind that makes it even more difficult for the mind never perceives any connection
us to represent ourselves as capable of arriv- between them. Perplexingly, however, these
ing at many beliefs that genuinely qualify as principles are not inconsistent, and uncover-
rationally justified. ing why Hume claims them to be so is a key
Harold Noonan and Galen Strawson both difficulty for interpreters of Hume’s views
explore what Hume calls the labyrinth of on the soul and the self. Noonan suggests
personal identity. Noonan considers various that Hume realizes that his account does
arguments in Hume against the Cartesian not explain our continuing belief in personal
conception of personal identity, against, that identity. One can accept Hume’s empiricist
is, the existence of an enduring self, identi- conclusions with respect to the external
cal from moment to moment and from day world and give up the notion that there is a
to day. Hume’s empiricism demands that we substance or substrata underlying the prop-
have an impression of such an entity, but this erties of bodies, but we cannot accept this
we do not have – all we find, on introspec- with respect to the self. Why not? Hume did
tion, is a bundle of variously related percep- not know.
tions. Hume’s ‘master-argument’ establishes Galen Strawson has a distinct account of
that all perceptions are logically capable of why Hume’s ‘hopes vanish’. Hume discov-
an independent existence. There is thus no ers – late in the day, in the Appendix to the
need for the ‘unintelligible chimera of sub- Treatise – that his whole empiricist philoso-
stance’ (THN / 222) in which prop- phy depends on a conception of the mind
erties must inhere. This is so for physical that his empiricism does not allow him to
things, such as wax – contra Descartes, wax have. His genetic account of our belief in
for Hume is just a collection of properties – an enduring self relies on the principles of
and for human beings: we do not require an association – it relies on the assumption that
enduring soul underlying our ever-changing we have a ‘Principle-Governed Mind’. This
properties. Hume thus criticizes a suppos- explains our belief in the self as well as our
ition of both materialists and immaterialists, belief in the external world and in causation.

9780826443595_Intro_Final_txt_print.indd 7 2/9/2012 9:37:58 PM


But such a conception of the mind goes Sandis also turns to interpretations of
beyond a loose association of distinct percep- Hume’s famous claim that ‘[r]eason is, and
tions. In order to legitimately ground one’s ought only to be the slave of the passions’
philosophy in such an account of the mind it (THN / 414) and argues that the
is required either that there is an observable received Humean theory of motivation is
real connection between the perceptions that unfounded. This is the view that an agent
make up the mind or grounds for claiming cannot be motivated by belief alone, but
that such perceptions inhere in some kind of only by a belief along with an appropri-
soul-substance. But Hume has argued against ately related desire. Sandis claims, though,
both possibilities. that such an account is not to be found in
A possible response here is to take Hume Hume. It is also suggested that Hume does
as sheltering in his scepticism: the essence not equate belief and opinions with judge-
of the mind is unknowable to us and thus it ment. Ideas and beliefs are distinguished by
cannot be this – the lack of knowledge of the their vivacity, and the vivacity of judgements
Principle-Governed Mind – that leads him to should be seen as lying somewhere between
despair. But, Strawson argues, such agnosti- that of ideas and beliefs. This is relevant to
cism cannot do the trick. Hume does need Hume’s account of morality: Hume does not
to, and does, assume a certain notion of the talk of moral judgements but, on Sandis’s
mind – a rule-governed one. He can perhaps account, this still allows Hume to have an
remain agnostic about just how it works, but account of moral beliefs and of their motiv-
he cannot be agnostic about its very exist- ational force.
ence – and its very existence is what is incom- James Harris turns to liberty and neces-
patible with Hume’s empiricism. Strawson sity, and to what Hume calls ‘the most con-
claims that Hume’s despair is a result of his tentious question of metaphysics, the most
acknowledgment of this deep inconsistency contentious science’ (EHU 8.23 / 95). Hume
in his philosophy. is often thought to be an advocate of an early
Constantine Sandis explores Hume’s version of what is now called compatibilism,
account of action and in so doing considers and it has been claimed that there is noth-
how reason, the will and the passions are ing distinctive about his position. Locke and
related. Hume’s account of action is an empir- Hobbes had also discussed this question and
ical one: we acquire knowledge of a person’s suggested compatibilist answers. Harris,
reasons for acting from careful observation however, argues that Hume is not just rehash-
of human behaviour. This ‘science of man’ ing their arguments.
grounds Hume’s History of England and Importantly, it is claimed, Hume is not a
the study of this work highlights how Hume determinist in the modern sense, unlike, for
sees character as playing a key motivating example, Hobbes. Determinism is a meta-
role in our behaviour. Further, the historian physical stance and Hume eschews metaphys-
is best placed to uncover the truth about ical questions. His claim is not that we have
human nature since he does not aspire to the reason to think that the laws of nature cannot
detached perspective of the philosopher, nor change – that they are determined; everyday
is he too close to his subjects and thus prone experience, rather, leads us to expect that
to bias or the distorting influence of his par- people behave in regular ways and we inter-
ticular interests and circumstances. act with them in light of these regularities.

9780826443595_Intro_Final_txt_print.indd 8 2/9/2012 9:37:58 PM


Harris also notes various changes in the extent to which he and other irreligious
emphasis between the Treatise discussion of thinkers of his time were forced to engage
this topic and that of the Enquiry. In the latter, in misdirection and linguistic contortions in
Hume more squarely targets metaphysicians. order to avoid social ostracism and the offi-
Once we clarify the nature of liberty and cial suppression of their writings. The authors
necessity, long-running metaphysical and, in of the three chapters in this anthology that
particular, religious disputes concerning God’s focus primarily on Hume and religion are
prescience and responsibility for evil will be therefore unanimous in presenting him as
undermined. Philosophy should ‘return, with a rigorous and intellectually honest thinker
suitable modesty, to her true and proper prov- who deploys a formidable set of arguments
ince, the examination of common life’ (EHU against any form of religious outlook based
8.36 / 103). Hume’s discussion of liberty and on the truth of theism or even a robust form
necessity is not a case of Hume engaging in of deism.
metaphysical debate, the question then arising Duncan Pritchard and Alasdair Richmond
of whether his contribution is original or not – investigate Hume’s notorious arguments on
he is, rather, agnostic about all such issues, his the topic of the credibility of testimony con-
discussion reflecting his empiricist attitude to cerning miracles. They are careful to locate
questions concerning the regularity of human these arguments within the broader frame-
behaviour and morality. work of Hume’s reservations about our abil-
At this point the contributors turn to the ity to justify expectations about the future
subject of Hume’s views on the truth and in a rational, non-circular manner and his
utility of religious beliefs. In his own era he pragmatic response to those sceptical wor-
was interpreted as attacking Christianity and ries. Although causal reasoning cannot be
all forms of theistic belief. However, his argu- supplied with a non-circular argumenta-
ments were frequently dismissed as incon- tive defence, it remains the case that human
sequential sophistries motivated not by a beings find such reasoning persuasive and
concern for the truth but by a desire to secure continue to use it, even after exposure to
personal notoriety and increase the sales of sceptical arguments, as a touchstone for
his books. Such an assessment of the force of assessing whether particular beliefs are ones
his arguments and his motivation for advan- they are content to endorse or ones that are
cing them is now wholly discredited. Yet the no more than mere foolishness. Consequently
recognition that he wrote on this particular Pritchard and Richmond construe Hume as
topic in good faith and with a commitment attempting to show that no testimony about
to seeking the truth and promoting human the occurrence of miraculous events capable
well-being has led some present-day com- of serving as the foundations of a system of
mentators to suggest that he was actually a religion has ever met the standards of dox-
defender of some philosophically purified astic acceptability that normally prevail in
form of theism that might even be compat- less contentious cases when we are weigh-
ible with the truth of Christianity. Such an ing human testimony concerning an alleged
interpretation seems to be based on nothing event against the implications of our obser-
more substantial than his invariable cour- vations of past natural regularities.
tesy when debating matters of religion and Pritchard and Richmond reject the suppo-
an almost inexcusable failure to recognize sition that any form of a priori conceptual

9780826443595_Intro_Final_txt_print.indd 9 2/9/2012 9:37:58 PM


argument forms part of Hume’s case against as confirmation of the truth of religious doc-
belief in miracles, and they also maintain trines or teachings.
that it is a mistake to construe him as arguing If Hume is right to maintain that reports
that the kind of regularity in experience that of alleged miracles fail to offer any genuine
would need to be interrupted in order for an support for the bold claims advanced by
event to qualify as a plausible candidate for religions about the ultimate nature of real-
being a miracle would be so well entrenched ity, where might such support be found?
and confirmed that no possible amount of Hume’s religious contemporaries placed
human testimony could render it appropriate great confidence in the probative value of
to believe that an interruption had occurred. the design argument, and Hume undertook
They emphasize that for Hume it is always a detailed examination of this argument in
a contingent matter whether the testimony his Dialogues concerning Natural Religion,
offered is weighty enough to overcome the which were published posthumously in
initial presumption that a hitherto well-con- 1779.
firmed natural regularity with no previously Andrew Pyle accordingly presents in his
known exceptions has not abruptly come to chapter an overview of the complex discus-
an end. Nevertheless the standards of dox- sion that occurs between Hume’s principal
astic acceptability we embrace in practical characters in the Dialogues, and he arrives at
contexts when we are making judgements the conclusion that this work was intended
in a careful and reflective manner are such to show that the design argument cannot
that this testimony needs to overcome an legitimately support theistic conclusions and
exceptionally high hurdle. Unless the plau- that a naturalistic explanation of the orderly
sibility that this testimony is mistaken or nature of the universe is, when judged by
deliberately deceitful is even lower than the everyday standards of good causal reason-
extremely low plausibility that attaches to ing, more acceptable than a theistic one. An
the supposition that a pervasive and well- interpretation of the Dialogues along these
tested regularity that has previously mani- lines might initially be thought to overlook
fested itself throughout all human history Hume’s repeated suggestions that our intel-
has been breached at a particular time and lectual faculties are wholly inadequate when
place, it is not appropriate for us to accept confronted by the task of arriving at satisfac-
that this testimony is correct. And although tory conclusions about such rarefied matters
testimony of this quality is at least conceiv- of inquiry. Those pronouncements appear to
able, Pritchard and Richmond hold that give very strong support to the conclusion
even when we assess Hume’s arguments that Hume holds that the only legitimate
from the perspective of Bayesian reasoning response to questions about the ultimate ori-
or the non-reductionist view that testimony gins of the universe is a stance of complete
can possess some independent credibility neutrality and suspension of judgement.
that does not ultimately derive from non- However, Pyle draws an important distinc-
testimonial sources, it is apparent that Hume tion between a commitment to a particular
manages to present a strong case for the hypothesis as more probable than all com-
conclusion that such exemplary testimony peting hypotheses with equivalently detailed
has never yet been forthcoming in the case content and a comparative judgement that
of any allegedly miraculous event presented a particular hypothesis is more likely to be


9780826443595_Intro_Final_txt_print.indd 10 2/9/2012 9:37:58 PM


true than some specified rival hypothesis. O’Connor emphasizes the role Hume allo-
He accepts that Hume is indeed inclined to cates in his causal story to the human dis-
maintain that no detailed hypothesis that position to anthropomorphize phenomena.
we can formulate about the ultimate origins People living in relatively sophisticated soci-
of the universe and the order it displays is eties, where information about the genuine
worthy of endorsement as an explanation hidden causes of otherwise puzzling phe-
that is more likely to be true than false. But nomena is fairly widespread, often engage
he maintains that Hume takes the view that in what we might term playful anthropo-
the empirical evidence, inadequate as it is in morphism as an amusement or a deliberately
terms of favouring a particular determinate chosen form of metaphor. O’Connor argues
theory as the most likely theory, does at least that Hume contrasts such playful anthro-
marginally favour a naturalistic account of pomorphism with a literal-minded anthro-
the universe as more plausible than the pomorphism that characteristically emerges
world-view represented by theism and con- when people have little or no grasp of the
ventional deism. true aetiology of striking or potentially dan-
As Hume was only too well aware, the gerous phenomena. In those circumstances
apparent paucity of good empirical evidence the human proclivity to think in terms of
for theism or conventional deism has not otherwise unexplained phenomena as arising
prevented the emergence of popular theis- from agency and intention exerts itself with
tic religions with vast numbers of nominal full force; and as no observable agents can
adherents. How can the widespread preva- be detected with the appropriate intentions
lence of this form of belief be explained if and purposes, the idea develops of multiple
it is not a response to evidence; and even if invisible and intelligent powers that have a
theistic religions potentially lack the virtue concern with human affairs. Such speculative
of offering a true description of the universe notions are, of course, theoretically distin-
and our place within it, could it be the case guishable from actual beliefs. But O’Connor
that the existence of such religions is a vital locates the mechanism that takes people from
bulwark of morality and an expression of the a spontaneous conception of invisible and
highest and most sublime aspects of human intelligent powers to belief in the existence
nature? of such powers in human fear and an acutely
David O’Connor investigates the account distressing sense of vulnerability. Once those
Hume provides of the psychological origins lively and pervasive passions are engaged, a
of religious belief, and he contends that this mere picture of the world is transformed into
account is a strongly deflationary one. Not a set of beliefs that guide people’s actions.
only does Hume describe in the Natural In particular, people attempt to relieve their
History of Religion a set of psychological helplessness and sense of vulnerability by
mechanisms that explain how religions can treating these hidden agents as susceptible to
arise and sustain themselves even if their core manipulation by flattery and supplication.
metaphysical and historical claims are both Ironically many theists would probably be
unwarranted and untrue, but he also presents happy to endorse this or some similar account
some of those mechanisms as dependent on of the origins of polytheism. O’Connor, how-
such unedifying aspects of human nature as ever, argues that Hume’s account of the psy-
ignorance, fear and servile self-abasement. chological genesis of theistic religion is every


9780826443595_Intro_Final_txt_print.indd 11 2/9/2012 9:37:58 PM


bit as subversive as his explanation of poly- those affected by that action, we are drawn
theism. Hume, in O’Connor’s judgement, to feel a certain moral sentiment of approba-
rejects as wholly inconclusive the supposed tion. The feeling of such a moral sentiment
evidence for theism from miracle reports constitutes, for Hume, a moral judgement.
and such arguments as the cosmological Given that moral judgements depend
argument and the design argument. And the on our emotional responses to others, they
emergence of theism from a background of must be, at least in some sense, subjective
antecedent polytheism is explained by Hume since they are not independent of our natu-
in terms of an attempt by people to ingratiate ral human responses to each other. Driver,
themselves with a particular invisible power however, explores how moral judgements
by assigning to that agent ever more impres- can nevertheless possess a kind of objectiv-
sive attributes and abilities in much the same ity in that moral truths are independent of
way that one might seek to curry favour with an individual’s particular responses to a
a murderous human despot by eulogizing his certain action. Such objectivity is supplied
or her non-existent qualities of wisdom, jus- by our ability to adopt the general point of
tice and benevolence. This practice of base view. We can ‘correct’ our sometimes mis-
flattery inevitably corrupts over time even guided moral judgements because we are
the judgement of the original flatterers, and able to adopt a perspective divorced from
Hume sees its impact on the beliefs of other our own, a perspective encompassing the
people in society, especially when aided by ‘narrow circle’ of those affected by a certain
education and religious instruction, as even action and not biased by our own concerns
more profound and pernicious. In this man- or interests. Can, then, a moral judgement be
ner, some previously negligible deity of com- true or false? Driver argues that it can, the
parable status to a petty human princeling is ultimate grounding for the truth of a moral
potentially exalted over many generations, judgement lying in the utility of the actions
without any assistance from cogent truth- that we judge to be virtuous from the general
oriented reasoning, into the supposedly point of view.
omnipotent, infinitely perfect and wholly Dan O’Brien continues to explore Hume’s
just creator of the entire universe. account of morality, focusing on his concep-
Of a piece with his attitude towards, and tion of virtue and vice. Virtues for Hume
arguments against, religion Hume provides a are those character traits that are useful and
secular moral theory, one in which there is agreeable to ourselves and to others, and
no place for God. Julia Driver turns to this thus people manifest many different kinds of
account and its grounding in the natural virtue and many different vices. Hume denies
emotional responses we have to the happiness that all virtues are innate and God-given and
and suffering of those around us. As O’Brien highlights the importance of artificial virtues,
and Hardin also go on to discuss in subse- traits that people in societies have developed
quent chapters, such emotional responses in order to aid the social interactions within
have their source in our sympathy with oth- those communities. We come to be able to
ers – sympathy, for Hume, seen as the capac- see certain traits as virtuous through sympa-
ity we have to share the emotions of our thizing with the effects that a person’s behav-
fellows. And, when the actions of a person iour has on those around them. Benevolence
lead us to feel pleasure, via sympathy with is virtuous because I resonate to the pleasure


9780826443595_Intro_Final_txt_print.indd 12 2/9/2012 9:37:58 PM


that the benevolent person’s actions bring behaviour that serves the interests of other
to his friends and acquaintances. Reason, people is primarily a product of psychologi-
however, also plays a role here, but only in cal mirroring and the conventions that arise
helping us to appreciate what sentiments we when self-interested agents of limited power
should feel if we are to be impartial in the are attempting to maximize their own ben-
requisite way. efits from repeated interactions with other
O’Brien goes on to explore how Hume similarly self-interested agents.
subverts the religious conception of virtue or Hardin describes psychological mirroring
what Hume calls the ‘monkish virtues’; ele- as an automatic response that people show
vating pride to his first natural virtue, a due to the actions and emotional states of other
sense of pride being agreeable to ourselves people. It is readily observable that human
and ultimately useful in our social engage- beings have a strong tendency to mimic
ments. That many traits are useful and agree- the behaviour of the people around them.
able is uncontroversial, but a distinction is However, it also seems to be the case that
also usually drawn between traits that have a most of us find other people’s observed emo-
moral dimension and those that do not: ben- tional states similarly infectious. Observing
evolence and compassion are of the former a person showing clear signs of distress or
kind; dexterity and wit, the latter. Hume, fear tends to give rise to analogous emo-
however, thinks that distinctions hereabouts tional states in the spectator. And behaviour
are not at all sharp and that moral virtues manifesting joy and gladness has at least
are not different in kind from other benefi- some tendency to raise the spirits of a person
cial ways of behaving. who observes such behaviour. Hardin credits
Russell Hardin moves the discussion Hume with being one of the first thinkers to
away from questions concerning the nature explore in any detail the implications of this
of moral judgement and what constitutes a phenomenon for human actions and choices.
good character to a consideration of Hume’s Given the existence of psychological mirror-
analysis of how self-interest can give rise ing, the psychological states of other people
to social conventions and organizations cannot be a matter of complete indifference
that promote public benefits. Hardin views to us. No matter how self-interested we hap-
Hume as deliberately eschewing the attempt pen to be, our own lives are more satisfac-
to show that moral principles are true or tory, all other things being equal, when the
rationally justified in favour of a scientific people around us are also faring well. And
investigation into how beings who are pre- this responsiveness to the psychological states
dominantly motivated by self-interest never- of other people is what Hardin identifies as
theless create institutions and social practices lying at the core of the Humean principle of
that serve the collective good. Numerous sympathy.
philosophers have implausibly purported to Sympathy alone, however, is an inadequate
show that altruistic behaviour is a funda- explanation for the range of circumstances in
mental dictate of reason or a requirement which people seem to accept some check on
imposed upon us by some divine lawgiver, their self-interest so that the interests of other
but Hume is seen by Hardin as adopting the people can be safeguarded or promoted.
radically different and substantially more Hardin accordingly attaches great importance
illuminating approach of explicating how to Hume’s exploration of the way in which


9780826443595_Intro_Final_txt_print.indd 13 2/9/2012 9:37:58 PM


repeated interaction between people comes to women throughout his life, relationships
shape social behaviour in ways that see collec- often based on mutual respect and shared
tive benefits emerging from self-interest. Even intellectual interests rather than transient
a purely self-interested agent needs to return sexual or romantic passion – although it is
favours if he or she is to have much chance also clear from her account that such passion
of securing the co-operation in the future of was certainly not wholly alien to Hume’s
people who are aware of past performances. character. Even more importantly Guimarães
And if we advance to a more sophisticated identifies Hume’s writings as showing a great
level, justice, in the sense of social stability willingness to deconstruct gender dichoto-
and good order, is something that we all have mies. In his History of England, female
some interest in promoting even if the institu- characters are frequently portrayed as active
tions and habits that maintain stability and agents endowed with energies, drives and
order sometimes prove inconvenient to us reasoning abilities that are at least equivalent
on particular occasions. Hardin accordingly to anything possessed by the men surround-
presents Hume as someone who succeeds in ing them. And Guimarães argues that when
setting before us a detailed account of the Hume is explicitly engaged in the study of
self-interested strategies that lead to the evo- human nature at a more theoretical level,
lution of some of the most salient social prac- his emphasis on human beings as embodied
tices and forms of organization that serve to mammalian animals responding to the influ-
enhance our collective well-being. ence of concrete conditions including social
One important aspect of any human soci- circumstances and personal relationships
ety is the relationship between the sexes, means that he avoids the trap of construct-
and Hume’s views on this relationship and ing an idealized account of our nature that
on gendered differences have aroused con- uncritically sees its essence as lying in such
siderable controversy. Some commentators allegedly masculine virtues as pure ration-
have accused him of acquiescing in and even ality and the suppression of the passions.
actively seeking to defend sexist forms of Moreover, Hume’s account of human reason
discrimination and oppression. Other read- and inference further subverts traditional
ers of his writings have, in contrast, seen him gender categories by presenting such reason-
as someone who seeks to enhance the status ing as founded in associations of ideas, our
of women and also wishes to see some of passions and the faculty of sympathy.
the values and psychological characteristics Guimarães also notes that when Hume is
traditionally associated with women dissemi- working within traditional gendered catego-
nated more widely throughout society. ries, he usually speaks in favour of the wider
Lívia Guimarães mounts a strong defence diffusion of supposedly feminine character-
of the view as elsewhere there should be an istics. Such virtues as tenderness, benevo-
accent on one of the i’s of Livia that it is the lence and mildness are not simply seen by
latter reading of Hume’s position that offers Hume as appropriate for women. Instead
the more illuminating perspective on his he argues that society, as a whole, would
attitude towards women and conventional benefit from these virtues being more wide-
distinctions between masculine and femi- spread amongst men as well. Hume read-
nine characteristics. She helpfully reminds ily acknowledges that the martial virtues of
us of Hume’s many close relationships with aggression, fierceness and intransigence have


9780826443595_Intro_Final_txt_print.indd 14 2/9/2012 9:37:59 PM


utility in primitive societies, where violence is our attention to the care that Hume took to
needed to maintain order and to repel inva- investigate this phenomenon by seeking out
sion and despotic oppression. But in more the best available data and using his exten-
civilized societies Hume sees other virtues sive acquaintance with classical authors to
as more effective at promoting the general compare Europe with the civilizations of the
well-being, and Guimarães draws our atten- ancient world. Moreover, Schabas maintains
tion to the fact that Hume frequently indi- that Hume was right to explain this rise in
cates that these less abrasive virtues are best wealth by invoking the combined impact of
spread throughout society by increasing the division of labour and the increased supply
opportunity for women to exert their influ- of silver coinage made possible by the mines
ence on men. Guimarães therefore concludes of the New World.
her chapter with the striking suggestion that Hume also emerges as equally astute in
Hume can be seen both as sketching out an his reflections on the consequences of such
ideal society that would constitute a femi- additional wealth for human happiness and
nist utopia and as recommending a greater welfare. Unlike conservative critics of wealth
emphasis on supposedly feminine virtues and and luxury who saw and often still affect to
attributes as an effective means of improving see such things as harbingers of moral decay,
existing societies. Hume held that the modern commercial
Margaret Schabas investigates Hume’s world and the opportunities that it gener-
economic thought. Schabas points out that ated had an improving effect on civil soci-
Hume differs from most present-day econo- ety and people’s characters. Schabas presents
mists by emphasizing the greater value to the him as arguing that trade and manufactur-
individual of mental well-being rather than ing promoted civility, gave a new impetus to
material wealth. Nevertheless Hume’s inter- learning and human ingenuity, and enhanced
est in all aspects of human nature and the liberty and equality. Hume saw these benign
social forces that shape people’s lives meant influences as most readily impinging on those
that he corresponded and published quite located in the middle ranks of society, but a
extensively on matters of economic policy flourishing middle class helped to bind all of
and theory. Indeed Schabas argues that the society together in ways that ultimately bene-
circulation and influence in the eighteenth fited everyone. So although Hume was fully
century of Hume’s economic essays entitles prepared to disparage the rapacious acqui-
him to be seen as one of the foremost econo- sition of expensive trinkets, Schabas locates
mists of his era. His pre-eminence amongst in his writings an ingenious account of how
British theorists was eventually to be usurped wealth indirectly promotes human happi-
by his friend Adam Smith. Schabas reminds ness. Hume’s predisposition towards Stoic
us, however, that this was not to happen until values meant that he viewed material posses-
the 1790s, despite the publication of Smith’s sions in themselves as being of little conse-
The Wealth of Nations in 1776 and Hume’s quence once the necessities of life had been
death in that same year. supplied, but the process of acquiring wealth
Schabas concentrates in her chapter on through trade and participation in manufac-
Hume’s response to the conspicuous rise in turing served the crucial role of giving people
the wealth of Western Europe in the seven- the opportunity to gain personal satisfaction
teenth and eighteenth centuries. She draws and a sense of purpose from the exertion


9780826443595_Intro_Final_txt_print.indd 15 2/9/2012 9:37:59 PM


of their mental and physical powers in an as such a judge. The thought here is that aes-
undertaking that was immediately appealing thetic pronouncements issuing from people
to them. As might be expected with Hume, who lack those attributes can be dismissed in
this confidence in the ameliorative powers of much the same way as the colour judgements
the commercial world is hedged around with of someone known to be suffering from a
substantial reservations. Schabas indicates fever or viewing an object in non-standard
that these reservations presciently included lighting conditions carry no weight with us
worries about the destructive power of pub- in respect of our assessments of the object’s
lic debt in the hands of politicians and the real colour. This, however, raises the question
prediction that the American colonies and of whose judgement is to be accepted if and
China would eventually eclipse Britain and when people who count as qualified judges
other European nations in terms of trade disagree. In some specific instances Hume
and manufacturing output. But Schabas is perfectly content to say that the disagree-
amply succeeds in showing that Hume’s case ment is irresoluble. But Rowe points out
for supposing that commerce and the pursuit that Hume is not always so accommodat-
of wealth can often improve people’s disposi- ing: sometimes he seeks a standard of taste
tions and moral character remains a useful that can override or correct the judgement of
antidote to the unthinking prejudice that qualified judges.
morality and personal development are best Rowe maintains that all Hume’s attempts
promoted by austerity and the eschewal of at explaining how such corrections can be
luxury. given legitimacy are unsuccessful. Hume
Mark Rowe, in contrast, is less sanguine sometimes appeals to rules of composition,
about the merits of Hume’s account of a but Rowe powerfully argues that these rules,
standard of taste in matters of aesthetic on a Humean account, can be nothing more
judgement. Rowe argues that Hume is con- than inductive generalizations that summa-
cerned to reconcile a form of subjectivism rize the characteristics displayed by works
about aesthetic taste with the supposition that people usually find pleasing and beau-
that some aesthetic judgements can be mis- tiful. Thus they lack the normative force
taken. The subjectivism is a product of that Hume requires. Similarly, Rowe rejects
Hume’s commitment to the view that aes- Hume’s alternative appeal to a consensus
thetic qualities are projected onto the world amongst qualified judges. Even if majority
rather than discovered in the world. But the opinion were against your personal verdict,
need to find some room for the concept of would it make sense for you, as a person
an error in aesthetic judgement arises from with the attributes requisite for being a
Hume’s conviction that some judgements of qualified judge, to treat your judgements as
aesthetic merit would be as perverse and as wrong simply because you are in a minority?
obviously illegitimate as some patently false Finding yourself in a minority might well give
pronouncements about physical objects and you grounds for reviewing your reactions to
such qualities as size and shape. a particular work or artistic performance
A part of the answer to Hume’s dilemma again. But being in a minority of qualified
lies in the concept of a qualified judge, and judges is not constitutive of being wrong in
Rowe enumerates the attributes that Hume your aesthetic judgement. Rowe concludes
regards as essential if someone is to be viewed that although Hume rightly sees the need to


9780826443595_Intro_Final_txt_print.indd 16 2/9/2012 9:37:59 PM


accommodate critical discussion when con- conclusions supported by real evidence. At

sidering questions of aesthetic merit, his view the same time the writer requires many of the
of what constitutes argument in that arena skills of the poet or dramatist: the truth must
is too impoverished to explain how genuine be shaped and ordered so that a reader read-
debate can take place and real discoveries ily enters into the narrative and engages with
can be made. the character and situation of the personages
Timothy Costelloe introduces us to Hume’s presented to him.
conception of ‘philosophical’ or ‘true’ history, Costelloe emphasizes that Hume sees his-
and his analysis of what Hume means by tory written in this manner as serving the
such history helps to bind together Hume’s crucially important function of laying out
more explicitly philosophical works and his the past before us so that it can serve as a
History of England. Some of Hume’s more guide to future conduct. Firstly, it allows dis-
malicious critics have accused him of effect- tant past events to be used by the scientist of
ively abandoning philosophy after the poor human nature as a means of confirming or
reception afforded the Treatise in order to refuting hypotheses about our mental mech-
pursue money and popular fame through the anisms and dispositions. Thus it provides the
alternative means of writing a best-selling philosopher, in his role as a psychological
history. However, Costelloe brings into focus anatomist, with the data he needs to guide
numerous important continuities between the and refine his conclusions. However, it also
philosophical project pursued in the Treatise serves the second function of improving our
and Hume’s aspirations for his History. moral judgements. People and events close to
Costelloe argues that Hume sees ‘philo- us in time and space are frequently assessed
sophical’ history as an attempt to combine through a prism of partiality that prevents
responsiveness to evidence, rather than the us from seeing how they strike other people,
promptings of partiality or the imagination, and important potential consequences have
with a reconstruction of the past that sus- often not yet had a chance to manifest them-
tains the reader’s interest and gives him or selves. In contrast, if we are reading a histori-
her a lively sense of the truth of the events cal narration of events that took place many
portrayed. A mere propagandist concentrates years ago that involved people not intimately
only on the second of these two tasks; but if connected to us, we have an opportunity
the author makes no attempt to select events to arrive at less biased and better-informed
and shape the narrative in a way that will moral judgements, a habit that can then be
appeal to the reader’s imagination and pow- transferred to situations in which our own
ers of empathy, then the resulting work will interests are at stake. Thus Costelloe main-
be entirely unreadable. Thus a writer of the tains that Hume regarded his philosophical
kind of history that Hume views as worthy and historical investigations as seamlessly
of a philosophical author needs to have the intertwined. Just as abstruse philosophy is
skills to weigh testimony carefully, a passion depicted in the first section of the Enquiry
for the pursuit of the truth that promotes concerning Human Understanding as guid-
impartiality and overrides any temptation ing us to a better understanding of human
to flatter influential patrons, and the capac- nature and easy philosophy is portrayed as
ity to keep the imagination in check so that inspiring us to act virtuously, so too ‘philo-
fanciful associative links do not crowd out sophical’ history continues the task of laying


9780826443595_Intro_Final_txt_print.indd 17 2/9/2012 9:37:59 PM


bare the hidden mechanisms of human action pattern of interpretation that stresses his rad-
while simultaneously depicting virtuous and ical empiricism and his affinities with Locke
vicious characters in such a light that we and Berkeley. Critics of empiricism have seen
sympathize with the former and are repelled Hume as providing the valuable service of
by the latter. exposing how radical empiricism ultimately
Paul Russell brings the volume to a con- collapses into incoherence, whereas philoso-
clusion with a discussion of changing trends phers of empiricist sympathies have often
in the interpretation of Hume’s philosophi- purported to find in Hume the inspiration for
cal position. Russell distinguishes between developing what they hoped would be a suc-
the interpretation of a philosopher’s position cessful version of empiricism that eschewed
and the legacy of that position. Interpretation all a priori metaphysical speculations. Russell
is a matter of arriving at an understanding is not inclined to deny that this interpretative
of a philosopher’s original aims and inten- tradition has inspired much important philo-
tions, whereas the legacy is constituted by sophical work, but he does deny that it offers
the reception of his or her views and their the resources required to construct an accu-
fruitfulness over time. The various compet- rate interpretation of the underlying nature
ing interpretations that might arise form of Hume’s philosophical stance.
part of that reception, but there need be The difficulty that Russell identifies is
no correlation between the dynamism of that it has become increasingly clear that
the interpretative framework and its philo- the empiricist elements of Hume’s thought
sophical fecundity. Similarly, a sensitive and co-exist with a naturalistic programme
well-balanced interpretation might reveal that involves the construction of an intri-
itself over time to be nothing more than cate science of human nature that purports
the accurate signposting of a barren philo- to be based on experience and experiment.
sophical cul-de-sac, whereas an interpreta- However, the interpretation of Hume as a
tion that is little more than a caricature of radical empiricist seems to have sceptical
a philosopher’s actual project might for- implications that are inconsistent with the
tuitously inspire subsequent philosophical development of such a science. Yet if we
developments of great value and independ- water down the empiricism in order to make
ent interest. Russell, however, warns against it more compatible with the positive side
the error of supposing that an important and of Hume’s philosophy, it remains the case
interesting legacy confirms the accuracy of that Hume’s philosophical writings appear
the interpretation that generated it. He also to contain an array of explicitly sceptical
points out that if we complacently allow an arguments that do not need to be embedded
interpretation of a philosopher’s views to go within a framework of radical empiricism in
unchallenged because it is linked to a valu- order to pose a serious challenge to his natu-
able legacy, we are in danger of forgoing ralistic project.
important philosophical developments that Russell strikingly sums up the situation as
might arise from reflection on some plausi- generating the worry that Hume’s philosoph-
ble alternative interpretation. ical outlook is ultimately broken-backed.
In the specific case of Hume, Russell The sceptical aspect of his philosophy, which
argues that much of the fruitfulness of his seems to be clearly present even if it is not to
legacy up to this point has arisen from a be construed as generated by a radical form


9780826443595_Intro_Final_txt_print.indd 18 2/9/2012 9:37:59 PM


of empiricism, does not initially appear to arguments with the more positive aspects of
cohere well with Hume’s science of human his philosophical position. Whether Russell
nature. Russell accordingly maintains that the is right to imply that this is both necessary
way forward is to place Hume’s philosophical and sufficient to permit such a reconcilia-
writings in a new interpretative framework, tion is not yet clear. However, other work by
one that sees Hume not as part of a trium- Russell has certainly undermined the suppo-
virate of British Empiricists or a follower sition that the contents of the Treatise lack a
of Newton or Hutcheson but as someone substantial connection to issues of religion.
who is actively attacking the metaphysical And it can safely be asserted that solving the
and moral foundations of Christianity as a puzzle of how to harmonize the sceptical
member of a partially concealed tradition of and positive sides of Hume’s philosophy is
‘speculative’ atheism. In Russell’s judgement, now widely acknowledged to be one of the
recognition of Hume’s irreligious inten- principal tasks that needs to be accomplished
tions as manifested even within the Treatise if we are ever to possess a truly satisfactory
provides the key to an account of Hume’s interpretation of all the essential elements of
writings that can reconcile his sceptical Hume’s philosophical perspective.


9780826443595_Intro_Final_txt_print.indd 19 2/9/2012 9:37:59 PM

Emilio Mazza

These are a few particulars, which may in . . . understanding, dignity of mind, would
perhaps appear trifling, but to me no bereave even a very good-natured, honest man
particulars seem trifling that relate to so of this honourable appellation. Who did ever
great a man (W. Cullen to J. Hunter, 17 say, except by way of irony, that such a one
September 1776)1 was a man of great virtue, but an egregious
blockhead?’ (EPM App. 4.2 / 314).4
1. ‘WAKE-MINDED’ Young Hume was troubled by a ‘weak-
ness’ of spirits; later on he would see a sig-
‘Our Davie’s a fine good-natured crater, but nificant relationship between ‘delicacy’ and
uncommon wake-minded’, Hume’s mother ‘weakness’ of the mind (LDH 1.17, 3; 1.397,
is supposed to have said in a piece of familial 214). The first as well as the last edition of
assessment.2 And with regard to Hume’s reli- his Essays open with ‘Of delicacy of taste
gious principles, his brother John ventured the and passion’, and only in 1772 does Hume
opinion: ‘My brother Davie is a good enough stop claiming a ‘very considerable connex-
sort of man, but rather narrow minded’.3 ion’ between these delicacies in the original
This latter judgement echoes in its choice of frame of the mind (E 603). His mother’s
words Hume’s own recollection that in Paris supposed saying has been discussed for 150
they ‘used to laugh at me for my narrow way years by those who seek to defend the repu-
of thinking in these particulars’ (LDH 2.273, tation of ‘one of the greatest philosophers of
484), and his description of Rousseau – ‘a any age, and the best friend to mankind’, as
very agreeable, amiable Man; but a great d’Holbach calls Hume, without contradict-
Humourist’ (LDH 2.13, 303; see LDH 2.130, ing a woman of ‘singular merit’, as Hume
381) – indicates that Hume shared with his calls his mother.5
brother a partiality for verbal sallies that com- At the end of his Life Hume celebrates
bined initial restrained praise with a less com- himself as ‘a man of mild dispositions, of
mendatory ending. We seem too to find in the command of temper, of an open, social,
Enquiry concerning the Principles of Morals, a and cheerful humour’, and his temper as
riposte on Hume’s part to his mother’s assess- ‘naturally cheerful and sanguine’ and not
ment of his character: ‘any remarkable defect ‘very irascible’ (MOL, LDH 1.1–3). In


9780826443595_Ch01_Final_txt_print.indd 20 2/9/2012 9:39:42 PM


1757, somewhat between jest and earnest, are the most ‘unfortunate’ books, and the
Hume says he is a ‘good-natured man of a Political Discourses a work ‘successful on the
bad character’ (LDH 1.264, 139) and also first publication’ and ‘well received abroad
(if the text be Hume’s) a ‘very good man, and at home’. The more sustained success
the constant purpose of whose life is to begins in the 1750s, when Hume discovers
do mischief’.6 The history of his writings symptoms of a ‘rising reputation’, including,
shows him as a man of ‘superior genius’, a for example, a ‘railing’ reaction of the clergy
quality that he does not even recognize in (MOL, LDH 1.2–4).
d’Alembert, who was simply a man of ‘supe-
rior parts’, even though after Paris Hume
considered him ‘with some few exceptions
(for there must always be some exceptions) 3. ‘NEVER TO REPLY TO ANY BODY’
. . . a better model of a virtuous and philo-
sophical character’ (LDH 2.110, 363). In about 20 years (1739–61) Hume publishes
almost every kind of writing: a Treatise,
its Appendix and Abstract; a Letter to a
friend and a True Account of the conduct of
2. MY OWN (UNSUCCESSFUL) another friend, the Essays, the Philosophical
WRITINGS Essays and the Enquiries; the Discourses and
the Dissertations; the History, the Natural
In 1734, when he begins ‘to despair of ever History and a Dialogue. He also receives
recovering’ from his ‘Disease of the Learned’, almost every kind of answer. In 1766 he
Hume wrote ‘a kind of History of my Life’ observes: ‘I could cover the Floor of a large
(LDH 1.13, 3; 1.17, 3); in 1776, when his life Room with Books and Pamphlets wrote
is really ‘despaired of’, he writes ‘the History against me, to none of which I ever made the
of my own Life’ or My own Life (LDH 2.318, least Reply, . . . from my Desire of Ease and
522A; LDH 2.323, 525). This short ‘funeral Tranquillity’ (LDH 2.92, 351).
oration of myself’, Hume says, contains ‘little With regard to the years 1749–51, in 1776
more than the History of my Writings’, since Hume declares he has ‘fixed’ and ‘inflexibly
‘almost all my life has been spent in liter- maintained’ a resolution ‘never to reply to
ary pursuits’ (MOL, LDH 1.7). The rhythm any body’ (MOL, LDH 1.3). He starts assert-
of the Life is the alternation of learning and ing this resolution in the second half of the
business, expectations and disappointments, 1750s, as a reaction to the ‘Warburtonian
which recalls that of action and repose in School’, but in 1760 he declares that he
the ‘Refinement in the Arts’ (E 270). Every formed it ‘in the beginning of my Life, that is,
disappointment is overcome by character, of my literary Life’ (LDH 1.320, 172), which
‘command of temper’ and ‘cheerful humour’ seems therefore to begin with a commitment
(MOL, LDH 1.7). Hume’s Life is also a his- to refuse any literary controversy. Like many
tory of the reception of his writings, where he official claims, however, this is not completely
commonly distinguishes between immediate reliable even though it does contain a sub-
and gradual success. The want of it is chiefly stantial admixture of truth. He often replies
measured by the standard of silence. The indirectly to his critics in his writings, and
Treatise and the first volume of the History sometimes he is even driven to the expedient


9780826443595_Ch01_Final_txt_print.indd 21 2/9/2012 9:39:43 PM


of explaining that he is giving an answer and both in the fact that he has never ‘preferred a
that it should be extended to different adver- request to one great Man, or even . . . [made]
saries, as in the 1775 ‘Advertisement’ to the advances of friendship to any of them’ and in
Enquiries (LDH 2.301, 509). the fact that he has nevertheless found him-
In 1757 someone suggests that he has self on good terms with such people in his
deliberately ‘so larded his Work with personal affairs, public business, and while
Irreligion’ that the first volume of the composing his History (MOL, LDH 1.5–6;
History ‘might sell’,7 and Hume observes LDH 1.113, 63; 1.295, 156; 1.355, 191;
that the few ‘Strokes of Irreligion’ are of 1.427–28, 232; 2.188, 422). As a man now
‘small Importance’, even though they are beyond middle age working for the Northern
likely ‘to encrease the sale’ (LDH 1.250, 132; Department, he finds that ‘to a Man of a lit-
1.256, 136). A few months later he allows erary turn, who has no great undertaking in
that he would accept the challenge to defend view, . . . public Business is the best Ressource
The Natural History of Religion against of his declining Years. Learning requires the
Warburton’s criticisms were he attacking his Ardor of Youth’ (LDH 2.385, 137). Thirty-
‘principal Topics’. As he tells the bookseller: three years before, in spring 1734, trying to
‘The Hopes of getting an Answer, might leave his distemper behind and working on
probably engage [Warburton] to give us the Treatise, he found ‘two things very good,
something farther of the same kind; which Business & Diversion’, and resolved ‘to seek
at least saves you the Expence of advertising’ out a more active life’, laying ‘aside for some
(LDH 1.265–67, 140). time’ his pretensions in learning (LDH 1.17,
Concerning his no-reply resolution, in 3; MOL, LDH 1.1).
1758 Hume still maintains that he ‘shall In 1767 the Earl of Rochford remembers
probably uphold it to the End of [his] life’. that Hume is ‘unfit for business’, and Hume
The Concise and Genuine Account of his dis- himself has already admitted that the office
pute with Rousseau recalls that Hume ‘hath of secretary requires ‘a Talent for speaking
seen his writings frequently censured with in public to which I was never accustomd’
bitterness . . . without ever giving an answer (LDH 1.519, 289).9 However, Hume’s essay
to his adversaries’, yet, in the case of the ‘Of Eloquence’ (1742) attacks, following
Rousseau imbroglio, the ‘circumstances’ were Swift and La Bruyère the ‘antient Prejudice
such as to draw Hume into a scandal, ‘in spite industriously propagated by the Dunces in
of his inclinations’. Consistent with them, he all Countries, That a Man of Genius is unfit
authorizes the editors to declare ‘that he will for Business’ (E 621), and the first Enquiry
never take the pen again on the subject’.8 (1748) claims that the accuracy of abstruse
philosophy is ‘subservient’ to every art or
profession: the politician, the lawyer and the
army general may take advantage from it
4. NOT UNFIT FOR BUSINESS: ‘THE (EHU 1.9 / 10). In part, at least, this sounds
ARMY IS TOO LATE’ like a defence of those aspects of his life and
career that were not directly connected to
Like Lucian in De mercede conductis and his literary and philosophical pursuits, for
Apologia pro mercede conductis, which Hume at various times he found himself taking on
first quotes in 1751–2, he takes satisfaction the roles of clerk for a merchant in Bristol


9780826443595_Ch01_Final_txt_print.indd 22 2/9/2012 9:39:43 PM


(1734), companion and tutor of a marquess dissipation; yet always returned to my closet
near St Albans (1745–6), secretary to a gen- with pleasure’ (LDH 1.451, 244).
eral and judge-advocate in Lorient (1746–7), In 1746 Hume receives an ‘unexpected’
secretary and aide-de-camp to the same invitation from St Clair to go with him as sec-
general in Vienna and Turin (1748), library- retary in his military expedition, which was
keeper in Edinburgh (1752–7), secretary to planned to be an attack on French Canada
the Embassy and Chargé des affaires in Paris but came to its conclusion on the coast of
(1763–6), and Under-Secretary of State for the Brittany (MOL, LDH 1.2; LDH 1.382, 206;
Northern Department in London (1767–8). 1.92, 51; NLH 24, 10). He arranges his
He also considers (but ultimately rejects) the ‘Departure for America’ (‘Such a Romantic
‘not agreeable’ life of the ‘Travelling Tutor’ Adventure, & such a Hurry’) with one box
(LDH 1.18, 3; 1.17–8, 24; 1.35–6, 14; 1. of books and one of paper in his trunk (LDH
57–8, 24; NLH 26, 10), even though he is 1.90, 50).10 Being asked whether he ‘would
often ‘mortally sick at sea’ (LDH 2.206, 432; incline to enter the Service’, he answers that
1.214, 105; also see LDH 1.114, 64; 2.95, at his years he could not ‘accept of a lower
352) and claims that ‘Shortness . . . is almost commission than a company’ (LDH 1.94,
the only agreeable Circumstance that can be 52). One year after he says that for the ‘Army
in a Voyage’ (LDH 1.105, 56). [it] is too late’ (NLH 26, 10). The expedi-
Every time he is enjoying his solitude tion is a ‘failure’, but it gives rise on Hume’s
Hume receives an ‘invitation’ he cannot refuse part to a beautiful letter to his brother, a
(MOL, LDH 1.2, 4, 5–6). According to the brief journal or hypomnema, a piece on ‘The
correspondence, his life is a permanent yearn- descent on the coast of Brittany in 1746’, and
ing for (philosophical) retreat and leisure, con- possibly an article (LDH 1.99, 54; 1.94–8,
tinuously thwarted by external circumstances, 53). The expedition also shapes Hume’s
leading him into some practical business: ‘I opinions about soldiers. Major Alexander
lived several years happy with my brother at Forbes, for example, is described as ‘a Man
Ninewells, and had not his marriage changed of the greatest Sense, Honour, Modesty,
a little the state of the family, I believe I should Mildness & Equality of Temper in the
have lived and died there,’ he says in 1759 World’: ‘His Learning was very great for a
about his own ‘reluctance to change places’, man of any Profession, but a Prodigy for a
even though in 1763 he has ‘so often changed’ Soldier. His Bravery had been try’d & was
his places of abode that he comes to think that unquestion’d.’ When Forbes kills himself as
‘as far as regards happiness, there is no great a result of anxiety and fear that he may have
difference among them’ (LDH 1.295, 156; been guilty of a dereliction of duty, Hume
1.415, 224; see also LDH 1.243, 128; 1.246, maintains that in the course of dying from
130; 1.531, 295; 2.189, 423). With regard to his self-administered injuries, he expressed
the years spent with General St Clair, the Life a ‘steady Contempt of Life’ and ‘determind
claims that ‘these were almost the only inter- philosophical Principles’. And after Hume
ruptions which my studies have received dur- has seen his friend die in front of him, it is
ing the course of my life’ (MOL, LDH 1.2–3), probably Hume who also undertakes the
even though 15 years before he has allowed: official duty of recording that one Dougal
‘I have frequently, in the course of my life, Steuart was made Captain ‘in room of Alexr
met with interruptions, from business and Forbes deceast’ (LDH 1.97, 53).11


9780826443595_Ch01_Final_txt_print.indd 23 2/9/2012 9:39:43 PM


In 1734 Hume had compared the soldier’s though soldiers have their exceptions, Hume
courage to the devotee’s devotion (LDH passed his St Clair years ‘agreeably, and in
1.21, 4). In 1748 he publishes ‘Of National good company’ (MOL, LDH 1.2-3): reliable
Characters’, where a few pages could be enti- officers, learned physicians, whist-players
tled ‘The Soldier and the Priest’. It is a double and humorous people who dedicate them-
reaction to his academic and military adven- selves to the ‘service of the Ladys’. Among
tures in Edinburgh and Lorient. In 1743 priests, on the contrary, ‘gaiety, much less
Hume reads Leechman’s sermon on prayer, the excesses of pleasure, is not permitted’
and sends him some remarks on argument (E 201n). He is ‘in the Army’ and he calls
and style, together with 22 small faults that it ‘our Family’ (LDH 1.97, 53; 1.132, 64).14
the author does not even take into consid- With these ‘friends or confidents’ – he says
eration. The sermon, Hume concludes, una- with Quintilian and Svetonius, or more sim-
voidably makes his religious author ‘a rank ply with Voiture – he can be free ‘in seriis et
Atheist’ (NLH 10–14, 6). Despite his youth- in jocis, – amici omnium horarum’ (in grave
ful claim that ‘there is nothing to be learnt and jocular manners, – friends of all hours)
from a Professor, which is not to be met with (LDH 1.102, 56).
in Books’,12 in 1744 Hume attempts in vain In 1747, when St Clair invites Hume to
to become professor at the University of go over to Flanders with him (LDH 1.108–9,
Edinburgh, and declares himself extremely 61), he has ‘a great curiosity to see a real
surprised that the ‘accusation of Heresy, campaign’, notwithstanding his fears of the
Deism, Scepticism, Atheism &c’ is supported ‘expense’ and looking ridiculous as a result
by the ‘Authority of Mr Hutcheson & even of ‘living in a Camp, without any Character
Mr Leechman’ (LDH 1.58, 24). & without any thing to do’ (NLH 23, 9).
In 1741 Leechman published another Nothing could be ‘more useful’ to his ‘histor-
sermon on the character of the priest. ‘Of ical projects’. Hume looks forward to pick-
National Character’ is also an answer to ing up a great ‘military knowledge’, by ‘living
him. Leechman claims we can never clearly in the General’s family, and being introduced
‘unvail’ to mankind their ‘hidden hypocrisy’, frequently to the Duke’s’ (ibid.). In 1748 he
nor justly contempt the devout worship- attends St Clair in his mission to Vienna and
pers by calling the outward displays of their Turin, notwithstanding an ‘infinite regret’ for
inward devotion ‘solemn grimaces, and hypo- leaving ‘stores of study & plans of thinking’
critical airs’;13 Hume replies that the clergy- (LDH 1.109, 61; 1.111, 62). In accordance
men ‘promote the spirit of superstition, by a with the opinions of Lucian, Bayle, Addison
continued grimace and hypocrisy’ and this and the Guardian, and following the advice
‘dissimulation often destroys’ their ‘candor’ contained in a volume by Polybius, which
(E 200n). Hume denounces their conceited he keeps in his hand (LDH 1.100, 54), he is
ambition, professional faction and perse- looking for ‘an opportunity of seeing Courts
cuting spirit. In contrast, soldiers are ‘lav- & Camps’:
ish and generous, as well as brave’, ‘candid,
honest, and undesigning’. Since ‘company’ is this knowledge may even turn to account
their sphere they can acquire ‘good breeding to me, as a man of letters, which I con-
and an openness of behaviour’ and a ‘con- fess has always been the sole object of my
siderable share’ of politeness (E 199). Even ambition. I have long had an intention,


9780826443595_Ch01_Final_txt_print.indd 24 2/9/2012 9:39:43 PM


in my riper years, of composing some that always, even from your earliest Years,
History; . . . some greater experience did most easily beset you’ (LDH 1.438, 237;
of the Operations of the Field, & the 2.353, App. C. V).
Intrigues of the Cabinet, will be requis- In Turin Hume becomes bored and sick.
ite, in order to enable me to speak with
Admiral John Forbes called him ‘the sleeping
judgement upon these subjects (LDH
philosopher’, someone says he was ‘affected
1.109, 61).
by a most violent Fever’, some other that he
‘received Extreme Unction in a dangerous ill-
St Clair arrives in Turin on the 8 May, Hume ness’. He hangs around with Lord Charlemont
and St Clair’s nephew, Sir Harry Erskine, and reads Montesquieu’s Esprit (LDH 1.133,
about eight days later; on 29 November 65).18 Consistent with his Treatise, and in the
1748 they all set out.15 The result of the mis- name of ‘sympathy’, he enjoys the pleasure
sion is a ‘long epistle’, which he calls a ‘sort and beauty of extended, fertile, cultivated
of Journal of our Travels’ (LDH 1.114, 64; plains. He wishes to make ‘a short Tour thro’
1.132, 64). Before the departure Hume opti- some of the chief Cities of Italy’, but appar-
mistically contrasts his situation with that of ently the Duke of Newcastle rejects the
the ‘severe’ Lord Marchmont, who, ‘entirely request.19 He does the accounts (as he did in
employed in the severer studies’, suddenly Bristol) and examines the Sardinian docu-
opens his eyes on a ‘fair nymph’ aged just 16 ments in the Commissary’s office.20 He writes
and marries her in a few days: ‘they say many St Clair’s official letters and copies them into
small fevers prevent a great one. Heaven be a letter book. He probably suggests passages
praised, that I have always liked the per- for St Clair’s letters (like the observation of
sons & company of the fair sex: For by that the historians that ‘Britain has commonly lost
means, I hope to escape such ridiculous pas- by Treaties what she gain’d by Arms’) and cer-
sions’ (LDH 1.110, 61). Ten days after his tainly receives suggestions for his future writ-
arrival he is already ‘troubled’ by an ‘indis- ings: ‘Of the Balance of Power’ discusses the
position’ connected with the ‘pretty women’ peace of Aix la Chapelle, and the dying Hume
of Turin. After two months he declares an is still remembering those inconceivably ‘good
‘attachment’ for a Countess of 24.16 The terms’ that France had granted to Britain.21
Turin-based Madame Duvernan anticipates Hume’s experience in Turin resumes that
the Parisian Madame de Boufflers and their begun in Lorient and prepares the way for
extrovert public reputations stand some- his 1760s appointments in Paris and London.
what in tension with Hume’s claim that he, General St Clair, Lord Hertford and his
like Mandeville’s perfect sociable benevolent brother General Conway all wanted Hume
man,17 took a ‘particular pleasure in the com- with them. St Clair ‘positively refused to
pany of modest women’ and had therefore accept of a Secretary from the Ministry’, and
no reason to be ‘displeased with the recep- Hume goes ‘along with him’; some 15 years
tion I met with from them’ (MOL, LDH later in 1763 Hertford is ‘resolved never to
1.7). In summer 1764 he reminds his rever- see, or do business with his Secretary, and
end friend Jardine that ‘A Man in Vogue will therefore desired [Hume] should attend him’
always have something to pretend to with the (LDH 1.111, 62; 1.421, 228). In March 1767
fair Sex’, and Jardine banters: ‘An inordinate Hume is ‘deeply immersed in study’, when
Love of the fair Sex . . . is one of those Sins, Hertford surprisingly urged him to ‘accept of


9780826443595_Ch01_Final_txt_print.indd 25 2/9/2012 9:39:44 PM


the office of Depute-Secretary of State under been jailed, he would have risked igno-
his brother’. He cannot refuse and sees him- miniously losing his life for having or not
self ‘embarked for some time in state affairs’. having drowned himself.24
Yet, he says, ‘I foresaw also that a place was
offered me of credit and confidence; that it It was the time of extravagant requests, like
connected me with General Conway’ (LDH that of the ‘Apulien Philosopher’ Vincenzo
2.123, 374; also see 1.511, 282). Hume says Maria Gaudio (1722–74). In January 1764
he feels like a ‘banished man in a strange he wrote to Hume asking him two questions
country’. He is not ‘hurry’d with Business’ ‘for the good of human kind’: ‘How many
and commonly attends on the Secretary and which physical and moral causes pro-
‘from ten to three’. He has ‘no more Business duce the variety and contrariety of opinions
than would be requisite for [his] Amusement’ among men?’; ‘How to reduce the sum of
in London (LDH 2.123, 374; 2.127, 377). evils and increase that of goods?’.25
Hume is not only employed in ‘cypher- When he is Under-Secretary in London,
ing and decyphering’: during his public ‘degenerated into a petty Statesman’ and
activities he does not forget his opinions. ‘entirely occupyed in Politics’ (LDH 2.128,
When the burden of diplomatic work at the 379), he meets another extravagant case: ‘one
embassy in Paris is falling entirely upon him Giraldi, an Italian Physician’. Giraldi, who
(LDH 511–12, 282–3), and friends start call- is in London and needs protection in Italy,
ing him ‘a man of Business’ (LDH 1.421n, addresses himself to Hume; Hume reporting
228),22 he saves from prison and death an to Lord Shelburne:
Englishman who attempted to kill himself
in the Seine. Marischal Keith congratulates [He] seems to me a man of sense and
Hume: ‘you have done many good works in learning, and whose orthodoxy has
your Ministerial functions, I am sure it was of consequence been brought under
great and I suppose just suspicions. . . .
one to save a pour fellow from the gallows,
It seems a Cardinal, in his absence, fell
who chose rather to drown than starve’.23
in love with his wife, and has taken
And Diderot has the complete story: her into keeping; and on the physician
expressing some displeasure at this treat-
They fished him out alive. They brought ment, his Eminence, who has great credit
him to the Grand Châtelet, and the in the Holy See, has threatend to have
Ambassador had to interpose his author- him put into the Inquisition . . . . He has
ity to prevent them from putting him to addressed himself to me, on the suppos-
death. Some days ago Mr. Hume told us ition, no doubt, that I woud sympathize
that no political negotiation had been with his cause. I conjure therefore your
more intriguing than this affaire and that Lordship, if there be any virtue, if there
he had been obliged to go twenty times to be any praise, if there be anything comely
see the first president before he could make or of good report, to save the poor her-
him understand that there was no article, etic from the flames . . . his case wou’d
in any of the treaties between France and puzzle Rhadamanthus himself: as a cuck-
England, that forbade an Englishman old, he ought to go to heaven; as a her-
from drowning himself in the Seine under etic to hell. But, without joking, his case
penalty of being hanged. And he added is worthy of compassion; and I recom-
that, if his compatriot had unfortunately mend it to your Lordship’s humanity.26


9780826443595_Ch01_Final_txt_print.indd 26 2/9/2012 9:39:44 PM


Some ‘fresh intelligence’ discovers to Hume of a large library’. At the beginning of the
that Giraldi ‘lives in intimacy with Gemino, 1760s ‘the copy-money given [him] by the
no great sign of his orthodoxy’. His project booksellers, much exceeded any thing for-
was to retire to the Island of Capri, which merly known in England’ and Hume is ‘not
Giraldi ‘represents as an earthly paradise’, only independent, but opulent’. In 1766 the
and – Hume concludes – ‘indeed the only Parisian Secretary returned to Edinburgh,
paradise he ever expects to go to’.27 ‘not richer, but with much more money, and
a much larger income’ than he left it. He
was now ‘desirous of trying what superfluity
could produce’. In 1769 the London Under-
5. MY OWN FORTUNE Secretary returned to Edinburgh ‘very opu-
lent’ (he ‘possessed a revenue of 1000 l. a
‘Money – says the Concise Account – is not year’) and with the double prospect of long
universally the chief object with mankind; enjoying his ‘ease’ and of seeing the increase
vanity weighs farther with some men’.28 Not of his ‘reputation’ (MOL, LDH 1.1–6).
entirely exempt from vanity, Hume never Thanks to Hertford’s family he really was, as
abandons the money that belongs to him he once wrote from Paris, ‘in the high Road
‘of right’, like the quarter salary from the to Riches’ and ‘in the high Road to Dignities’
Annandale Estate and the half-pay military (NLH 78, 38; LDH 1.421, 228).
pension from the Treasury: after more than
15 years he is still fighting for it. But he is
also ready to retract his application at the
Advocates’ Library, retain the office and give 6. STRIKE OUT STERNE:
a friend a bond of annuity for the salary. In FASHION IN PARIS
1747 he calls himself ‘a good Oeconomist’
(NLH 26, 10). Riches are valuable ‘at all Hume was in Paris, Reims and La Flèche
times, and to all men’ (E 276), and in his in the 1730s, Paris in 1748 and Paris again
short Life he spends some words celebrating in the 1760s. He constantly saw himself
his income. through the French looking-glass: the first
He says he was ‘of a good family’, but ‘not philosophical readings and the successful
rich’. As a younger brother, his ‘fortune’ was French translations of his writings (in 1761
‘very slender’ and therefore unsuitable to his the Essais Philosophiques earn themselves a
literary plan of life. So he laid down a rule: place in the Index Librorum Prohibitorum),
‘to make a very rigid frugality supply my the Embassy, the Court, the Great Ladies
deficiency of fortune’. In 1745 his Annandale (Madame de Boufflers) and the Philosophes
appointments made a ‘considerable accession (Rousseau). In 1745 Hume first expresses
to [his] small fortune’; in 1746–8 the St Clair the slightly melancholy intention of retiring
appointments earned him a ‘fortune’ that he to the South of France (NLH 17, 7). In the
calls ‘independent’. He wanted ‘to maintain Life he remembers living in Paris as a ‘real
unimpaired [his] independency’ and he is satisfaction’: ‘I thought once of settling there
now ‘master of near a thousand pounds’. In for life’ (MOL, LDH 1.6). Everyone affects
the 1750s the Faculty of Advocates gave him to consider him ‘one of the greatest geniuses
‘little or no emolument’ but the ‘command in the world’ (LDH 1.410, 223), since in


9780826443595_Ch01_Final_txt_print.indd 27 2/9/2012 9:39:44 PM


Paris, unlike London, a man of letters ‘meets Sterne unveils Hezekiah-Hertford’s ‘vanity’
immediatly with Regard & Attention’ (LDH and ‘ostentation’. Later on, at Hertford’s
1.497, 272). table, Sterne had a dispute with Hume (a ‘lit-
‘Anglomania was the manner of the place,’ tle pleasant sparring’, he says). In his sermon
Charlemont observes, and ‘Hume’s Fashion’ Sterne had celebrated integrity and miracles,
was ‘truely rediculous’: ‘no Lady’s Toilet and blamed pride and hypocrisy. At dinner
was compleat without Hume’s attendante.’ ‘David was disposed to make a little merry
Walpole is more concise: ‘Mr. Hume is fash- with the Parson; and, in return, the Parson
ion itself.’29 Indeed, he was more celebrated was equally disposed to make a little mirth
for his name rather than his writings, for with the Infidel’. Sterne concludes: ‘it is this
his economical, historical and anti-religious amiable turn of character, that has given more
writings rather than his philosophical opin- consequence and force to his scepticism, than
ions, and for his general opinions instead all the arguments of his sophistry.’33
of his precise arguments. The French mode At the end of 1765 Sterne publishes his
entailed ‘excessive civilities’ (MOL, LDH Sermons with a probably less ‘unlucky’ and
1.6), but what was ‘at first oppressive’ in offensive version of ‘Hezekiah’. He is ready to
two months ultimately sat ‘more easy’, espe- ‘quarrel’ with Hume by calling him a ‘deist’,
cially as he gradually recovered the ‘facility’ if he will not add his name to the ‘most splen-
of speaking the language (LDH 1.417, 225; did list’ of subscribers. The Sermons came
1.414, 224; 1.498–9, 272).30 The Life sums out, but Hume’s name was not in the list. In
up: ‘Those who have not seen the strange 1767 Hume recalls the ‘usual extravagance’
effects of modes, will never imagine the of Sterne’s productions (NLH 160, 80), and
reception I met with at Paris’. And Hume in the Sentimental Journey Sterne plays with
reports, in a remark that he was later to Hume the historian, his ‘excellent heart’
strike out, that ‘Dr Sterne told me, that he and bad knowledge of French. When Sterne
saw I was [celebrated in town] in the same dies, Hume subscribes five guineas for his
manner that he himself had been in London: widow.34 In 1773 Hume detects in Brydone’s
But he added, that his Vogue lasted only one Tour through Sicily and Malta ‘some Levities,
Winter’ (MOL, LDH 1.6).31 too much in the Shandean Style’, which he
In 1762 Sterne does not worship the advises the author to ‘obliterate’. He also says
French goddesses, but, he says, he has ‘con- that Tristram Shandy is ‘the best Book, that
verted many unto Shandeism’. In 1764 he has been writ by any Englishman these thirty
preached a sermon deemed ‘offensive’ (he Years . . . , bad as it is’ (LDH 2.269, 482).
calls it ‘innocent’) at the Embassy Chapel. Three years later he first writes and then
Hertford has just furnished the new and strikes Sterne’s name out of his Life.
‘magnificent’ Hôtel de Brancas, which gave
‘the subject of conversation to the polite
circles of Paris, for a fortnight at least’.32
Sterne preaches on the Book of Kings and 7. LIFELONG LUCIAN AND
Hezekiah, who foolishly showed all the pre- THE IRISH SKYTHS
cious things that were in his house; even his
wives and concubines, adds the preacher. ‘Lucien est votre auteur favori, et . . . je
Behind ‘urbanity or the etiquette of courts’, l’aime bien autant que vous’ (‘Lucian is


9780826443595_Ch01_Final_txt_print.indd 28 2/9/2012 9:39:44 PM


your favourite author, and . . . I love him as cyphers was part of Hume’s official duties in
much as you do’), Morellet reminds Hume both Turin and Paris. Murphy was also the
in 1766.35 Lucian follows Hume through- editor of The Select Dialogues of Lucian, first
out his literary career. In 1742 he allows that printed in 1744. In 1767 Hume compares
‘some Dialogues’ of Lucian are among the him to the ‘Royal philosopher Anacharsis’.
few excellent pieces of pleasantry in ancient Murphy usually calls himself ‘Ô Murraghoo
literature (E 134). The explosion of Lucian Rex’, Anacharsis is one of Lucian’s dialogues
occurs in the second half of the 1740s. ‘The and the name of a character in Scytha sive
Sceptic’ (1753 version) suggests that we can hospes.
improve our mental disposition by reading In 1765 Hume had refused to go to
the ‘entertaining moralists’ and engaging Ireland with Hertford: the Dubliners and
with the ‘imagination of Lucian’, if nature the Londoners did not want the Scottish phi-
has endowed us with a ‘favourable’ tem- losopher to make such a visit. Hertford had
per (E 179n). Moreover, the moral Enquiry prepared him an apartment in the Castle
assesses Lucian as ‘licentious with regard to of Dublin, but Hume thought it ‘not worth
pleasure’ but a ‘very moral writer’ in other while’: ‘It is like Stepping out of Light into
respects, and accordingly regards it as highly Darkness to exchange Paris for Dublin’
significant that he ‘cannot, sometimes, talk (LDH, 1.514, 285). In Ireland the philoso-
of virtue, so much boasted, without betray- pher and historian was ‘excessively disliked’.
ing symptoms of spleen and irony’. In Great It will be ‘an Age or two at least’ before the
Britain, adds Lucianic Hume, such a ‘contin- Irish can perceive his doctrines, and ‘perhaps
ued ostentation’ of public spirit and benevo- an age or two more’ before they can relish
lence inclines men of the world ‘to discover a them, writes Chaplain Trail: ‘I could almost
sullen incredulity on the head of those moral as soon promise Antichrist himself a welcome
endowments’ (EPM 6.21 / 242). Reception.’36
In the first Enquiry, where he laments the Possibly alluding to Hume’s account of
‘harsh winds of calumny and persecution’ the ‘most barbarous’ cruelty allegedly perpe-
directed against philosophy, Hume bitterly trated by the Irish during the ‘universal mas-
observes: ‘it does not always happen, that sacre’ of the English in 1641, where ‘[n]o age,
every Alexander meets with a Lucian, ready to no sex, no condition was spared’ (H 5.55,
expose and detect his impostures’ (EHU 11.2 / 341), Murphy says in June 1767 that Hume
132–3; EHU 10.23 / 121). In all antiquity, says considers the Irish ‘Savages’ because they ‘eat
the ‘Populousness of ancient nations’, there Human Flesh when [they] can get it good’.
is not a philosopher ‘less superstitious’ than The native Irish, adds Murphy, are ‘provd by
Lucian (and Cicero). The ‘agreeable’ Lucian, History to be Scythians by Descent, or rather
says the Natural History, had ‘employed the . . . Skyths, which word has been corrupted
whole force of his wit and satire against the into Scots’.37 In a swift Lucianic style Murphy
national religion’ (E 463n; NHR 12.174). invites Hume to Ireland, ensuring him he will
Morellet is not the only translator of be treated ‘as safe, as kindly . . . as ever [he]
Lucian with whom Hume was acquainted. was in Paris, or Edenburgh’:
In Turin he met Edward Murphy (1707–77).
Murphy’s repeated ‘grand query’ to Hume We do not devour Strangers who visit us
concerns a cypher he invented, and the use of as Friends; not even such as, we know,


9780826443595_Ch01_Final_txt_print.indd 29 2/9/2012 9:39:44 PM


come to rob us. If you will not send me ‘all [his] manuscripts . . . desiring him to
a good Answer to my Query, I will go at publish [the] Dialogues’. Hume reserves him-
the Head of my Mighty Men and extir- self the power to alter his will at any time,
pate your Nation. Yea, we would eat you ‘even in death-bed’.43 And so he does. He first
all up, Man, Woman, and Child; but that
leaves ‘entirely’ to Smith’s discretion ‘at what
we are a little nice about the Quality of
time . . . , or whether’ to publish the Dialogues
our Fleshmeat. I never saw a Piece of a
Scotchman (though we have a whole (LDH 2.316–18, 522–522A). Smith accord-
Province stock’d with them) at any gen- ingly thinks he has persuaded Hume to
teel Table here.38 allow him to leave the Dialogues concerning
Natural Religion unpublished if he views that
as advisable. Hume feels that his own death
The pretended ‘Reverend Murphy’, who in is imminent, and begins to think in terms of
Rome bought a papal plenary indulgence printing a ‘small edition’ and giving his edi-
for three crowns and made some remarks on tor Strahan the ‘literary property’ (LDH
‘the pope and his fellow jugglers’, denounces 2.323, 525).44 In a first codicil he makes
Lucian’s ‘entire want of Candour, while he Strahan ‘entirely Master’ of his manuscripts
talks against the Christian Religion’, yet, (LDH 2.325, 527): the Dialogues must be
he adds, ‘it is impossible not to admire his ‘printed and published any time within two
matchless Abilities’.39 Murphy’s translation Years after [his] Death’, and the Life ‘pre-
of Lucian seems to have been a (Greek) text- fixed to the first Edition of [his] Works’.45 In
book at Trinity.40 Three months before dying a second codicil he ordains: ‘if my Dialogues
the ageing Hume invites his nephew David . . . be not publisht within two Years and a
to read Lucian ‘sometimes’ and not to forget half after my Death, as also the Account
his Greek, to mix the volumes of ‘taste and of my Life, the Property shall return to my
imagination’ with ‘more serious reading’, as Nephew’.46 Smith criticizes this ‘unnecessary
the young Hume used to read at his pleasure clause’,47 and Hume leaves Smith a ‘security’
‘sometimes a Philosopher, sometimes a Poet’.41 copy (LDH 2.334, 538). Finally, two days
before dying, he informs Smith and Strahan
that he is leaving his nephew the ‘property of
the Manuscript in case by any accident [to
8. HUME AND SMITH: A LIVING Strahan’s Life] it should not be published
SUMMER DIALOGUE within three years after [his] decease’ (LDH
2.336, 540).
Lucian’s writings have something to say Smith is trying to move away from the
about Hume’s death and legacy, the publica- Dialogues: ‘If you give me leave I will add a
tion of Hume’s Life and Dialogues, and the few lines to your account of your own life.’
role taken by Adam Smith: an alternation It would make ‘no disagreeable part of the
of intentions and second thoughts, trust and history’ to relate Hume’s ‘want of an excuse
worries, will and codicils. In 1773 Smith is to make to Charon, the excuse you at last
not well and leaves Hume ‘all [his] literary thought of, and the very bad reception which
papers’ and the publication of the ‘history of Charon was likely to give it’. Inspired by a
the Astronomical Systems’;42 on 4 January preceding letter to Wedderburn, Smith wants
1776 Hume is seriously sick and leaves Smith to celebrate Hume’s ‘steady cheerfulness’


9780826443595_Ch01_Final_txt_print.indd 30 2/9/2012 9:39:44 PM


towards death.48 In the January will Hume to obtain a delay from Charon.55 As a writer
had asked Smith to take the pains of ‘correct- very busy in ‘correcting [his] works for a new
ing and publishing’ the Dialogues; now Smith edition’, Hume asks for ‘a little time’ to see
offers to publish an Addition to Hume’s Life ‘how the public receives the alterations’; as
and to ‘correct the Sheets of the new edition an anti-religious writer ‘endeavouring to
of [his] works’, the Dialogues excepted.49 open the eyes of people’ he asks for ‘a few
Hume answers Smith that he is ‘too good in years longer’ to have ‘the satisfaction of
thinking any trifles that concern [him] are so seeing the downfall of some of the prevail-
much worth of [his] attention’ and gives him ing systems of superstition’. Charon is not
‘entirely liberty to make what Additions [he] convinced, Hume wants ‘a lease for so long
please[s] to the account of [his] Life’ (LDH a term’: ‘there will be no end’ of correcting
2.336, 540). Strahan tells Smith that the and opening people’s eyes will take ‘many
Addition will be ‘highly proper’, like ‘every hundred years’.56 By the way, Lucian – says
particular respecting that great and good the Enquiry – ‘entirely opened the eyes of
man’.50 At the beginning of October Smith’s mankind’ (EHU 10.23 / 121).
Addition is ready: ‘I think there is a propri- Hume, says Smith, ‘diverts himself with
ety in addressing it as a letter to Mr Strahan correcting his own works’, he makes ‘many
to whom [Hume] has left the care of his very proper corrections, chiefly in what con-
works.’51 Acknowledging that Smith’s narra- cerns the language’, and Smith is ready to
tive is the consequence of his ‘request’ and ‘revise the sheets and Authenticate its being
Hume’s ‘approbation’ of it, Hume’s brother according to his last corrections’, as he has
‘much’ approves it and suggests a few altera- ‘promised’ to Hume.57 In the Addition Smith
tions, consistent with the ‘short and simple prepares the excuse of correcting by telling
a manner’ of the Life.52 Smith adopts his the reader that Hume ‘continue[s] to divert
remarks and sends Strahan his ‘small addi- himself, as usual, with correcting his own
tion’ to Hume’s ‘small piece’ (LDH 2.318, works for a new edition’.58 This agrees with
522A; 2.323, 525), and Strahan likes it Hume’s self-representation as an author
‘exceedingly’.53 In 1777, contrary to Hume’s extremely ‘anxious of Correctness’ (LDH
dispositions, Strahan publishes the Life and 1.175, 82; 2.304, 511; see ibid. 1.38, 16;
the Addition ‘separately’ from Hume’s work. 1.175, 82; 2.239, 455). ‘There is no End of
The Dialogues are published in 1779 (three correcting’, proclaims Hume in 1763, and in
years after his death) by Hume’s nephew, 1771, ‘an Author may correct his works, as
possibly with some help from Strahan. long as he lives’ (LDH 1.379, 203; 2.246–7,
According to friends and doctors, in his last 457). As he tells the printer, ‘I am perhaps the
months Hume ‘amuses himself’ with reading only Author . . . who gratutiously employ’d
(classic) ‘amuseing books’.54 And in his last great Industry in correcting a Work, of which
days he is reading ‘the dialogues of Lucian’, he has fully alienated the Property’ (LDH
the ‘Dialogues of the Dead’ and ‘the dialogue 2.239, 455).
entitled Kataplu’. On 8 August 1776 Hume Hume, who finds ‘curious’ that an author
has a Lucianic conversation with himself. By could have no patience to read over his pub-
mixing up Megaphentes’ excuses (‘Cataplus’) lished works (NLH 62, 31; LDH 2.31, 314),
with Socrates’ final resistance (Dialogues of recalls in the 1770s a ‘saying’ of Jean-Baptiste
the Dead), he invents ‘several jocular excuses’ Rousseau the poet and offers his own version:


9780826443595_Ch01_Final_txt_print.indd 31 2/9/2012 9:39:44 PM


‘a man might spend his whole Life in cor- any further Improvements. This is some
recting one small Volume, and yet have inac- small Satisfaction to me in my present
curacies in it’ (LDH 2.243, 456; 2.304, 511; Situation; . . . it is almost the only one
2.250, 461).59 In 1742 he makes a mistake in that my Writings ever afforded me: For as
to any suitable Returns of Approbation
quoting the aforementioned Rousseau (‘C’est
from the Public . . . they are yet to come
la politesse d’un Suisse / Dans la Hollande
(LDH 2.322, 525).
civilisé’), and six years later he readily cor-
rects it (‘En Hollande civilisé’).60 Most of
his corrections ‘fall upon the Style’, he easily In his Lucianic dialogue Hume is concerned
allows, and locating and eliminating inaccur- with the effect of his writings on religion. In
acies gives him a ‘sensible pleasure’ (LDH his Addition Smith tries to soften the effect
2.151, 394; 2.188, 422; 2.243, 456; 2.250, of this imaginary conversation with the
461). At the end of July 1776 he employs mythical ferryman: he moderates the lan-
himself in correcting the sixth volume of guage and introduces the excuse that Hume
the History and the moral Enquiry. It is the desires yet a further opportunity to correct
trifling occupation of a dying author: his manuscripts. According to Smith’s letter
to Wedderburn, Hume is dying with ‘great
You will wonder, that, in my present chearfulness . . . and more real resignation
Situation I employ myself about such to the necessary course of things, than any
Trifles, and you may compare me Whining Christian ever dyed with pretended
to the modern Greeks, who, while resignation to the will of God’; according
Constantinople was besieged by the to the Addition, he is simply dying with ‘the
Turks and they themselves were threat- utmost cheerfulness, and the most perfect
ened with total Destruction, occupyed complacency and resignation’.62 Hume’s
themselves entirely in Disputes concern- ‘pleasure of seeing the churches shut up, and
ing the Procession of the holy Ghost. the Clergy sent about their business’ is turned
Such is the Effect of long Habit! (LDH
into that of ‘seeing the downfall of some of
2.329, 531)
the prevailing systems of superstition’.63
Hume’s brother remarks that Hume did
Strahan takes these lines as a ‘living Evidence’ not say ‘I am dying as fast as my worst
that we are ‘much interested in what is to enemies could wish’, as Smith writes in
pass after our Deaths’.61 In his last letter the first draft of the Addition, but ‘as my
to Strahan, concerning the moral Enquiry, enemies, if I have any, could wish’.64 Smith
Hume announces: ‘This is the last Correction accepts the correction and adds that Hume
I shall probably trouble you with: . . . all shall has ‘no enemies upon whom he wished to
be over with me in a very little time.’ (LDH revenge himself’.65 This fits with My Own
2.331–32, 537) Hume’s excuses in Smith’s Life, where Hume (ironically and by a
Addition sound like a little pleasantry about marginal addition) declares that he is ‘little
his own final remarks: susceptible of enmity’ and has never been
attacked by the ‘baleful tooth’ of ‘calumny’,
I have made [my new edition] extremely even though he ‘wantonly’ exposed himself
correct; . . . if I were to live twenty Years to ‘the rage of both civil and religious fac-
longer, I shoud never be able to give it tions’ (MOL, LDH 1.7). The Life sounds


9780826443595_Ch01_Final_txt_print.indd 32 2/9/2012 9:39:45 PM


like the Manifesto of the Humean Party: Volpone, talk of childless rich old men and
‘English, Scotch, and Irish, Whig and Tory, parasitical clients rather than real friends:
churchman and sectary, freethinker and reli- perfect friendship does not contemplate
gionist, patriot and courtier, united in their money, as Lucian remembers in ‘Toxaris’
rage against [me]’ (MOL, LDH 1.4; also see and Montaigne in ‘L’Amitié’. In Hume’s ‘A
NLH 194, 107). Before publishing the third Dialogue’, ‘by his Will’ Usbek makes his ‘inti-
Book of the Treatise, Hume had already mate Friend’ Alcheic ‘his Heir to a consider-
realized that the clergy are ‘always Enemys able Part of his Fortune’ (EPM Dial. 8 / 326).
to Innovations in Philosophy’, and only after By his own will Hume leaves 200 pounds
the publication of the History does he rather sterling to his ‘friends’ Ferguson, d’Alembert
belatedly resolve ‘to be more cautious than and Smith. D’Alembert accepts the legacy.
formerly in creating myself Enemies’ (LDH He had helped Hume with the publication of
1.37, 15; 1.352, 189).66 the original French version of A Concise and
As Strahan realizes and Smith endeav- Genuine Account of the Dispute between
ours to forget, Hume shows an ‘extreme Mr. Hume and Mr. Rousseau, and Hume
solicitude’ to publish the Dialogues.67 He has had even asked Smith to tell d’Alembert he
been accused of being ‘as great an atheist as had made him ‘absolute Master to retrench
Bolingbroke’, even though Bolingbroke’s vol- or alter what he thinks proper’ (LDH 2.83,
umes contain ‘so little Variety & Instruction’ 348).69 Smith did not help Hume with the
and the clergy have ‘no Reason’ to be correction and publication of the Dialogues:
‘enrag’d against him’ (LDH 1.168, 78; 1.209, he discharges the legacy because he cannot
101; 1.214, 105). Yet, for the sake of his ‘with honour accept it’.70 Hume had ‘hith-
Dialogues, Hume does not scruple to com- erto forborne’ to publish the Dialogues,
pare himself to Bolingbroke. And he tells both desiring to ‘keep remote from all Clamour’
Smith and Strahan that Bolingbroke’s editor (LDH 2.323, 525); Smith is ‘uneasy about
was not ‘any wise hurt by his Publication’, the clamour’ they could excite. Worried that
as ‘he always justify’d himself by his sacred they could ruin his tranquillity, he tries to
Regard to the Will of a dead Friend’. If he make Strahan apprehensive that they could
leaves Strahan the Dialogues ‘by Will’, his ruin his interest.71 Hume assures Smith that
‘executing the Desire of a dead Friend, will the Life is a ‘very inoffensive’ piece (when
render the publication still more excusable’ compared to the Dialogues), which ‘will
(LDH 2.316, 522; 2.324, 525). According to be thought curious and entertaining’ (LDH
the last codicil, the ‘duty’ of his nephew in 2.318, 522A; 2.323, 525); Smith agrees, and
publishing the Dialogues ‘as the last Request is later offended by the reception accorded
of his Uncle, must be approved of by all the his own Addition: ‘a single, and as I thought
World’.68 a very harmless Sheet of paper, . . . concern-
In the 1750s Hume is extracting from the ing the death of our late friend Mr Hume,
classics ‘what serv’d’ concerning the ancients brought upon me ten times more abuse
(LDH 1.152–3, 71): from Lucian, he says, than the very violent attack I had made
we may ‘gather’ that ‘leaving great sums of upon the whole commercial system of Great
money to friends’ was a common practice Britain’.72
in Greece and Rome (E 400n), even though In 1747 Hume says he could not see ‘what
the Dialogues of the Dead, like Jonson’s bad consequences follow, in the present


9780826443595_Ch01_Final_txt_print.indd 33 2/9/2012 9:39:45 PM


age, from the character of an infidel’ (LDH says Strahan, ‘knew [Hume] well, and loved
1.106, 58); 30 years afterwards Horne vitu- him much’, and it ultimately falls to Smith
peratively attacks the ‘foolish and insens- to tell Strahan and the world of the death of
ible’ Hume for reading Lucian and inventing ‘our most excellent, and never to be forgot-
droll conversations with Charon. Hume’s ten friend’. After 11 years he also unforgiv-
excuses are an attack on ‘superstition’, and ably celebrates ‘the abilities and virtues of the
‘we all know . . . against what Religion his never to be forgotten Dr Hutcheson’.76
shafts are levelled, under that name’. Smith
wished that the Addition could be agree-
able; Horne thinks that he who can reflect
‘with complacency’ on Hume, amusing him- 9. STUCK IN A BOG (FISHWOMEN
self at his death ‘may smile over Babylon in FOR THEOLOGIANS)
Hume was reading Lucian. ‘It is an idle There is an anecdote about one of Hume’s
thing in us,’ he writes in June, 1776, ‘to be misadventures in Edinburgh, which much
concerned about any thing that shall hap- amused the historian Sir Leslie Stephen and
pen after our Death’; yet, he adds, ‘I often his daughter Virginia Woolf, and reminds
regretted that a Piece, for which I had a par- us of Thales’ tumble and the laughter of the
ticular Partiality, should run any hazard of Thracian maid.77 It agrees with the pagan
being suppressed after my Decease’ (LDH frolic tone of the Natural History and says
2.325–6, 527). Again Strahan takes it as a something about Hume’s style, his attitude
sign that ‘our Existence will be protracted towards religion, and his not-so-serious pro-
beyond this life’.74 With regard to the nouncements on women, monks and super-
Enquiry, in 1754 Hume had allowed: ‘I am stition (NHR 12.144). It also says something
willing to be instructed by the Public; tho’ about the distinction between popular and
human Life is so short that I despair of ever pretended philosophical religion, and some
seeing the Decision’; in 1776 he laments: ‘I Humean resolutions to ‘keep [him]self in
shall not live to see any Justice done to me’ a proper disposition for saying the Lord’s
(LDH 1.187, 91; 2.322, 525). ‘If I live a few Prayer, whenever [he] shall find space enough
Years longer,’ Hume tells Smith, ‘I shall pub- for it’ and to ‘proceed directly to attack the
lish [the Dialogues] myself’ (LDH 2.316, Lord’s Prayer & the ten Commandments &
522); and Smith tells Strahan: ‘I could have the single Cat’ (LDH 1.148, 70; NLH 43,
wished [they] had remained in Manuscript 25). Finally, it says something about his real
to be communicated only to a few people expectations as regards the effect of his work:
. . . [they] never should have been published ‘this did not convert the generality of man-
in my lifetime.’ In 1763 Elliot and Blair had kind from so absurd a faith; for when will
expressed the same sentiments.75 the people be reasonable?’ (NHR 12.173).
In his last days Hume was ‘revising’ the Hume is engaged in building a ‘small
Dialogues (LDH 2. 334, 538), but despite these House, I mean [he says] a large House for an
exertions on the part of a dying man his good Author’ (he did finish it), which is the ‘second
friend Smith remained determined to have great Operation of human Life’ (LDH 2.232,
nothing to do with the publication of what he 451; 2.235, 453). As always, he is out of
saw as so incendiary a piece of writing. Smith, the common road:


9780826443595_Ch01_Final_txt_print.indd 34 2/9/2012 9:39:45 PM


On his daily visits to inspect the work, 1846), vol. 1, p. 294n1; S.J. Pratt, Supplement
he was in the habit of taking a short to the Life of David Hume (London: J. Bew,
cut across what was then a swamp . . . 1777), pp. 33–4.
he made a slip, fell over, and stuck fast Hume seems to reflect on La Rochefoucauld’s
maxim: ‘everybody speaks well of his own
in the bog. Observing some Newhaven
heart, nobody dares to speak well of his own
fishwomen passing with their ‘creels’, he understanding’ (La Rochefoucauld, Maximes
called aloud to them for help; but, when (Paris: Garnier, 1991), p. 309.
they came up, and recognised the wicked 5
Letters of Eminent Persons addressed to
unbeliever David Hume, they refused David Hume, ed. J.H. Burton (Edinburgh:
any assistance, unless he first repeated, in W. Blackwood, 1849), p. 252.
a solemn tone, the Lord’s Prayer. This he 6
Burton, Life, vol. 1, p. 226.
did, without pause or blunder, and was J. Brown, An Estimate of the Manners and
extricated accordingly. He used to tell Principles of the Times, 2nd edn (London: L.
this story himself with great glee, declar- David and C. Reymers, 1757), p. 57.
E.C. Mossner, ‘New Hume Letters to Lord
ing that the Edinburgh fishwives were
Elibank, 1748–1776’, Texas Studies in
the most acute theologians he had ever Literature and Language 4 (1962), p. 445;
encountered.78 A Concise and Genuine Account of the
Dispute between Mr. Hume and Mr. Rousseau
(London: T. Becket and P.A. De Hondt, 1766),
Even when getting out of his bog, the sceptic
pp. iii, iv, viii.
will always be ‘the first to join in the laugh 9
Horace Walpole’s Correspondence with
against himself’ (EHU 12.23 / 160). Madame Du Deffand and Wiart, ed. W.S.
Lewis and W.H. Smith (New Haven: Yale
University Press, 1970), pp. 253, 266.
BL, Add. MS 36638, f. 22r.
NLS, MS 25692, f. 33v.
E.C. Mossner, ‘Hume at La Flèche, 1735:
Early Responses to Hume’s Life and Reputation, An Unpublished Letter’, University of Texas
2 vols, 2nd edn, ed. J. Fieser (Bristol: Thoemmes Studies in English 37 (1958), pp. 30–3,
Continuum, 2005), vol. 1, p. 294. p. 32.
These alleged words are often treated as an 13
W. Leechman, The Temper, Character, and
indication that Hume’s mother viewed him Duty of a Minister of the Gospel, 5th edn
as weak-minded; other interpreters main- (Glasgow: R. Foulis, 1749), pp. 34, 41.
tain that the sense of the words in the local 14
J.C. Hilson, ‘More Unpublished Letters of
‘Vernacular’ of the time was such that his David Hume’, Forum for Modern Language
mother was drawing a contrast between Studies 6(4) (1970), pp. 315–26, p. 321.
his good nature and his ‘uncommon acute- 15
LDH 1.131, 64; NLS, MS 25708, f. 36v; MS
ness’ (see H. Calderwood, David Hume 25703, f. 187v.
(Edinburgh: Oliphant Anderson & Ferrier, 16
NLS, MS 25703, ff. 188r, 212v.
1898), p. 14; D.F. Norton, ‘An Introduction to 17
B. Mandeville, The Fable of the Bees, ed.
Hume’s Thought’, in D.F. Norton (ed.), The P. Harth (London: Penguin, 1989), p. 341.
Cambridge Companion to Hume (Cambridge: 18
NLS, MS 25703, 188r; ff. 210r, 212v; Early
Cambridge University Press, 1993), pp. Responses to Hume’s Life and Reputation, ed.
1–32, p. 2; E. Mazza, ‘La mamma di Hume. J. Fieser, vol. 2, p. 203; Charlemont, ‘Anecdotes
Interpretazioni di un detto apocrifo’, in Il of Hume’, RIA, Charlemont MSS 12.R.7,
mestiere di studiare e insegnare filosofia f. 518v.
(Milano: Wise, 2000), pp. 93–152). 19
NLS, MS 25708, ff. 40v; LDH 1.132, 64.
J.H. Burton, Life and Correspondence of 20
NLS, MS 25708, ff. 46rv, 50v.
David Hume, 2 vols (Edinburgh: W. Tait, 21
NLS, MS 25708, ff. 33r, 41v; E 2.7.15, 339;


9780826443595_Ch01_Final_txt_print.indd 35 2/9/2012 9:39:45 PM


Home, J., ‘Diary of a journey with Hume to Archiwum Historii Filozofii I Mysli Spolecznej
Bath’, in Early Responses, vol. 1, p. 284; NLS, 9:5 (1963), p. 138; LDH 1.10, 1.
MS 25708, f. 33r. The Correspondence of Adam Smith, ed. E.C.
Letters of Eminent Persons, p. 165. Mossner and I.S. Ross (Indianapolis: Liberty,
Ibid., p. 67. 1987), p. 168.
24 43
D. Diderot, Correspondance V, ed. G. Roth, D. Hume, ‘Disposition and Settlement’, 4
trans. E. Mazza (Paris: Les Editions de Minuit, January 1776, in ‘Record of Testaments
1959), p. 132. for the year 1781’ (8 March 1781), NAS,
NLS, MS 23163, f. 100. CC8/8/125, ff. 863, 865.
26 44
NLS, Acc 11139, f. 24r. The Correspondence of Adam Smith, p. 211.
27 45
NLS, Acc 11139, f. 25r. D. Hume, ‘Codicil to my Will’ (7 August
A Concise and Genuine Account, p. 90. 1776) (1st codicil: ‘In my latter Will and
Charlemont, ‘Anecdotes of Hume’, f. 521; Disposition’; 2nd codicil: ‘I also ordain’), NLS,
Horace Walpole’s Correspondence with MS 23159, item 24, f. 17; see LDH 2.318,
Hannah More et al., ed. W.S. Lewis et al. 522A; 2.323, 525.
(New Haven: Yale University Press, 1961), 46
NLS, MS 23159, item 24, f. 17; LDH 2.323, 525.
pp. 47, 49. 47
The Correspondence of Adam Smith, pp. 199,
Mossner, ‘New Hume Letters to Lord Elibank’, 206, 211.
p. 452. 48
Ibid., pp. 203–4, 206.
NLS, MS 23159, item 23, f. 10. 49
Hume, ‘Disposition and Settlement’, f. 863;
L. Sterne, Letters, ed. L.P. Curtis (Oxford: The Correspondence of Adam Smith, p. 206;
Clarendon, 1967), pp. 157, 219. see ibid., pp. 203, 219.
Sterne, Letters, p. 218; L. Sterne, ‘The Case of 50
The Correspondence of Adam Smith, p. 212.
Hezekiah and the Messengers’, The Sermons of 51
Ibid., pp. 214, 216.
Mr. Yorick, vol. 3 (London: T. Becket and P.A. 52
Ibid., pp. 206, 208, 214–15.
De Hondt, 1766), pp. 30, 42. 53
Ibid., pp. 216, 221, 222.
Sterne, Letters, pp. 219, 239, 445 (also see 54
Ibid., pp. 207, 218; J. Home, ‘Diary of a
pp. ivi, 235, 254); L. Sterne, A Sentimental Journey with Hume to Bath’, p. 281.
Journey through France and Italy by Mr. 55
The Correspondence of Adam Smith, pp. 203,
Yorick (London: T. Becket and P.A. De Hondt, 219; Correspondence of Dr William Hunter,
1768), pp. 92–3. vol. 2, p. 226.
Lettres d’André Morellet. Tome I: 1759–1785, 56
The Correspondence of Adam Smith, p. 219.
ed. D. Medlin et al. (Oxford: The Voltaire 57
Ibid., pp. 204, 206, 211.
Foundation, 1991), p. 72. 58
Ibid., p. 218.
The Correspondence of Dr William Hunter, 59
J.-B. Rousseau, ‘Preface’ (1712) to Œuvres
1740–1783, ed. C.H. Brock, 2 vols (London: diverses, 2 vols (London: J. Tonson and J.
Pickering & Chatto, 2008), vol. 1, pp. 229, Watts, 1723), vol. 1, p. iv.
230–1. 60
E 127. The initial error and subsequent correc-
Letters of Eminent Persons, pp. 171–72; see tion are displayed in D. Hume, Essays, Moral
J. Curry, An Historical Review of the Civil and Political (Edinburgh: A. Kincaid, 1742),
Wars in Ireland, 2 vols (Dublin: L. White, p. 80, and Essays, Moral and Political, 3rd
1786), vol. 1, p. 215n. edn, corrected (London: A. Millar; Edinburgh:
NLS, MS 23156, f. 77. A. Kincaid, 1748), p. 177.
Dublin, Trinity College, 543/2/11 /E.2.19; 61
NLS, MS 23157, item 69, f. 294.
E. Murphy, The Select Dialogues of Lucian 62
The Correspondence of Adam Smith, pp. 203,
(London and Dublin: Edward Exshaw, 1746), 218 (see also p. 206).
pp. xi–xii. 63
Ibid., pp. 204, 219.
Notes and Queries 3 (8 April 1899), 64
Ibid., p. 215 (see Correspondence of
p. 262a. Dr William Hunter, vol. 2, p. 226).
T. Kozanecki, ‘Dawida Hume’A Nieznane Listy 65
The Correspondence of Adam Smith, pp. 215,
W Zbiorach Mezeum Czartoryskich (Polska)’, 219.


9780826443595_Ch01_Final_txt_print.indd 36 2/9/2012 9:39:45 PM


66 75
See Burton, Life, vol. 1, p. 226. The Correspondence of Adam Smith, p. 211;
The Correspondence of Adam Smith, p. 212. LDH 1.380, 203; NLH 71, 35; 72, 36; NLS
NLS, MS 2319, item 24, f. 17. MS 23153, 51, f. 157.
69 76
Letters of Eminent Persons, p. 252; The Correspondence of Adam Smith, pp. 220, 309.
J.C. Hilson and J.V. Price, ‘Hume and L. Stephen, ‘Hume’, in S. Lee (ed.),
Friends, 1756 and 1766: Two New Letters’, Dictionary of National Biography (London:
The Yearbook of English Studies 7 (1977), Smith, Elder, 1891), vol. 28, p. 223b; V.
pp. 121–7, p. 126. Woolf, To the Lighthouse (1927) (London:
The Correspondence of Adam Smith, pp. 209, Grafton, 1988), pp. 62, 70; G.L. Strachey,
214, 210, 215. ‘Hume’ (1928), in Portraits in Miniature and
Ibid., pp. 211, 216–17. Other Essays (London: Chatto & Windus,
Ibid., p. 251. 1931), pp. 151–52.
73 78
G. Horne, A Letter to Adam Smith LL.D. On Selections from the Family Papers Preserved
the Life, Death and Philosophy of his Friend at Caldwell, pt II, vol. 2 (Glasgow: 1854),
David Hume (Oxford: Clarendon Press, 1777), pp. 177–8n1; see The Scotch Haggis
pp. 9–11, 12–13, 29–30. (Edinburgh: D. Webster and Son, 1822), p. 77;
NLS, MS 23157, item 69, f. 297. Burton, Life (1846), vol. 2, p. 458.


9780826443595_Ch01_Final_txt_print.indd 37 2/9/2012 9:39:46 PM

Tom Seppalainen and Angela Coventry


Vivacity, the ‘liveliness’ of perceptions, is The copy principle connects ideas to impres-
central to Hume’s epistemology. Hume sions. The actual process is left without much
equated belief with vivid ideas. Vivacity is a clarification by Hume. We know it is a causal
conscious quality and so believable ideas are one and that impressions are causally prior
felt to be lively. Hume’s empiricism revolves to ideas. We also know that causation holds
around a phenomenological, inner epistem- only for concrete particulars. These ideas,
ology.1 Through copying, Hume bases viv- together with Hume’s general scepticism
acity in impressions. Sensory vivacity also about the intelligibility of causal relations,
concerns liveliness or patterns of change. suggest that a lack of insight into the process
Through learnt skilful use, vivacity tracks is to be expected. From Hume’s standpoint,
change specific to intentional sense-percep- there are not many empirically founded con-
tual experience consisting in Hume’s ‘coher- ceptual insights to share. Although such a
ent and constant’ complex impressions. state of affairs is common for fundamental
Copying, in turn, communicates vivacity to posits of scientific theories including those
ideas where it becomes an indicator of the of psychology, it has caused much dismay in
believability of ideas. Hume’s copying con- philosophy and for a good reason. At stra-
cerns then the causation of conscious skills tegic philosophical occasions, when Hume
required for the identification of empiric- challenges the intelligibility of commonly
ally warranted structures. Copying allows accepted philosophical beliefs, he introduces
Hume to combine a radically externalist copying as a test or justificatory principle
empiricism with a phenomenological inner (see, for example, EHU 1.13 / 13). Copying
epistemology. is Hume’s distinct operationalization of


9780826443595_Ch02_Final_txt_print.indd 38 2/9/2012 9:39:33 PM


empiricism; the paucity of information on Once we articulate the targets of copying, we

the process leaves Hume’s empiricism with can form an exact judgement about Hume’s
a ‘conceptual’ problem. This is the context externalist empiricism.
for our reassessment of the copy principle. Prima facie, copying indicates externalism.
Despite the conceptual problem, copying is Copies share significant properties with their
the centerpiece of Hume’s empiricism. Why originals. Think of scale models and photo-
this is so can be best explained in program- copies. Both are results of copying and both
matic terms. appear like their originals. The terminology
The distinguishing features of the few Hume uses indicates a commitment to this
broad epistemological orientations are strong externalist form of copying. Ideas for
clear.2 Empiricism embodies an externalist him are ‘faint’ versions of impressions (THN
explanatory strategy for whatever it seeks to / 1; EHU 2.5 / 19). Copies can also
explain or clarify. In Hume’s case, the copy function like their originals and not just seem
relationship establishes an externalist intel- or appear like them. Often they do both. This
lectual strategy for the cognitive realm of point echoes through the history of debates
ideas through the sensory realm of impres- between internalists and externalists. One
sions. Externalism differs from internalist philosopher’s empiricism is another’s nativ-
paradigms of epistemology such as rational- ism when it is not specified whether proper-
ism and constructivism. These emphasize the ties or processes or both are shared across
role of inherent, innate, or just a priori pos- systems. The historical reminder is important
its for cognitive and behavioural outcomes. in the case of Hume. He uses potentially
Empiricism minimizes the significance of misleading terminology according to which
such posits even if the mind is not construed ‘perceptions of the mind’ are ‘objects.’ But
as a Lockean tabula rasa. Epistemological copying need not be limited to the passing of
internalism is not the only relevant contrast. objects and their properties to cognitions for
Externalism comes also in a form that cou- processing. All kinds of ‘entities’ can be cop-
ples the mind (or parts of it) to the world ied, including functions or activity types. We
or environment (or parts of it). This type pose the question about the targets of copy-
should be called ‘objectivism’. Hume’s ‘exter- ing in an open-minded way.
nalist’ empiricist epistemology concerns the According to collective wisdom, copy-
role of the sensory environment for acts of ing indicates Hume’s ‘meaning empiricism’.
cognition. Copying is a relation between two According to this ‘semantic hermeneutics’,
mental domains and does not entail object- linguistic meaning is the target of copying.
ivism. In Hume’s case, objectivism follows The interpretation is codified in slogans:
only if the sensory system is explained by copying gives the ‘empirical cash value’ of
factors of the environment, ones external terms.3 Hume’s concern with the meaning
to it. We leave open the issue of objectiv- of linguistic expressions has textual support.
ism. An externalist hypothesis about Hume’s He claims to apply the copy principle to
empiricism leads to a straightforward heur- philosophical terms (EHU 2.9 / 22). Despite
istic for clarifying copying. Since the pro- this, the semantic hermeneutics distorts the
cess is unclear, we must focus on its targets. causal nature of copying. Copying becomes


9780826443595_Ch02_Final_txt_print.indd 39 2/9/2012 9:39:33 PM


interpreted in analytic or definitional terms are copied, senses and cognitions resemble in
such as operational definitions. This is not terms of their structural features. This holds
Hume but empiricism of a much later ori- even if no ‘abstract’ structure ever occurs in
gin. And semantic hermeneutics is not just the systems independently, without ‘objects’
bad historiography. It has led to scathing cri- or contentful complex information. Structure
tiques of Hume. Although we believe in his- need not be independent for it to be a psy-
tory serving current concerns, these critiques chologically and epistemologically critical
should by now be seen as vitiating any such factor. It only needs to occur robustly across
edifying function for the semantic hermen- circumstances, have a cognitive function and
eutics. To us, they raise only one question. be knowable. The robust presence of particu-
Why study Hume if his central causal prin- lar complex impressions is enough for the
ciple is without merit, theoretical function, first condition. The latter two conditions will
or promise? For us, copying targets informa- be taken up next.
tion of the senses and information comes in Copied structure opens up the possibility
many forms, including semantic ones. Let us for further targets of copying. Structure itself
start with Hume’s ideas about the targets of has an origin. This may be similar for impres-
copying that fit clearly under the heading of sions and ideas, at least under some circum-
information and then turn to the less obvious stances. Generally, if complex structure is the
informational targets. end result – effect – of some sort of activity –
Hume’s senses include atomic elements or its cause – such activity may itself be copied.
‘simple impressions’. Any particular impres- More specifically, if the structured informa-
sions of sight, smell, taste, sound, touch, tion in the senses is a function of activity,
pleasure or pain, are copied to the realm of Hume’s system allows for its copying. Since
ideas as simple ideas. But this is not all. Most Hume’s psychology is a ‘genetic’ or develop-
information is in a form more complex or mental theory this is a genuine possibility.
structured than atomic perceptions and this We believe that it is more than a theoretical
is also the case for Hume. Both senses and possibility.
cognitions contain complex impressions and The fact that Hume never discussed the
ideas, respectively. Some of this structured copying of activities should not prejudice this
information of cognition owes its origin to theoretical option. A direct treatment is pro-
copying but not all of it. Hume does not say hibited by Hume’s scepticism about the intel-
whether all structured information in the ligibility of all causal relations qua relations.
senses is copied to cognitions; it is clear that It applies to activity. For this reason, Hume
‘cognitive’ processes of imagination generate included in his psychological taxonomy an
complex information that is novel relative to inherently action-related quality, ‘vivacity’,
the senses. It is novel since no such specific the ‘liveliness’ of perceptions. Hume’s syno-
complex structures exist in the senses – even nyms, ‘strength’, ‘violence’ and ‘vigour’, are
if the elemental information does. The idea all action-related terms. To explore the copy-
of a unicorn is the classic exemplar. So far ing of activity we must address the copying
everything is clear and uncontroversial. of vivacity.
The copying of even some complex impres- Hume never clearly mentions the copying
sions shows that copying also targets struc- of vivacity. It is obvious in the light of his
ture. Given that some complex impressions methodology and theory. Targets of copying


9780826443595_Ch02_Final_txt_print.indd 40 2/9/2012 9:39:33 PM


can be explored through a resemblance- role. Hume uses relative vivacity in psycho-
based interpretative methodology. This logical taxonomy to classify perceptions into
coheres fully with Hume’s theory and use. ideas and impressions, distinguish memory
For Hume, ideas are not only caused by but from other ideas, and distinguish beliefs
also resemble impressions (THN / 4). from mere fictions of the imagination. Thus,
Copying results in a resemblance between through vivacity, information about mental
impressions and ideas as ‘objects’. Moreover, activity acquires a fundamental taxonomic
if ideas are particular images, then ideas are role in Hume’s psychology. Even if we cannot
literal ontological copies. Copying can also know activity itself, we can know its quality,
concern activities and still be operational- vivacity. Vivacity is the knowable indicator
ized in terms of resemblance. And finally, of activity. But this is not all.
copying is an associative causal process yet Hume equated the central epistemologi-
unique in being based in the ‘natural’ rela- cal notion, belief, an idea type, with vivacity
tion of resemblance instead of contiguity. (THN / 106). This has caused much
Importantly, a resemblance methodology can puzzlement in the secondary literature.4
be used in a two-way fashion. Facts concern- Much of it can be resolved by articulating
ing the realm of impressions can be used to the informational nature of vivacity. ‘Belief’
interpret the realm of ideas. Nothing prohib- is an ambiguous term and combines informa-
its using resemblance in reverse, in theorizing tion with attitude. When we believe some X
and interpretation. Hume arguably did both we ‘have’ the informational content X ‘in our
if only because ordinary consciousness is mind’. But we also ‘have’ a specific doxastic
ambiguous between the presence and contri- attitude towards the content (THN
bution of ideas and impressions. Let us apply / 94). ‘Belief’ is normally used for both. In
this methodology to vivacity. our analysis, ‘belief’ refers to the former and
Impressions and ideas resemble in terms of ‘believability’ to the latter, doxastic attitude.
vivacity in the way that ideas and impressions For Hume, vivacity is a quality of perceptual
generally resemble. Idea copies are ‘faded’ objects instead of the object as such. Thus,
versions of impressions also with respect vivacity should not be identified with belief
to the quality of vivacity. Clearly, vivacity as such. Instead, vivacity concerns believ-
is a target of copying. But is it not already ability. Believability is based in the ‘lively’
included in the general idea of a faded object- or active dimension of ideas. Since vivacity
copy? The question points at the information is copied, the doxastic attitude of believ-
carried by vivacity. That is of a kind no indi- ability is copied. The copying of both vivac-
vidual ‘perceptual object’ can contain. Let us ity, a quality of perceptual objects, and the
see why. informational content of perceptual objects
For Hume, vivacity is a ‘quality’ of per- couples belief-ideas tightly with the senses.
ceptions. This does not mean that vivacity is This is understandable for an empiricist.
an ordinary ‘inherent’ property. It cannot be For rationalists, ‘judgement’ stands behind
transferred to the realm of ideas with a copied ‘believability’ as an independent internal
object. As liveliness, ‘vivacity’ is a quality of epistemological process. In Hume’s external-
a ‘series’ of impressions and of the changes ism, both the doxastic attitude and (much of)
between them. Slow change is not lively. As a the contents of belief are copied to cognitions
quality of change, liveliness has an indicator from the realm of the senses.


9780826443595_Ch02_Final_txt_print.indd 41 2/9/2012 9:39:33 PM


Above we dubbed vivacity an ‘indicator’ Descartes’s methodology of clear and dis-

for psychological taxonomy. The idea applies tinct ideas is the famous ‘inner’ or phenom-
also to belief. The quality of vivacity is an enological epistemology. Despite the ‘Rules
indicator of believability. Vivid ideas cause the for the Direction of the Mind’,5 the conscious
doxastic attitude. This shows that structured skills of thinking needed to attain clear and
information has a central cognitive role for distinct ideas remain unsatisfactory. The
Hume. Through its origins, that role extends flavour of a pseudo-methodology, a mere
to the activities responsible for structure. The ‘inner gaze’ without a clear learning history,
second condition for taking complex struc- remains. The reason is in the innate nature of
ture as a target of copying is satisfied. To Cartesian epistemologically appropriate ideas.
complete the sketch, we must address knowl- Descartes’s metaphysics of consciousness com-
edge of structure and its origins. plements the epistemology. Consciousness is
Vivacity is a feeling for Hume. This is not independently in all mental systems, govern-
surprising. All central posits of his mental ing and unifying them. Hume’s metaphysics
mechanism have a conscious dimension. Yet is externalist – ‘genetic’ or developmental and
vivacity is unique among these. It is a quali- his ‘inner’ epistemology complements it. His
tative indicator of activity. Hume’s scepti- emphasis on ‘experience’ and external causal
cism about ‘relations themselves’ covers all factors opens the possibility that epistemo-
activities. It also spans all methods includ- logically relevant conscious skills are learnt
ing ‘inner’ ones. The evidentiary limits of through a specific causal history.
phenomenology for activities are severe and Vivacity has a central role in the specific
comprehensive. No wonder Hume aspired causal history of skills of conscious activ-
to be the Newton of the Mind (THN Intro. ity. Tracking change in perceptions through
7–8 / xvi–xvii). He professed no comprehen- vivacity is of epistemological use only if
sive phenomenological epistemology for the vivacity discriminates among different types
‘metaphysical’ nature of mental associative of change. More specifically, vivacity allows
relations that make the mind move; with for distinguishing beliefs from mere concep-
his ‘non fingo’, Newton professed the same tions only if we are able to ‘read’ change.
for all observational methodologies con- How should this skill be conceptualized?
cerning force or gravitational ‘associations’ Copying makes it possible that we learn from
that make objects with mass move. Vivacity vivacity within the sensory realm and then
remains ‘only’ an indicator of activity. There use it within the realm of cognition. So what
is a further limit. All perceptual objects are does sensory vivacity teach us?
vivid, yet, for Hume, not all of them are Sensory vivacity also tracks change. One
believable. Vivacity is only ‘effects-based’ ‘type’ of sensory change has a clear cogni-
information about activities, yet sufficient tive function for Hume. Complex impres-
to discriminate among critical activity types. sions that have the ‘qualities’ of coherence
Vivacity is a conscious indicator whose use and constancy are involved in the causation
or ‘indicator function’ must be learned. It of the belief in an external world (THN
is a skill of conscious activity or conscious–20 / 194–5). And this belief-idea, in
skill, for short. The idea of a learnt conscious turn, is conscious information about the exist-
epistemic skill is best introduced through its ence of a world independent of us. Thus, only
alternative, Cartesianism. some complex impressions through copying


9780826443595_Ch02_Final_txt_print.indd 42 2/9/2012 9:39:34 PM


partake in the causation of the intentionality elements through structure to activities and
or ‘about-ness’ of ideas. Through their phe- conscious qualities that track them consti-
nomenological nature and this causal role, tutes the targets of Humean copying. As a
coherence and constancy themselves are con- result, copying grounds Hume’s epistemol-
scious sensory information about the world. ogy causally and creates an externally gov-
The skilful use of vivacity is learnt in this con- erned inner phenomenological epistemology.
text. The correlation between different sense- To bolster the idea that copying communi-
perceptual outcomes, coherent and constant cates to ideas a conscious indicator of believ-
complex impressions and their opposites, and ability, for the remainder of our chapter we
different profiles of sensory vivacity grounds focus on the secondary literature. We analyse
vivacity’s indicator function in the senses. The four interpretations of vivacity directly rel-
conscious skill is learned in the senses. More evant to our sense-based account. Each has a
specifically, in the copied form that indicates kernel of truth yet each fails in an instructive
believability, vivacity informs about those way. The critical analysis is intended both to
changes in ideas that are relevant for empir- clarify and support our position. If it fails to
ically warranted believability. In the senses, do so, a critical review of the secondary lit-
such changes cause sense-perceptual inten- erature stands on its own and fits well with
tionality. They create complex impressions the spirit of a reassessment.
with constancy and coherence. In general,
Hume’s felt indicator of vivacity has then the
function to track mental actions relevant for
the creation of an empirical world, one with 3. VIVACITY AS ‘QUALIA’
structured stability. When copied to ideas, it
grounds beliefs and demarcates beliefs from Much of the secondary literature on vivac-
mere imaginings. ity fits under the theoretical concept from
Let us take stock. Hume’s ideas and current philosophy of mind, ‘qualia’. The
impressions are systems that are causally ‘qualia-interpretation’ of vivacity encom-
coupled by the copy-relation. Copying tar- passes three dimensions: ‘folk psychological’,
gets information in the sensory system. This ‘epistemological’ and ‘metaphysical’.6 Let us
information is based on structure, elements, briefly analyse these and then turn to the sec-
properties and relations. Much of the struc- ondary literature on vivacity.
ture is based in activities since much of it In the folk psychological sense, qualia
is not hardwired or innate. Moreover, both refer to the qualitative, ‘felt’ content of
systems are mental. Mental activities are not conscious states. The milky blueness of the
knowable but their effects are. Vivacity is a northern sky on a clear spring day would
conscious quality which tracks such effects be an example. In its epistemological sense,
and profiles of change and it informs about qualia constitute the ‘sensuous’ foundation
skilful mental activities when specific percep- of knowledge. The qualia epistemology of
tual outcomes correlate with specific profiles empiricists coincides with the Cartesian epis-
of change. This indicator of skilful mental temological ‘marks of the mind’.7 The third
activity is copied to cognitions along with dimension concerns the intrinsic nature of
the rest of the information in the senses. The qualia.8 Metaphysically, qualia are non-
broad informational matrix ranging from relational properties. Together, the three


9780826443595_Ch02_Final_txt_print.indd 43 2/9/2012 9:39:34 PM


dimensions form the ‘internalist’ Cartesian Empiricist epistemology bottoms out in

picture of the mind: intrinsic qualia are dir- perceptual theory. The transparent, indubi-
ectly known, without inference and represen- table epistemology of qualia rests on their
tations, and the object of such knowledge is metaphysical and phenomenological intrin-
an intrinsic quality, fully internal to the mind sicality. This is evident in the jargon used for
and independent of anything external to it. knowledge of qualia. They are ‘felt’, ‘had’, or
For Hume, vivacity varies in strength ‘acquainted with’ instead of explored and
across mental state types. ‘Vivacity’ may be represented. The epistemological puzzles
taken to refer to the strength in folk psy- indicate the same. Qualia foundationalism
chologically understood qualia. Bennett’s inherits Descartes’s ‘problem(s) of knowl-
and Stroud’s interpretation of vivacity as edge’ about anything external to inner qualia
‘phenomenological intensity’ does just that.9 including the world. Empiricist perceptual
For them, vivacity is Hume’s description of theory has a symmetrical problem: sense-
a directly phenomenologically known varia- perceptual consciousness is intentional,
tion in intensity. For example, if the quale is about the world, yet qualia are intrinsic, not
the colour red, ‘vivacity’ describes its vary- about anything. The ‘problem of perception’
ing intensity so that colour impressions are concerns how sense-perception ‘reaches’ to
highly vivid and memory colours less so. The an external world from intrinsic qualia. An
interpretation gains support from Hume’s example clarifies this model of the senses and
phenomenological methodology and specific how it influences Hume interpretation.
remarks about the uniform nature of vivacity Price describes sense data in the experi-
across perceptions: ‘the idea of red we form ence of a tomato as ‘a red patch of a round
in the dark and that impression which strikes and somewhat bulgy shape’.11 Qualia,
our eyes in sun-shine differ in degree, not in sense data, hues and shapes are the given
nature’ (THN / 3). foundation for his empiricist epistemology.
The qualia interpretation does not just But they are also the given for his percep-
concern the folk psychology. Qualia are a tual theory – ‘something from which all
close cousin of the epistemological primitives our theories of perception ought to start,
of neo-classical British empiricism. Their however much they may diverge later’.12
‘sense data’ and ‘the given’ characterize the For ‘Humean’ empiricists such as Price,
foundational epistemological role of ‘sensu- intrinsic qualia form the elements of
ous’ experience. Price, for example, defines sense-perception; they are the ‘sensations’
sense data as that which cannot be doubted of sensory-psychologists. The intentional
when having a sensory-experience and iden- sense-perception of a tomato is created out
tifies them with Humean impressions.10 And of non-intentional hue and shape qualia.
since all twentieth-century British neo-classi- How this occurs, however, is a problem in
cal empiricists are typically seen as Humeans, the philosophy of perception. When qualia
the qualia-interpretation results in historical are identified with sensations of perceptual
continuity. More specifically, as an epistemo- theory, they are typically also phenomeno-
logical qualia-concept, vivacity indicates the logically non-representational/intentional
foundational role of the senses. Yet common states. Vivacity, in the qualia-interpretation,
epistemological inclinations do not exhaust inherits the metaphysics and phenomenol-
the qualia-interpretation. ogy of sensory intrinsicality.13


9780826443595_Ch02_Final_txt_print.indd 44 2/9/2012 9:39:34 PM


Sensory vivacity is central for the qualia gaps in sensory information and to create a
interpretation so let us focus on it for criti- belief in an external world.17 And given that
cal purposes. The qualia model of senses is beliefs for Hume are phenomenologically
programmatic to neo-classical empiricism. vivid ideas, Price’s gap filler ‘hypothesis’ also
But is it Hume’s sensory metaphysics? In psy- concerns perceptual phenomenology. Our
chology, as elsewhere, Hume emphasizes the awareness of a ‘smooth and continuous’
‘non-relational’ and thus intrinsic aspects of world requires intra-mental higher-order
perceptions: ‘every distinct perception, which ‘gap fillers’.
enters into the composition of the mind, is a Contrary to the qualia account, we believe
distinct existence, and is different, and distin- that Hume’s complex impressions have
guishable from every other perception’ and intentional information and that Hume gives
that there are no ‘two distinct impressions, a causal explanation of it. His central theor-
which are inseparably conjoin’d’ (THN etical sense-perceptual concepts are two con- / 259, / 66). In his theory of scious qualities, constancy and coherence’
senses, Hume emphasizes that simple impres- (THN–20 / 194–5). Information
sions, colour, taste, smell and touch, ‘admit of on the constancy and coherence of sensory
no distinction or separation’ (THN / objects concerns complex impressions.18
7–8). These are typically also interpreted as These qualities causally influence imagin-
indivisible perceptual ‘atoms or corpuscles’ ation to form similar qualities in ideas,
that make up the idea of space (THN ‘continued and distinct existing objects’.
/ 30).14 Simple impressions are clearly non- Imagination does not ‘supply’ the latter qual-
intentional. They lack relations and thus ities central to the belief-idea in the exter-
structure, including spatial structure. Because nal world. Instead, imagination processes
of the intrinsic nature of simple impressions, complex sensory information that already
Hume’s theory of the senses is often seen as is coherent and constant. Price’s dynamical
the paradigmatic ‘atomistic’ one.15 metaphor of cognitive processes ‘working
But are Hume’s complex sensory impres- upon data’ is then misleading. First, it dimin-
sions similarly intrinsic? The common empiri- ishes the information in complex impres-
cist interpretation of Hume’s account of the sions. Second, it overlooks the similarity in
origin of clearly intentional perceptions, information between ideas and impressions,
belief-ideas, shows that complex impres- the similarity of the quality pairs, ‘continued
sions are not seen to add much to atomic and distinct’ and ‘coherent and constant’.
qualia. Price’s classic account of Hume’s Hume relies on this similarity in explan-
theory of the external world, the belief-idea ation. According to him, the belief in exter-
of a separate and independent world, is an nal objects ‘must arise from a concurrence’
example.16 For Price, Hume’s ‘fleeting, per- of some ‘qualities’ of ‘certain impressions’
ishing impressions’ form a ‘gappy’ sequence combined with ‘qualities of the imagination’
of sense data (THN / 195). Even (THN / 194; emphasis added). In
complex impressions give only ‘gappy’ Hume’s full account, idea-intentionality and
information. Price thinks that higher-order belief in external bodies is caused by constant
cognitive processes of the imagination sup- and coherent complex impressions. Copying
ply the needed ‘gap-fillers’. He concludes ‘communicates’ causally similar qualities to
that Hume’s mind ‘postulates’ to fill in the ideas, viz. distinct and continuous objects.


9780826443595_Ch02_Final_txt_print.indd 45 2/9/2012 9:39:34 PM


The similarity encompasses the felt inten- ‘stable experience’. It is created from change
tional object-involving qualities of both per- in impressions. The causally central facts of
ception-types. Principles of the imagination change include the variation in stability of
are also involved in the causal process but complex impressions and the fact that stable
these cause identity in the intentional objects ones do not just pop inexplicably into exist-
of ideas. In Hume’s own words, these innate ence. The additional causal posit – vivac-
processes ‘render the uniformity as compleat ity – tracks change and informs about stable,
as possible’ once the mind is ‘in the train of coherent and constant change. Since this
observing uniformity among objects’, and the type of change is necessary and sufficient
postulation of continued existence ‘suffices’ for sensory intentionality, vivacity is con-
for this purpose (THN / 197–8). scious information on intentionality. And,
Hume’s account differs significantly from once copied, vivacity causes the association
the Cartesian-Kantian nativist approach to of ideas with an ‘external world’ instead of
cognition. For Hume, internal imaginative internal (causal) fancy. Vivacity is a quality
processes exaggerate the intentionality of indicative of warranted intentional objects,
ideas. Ideas are neither a cognitive achieve- sense-based ones.22
ment nor a priori. Moreover, the causes of Let us conclude with a summary. The
exaggerated ‘postulation’ extend beyond the qualia-interpretation leaves vivacity without
mind to language – as they did for Hume’s an epistemological function. At best, vivacity
nominalist predecessors. But whatever their informs about inner qualitative change. But
causation, the felt ‘distinct and continued’ such change does not inform even about the
objects of ideas – Hume’s pun on Descartes’s existence of a world external to the qualia
distinct and clear idea-objects – are only simi- let alone distinguish beliefs from imaginative
lar to sensory objects. Since intentional objects creations. But for Hume, vivacity informs
and, at times, their causes differ for impres- about the lawlike or constant and coherent
sions and ideas, vivacity is also included in change of intentional sensory objects. This
Hume’s causal model. Let us see how. function is copied to the idea-realm where
Both of the qualities of constancy and vivacity informs about believable ideas
coherence pertain to sensory change. For through their (intentional) objects’ vivacity.
Hume, coherence is directly about the change The epistemological emptiness of the qualia-
of complex impressions. When their change interpretation follows from a mistaken view
profile is lawlike, complex sensory informa- of Hume’s phenomenology and metaphysics
tion is ‘coherent’.19 The quality of constancy of the senses. Constant and coherent com-
also concerns change. Hume characterizes plex impressions are both metaphysically
it in terms of the ‘order’ of appearance of and phenomenologically intentional.
objects with respect to changes in the per-
ceiver such as his or her head movements.20
This is a mere illustration of the perpet-
ual change in circumstances of perception. 4. VIVACITY AS PRESENTEDNESS
Despite these changes, the objects remain
constant yet ‘only’ so instead of identical.21 We are not alone in underscoring the inten-
For Hume, lawlike change is the key tional nature of Hume’s senses. In this section,
causal factor in sensory intentionality and we analyse Dauer’s intentional interpretation


9780826443595_Ch02_Final_txt_print.indd 46 2/9/2012 9:39:34 PM


of vivacity as the ‘what it’s like of conscious- aspect of seen colour. We will at times use
ness’. For Dauer, vivacity is Hume’s way of the term ‘presentational-representedness’ to
describing what it is like to undergo sense-per- underscore the intentional nature of Hume’s
ception, remembering and believing. Although senses for Dauer.
vivacity applies broadly in phenomenology, Presentational-representedness causes prob-
the core meaning emerges from sensory lems both for the intended phenomenological
experience. Dauer defines sensory vivacity in sense of vivacity and the categorical distinc-
terms of a ‘sense of presentedness’.23 Sensory tion between impressions and ideas. Bats and
consciousness presents whereas thinking rep- humans differ in sensory phenomenology
resents and the two, for Dauer, can ‘never be yet their vivid sensory states are about the
confused’.24 Thus vivacity demarcates cat- same intentional object, shape. We will label
egorically between impressions and ideas and it ‘shape’ to underscore its phenomenologi-
gives the unique phenomenology of sensing. cal opaqueness. ‘Shape’ cannot reduce to the
Let us look at the details. ‘transparent’ qualitative shape or colour of
Dauer develops the sensory phenomenol- human visual experience because bats access
ogy of presentedness through the ‘what it’s ‘shape’ too without experiencing colours. The
like’ concept popularized by Nagel.25 Dauer opaque phenomenology undermines Dauer’s
interprets Nagel’s scenario about the dif- distinction between representational ideas
ferent subjective points of view of bats and and presentational impressions. For Dauer,
humans in terms of shape perception. For representations are many-one ‘mappings’.
bats that echolocate, sounds present shapes. But both ‘presentations’ and ‘representations’
That is their vivid experience or ‘what it’s are in a many-one relationship to a common
like’ of sensory consciousness. Humans dif- object. ‘Shapes’ are presented by sounds and
fer subjectively since colours present shapes. colours; ‘shapes’, for Dauer, are represented
For Dauer, it is ‘an essential feature of visual by equations and sonar. What is the differ-
experience’ that ‘colours give us shape’.26 The ence between the two many-one mappings?
presentation of ‘shape in colours’ constitutes Since presentationally represented objects
Humean (visual) vivacity for humans. are opaque in phenomenology, Dauer cannot
In Dauer’s account, Humean sensory phe- ground the difference in a phenomenology
nomenology is intentional. It is ‘about’ some- of the given. The problem applies directly to
thing. Colour impressions do not just ‘give’ vivacity.
colour as qualia would but ‘give’ shape. For Dauer, Humean senses do not reduce
Colours are about shapes. Bat phenomen- to qualia. Vivid colours, for example, do not
ology is also intentional, about shape. Dauer’s reduce to ‘phenomenologically intense’ col-
presentations of shape occur across different our qualia. Vivacity is an intentional ‘quality’.
types of sensory experience. This state of Dauer also distinguishes presentational-rep-
affairs, in turn, is ‘definitional’ for intention- resentations from representations in terms
ality. Intentionality is ‘aspectual’: intentional of the former’s vivacity. Yet both are inten-
states reach out to their objects ‘under this tional. Dauer cannot base this distinction in
aspect rather than another’.27 Bats’ sensory either qualia or intentionality. As a result,
states reach out to the ‘object’ of shape under we do not know what role qualia have in an
the aspect of heard sound; humans’ sensory intentional sensory phenomenology, or vice
states reach out to that object under the versa. Dauer’s predicament is common for


9780826443595_Ch02_Final_txt_print.indd 47 2/9/2012 9:39:34 PM


intentional accounts of sense-perception. The play a role in intentional sense experience if

relationship, if any, between the phenomeno- changing colour impressions have a role in
logically given and representational object/ the creation of constant and coherent com-
content is the core puzzle for such accounts. plex impressions. It is very likely that col-
Without an answer to it, colours could be ours have a systematic role in this. Sensed
seen to ‘give’ shape non-phenomenologically. objects have systematic colour-profiles of
Indicator-theories of sensory content do just change. Such ‘coherent and constant’ change
that. Colours serve the function of shape per- concerns change in the colour of objects
ception without being ‘felt’.28 But a merely relative to our action and changes in the
‘indicated’ intentional object is not a given in objects and their perceptual environment.29
phenomenology just as a represented object For Hume, colours have a role in ‘presented-
need not be felt in its representation. Both ness’ of sense-perception because they have
are opaque representations. a role in the causation of intentional sensory
Is there a way to resolve Dauer’s prob- objects.
lem in order to advance a representational- What is the role and nature of vivacity
intentional account of Humean senses that for intentionality? Vivacity is an indicator of
includes vivacity? Dauer’s predicament rests change and lawlike profiles of change make
on the specificity of the representational intentional sensory objects in the first place.
object, ‘shape’, and representer, colour, Vivacity has a central causal role. But what is
together with the opaqueness of phenom- the phenomenological nature of vivacity rel-
enology. This potentially initiates a problem- ative to philosophical terms of art such as the
atic slide towards indicator theories. ‘Shape’ ‘given’? Dauer equates vivacity with a given
could though refer to the felt world-involv- sense-perceptual presentedness. But liveli-
ing feature of sensory experience as such. If ness cannot characterize momentary experi-
so, ‘shape’ is neither a specific type of quale ences, states or ‘perceptual objects’. Vivacity
nor an opaquely indicated domain. Instead, pertains to change and characterizes events.
‘shape’ refers to the intentional ‘feeling’ of If ‘the given’ excludes definitionally the phe-
vision. If this is Dauer’s view of Humean nomenology of events, vivacity cannot be
vivacity, it is promising, although it is not given. But to understand Hume we must find
an accurate account of Hume’s view. Let us room for an event phenomenology. The con-
see how colours might play a role in vivid or scious qualities of constancy and coherence
intentional sensory experience according to of complex impressions themselves charac-
Hume’s theory of senses. terize change. Intentional sensory objects are
Colours qua simple impressions do not created from such change and are always in
‘give us’ anything in phenomenology. Simple a process of making. Nothing in an experien-
impressions are non-intentional qualia. tial state corresponds to the quality of vivac-
Not even all Humean ‘coloured’ complex ity; nothing in experience corresponds to a
impressions ‘give’ us anything. For example, fixed sense-perceptual object. As Noë, the
a still and unchanging or ‘uniform’ colour enactivist perceptual theorist, puts it, there
mosaic will lack intentional appearance. are no ‘snapshot pictures’ in experience.30
This is because, for Hume, only complex Our Hume would agree.
impressions with felt coherent and constant Still, ‘presentedness’ does capture a fun-
change-profiles appear intentional. Colours damental feature of sensory experience.


9780826443595_Ch02_Final_txt_print.indd 48 2/9/2012 9:39:35 PM


To articulate this ‘given’, we must bring behind intentional objects overlap only for
into view the change-oriented nature of the ‘sensing and thinking’. Contentful ideas have
presented world. Noë refers to it as a ‘felt many causes. In epistemology, Hume is con-
presence’.31 Sensory awareness of a world cerned with errant causes, imaginings and lin-
concerns the availability of further detail guistic habits, and their fictitious idea-effects.
about its objects. We feel this and this is Equating believability with an indicator cop-
because of change in the sensed objects and ied from the senses addresses this concern.
their circumstances. Our Hume would agree. Whether vivacity should be seen as a ‘given’
He would underscore that the ‘felt presence’ according to the conceptual habits of some
is a ‘given’ only once we have learnt the later empiricist movements is far less import-
indicator function of vivacity with respect ant than knowing what vivacity teaches us.
to profiles of change in sensory objects. We For Hume, it informs us of the sensory origin
are accustomed to ‘just feel’ how the sensed of some representational objects and teaches
world will change according to learnt pro- us which ideas are worthy of belief.
files of change. We learn associations and feel
them. A better word for this phenomenology
is a felt ‘giving’. As a ‘given’, vivacity must be
seen in intentional terms, as being about the 5. VIVACITY AS VERISIMILITUDE
further information available to us. That is
what vivacity tracks and what it has taught In this section, we analyse another intentional
our sensory phenomenology.32 interpretation of vivacity, Waxman’s ‘verisi-
Our account of ‘the giving’ in vivac- militude’. The term was originally used by
ity applies to ideas in Hume’s comparative Popperians in philosophy of science to refer
sense. The intentional objects of ‘idea-events’ to the ‘truth-likeness’ of scientific theories.
change less and for several reasons. Cognition But Waxman’s ‘verisimilitude’ is not a concept
is ‘off-line’ and not open to external change of metaphysics or even an epistemological
in the manner ‘on-line’ sense-perception is. judgement of representations. Waxman uses
Sense-perception allows us to access informa- it to describe the ‘truth-likeness’ of Humean
tion from the world instead of just (copied) perceptions. For Waxman, ‘before there can
memories. Moreover, ideas traffic in ‘ficti- be belief in the existence of any perception,
tious’ categories based in the philosophical consciousness must respond to it with the
relation of identity. Extra-sensory, intramen- feeling that it is real’.33 Verisimilitude is this
tal processes ‘reify’ their intentional objects phenomenological ‘reality-likeness’ of per-
in terms of identity, whose effects are not felt ceptions, their felt ‘presence and reality’.34
in the senses – as each of us can verify by Let us first explore Waxman’s idea by com-
attending carefully to the perpetual change paring it with the two earlier interpretations
of sensory objects. For these reasons, ideas and then turn to a critical analysis.
lose in vivacity to the ‘constant and coherent’ Waxman’s verisimilitude reading differs
intentional sensory objects. from the qualia-interpretation. Verisimilar
Hume’s epistemological message with viv- perceptions are a response to qualia, whereby
acity is that a fully empirical, impression- the intrinsic states become about reality.
related indicator applies to ideas. Hume’s Waxman’s interpretation of vivacity is simi-
empiricism is externalist. The causal processes lar to Dauer’s ‘presentedness’. Yet Waxman


9780826443595_Ch02_Final_txt_print.indd 49 2/9/2012 9:39:35 PM


overcomes the ambiguity about the object something ‘attending perceptions’.39 Waxman
that is presented in phenomenology and the concludes that vivacity is an ‘original div-
un-Humean narrow applicability of pre- ision’ of consciousness for Hume.40
sentedness only to impressions. Waxman’s Waxman’s analysis of vivacity incorpo-
account underscores the general world- rates the causal history of the relevant phe-
involving nature of vivacity for the phenom- nomenology. Waxman’s Hume is a nativist:
enology of both ideas and impressions. He ‘one could in principle distinguish impres-
distinguishes vivacity from Humean ‘percep- sions and ideas from the first moment of
tual objects’. The latter are mere qualia or one’s conscious life.’41 This nativism is based
‘appearances’.35 Verisimilitude is not their in an innate vivacity. Nativism entails that
quality (of intentionality) but a ‘quality of intentionality is not caused by sensory expe-
consciousness’. As a result, Waxman would rience or intramental cognitive processes
disagree with Dauer’s contention that, in vis- such as judgement. Or more generally, since
ual phenomenology, colours ‘give’ or ‘present’ vivacity is not a quality of perceptual objects,
us with shape. Colours as perceptual objects it cannot be affected by the causes that
are qualia that present or ‘give’ nothing. affect Humean perceptual objects. Nativism
Once colours are verisimilar, they are inten- also influences Waxman’s phenomenology.
tional but not about any specific object (such Vivacity is an intentional ‘frame’ of phenom-
as shape) nor opaquely so (as they would be enology and cannot vary across specific types
if they were about ‘shape’). Verisimilar visual of intentional objects.
experience presents a full-blown, intentional, Let us turn to critical analysis. Waxman
coloured reality. pairs nativism with naturalism. By leav-
To achieve interpretative overlap and ing verisimilitude with ‘human nature’ and
conceptual coherence, it is best to analyse by conceptualizing it as having a ‘primitive
Waxman’s distinction between qualities of character’ that ‘applies to already animals
consciousness and perceptual objects as two and infants’, Waxman intends to do justice to
different ‘styles’ of consciousness. In one, Hume’s naturalism.42 But, historically, nativ-
perceptions are merely ‘present’ as qualia to ism is associated with the anti-naturalism of
consciousness.36 For Waxman, this style of the rationalists. Not surprisingly, Waxman’s
consciousness is not vivid. The other style nativist reading contradicts Hume’s causal
is. Consciousness ‘responds’ to idea and theory of intentionality. Let us briefly revisit
impression qualia with a feeling of reality Hume’s account to see why.
or verisimilitude. These perceptual objects Waxman ignores the distinction between
are intentional. For Waxman, both styles simple and complex impressions and their
of consciousness are a matter of ‘human qualities. Simple impressions have no intra-
nature’ or ‘reflect its constitution’.37 In other systemic causal history for Hume. If we access
words, both are innate. The terms Waxman these in phenomenology, they are qualia.
uses for the role of consciousness in imbuing Simple impressions are ‘original’ with respect
Humean qualia with verisimilitude – ‘act of to Hume’s theoretical psychology and maybe
the mind’, a ‘regarding’, a ‘stance’ – are then to his account of consciousness. Neither is true
misleading in their causal, process perspec- of all complex impressions. First, all complex
tive.38 All Humean perceptual objects acquire impressions have a causal history in process-
vivacity just ‘in’ consciousness, so vivacity is ing. Much of this falls below the threshold


9780826443595_Ch02_Final_txt_print.indd 50 2/9/2012 9:39:35 PM


of consciousness with respect to ‘felt’ indi- The heuristic leaves the functionalist inter-
cators. To the extent that it does, the causal pretation straddling the fence between
origins of complexes have no corresponding externalism and internalism. Since our own
original divisions in ‘qualities’ of conscious- view underscores mental functioning, a few
ness. But this is not true of the causation of remarks on it are in order to bring our ana-
constant and coherent complex impressions. lysis of the secondary literature to a close.
Such structures are caused by constant and For functionalists, vivacity concerns the
coherent change. It is reflected in conscious- system-wide significance of perceptions, their
ness through effects, constant and coherent ‘force’ in influencing other perceptions.44
sensory objects and vivacity. Furthermore, Vivacity is the intramental ‘functional role’
these qualities, together with copying, cause of perceptions. And such roles vary since
the belief in a separate and distinct external perceptions differ in their degree of inter-
world. And this idea, in turn, has precisely connectedness and their associated felt
the phenomenological identity Waxman ‘force’. Functionalists have an explanation of
refers to with ‘verisimilitude’, the felt pres- Hume’s identification of belief with vivacity.
ence of an external reality. As a consequence, Believable ideas have interconnections and a
neither verisimilar ideas nor verisimilar com- felt influence on other ideas, whereas mere
plex impressions are phenomenologically ideas do not.
‘original’ or causally innate. Both have a Sensory vivacity is problematic for func-
causal history that extends to our conscious- tionalists and especially when impressions are
ness of them. And when we add that Hume’s seen as non-relational qualia. In her seminal
account also concerns specific intentional functionalist interpretation, Govier solves the
objects and not just the intentional ‘frame’ of problem by dividing Hume’s vivacity-terms
phenomenology, we see that Waxman’s inter- into two categories.45 ‘Forceful’ and ‘vigor-
pretation misses its mark also in the role of ous’ are functional terms. ‘Vivid’ and ‘lively’
learning. The indicator function of sensory are not but describe the clarity and amount
vivacity with respect to specific coherent and of detail of a perception. Govier argues that
constant objects is learnt. Only through a the functional concept applies to ideas and
successful causal learning-history can copied that sensory vivacity concerns only clarity
vivacity ground the epistemology of believ- and detail.
ability within ideas. Epistemologically war- Govier’s Hume turns out to be a coher-
ranted, verisimilar consciousness is a learnt entist instead of an empiricist. Vivacity of
style of consciousness without an influence beliefs has nothing to do with vivid impres-
from unnaturalistic ‘judgemental’ processes. sions. Copied sensory vivacity concerns only
‘informational quality’ – clarity and detail –
which is without epistemological significance
for Govier. Hume’s epistemology rests fully
6. VIVACITY AS FUNCTIONAL ROLE on the interconnections among ideas. This is
a distorted picture. Yet Govier’s interpreta-
Vivacity is a copied epistemological indicator. tion can be diagnosed in a manner that sup-
In accordance with this idea, we have ignored ports Hume’s central epistemological idea,
interpretations of vivacity that emphasize the relationship between sensory informa-
the internal processing of perceptions.43 tion and cognitive function.


9780826443595_Ch02_Final_txt_print.indd 51 2/9/2012 9:39:35 PM


Only intentional complex impressions becomes a logical extension of his ‘genetic’

have Govier’s informational qualities of psychology of sensory intentionality. But
detail and clarity. Only a specific type of coherence can make myths. A historical
change in impressions, a specific ‘functional hypothesis must also be fruitful. Our con-
role’, causes impressions with both detail cluding remarks are intended to achieve just
and clarity. The phenomenology of ‘felt pres- that, offer ideas for a reassessment of Hume’s
ence’ or ‘giving’ shows that sensory objects place in the history of philosophy.
take clear form through detail. Sensory ‘Idea’ is the epoch-making notion of pre-
information integrates functional profiles Kantian, ‘early-modern’ philosophy. Hume’s
with informational quality. This allows for philosophy is one theory of ideas. History
an inner epistemology for ideas: through has not dealt kindly with the ‘idea-idea’.
copying, detail and clarity become criteria Linguistic and/or logically structured propos-
for the intentional objects of ideas which itions replaced ideas without a trace from the
themselves are an effect not just of copying analytic tools of epistemologists. Ordinary
but also of ‘independent’ intramental cogni- languages, forms of life, and other holistic
tive functioning. The skill of thinking and epistemologies buried ideas under mangled
internal functioning is itself governed by practices. ‘Ideas’ are so thoroughly discred-
skilful sensory functioning through shared ited that addressing Hume’s version appears
criteria for the intentional objects of both. of antiquarian interest, at best. But the cur-
Vivacity is a ‘functional role’ indicator but rent wisdom is itself importantly historical:
an empirical one, without a trace of system- critics of early empiricism learnt their early
wide coherentist epistemological functions. empiricism from much later empiricists. If
Unintentionally, Govier gives evidence in only for this reason, Hume’s ‘ideas’ deserve
favour of our integrated interpretation.46 In a fresh look.
seventeenth-century English, her term for Hume’s empiricist predecessors never
clarity and detail, ‘vivid’, meant ‘lively’, her recovered ideas from the Cartesian men-
term for functional role. Perhaps Hume’s tal depths. But Hume did and by meeting
use of ‘vivid’ caused a natural association Descartes on his turf. Hume too articulated
between ideas of change and stability in the an ‘inner methodology’ for ideas. Cartesian
minds of his audience? clear and distinct ideas were replaced with
vivid ideas. This indicator of believable
ideas was copied from the senses. Hume’s
externalist epistemological internalism was
7. CONCLUDING REMARKS: usable. Fictitious ideas including dreams are
HUME’S DEVELOPMENT OF felt to be unbelievable, phenomenologically,
THE THEORY OF IDEAS through idiosyncracies in vivacity. Fictions
can be identified ‘internally’, with vivacity
Our interpretation has followed a traditional of copied memories and idea-acts – without
internalist historiography. Its goal is coher- further experience. Hume thought that the
ence. Hume’s theory becomes coherent when inner methodology could even fight ‘custom’
vivacity is understood as a phenomenologi- and its imaginative necessary connections.47
cal indicator of an ‘inner methodology’ cop- The ‘internal arm’ of Hume’s empiricism
ied from the senses; Hume’s epistemology was decisive on the ‘problem of the external


9780826443595_Ch02_Final_txt_print.indd 52 2/9/2012 9:39:35 PM


world’ and with facts closest to home, con- Cartesian internal depths even if he stopped
sciousness of change in perceptions. short of turning them into logical, social arte-
Despite their overlap, Hume did not facts. But did he drag ideas to the world? His
extend Descartes’s project. To Hume’s detri- perceptions constitute an ‘empirical world’.
ment, it was Descartes’s project that through But this is not based in transcendental rea-
neo-Kantians won a decisive (interim) vic- sons of a priori cognitive achievements. The
tory in the competition over the sociological innate a priori categories of Kant’s nativist-
aspects of epistemology. Cartesian-Kantian empiricism ‘schematized’ in the senses are
epistemology focuses on ‘form’, representa- not a limit for or origin of Hume’s empirical
tional frameworks. For understanding Hume, world. It is a labile causal interface of per-
the differences between the form of ideal- ception and action. Complex structure is cre-
ized languages and clear and distinct ideas ated from change and its only resting place is
is irrelevant. Hume had no epistemology of coherent and constant change accessed from
form because he believed no representational the world, as such, unknowable.
framework existed nor was required for the Hume was then an empiricist about the
mind to know. Abstract ideas are fictions nature of empiricism, unlike Kant and the
because no abstract mental representational logical positivists. In this ‘naturalism’, Hume
vehicle exists. Without one, information can- is reminiscent of today’s epistemologists
not be structured by identity and difference, who deny ‘first philosophy’. He never gave
whether codified in clear and distinct ideas a ‘first version’ of his empiricism either. Yet
or analytic definitions. the modern naturalism is based in Quine’s
His disregard of form entails that the text- view of holistic knowledge, one that ignores
book image of Hume as the forefather of log- psychology and the knower’s awareness.
ical positivism is misleading. His structured Modern naturalism is eliminativist in its
knowledge is inexpressible through defini- externalism: a knower need not know that
tions, whether sensation, observation or he or she knows to know. Justification exists
operation language-based. The reason is that below and beyond the radar of the knower,
his knowledge-mechanism does not follow in phylogenetic processes (evolutionary epis-
a syntax or grammar. Thus his empiricism temology), in the social epistemology of
does not rest on empirical atomic proposi- scientific practices (Quine and Putnam’s div-
tions – Tractarian or Russellian logical atoms ision of linguistic labour), and unconscious
that mirror empirical facts – that turn into neural processes (Churchlands’s eliminativ-
complex propositions by truth-functional ism). Each demands that the agent defers
connectives. Hume’s mechanism is not rule- justification to external forces and loses
governed at all because it is causal; it is touch with the felt dimension of knowledge
governed by all of the causally robust dimen- to know.
sions of acts of experience. No logic governs In Hume’s naturalism, only the phenom-
these causal principles; complex information enological afterglow of associations can cre-
is a contingent matter. Nor can it be stud- ate justified beliefs and an epistemic agent.
ied in the abstract or normatively as gram- He never let go of the role of the subject’s
mar and logic are. In short and unlike that learnt skilful awareness for knowledge. In its
of the logical positivist, Hume’s empiricism is developmental, individualistic and phenom-
naturalistic. Still, he did drag ‘ideas’ from the enological orientations, Hume’s account is


9780826443595_Ch02_Final_txt_print.indd 53 2/9/2012 9:39:35 PM


too limited to solve many of today’s burn- ‘Empiricism about Meanings’, in P. Millican
ing epistemological problems. The contrast (ed.), Reading Hume on Human Understanding
(Oxford: Oxford University Press, 2002),
in naturalisms appears to us then as much
chap. 3; A. Flew, David Hume: Philosopher
historical as it is substantive. Historically, of Moral Science (Oxford: Blackwell, 1986),
Hume’s epistemology is of and for the pp. 22–3 and D. Pears, Hume’s System (Oxford:
Enlightenment. From the Enlightenment Oxford University Press, 1990), pt 1. More
perspective, current naturalism is not only recent advocates of a semantic interpretation
of copying focus on ‘concepts’, stored mental
psychologically implausible but socially irre-
representations that are used in thinking or
sponsible. Substantially, if the Enlightenment cognition more broadly (see, for example,
philosophers tackled problems still rele- J. Fodor, Hume Variations (Oxford: Clarendon
vant today, Hume offers a systematic nat- Press, 2003). Although this idea has important
uralistic empiricism for addressing them. In differences from the other one, we will not
differentiate the two on this occasion. The main
both cases, our reassessment of his copying
reason is that the overall outcome for the ‘fate’
should help to see the history of philosophy of copying has been largely the same here as it
and especially empiricism in a novel, fruitful has in the more directly linguistically-oriented
manner. semantic traditions. For an exception, however,
see J. Prinz, Furnishing the Mind: Concepts and
Their Perceptual Basis (Cambridge, MA: MIT
Press, 2002), chap. 5.
For an overview of the main puzzlements that
surround Hume on belief see M. Gorman,
A comprehensive analysis would also cover ‘Hume’s Theory of Belief’, Hume Studies 19(1)
Hume’s other quality, ‘facility’ (THN / (1993), pp. 89–102. No doubt the fuel to this
146; / 422; EHU 4.2 / 25). We will not scholarly fire is provided by Hume’s own lin-
explore facility on this occasion, but merely guistic dissatisfaction with the account of belief
remark that facility is clearly dependent on in the Appendix to the Treatise, admitting that
vivacity since no ease of transition within he is at a ‘loss for terms to express [his] mean-
perceptions can occur without identification of ing’ (THN / 628).
transitions or patterns of change. R. Descartes, Rules for the Direction of the
See P. Godfrey-Smith, Complexity and the Mind [1628] in J. Cottingham et al. (eds), The
Function of Mind in Nature (Cambridge: Philosophical Writings of Descartes, vol. 1
Cambridge University Press, 1996), chap. 2. (Cambridge: Cambridge University Press,
See H.H. Price, ‘The Permanent Significance 1984–91).
of Hume’s Philosophy’, in A. Sesonske and These three dimensions map onto Dennett’s
N. Fleming (eds), Human Understanding: influential analysis although he does not label
Studies in the Philosophy of David Hume them; see ‘Quining Qualia’ in W. Lycan (ed.)
(Wadsworth Publishing, 1968), p. 7. See Mind and Cognition: A Reader (Cambridge,
also A. Rosenberg, ‘Hume’s Philosophy of MA: MIT Press, 1990), pp. 519–47.
Science’, in D.F. Norton (ed.), The Cambridge Originally in A. Marcel and E. Bisiach (eds),
Companion to Hume, 2nd edn (Cambridge: Consciousness in Modern Science (Oxford:
Cambridge University Press, 1993), p. 66; Oxford University Press, 1988).
G. Dicker, Hume’s Epistemology and Dennett includes Cartesian marks such as pri-
Metaphysics: An Introduction (London: vacy and direct knowledge in ‘Quining Qualia’,
Routledge, 1998), p. 11ff.; J. Bennett, Locke, p. 519 (see also, e.g., J. Kim, Philosophy of
Berkeley and Hume: Central Themes (Oxford: Mind, 2nd edn (Boulder, CO: Westview, 2006),
Clarendon Press, 1971), chap. 9; Learning pp. 18–22).
from Six Philosophers, vol. 2 (Oxford: Dennett lists other key ideas used to charac-
Oxford University Press, 2001), chap. 32; and terize the ‘intrinsicality’ of qualia: intrinsic


9780826443595_Ch02_Final_txt_print.indd 54 2/9/2012 9:39:35 PM


qualia are non-relational properties – dynami- or a bed, table, ‘books and papers’, ‘sun or
cally – they do not change depending on the ocean’. Hume also gives the scenario where he
experience’s relation to other things, and, syn- leaves the room for an hour and finds when he
onymously or interchangeably, intrinsic proper- comes back that the fire is not in the same way
ties are ‘simple’, ‘homogenous’, ‘unanalyzable’ as he left it (THN–19 / 194–5). These
and/or ‘atomic’ (‘Quining Qualia’, p. 519). illustrations all involve complex impressions
See Bennett’s Locke, Berkeley and Hume, with clear intentional objects even if processes.
p. 225; Stroud’s Hume (London: Routledge Re-entering the room he finds the fire not in
& Kegan Paul, 1977), p. 29; J. Broughton, the same way as he left it, but he is accustomed
‘Impressions and Ideas’ in S. Traiger (ed.), The in other instances to see a ‘like alteration
Blackwell Guide to Hume’s Treatise (Oxford: produced in a like time, whether [he] is absent,
Blackwell, 2006), p. 45; Dicker’s Hume’s present, near or remote’ so this coherence of
Metaphysics, p. 5 and D. Owen, Hume’s ‘their changes is one of the characteristics of
Reason (Oxford: Oxford University Press, external objects’ (THN / 195).
1999), p. 73. In the case of constancy he writes: ‘Those
See H.H. Price, Perception (London: Methuen mountains, and houses, and trees, which lie at
& Co., 1932), pp. 3, 19. See also Bennett’s present under my eye, have always appear’d to
Locke, Berkeley and Hume, p. 222 and Flew’s me in the same order; and when I lose sight of
David Hume, p. 26. them by shutting my eyes or turning my head,
Price, Perception, p. 3. I soon after find them return upon me without
Ibid., p. 19. the least alteration. My bed and table, my
Since vivacity attaches to all Humean percep- books and papers, present themselves in the
tions, according to the qualia-interpretation, same uniform manner, and change not upon
all mental states have a non-representational account of any interruption in my seeing or
dimension since all of them have a qualia perceiving them’ (THN / 195).
dimension of vivacity. See, for example, A. Noë, Action in Perception
D. Raynor, ‘Minima Sensibilia in Berkeley and (Cambridge, MA: MIT Press, 2004).
Hume’, Dialogue 19 (1980), pp. 196–200; The idea that vivacity informs about stable
D. Garrett, Cognition and Commitment complex impressions creates conceptual uni-
in Hume’s Philosophy (Oxford: Oxford formity among the many alternative terms
University Press, 1997), pp. 60–2; and Hume uses for ‘vivacity’ such as strength,
W. Waxman, Hume’s Theory of Consciousness violence, liveliness, vigour, firmness, steadiness
(Cambridge: Cambridge University Press, and solidity. From an ordinary language point
1994), pp. 44–6. of view, the list contains unrelated and even
T. Holden, ‘Infinite Divisibility and Actual contradictory terms. For example, ‘solidity’
Parts’, Hume Studies 28(1) (2002), pp. 3–26, seems unrelated if not antagonistic to ‘live-
p. 4; M. Frasca-Spada, ‘Simple Perceptions in liness’. Solid things such as statues are inert,
Hume’s Treatise’, in E. Mazza and E. Ronchetti unmoving and generally lacking in liveliness.
(eds), New Essays on David Hume (Milan: But from our interpretative perspective, the
FrancoAngeli, 2007), p. 42. two terms are naturally connected. ‘Solid’ com-
H.H. Price, Hume’s Theory of the External plex impressions are ones whose intentional
World (Oxford: Clarendon Press, 1940). objects remain constant despite change and
Ibid., chap. 2. For other qualia accounts on because they display a robust ‘lively’ profile
this score see Stroud’s Hume, chap. 5 and of change. The object remains stable amidst
Dicker’s Hume’s Metaphysics, p. 169. change by emerging from lawlike change.
Constancy and coherence cannot be qualities The solid objects are also found out to be so
of simple impressions for they are ‘fleeting’, through our own action which, from a sensory
‘perishing existences and appear as such’ end point, consists of ‘liveliness’ or patterned
(THN / 195). That these qualities are change in impressions. In short, solidity and
of complex impressions is clear from Hume’s ‘firmness’ emerge from liveliness or vivacity in
examples such as mountains, houses and trees, the same way that constancy amidst change


9780826443595_Ch02_Final_txt_print.indd 55 2/9/2012 9:39:36 PM


and coherence of change complement each Ibid.
other. We leave it to the reader to work out Hume’s distinction between present in fact and
how other apparently unrelated if not antagon- present in power can be easily used to articu-
istic terms Hume uses to characterize vivacity late this phenomenology and the dynamic
indicate the dynamic between stability and between change and stability. Hume does so
change. That Hume’s apparently unrelated lin- himself for ideas; see THN / 20.
guistic descriptors form a coherent whole lends Waxman supports the verisimilitude reading
support to our interpretative perspective. from passages where Hume talks about ‘acts
F.W. Dauer, ‘Force and Vivacity in the Treatise of the mind’ that have a reality-revealing phe-
and the Enquiry’, Hume Studies 25(1&2) nomenological indicator quality; see Hume’s
(1999), pp. 83–100, p. 86. Theory of Consciousness, p. 33ff.
24 34
Ibid. Ibid., p. 34.
25 35
T. Nagel, ‘What Is It Like to Be a Bat?’, Ibid., pp. 38–9.
Philosophical Review (1974), pp. 435–50. Ibid., p. 34.
26 37
Dauer, ‘Force and Vivacity’, p. 86. Ibid., p. 35 and also Waxman’s ‘Impressions
G. Graham, Philosophy of Mind (Cambridge: and Ideas: Vivacity as Verisimilitude’, Hume
Cambridge University Press, 1993), p. 136. Studies 19(1) (1993), pp. 75–88, p. 77.
28 38
In indicator theories, sensory content is identi- Waxman, Hume’s Theory of Consciousness,
fied with representational content. It has no pp. 33–5, 38, and ‘Impressions and Ideas’,
phenomenological ‘intrinsic’ features. In some pp. 77, 81.
versions, we just ‘see through’ experience to its Waxman, Hume’s Theory of Consciousness, p. 42.
representational content instead of accessing Ibid., p. 27.
qualia (e.g. G. Harman, ‘The Intrinsic Quality Waxman, ‘Impressions and Ideas’, p. 75.
of Experience’, in Philosophical Perspectives 4 Waxman, Hume’s Theory of Consciousness,
(1990), pp. 31–52). In Dretske’s version, col- pp. 38, 42.
ours can represent surface spectral reflectance For example, Stevenson’s account of conscious-
without it being transparent in phenomenology ness as an impression of reflection in ‘Humean
(except in colour constancy); see F. Dretske, Self-consciousness Explained’, Hume Studies
Naturalizing the Mind (Cambridge, MA: 24 (1998), pp. 95–130.
MIT Press, 1995). We can ‘see’ nothing about See, for example, S. Everson, ‘The Difference
the object represented but still our content between Feeling and Thinking’, Mind 97,
co-varies with and provides information on it. (1988), pp. 401–13.
All versions are problematic both for Dauer’s T. Govier, ‘Variations on Force and Vivacity
idea of a phenomenologically given vivacity in Hume’, Philosophical Quarterly 22 (1972),
and the distinction between impressions and pp. 44–52, p. 45ff.
ideas in terms of intentionality or vivacity. Govier, ‘Variations’, p. 46.
29 47
For details, see, for example, Noë’s Action in For learning contexts geared towards novelty
Perception and J. Broackes, ‘The Autonomy of and general applicability, such as those of
Colour’, in K. Lennon and D. Charles (eds), science, the inner arm had to be refined with
Reduction, Explanation, and Realism (Oxford: a codified and more rigorous external arm,
Oxford University Press, 1992), pp. 421–65. in essence, Baconian or ‘Millian’ methods of
See Noë’s Action in Perception, p. 35. causal inference (THN 1.3.15 / 173ff.).


9780826443595_Ch02_Final_txt_print.indd 56 2/9/2012 9:39:36 PM

Peter Millican

‘Is Hume a sceptic about induction?’ This on the other side, appeal is made to Hume’s
might seem to be a fairly straightforward writings as a whole – including the Treatise,
question, but its appearance is misleading, and Essays, Enquiries, Dissertations, History
the appropriate response is not to give a direct and Dialogues – which display a clear com-
answer, but instead to move to a more funda- mitment to induction, and even reveal their
mental question which is suggested by Hume author to be a fervent advocate of inductive
himself at the beginning of his definitive dis- science. The evidence on each side is then
cussion of scepticism in Enquiry Section 12: judiciously weighed, and an appropriate con-
‘What is meant by a sceptic?’ (EHU 12.2 / clusion drawn depending on which way the
159). His point here is that ‘sceptic’ can mean balance falls. But this whole procedure is mis-
many things, and what counts as ‘sceptical’ directed, because once we recognize the varie-
will often depend on the relevant contrast. ties of scepticism, it becomes clear that these
Someone who is sceptical about morality or two bodies of evidence are not in conflict.
the existence of God, for example, need not
be sceptical about the external world. And
someone who is sceptical about the rational
basis of inductive inference need not be scep- 1. A SCEPTICAL ARGUMENT, WITH
tical at all – in the sense of dismissive or crit- A NON-SCEPTICAL OUTCOME
ical – about the practice itself.
This crucial point about the varieties of In this chapter, I shall maintain that Hume’s
scepticism is often overlooked in discussions argument concerning induction is indeed a
of Hume on induction, generating a great sceptical argument, in the sense of showing
deal of misunderstanding. Commonly, the that inductive extrapolation from observed
debate will be framed in terms of a simple to unobserved lacks any independent rational
contest between ‘sceptical’ and ‘non-scep- warrant. To avoid any misunderstanding
tical’ interpretations. Then on the one side, on the way, however, it will help to be clear
a case is made drawing on Hume’s famous from the start that this is entirely compatible
negative argument which apparently denies with his wholehearted endorsement of such
induction any basis in ‘reason’.1 Meanwhile, extrapolation as the only legitimate method


9780826443595_Ch03_Final_txt_print.indd 57 2/9/2012 9:38:48 PM


for reaching conclusions about ‘any matter then we have no good reason for supposing
of fact, which lies beyond the testimony of that human life will indeed perish in these cir-
sense or memory’ (EHU 12.22 / 159). The cumstances. But Hume suggests that even the
two may initially seem incompatible, but if Pyrrhonist – whatever his theoretical commit-
so, this is because we are taking for granted ments – will be quite unable to insulate him-
that a method of inference is to be relied upon self from such common-sense beliefs: ‘Nature
only if it can be given an independent rational is always too strong for principle. . . . the first
warrant. And one of the central messages of and most trivial event in life will put to flight
Hume’s philosophy is that this assumption is all his doubts and scruples, and leave him the
itself a rationalist prejudice that we should same, in every point of action and specula-
discard, even though it is shared by both tion’ with the rest of us (EHU 12.23 / 160).
the Cartesian dogmatist and the extreme Hume cannot, of course, prove that
‘Pyrrhonian’ sceptic. In the contest between putting total scepticism into practice will lead
those two extremes, the Pyrrhonist ‘seems inevitably to disaster, at least not to the sat-
to have ample matter of triumph’ while he isfaction of the Pyrrhonist who consistently
‘justly’ urges Hume’s own ‘sceptical doubts’ of refrains from induction. Nor can he prove
Enquiry 4 (the famous argument which is then that common life will always trump scepti-
summarized at EHU 12.22 / 159). However, cal principle. But if in fact Hume’s inductive
the appropriate response, as Hume himself conclusions about human psychology are
explains, is not to follow the dogmatist in correct, then he does not need to prove these
vainly attempting to challenge the argument points to any such opponent:
that yields these doubts, but rather to ask the
Pyrrhonist: ‘What his meaning is? And what Nature, by an absolute and uncontroula-
he proposes by all these curious researches?’ ble necessity has determin’d us to judge
(EHU 12.23 / 159). What, after all, does he as well as to breathe and feel; nor can
we any more forbear [making inductive
really expect us to do in response to this scep-
inferences], than we can hinder ourselves
tical argument, even if we fully accept it? Is
from thinking as long as we are awake,
he seriously proposing that we should stop or seeing the surrounding bodies, when
drawing inferences about the unobserved? we turn our eyes towards them in broad
That would obviously be absurd: sunshine. Whoever has taken the pains to
refute the cavils of this total scepticism,
a Pyrrhonian . . . must acknowledge, if has really disputed without an antagonist,
he will acknowledge any thing, that all and endeavour’d by arguments to estab-
human life must perish, were his princi- lish a faculty, which nature has anteced-
ples universally and steadily to prevail. ently implanted in the mind, and render’d
All discourse, all action would immedi- unavoidable. (THN / 183)
ately cease; and men remain in a total
lethargy, till the necessities of nature,
unsatisfied, put an end to their miserable So if in fact the sceptic’s doubts will be spon-
existence. (EHU 12.23 / 160) taneously ‘put to flight’ as soon as common
life intrudes, then Hume’s point is practic-
Theoretically, the Pyrrhonist might try to ally successful even if theoretically unproved.
deny any such disastrous consequences, on And recall again that Hume himself need
the ground that if induction is unwarranted, not be committed to accepting only what is


9780826443595_Ch03_Final_txt_print.indd 58 2/9/2012 9:38:49 PM


theoretically provable – that is the very preju- Such antecedent scepticism is utterly unwork-
dice which he is aiming to undermine. able, because in refusing to trust our facul-
Hume’s subtle approach to scepticism ties from the start, we are denying ourselves
is made harder to appreciate by the vigour the only tools that could possibly provide
and rhetoric of some of his negative argu- any solution. The proper alternative, Hume
ments and conclusions (especially in the seems to be saying, is to accord our faculties
Treatise, where his ultimate position on some initial default authority, and to resort
scepticism remains relatively obscure), but to practical scepticism about them only ‘con-
also, I suspect, by the widespread tradition sequent to science and enquiry’, in the event
of approaching scepticism initially through that those investigations reveal their ‘falla-
Descartes’s Meditations. Descartes sees the ciousness’ or ‘unfitness’ (EHU 12.5 / 150).
sceptic as an opponent to be refuted out- Thus the onus is shifted onto the sceptic to
right, through rational argument of such give reasons for mistrusting our faculties,
overwhelming force as to be immune to any and in the case of induction, that onus is at
possible doubt. He thus takes on the onus of most only partially fulfilled. Admittedly,
providing an ultimate justification of human
reason, with any ineradicable doubt telling The sceptic . . . seems to have ample mat-
in favour of his sceptical opponent. Hume ter of triumph; while he justly insists,
succinctly points out the fundamental flaw that all our evidence for any matter of
in this approach immediately after having fact, which lies beyond the testimony of
raised the question ‘What is meant by a scep- sense or memory, is derived entirely from
tic?’ at the beginning of Enquiry Section 12: the relation of cause and effect; that we
have no other idea of this relation than
There is a species of scepticism, anteced- that of two objects, which have been
ent to all study and philosophy, which frequently conjoined together;2 that we
is much inculcated by Des Cartes . . . have no argument to convince us, that
It recommends an universal doubt . . . objects, which have, in our experience,
of our very faculties; of whose verac- been frequently conjoined, will likewise,
ity . . . we must assure ourselves, by a in other instances, be conjoined in the
chain of reasoning, deduced from some same manner; and that nothing leads us
original principle, which cannot possibly to this inference but custom or a certain
be fallacious or deceitful. But neither is instinct of our nature; which it is indeed
there any such original principle, which difficult to resist, but which, like other
has a prerogative above others, that are instincts, may be fallacious and deceitful.
self-evident and convincing: Or if there (EHU 12.22 / 159)
were, could we advance a step beyond
it, but by the use of those very faculties, But this result – as we have seen – gives no
of which we are supposed to be already practical basis for scepticism. Certainly it
diffident. The Cartesian doubt, there-
raises a ground for theoretical concern, and
fore, were it ever possible to be attained
highlights ‘the whimsical condition of man-
by any human creature (as it plainly is
not) would be entirely incurable; and no kind, who must act and reason and believe;
reasoning could ever bring us to a state though they are not able, by their most
of assurance and conviction upon any diligent enquiry, to satisfy themselves con-
subject. (EHU 12.3 / 149–50) cerning the foundation of these operations’


9780826443595_Ch03_Final_txt_print.indd 59 2/9/2012 9:38:49 PM


(EHU 12.23 / 160). But unless we are in the for his notion of ‘reason’ and for the rational
grip of the rationalist prejudice that Hume status of inductive inference. These issues
rejects, we should not see this lack of theoret- are far from straightforward, partly because
ical satisfaction as sufficient reason to aban- the argument appears three times in Hume’s
don our only respectable method of inference works, with many differences between the
about the unobserved. That would be to take three presentations – some of them highly
the sceptical considerations to a ridiculous significant – and clear evidence of a system-
(and anyway unachievable) extreme. Instead, atic development in his views. But for our
the appropriate response is less dramatic but purposes, it will be enough here just to high-
far more valuable: to recognize our ‘whimsi- light the most salient points.
cal condition’ as a ground for modesty about
the depth and extent of our powers, and to 2.1 THE ARGUMENT OF THE TREATISE
adopt a ‘mitigated scepticism’ which is cor-
respondingly diffident and cautious (EHU In the Treatise, the famous argument occurs
12.24 / 161–2), and which confines our atten- within the context of Hume’s rather ram-
tion to the subjects of common life, ‘avoiding bling search for the origin of the idea of
distant and high enquiries’: necessary connexion, which he has previ-
ously (THN / 77) identified as the
While we cannot give a satisfactory rea- key component of our idea of causation. Not
son, why we believe, after a thousand having ‘any certain view or design’ on how
experiments, that a stone will fall, or to trace the impression(s) that could account
fire burn; can we ever satisfy ourselves for this crucial idea, he sets off to ‘beat about
concerning any determination, which we all the neighbouring fields’ in the hope that
may form, with regard to the origin of something will turn up (THN–13
worlds, and the situation of nature, from, / 77–8). His first such ‘field’ concerns the
and to eternity? (EHU 12.25 / 162) basis of the Causal Maxim ‘that whatever
begins to exist, must have a cause of exist-
This sentence is Hume’s last word on the ence’ (THN / 78), but after conclud-
question of inductive scepticism, and rep- ing that this Maxim cannot be ‘intuitively or
resents the conclusion of a coherent line of demonstratively certain’,3 he quickly moves
thought which can be traced from the begin- on to a related question, ‘Why we conclude,
ning of Enquiry Section 12, his most clear and that such particular causes must necessar-
explicit – and repeatedly refined – treatment ily have such particular effects, and why we
of scepticism as a whole. So far, then, we have form an inference from one to another?’
a clear outline of his mature position. (THN–9 / 82). He soon narrows his
focus onto what he considers the paradigm
case of a causal inference, from a sensory
impression of one ‘object’ (for example, we
2. HUME’S SCEPTICAL ARGUMENT see a flame), to forming a belief – a lively
idea – of its effect or cause (for example, we
The main aim of this chapter is to understand expect heat). He then analyses such an infer-
the logic and significance of Hume’s famous ence into its component parts: ‘First, The ori-
argument, and in particular its implications ginal impression. Secondly, The transition


9780826443595_Ch03_Final_txt_print.indd 60 2/9/2012 9:38:49 PM


to the idea of the connected cause or effect. in the Treatise but disappears from his later
Thirdly, The nature and qualities of that writings.5
idea.’ (THN / 84). The remainder of Hume has now established one of the
Section 1.3.5 discusses the first component, most important results of his philosophy:
then 1.3.6, entitled ‘Of the Inference from the ‘’Tis . . . by experience only, that we can
Impression to the Idea’, comes to the second infer the existence of one object from that of
component, the causal inference itself.4 another.’ (THN / 87). And he imme-
Hume’s first move, in discussing this diately goes on to explain that the kind of
paradigm causal inference, is to insist that it experience which prompts such a causal
cannot be made a priori, simply from obser- inference is repeated conjunctions of pairs of
vation of the cause: ‘objects . . . in a regular order of contiguity
and succession’. Where we have repeatedly
There is no object, which implies the seen A closely followed by B, ‘we call the
existence of any other if we consider one cause and the other effect, and infer the
these objects in themselves, and never existence of the one from that of the other’.
look beyond the ideas which we form of Hume enthusiastically trumpets this relation
them. Such an inference . . . wou’d imply of constant conjunction as the sought-for key
the absolute contradiction and impossi-
to the crucial notion of necessary connexion,
bility of conceiving any thing different.
with a clear allusion back from THN
But as all distinct ideas are separable,
’tis evident there can be no impossibil- / 87 to / 77,6 and he celebrates the
ity of that kind. When we pass from a progress of his rambling journey of discov-
present impression to the idea of any ery. Admittedly there is still some way to go,
object, we might possibly have separated because mere repetition of conjunctions does
the idea from the impression, and have not seem to generate ‘any new original idea,
substituted any other idea in its room. such as that of a necessary connexion’. But
(THN / 86–7) the line of investigation seems clear:

Here Hume is appealing to the principle that having found, that after the discovery of
if an inference is to be a priori, there must the constant conjunction of any objects,
be an absolute contradiction and impossi- we always draw an inference from one
object to another, we shall now examine
bility of conceiving things as turning out
the nature of that inference . . . Perhaps
differently: an a priori inference has to yield
’twill appear in the end, that the neces-
total certainty. He also seems to be taking for sary connexion depends on the inference,
granted that such a contradiction in concep- instead of the inference’s depending on the
tion implies a contradiction in fact, which is necessary connexion. (THN / 88)
closely related to his Conceivability Principle
that whatever we conceive is possible (this This last sentence provides an elegant
makes a more explicit entrance shortly, at epitome of the link between Hume’s theories
THN / 89). Note also his appeal to of induction and causation, anticipating the
what is commonly called his Separability eventual outcome of his quest for the elusive
Principle, that ‘all distinct ideas are separ- impression of necessary connexion (which
able’ (cf. THN / 10, / 18–19, will come much later, at THN / / 79–80), which plays a major role 164–5). For present purposes, however, we


9780826443595_Ch03_Final_txt_print.indd 61 2/9/2012 9:38:49 PM


can forget about that quest, and focus on the Conceivability Principle shows it cannot be:
nature of inductive inference. ‘We can at least conceive a change in the
Having established that causal inference course of nature; which sufficiently proves,
‘from the impression to the idea’ (e.g. from that such a change is not absolutely impos-
seeing A to expecting B) depends on experi- sible. To form a clear idea of any thing, is an
ence, Hume goes on to pose the central ques- undeniable argument for its possibility, and
tion that his argument aims to answer, namely is alone a refutation of any pretended dem-
which mental faculty is responsible for the onstration against it.’ (THN / 89). As
inference: ‘the next question is, Whether for probable arguments (that is, arguments in
experience produces the idea by means of the which we draw conclusions – typically about
understanding or imagination; whether we things in the world of our everyday experi-
are determin’d by reason to make the tran- ence – with less than total certainty), these
sition, or by a certain association and rela- must be based on causal relations, because
tion of perceptions?’ (THN / 88–9). causation is ‘The only . . . relation of objects
If the faculty of reason were responsible, . . . on which we can found a just inference
Hume says, this would have to be on the basis from one object to another’ (THN /
of an assumption of similarity between past 89).9 But Hume has just argued that causal
and future, commonly called his Uniformity inference is ‘founded on the presumption of a
Principle: ‘If reason determin’d us, it wou’d resemblance betwixt those objects, of which
proceed upon that principle, that instances, of we have had experience, and those, of which
which we have had no experience, must resem- we have had none’ (an argument that he reca-
ble those, of which we have had experience, pitulates at THN–7 / 89–90, echoing
and that the course of nature continues always the discussion of THN–4 / 82–9).10
uniformly the same.’ (THN / 89). So And since probable inference relies on causal
the next stage is to see whether there is any relations, ‘’tis impossible this presumption
argument by which reason could establish [of the Uniformity Principle] can arise from
this principle, and if there is not, then Hume probability’, on pain of circularity.11 So nei-
will conclude that reason cannot be the basis ther demonstrative nor probable arguments
for our inductive inferences. can provide any solid basis for the Uniformity
Following the standard categoriza- Principle, and Hume quickly concludes that
tion deriving from John Locke,7 just two reason cannot be responsible for causal
types of argument are potentially available, inference:12
demonstrative and probable, and Hume
now eliminates each in turn. First, demon- Thus not only our reason fails us in the
strative arguments proceed with absolute discovery of the ultimate connexion of
certainty based on self-evident (‘intuitive’) causes and effects, but even after experi-
ence has inform’d us of their constant con-
relationships between the ideas concerned;
junction, ’tis impossible for us to satisfy
these sorts of argument are capable of yield-
ourselves by our reason, why we shou’d
ing ‘knowledge’ in the strict sense, and are extend that experience beyond those
mostly confined to mathematics.8 But no such particular instances, which have fallen
argument can possibly prove the Uniformity under our observation. We suppose, but
Principle, because that would mean the prin- are never able to prove, that there must
ciple is absolutely guaranteed, which the be a resemblance betwixt those objects,


9780826443595_Ch03_Final_txt_print.indd 62 2/9/2012 9:38:49 PM


of which we have had experience, and afford us a reason for drawing a conclu-
those which lie beyond the reach of our sion beyond it; and, that even after the
discovery. (THN / 91–2) observation of the frequent or constant
conjunction of objects, we have no rea-
son to draw any inference concerning any
Instead, such inference must derive from asso-
object beyond those of which we have had
ciative principles in the imagination (THN
experience; . . . and this will throw them / 92), and in particular, from a mech- so loose from all common systems, that
anism which Hume calls custom (e.g. THN they will make no difficulty of receiving / 97, / 102) or habit (e.g. any, which may appear the most extraor-
THN / 118). Experience of constant dinary. (THN / 139)
conjunction between A and B establishes an
associative connexion between them, mak- But this again is within a context where his
ing our mind habitually move easily from aim is to develop his theory of belief, now
the idea of one to the idea of the other. When focusing on inferences involving probability
we then see an A, the ‘force and vivacity’ of where the relevant past conjunctions are not
that sense impression is transferred through constant.
the associative link to our idea of B, enliven- Books 1 and 2 of the Treatise were pub-
ing it into a belief. Hume accordingly goes on lished at the end of January 1739, but well
to define a belief as ‘a lively idea related before the end of that year, Hume seems to
to or associated with a present impres- have radically reassessed the significance of
sion’ (THN / 96), and to expand on his philosophy. By then he had written his
this theory of belief formation over the sub- Abstract of the Treatise, which appeared
sequent sections. in print in March 1740, and which devotes
8 paragraphs out of 35 (paragraphs 8 and
2.2 FROM THE TREATISE TO THE ABSTRACT 10–16) to the famous argument. From being
a very small part of a much larger system,
Given the fame that it has subsequently suddenly it becomes the prime focus of
enjoyed, Hume’s argument in Treatise 1.3.6 his philosophy, as it remained in the first
is surprisingly inconspicuous. It occurs within Enquiry of 1748, which can indeed be seen
a detour (at THN / 82) from a ram- as mainly constructed around the argument
ble through fields (THN / 77–8); the and its implications.
core of it occupies only six fairly short para- The declared purpose of the argument in
graphs (1–2 / 86–7 and 4–7 / 88–90); and its the Abstract is to understand ‘all reasonings
primary role seems to be to identify custom concerning matter of fact’ (Abs. 8 / 649),
as the ground of causal belief – as a compo- rather than limiting discussion to the para-
nent in Hume’s larger theory of belief – rather digm case of a causal inference – ‘the infer-
than to emphasize its own apparently scep- ence from the impression to the idea’ – which
tical conclusion. He does later remark on the had been the topic of Treatise 1.3.6. But
striking nature of this conclusion:13 Hume then immediately states that all such
factual reasonings (to coin a shorthand term)
Let men be once fully perswaded of these ‘are founded on the relation of cause and
two principles, that there is nothing in effect’, thus making clear that causal infer-
any object, consider’d in itself, which can ence is still the focus. However, this initial


9780826443595_Ch03_Final_txt_print.indd 63 2/9/2012 9:38:50 PM


move is helpful in both emphasizing the gen- forthrightly elsewhere: ‘that to consider the
erality of the argument and also streamlining matter a priori, any thing may produce any
it, avoiding the need for the recapitulation of thing’ (THN / 247, cf. /
his treatment of causal reasoning which had 173, EHU 12.29 / 164).
occupied THN–7 / 89–90. Now, in So experience is necessary to ground any
proving that all causal reasoning presupposes causal inference (and hence any inference
the Uniformity Principle, he will have proved ‘concerning matter of fact’). And Hume goes
at the same time that ‘all reasoning concern- on to explain that the type of experience rele-
ing matter of fact’ – and hence all probable vant to his thought-experiment would be of
reasoning – has such a dependence.14 ‘several instances’ (Abs. 12 / 651) in which
To facilitate discussion, Hume introduces Adam saw the collision of one ball into
the simple example of one billiard ball strik- another followed by motion in the second
ing another and causing it to move (Abs. 9–10 ball. Such experience would condition him
/ 649–50). He then presents a vivid thought- ‘to form a conclusion suitable to his past
experiment, imagining the first man Adam, experience’, and thus to expect more of the
newly created by God, and confronted with same. ‘It follows, then, that all reasonings
such an imminent collision: concerning cause and effect, are founded
on experience, and that all reasonings from
without experience, he would never be experience are founded on the supposition,
able to infer motion in the second ball that the course of nature will continue uni-
from the motion and impulse of the formly the same.’ (Abs. 13 / 651). So as in
first. It is not any thing that reason sees
the Treatise, we reach Hume’s Uniformity
in the cause, which makes us infer the
Principle, and he now proceeds accordingly
effect. Such an inference, were it possi-
ble, would amount to a demonstration, to consider what rational basis this principle
as being founded merely on the compari- could be given:
son of ideas. But no inference from cause
to effect amounts to a demonstration. ’Tis evident, that Adam . . . would never
Of which there is this evident proof. The have been able to demonstrate, that the
mind can always conceive any effect to course of nature must continue uni-
follow from any cause, and indeed any formly the same, and that the future
event to follow upon another: what- must be conformable to the past. What
ever we conceive is possible, at least in a is possible can never be demonstrated to
metaphysical sense: but wherever a dem- be false; and ’tis possible the course of
onstration takes place, the contrary is nature may change, since we can con-
impossible, and implies a contradiction. ceive such a change. (Abs. 14 / 651)
There is no demonstration, therefore,
for any conjunction of cause and effect. As in the Treatise, we have an appeal to
(Abs. 11 / 650–1) the Conceivability Principle to show that a
change in the course of nature is possible,
Compared to the equivalent passage in the which in turn implies that uniformity cannot
Treatise (THN / 86–7), this is clearer be demonstrated.
and more straightforward, proving by dir-
ect appeal to the Conceivability Principle Nay, . . . [Adam] could not so much as
a general lesson which he states even more prove by any probable arguments, that the


9780826443595_Ch03_Final_txt_print.indd 64 2/9/2012 9:38:50 PM


future must be conformable to the past. how the world is, and so can be known (if at
All probable arguments are built on the all) only through experience. Some matters of
supposition, that there is this conformity fact we learn directly by perception, and can
betwixt the future and the past, and there- later recall.16 But what of the rest? Hume sets
fore can never prove it. This conformity is
himself to address this key question: ‘what
a matter of fact, and if it must be proved,
is the nature of that evidence, which assures
will admit of no proof but from experi-
ence. But our experience in the past can us of any real existence and matter of fact,
be a proof of nothing for the future, but beyond the present testimony of our senses,
upon a supposition, that there is a resem- or the records of our memory’ (EHU 4.3 /
blance betwixt them. This therefore is a 26)? On what basis do we infer from what we
point, which can admit of no proof at all, perceive and remember, to conclusions about
and which we take for granted without further, unobserved, matters of fact?
any proof. (Abs. 14 / 651–2) Hume calls such inferences ‘reasonings
concerning matter of fact’ (EHU 4.4 / 26),
Here the logical circularity of attempting to a term we saw introduced just once in the
give a probable argument for the Uniformity Abstract but which now becomes his stand-
Principle is more explicitly spelled out than ard way of referring to what he had pre-
in the Treatise. With both demonstrative and viously called ‘probable arguments’. The
probable argument eliminated, Hume briskly reason for this terminological adjustment
concludes that ‘We are determined by custom seems to be to avoid the infelicity of calling
alone to suppose the future conformable to the such inferences merely ‘probable’ even when
past. . . . ’Tis not, therefore, reason, which is the they are based on vast and totally uniform
guide of life, but custom.’ (Abs. 15–16 / 652).15 past experience that yields complete ‘moral
certainty’ (that is, practical assurance). In a
2.3 THE ARGUMENT OF THE ENQUIRY footnote to the heading of Section 6, Hume
will accordingly draw a distinction – within
In the Enquiry concerning Human the class of ‘reasonings concerning matter of
Understanding of 1748, the famous negative fact’ – between probabilities and proofs, the
argument occupies virtually all of Section 4, latter being ‘such arguments from experience
with the positive account in terms of custom as leave no room for doubt or opposition’, as
appearing in Section 5. Compared with the when we conclude that ‘all men must die, or
versions in the Treatise and Abstract, the that the sun will rise to-morrow’.17
argument is clarified and greatly expanded, In Enquiry 4, the famous argument now
leaving little doubt that Hume considers this proceeds much as it had in the Abstract, albeit
his definitive presentation. greatly filled out. The appendix to this chap-
Section 4 starts with an important dis- ter lays out a structure diagram involving 20
tinction now commonly known as ‘Hume’s stages,18 with the stages numbered accord-
Fork’, between relations of ideas – that is, ing to the logic of the argument. The same
propositions (notably from mathematics) numbers will be followed here, within square
that can be known to be true a priori, just by brackets, to enable easy cross-referencing.
examining and reasoning with the ideas con- First, we learn that [2] ‘All reasonings concern-
cerned – and matters of fact – that is, prop- ing matter of fact seem to be founded on the
ositions whose truth or falsehood depends on relation of Cause and Effect’ (EHU 4.4 / 26),


9780826443595_Ch03_Final_txt_print.indd 65 2/9/2012 9:38:50 PM


since [1] ‘By means of that relation alone we discovered in the cause, and the first
can go beyond the evidence of our memory invention or conception of it, a priori,
and senses.’ As in the Abstract, starting in this must be entirely arbitrary. And even after
way has the virtue of streamlining the argu- it is suggested, the conjunction of it with
the cause must appear equally arbitrary;
ment that follows, so that conclusions to be
since there are always many other effects,
drawn about causal reasoning will automatic-
which, to reason, must seem fully as con-
ally apply to the entire class of factual reason- sistent and natural. (EHU 4.11 / 30)
ing. The first of these conclusions, as before, is
that [5] all knowledge of causal relations must
be founded on experience: ‘the knowledge Note the strong emphasis on arbitrariness,
of this relation [i.e. causation] is not, in any making clear that it is not just the conceiv-
instance, attained by reasonings a priori; but ability – or mere theoretical possibility – of
arises entirely from experience, when we find, alternative outcomes which makes any a pri-
that any particular objects are constantly con- ori inference from cause to effect impossible;
joined with each other.’ (EHU 4.6 / 27). Again it is the fact that from an a priori point of
we get a thought-experiment involving Adam, view, there is nothing to suggest one outcome
but this time with water and fire, illustrat- over another.19
ing the general truth that [3] ‘No object ever If causal relations cannot be known a pri-
discovers [i.e. reveals], by the qualities which ori, then factual inference cannot be a pri-
appear to the senses, either the causes which ori either (given [2] that factual inference is
produced it, or the effects which will arise founded on causation). [6] ‘In vain, therefore,
from it’. This is relatively easy to see when the should we pretend to determine any single
phenomena are untypical or unfamiliar, such event . . . without the assistance of observation
as the unexpected adhesion between smooth and experience.’ Hume now brings Part 1 of
slabs of marble, the explosion of gunpow- Section 4 to a close, with two very important
der, or the powers of a (magnetic) lodestone, corollaries for his philosophy of science. The
where we have no temptation to imagine first is that since we cannot aspire to a priori
that we could have predicted these effects insight into why things work as they do, the
in advance (EHU 4.7 / 28). But with com- appropriate ambition for science is instead
monplace occurrences, such as the impact of to aim more modestly for systematization
billiard balls (EHU 4.8 / 28–9), we might sup- of those cause and effect relationships that
pose that the effect was foreseeable a priori. experience reveals: ‘to reduce the princi-
To prove that this is an illusion, Hume asks us ples, productive of natural phænomena, to
to imagine how we could possibly proceed to a greater simplicity, and to resolve the many
make such an a priori inference, arguing that particular effects into a few general causes,
we could not, on the grounds that the effect is by means of reasonings from analogy, experi-
a quite distinct event from the cause (EHU 4.9 ence, and observation.’ (EHU 4.12 / 30). Then
/ 29), while many different possible effects follows Hume’s most explicit account of
are equally conceivable (EHU 4.10 / 29–30). applied mathematics (which he calls ‘mixed
Summing up [4]: mathematics’), emphasizing that although
mathematical relationships are a priori, the
every effect is a distinct event from laws through which they are applied to the
its cause. It could not, therefore, be world – his example is the Newtonian law of


9780826443595_Ch03_Final_txt_print.indd 66 2/9/2012 9:38:50 PM


conservation of momentum – remain unam- This passage seems to be saying that [7] when
biguously a posteriori: ‘the discovery of the we draw conclusions from past experience,
law itself is owing merely to experience, and we presuppose a resemblance between the
all the abstract reasonings in the world could observed and the unobserved, extrapolating
never lead us one step towards the know- from one to the other.22 Later, when appar-
ledge of it’ (EHU 4.13 / 31).20 ently referring back to this passage, Hume
Part 2 starts by summarizing Hume’s confirms such a reading: ‘We have said, . . .
results so far, and anticipating his eventual that all our experimental conclusions pro-
conclusion [20]: ceed upon the supposition, that the future
will be conformable to the past.’ (EHU 4.19
When it is asked, What is the nature / 35). So his ‘main question’ at EHU 4.16 /
of all our reasonings concerning mat- 34 concerns, in effect, the foundation of the
ter of fact? the proper answer seems to Uniformity Principle.23 He repeats (cf. EHU
be, that they are founded on the rela-
4.6 / 27) that [3] ‘there is no known con-
tion of cause and effect. When again it
nexion between the sensible qualities and
is asked, What is the foundation of all
our reasonings and conclusions concern- the secret powers’ of any object, and infers
ing that relation? it may be replied in one from this that [9] ‘the mind is not led to form
word, Experience. But if we still carry such a conclusion concerning their con-
on our sifting humour, and ask, What is stant and regular conjunction, by any thing
the foundation of all conclusions from which it knows of their nature’ (EHU 4.16
experience? this implies a new question / 33). So the Uniformity Principle cannot be
. . . I shall content myself, in this section, established on the basis of anything that we
with an easy task, and shall pretend [i.e. learn directly through sense perception, in
aspire] only to give a negative answer to which case [10] any foundation for it will
the question here proposed. I say then,
have to draw on past experience, which for
that, even after we have experience of
the sake of the argument can here be taken
the operations of cause and effect, our
conclusions from that experience are not as infallible: ‘As to past Experience, it can be
founded on reasoning, or any process of allowed to give direct and certain informa-
the understanding. (EHU 4.14–15 / 32) tion of those precise objects only, and that
precise period of time, which fell under its
Having established that experience is required cognizance . . . . (EHU 4.16 / 33). The ‘main
for any factual inference, Hume goes on to question’ is then urged: how to justify the
explain how experience plays that role: step from past experience to the assumption
of future resemblance?
we always presume, when we see like sen-
sible qualities, that they have like secret These two propositions are far from
powers,21 and expect, that effects, simi- being the same, I have found that such
lar to those which we have experienced, an object has always been attended with
will follow from them. . . . But why [past] such an effect, and I foresee, that other
experience should be extended to future objects, which are, in appearance, simi-
times, and to other objects, which for lar, will be attended with similar effects.
aught we know, may be only in appear- I shall allow, if you please, that the one
ance similar; this is the main question . . . proposition may justly be inferred from
(EHU 4.16 / 33–4; emphasis added) the other: I know in fact, that it always is


9780826443595_Ch03_Final_txt_print.indd 67 2/9/2012 9:38:50 PM


inferred. But if you insist, that the infer- can never be proved false by any demon-
ence is made by a chain of reasoning, I strative argument or abstract reasoning a
desire you to produce that reasoning. priori. (EHU 4.18 / 35)
The connexion between these propos-
itions is not intuitive. There is required
As in the Treatise and Abstract, Hume appeals
a medium, which may enable the mind
to the Conceivability Principle, though
to draw such an inference, if indeed it
be drawn by reasoning and argument. slightly differently: here he expresses it as the
(EHU 4.16 / 34) principle that what is conceivable implies no
contradiction, rather than saying that what is
So because [11] the inference from past conceivable is possible.26 Moving on now to
experience to future resemblance is not intui- probability:
tive (i.e. not immediately self-evident), [12]
there must be some medium, some ‘connect- [16] If we be, therefore, engaged by
ing proposition or intermediate step’ (EHU arguments to put trust in past experi-
4.17 / 34) if indeed the inference is ‘drawn by ence, and make it the standard of our
reasoning and argument’.24 future judgment, these arguments must
The long paragraph that we have just be probable only, or such as regard mat-
been discussing (EHU 4.16 / 32–4) includes ter of fact and real existence, . . . But
steps that have no parallel in the Treatise . . . there is no argument of this kind,
. . . We have said, that [2] all arguments
and Abstract, where, as we saw, Hume sim-
concerning existence are founded on the
ply takes for granted that if the Uniformity
relation of cause and effect; that [5] our
Principle is to be rationally well founded, knowledge of that relation is derived
then this must be on the basis of some chain entirely from experience; and that [7] all
of reasoning, either demonstrative or prob- our experimental conclusions proceed
able. Here in the Enquiry, he explicitly rules upon the supposition, that the future
out both sense experience and intuition as will be conformable to the past. [17] To
sources of foundation for the Uniformity endeavour, therefore, the proof of this
Principle, and only then comes to consider last supposition by probable arguments,
demonstration and probability, which are in or arguments regarding existence, must
turn dismissed in the familiar way, but again be evidently going in a circle, and taking
that for granted, which is the very point
with the structure of the argument made
in question. (EHU 4.19 / 35–6)
somewhat more explicit:

[13] All reasonings may be divided into Note in passing how Hume just assumes
two kinds, namely demonstrative reason- here some obvious inferences, linking [2]
ing, or that concerning relations of ideas, with [5] to deduce that [6] all factual infer-
and moral reasoning, or that concerning ences (‘probable arguments’, ‘arguments
matter of fact and existence.25 [15] That concerning existence’) are founded on
there are no demonstrative arguments
experience, and then combining this with
in the case, seems evident; since [14] it
[7] to deduce in turn that [8] all factual
implies no contradiction, that the course
of nature may change, . . . Now what- inferences ‘proceed upon the supposition’
ever is intelligible, and can be distinctly of the Uniformity Principle.27 He also now
conceived, implies no contradiction, and leaves the reader to piece together the final


9780826443595_Ch03_Final_txt_print.indd 68 2/9/2012 9:38:50 PM


stages of his argument.28 First, that since the (A) The argument concerns all inferences
Uniformity Principle cannot be established to matters of fact that we have not
by either demonstrative or factual inference, observed: what the Enquiry calls ‘rea-
it follows that [18] there is no good argu- sonings concerning matter of fact’ (here
factual inferences for short). Although
ment for the Uniformity Principle. Secondly,
the Treatise version starts with a nar-
that therefore (given [12]),29 it follows that
rower focus on causal inference ‘from
[19] the Uniformity Principle cannot be the impression to the idea’, it later
founded on reason, and finally, that since requires the lemma that all factual
[8] all factual inferences are founded on the inferences are based on causal relations
Uniformity Principle, it follows that [20] no (stated at THN / 89). So the
factual inference (i.e. no ‘reasoning concern- argument is improved both structur-
ing matter of fact and existence’) is founded ally and philosophically by starting with
on reason. Hume had anticipated this con- all factual inferences, as in the Abstract
clusion at EHU 4.15, quoted earlier:30 ‘I say and the Enquiry, and then deriving this
then, that, even after we have experience lemma as its first main stage (Abs. 8 /
649; EHU 4.4 / 26–7).
of the operations of cause and effect, our
(B) Hume next argues that causal relations
conclusions from that experience are not
cannot be known a priori, and hence
founded on reasoning, or any process of the are discoverable only through experi-
understanding.’ (EHU 4.15 / 32). Also in the ence (THN / 86–7, Abs. 9–11
following section – most of which is devoted / 649–51; EHU 4.6–11 / 27–30). This
to sketching his theory of belief as based on is a major principle of his philosophy,
‘Custom or Habit’ (EHU 5.5 / 43) – Hume wielded significantly elsewhere (e.g.
refers back to this argument and states its THN / 173, / 247–8;
conclusion explicitly, once purely negatively EHU 12.29 / 164).
and once alluding to his positive theory: (C) From this principle, together with the
‘we . . . conclude . . . in the foregoing sec- lemma from (A), Hume concludes that
all factual inferences are founded on
tion, that, in all reasonings from experience,
experience, the relevant experience
there is a step taken by the mind, which is
being of those constant conjunctions
not supported by any argument or process through which we discover causal rela-
of the understanding; . . .’ (EHU 5.2 / 41); tionships (THN / 87, Abs. 12 /
‘All belief of matter of fact or real existence 651; EHU 4.16 / 33).
. . . [is due merely to] . . . a species of nat- (D) Factual inferences thus involve extrapo-
ural instincts, which no reasoning or pro- lation from observed to unobserved,
cess of the thought and understanding is based on an assumption of resem-
able, either to produce, or to prevent.’ (EHU blance between the two. Initially in the
5.8 / 46–7). Treatise, Hume seems to suggest that
such an assumption of resemblance –
commonly called his Uniformity
Principle (UP) – would be necessarily
implicated only if reason were respon-
sible for the inference (THN /
We can now distil the essence of Hume’s 88–9). But his settled view, expressed in
argument from these three different presen- all three works (see note 10 above), is
tations, into eight main stages: that UP is presupposed by all factual


9780826443595_Ch03_Final_txt_print.indd 69 2/9/2012 9:38:51 PM


inferences,31 simply in virtue of their This [resemblance between past and

taking for granted a resemblance future] is a point, which can admit of
between observed and unobserved. no proof at all, and which we take for
(E) Hume now proceeds to investigate criti- granted without proof. (Abs. 14 / 652)
cally the basis of UP itself. In the Treatise
(THN / 88–9) and Abstract it is not reasoning which engages us
(Abs. 14 / 651–2), he appears to assume to suppose the past resembling the
immediately that any foundation in reason future, and to expect similar effects
would have to derive from some demon- from causes, which are, to appear-
strative (i.e. deductive) or probable (i.e. ance, similar. (EHU 4.23 / 39)
factual) inference. In the Enquiry, how-
ever – which hugely expands this part of (H) Since UP is presupposed by all factual
the argument from the cursory treatment inferences (D), and UP has no foundation
in the earlier works – he considers demon- in reason (G), Hume finally concludes
strative and factual inference only after that factual inference itself has no foun-
first (EHU 4.16 / 32–4) explicitly ruling dation in reason. Again he expresses this
out any foundation in sensory awareness conclusion in various ways (and note
of objects’ powers, or in immediate intu- here the narrower focus of the Treatise on
ition (i.e. self-evidence).32 causal inference ‘from the impression to
(F) Any demonstrative argument for UP is the idea’, as pointed out at (A) above):
ruled out because a change in the course
When the mind . . . passes from the idea
of nature is clearly conceivable and there-
or impression of one object to the idea
fore possible (THN / 89; Abs.
or belief of another, it is not determin’d
14 / 651–2, EHU 4.18 / 35). Any factual
by reason (THN / 92)
argument for UP is ruled out because,
as already established at (D), such argu- ’Tis not, therefore, reason, which is the
ments inevitably presuppose UP, and guide of life, but custom. That alone
hence any purported factual inference determines the mind . . . to suppose
to UP would be viciously circular (THN the future conformable to the past. / 89–90; Abs. 14 / 651–2; EHU However easy this step may seem,
4.19 / 35–6). reason would never, to all eternity, be
(G) The upshot of this critical investigation able to make it. (Abs. 16 / 652)
is that UP has no satisfactory founda-
I say then, that, . . . our conclusions
tion in reason, though Hume expresses
from . . . experience are not founded
this in various ways:
on reasoning, or any process of the
’tis impossible to satisfy ourselves by understanding. (EHU 4.15 / 32)
our reason, why we shou’d extend
in all reasonings from experience, there
[our] experience beyond those par-
is a step taken by the mind, which is not
ticular instances, which have fallen
supported by any argument or process
under our observation. We suppose,
of the understanding (EHU 5.2 / 41)
but are never able to prove, that there
must be a resemblance betwixt those
objects, of which we have had experi- Note also that two of these quotations – from
ence, and those which lie beyond Abs. 16 / 652 and EHU 5.2 / 41 – could just
the reach of our discovery. (THN as appropriately have been cited as illustra- / 91–2) tions of (G), because both refer to that ‘step’


9780826443595_Ch03_Final_txt_print.indd 70 2/9/2012 9:38:51 PM


which is precisely the presupposition of the glosses the conclusion of the argument in
Uniformity Principle. Since factual inference apparently very negative terms, as showing
operates by extrapolation from past to future, that ‘we have no reason’ to draw any factual
Hume takes it to be obvious that the founda- inference (THN / 139), and that
tion of such inference must be the same as the ‘we cannot give a satisfactory reason, why we
foundation of the principle of extrapolation. believe, after a thousand experiments, that a
Hence he does not consistently distinguish stone will fall, or fire burn’ (EHU 12.25 / 162).
between (G) and (H), making the last stages In this light, it seems entirely appropriate that
of his argument less explicit than one might he should entitle Enquiry Section 4 ‘Sceptical
wish (cf. the end of section 2.3 above). Doubts concerning the Operations of the
Understanding’, and describe it as appearing
to give the sceptic ‘ample matter of triumph’
(EHU 12.22 / 159).
3. THE NATURE OF HUME’S As discussed earlier, however, the issue
SCEPTICAL CONCLUSION of Hume’s inductive ‘scepticism’ is not so
straightforward, and it is far from clear that
Hume usually expresses the conclusion of his he sees the acknowledged incapacity of rea-
famous argument in a way that seems to imply son to ‘prove’ or ‘support’ the Uniformity
some incapacity on the part of human reason. Principle as any sort of genuine problem.
The Uniformity Principle is something that we Certainly he does not infer from it (either in
‘are never able to prove’ (THN / 92), the Treatise, the Abstract or the Enquiry) that
and which indeed ‘can admit of no proof at induction is unreasonable in any pragmatic
all’ (Abs. 14 / 652). Because of this, ‘’tis impos- sense. And indeed the line of thought sketched
sible to satisfy ourselves by our reason’ (THN in section 1 above, drawing on Section 12 of / 91) concerning the inferential step the Enquiry, somewhat suggests that he con-
from past to future, a step which ‘reason siders it inevitable that our most basic prin-
would never, to all eternity, be able to make’ ciples of inference – precisely because they
(Abs. 16 / 652) and ‘which is not supported by are so basic – will lack any ultimate justifica-
any argument or process of the understand- tion beyond their fundamental place in our
ing’ (EHU 5.2 / 41). Hume also frequently mental economy. That being so, the central
uses similar terms within the argument itself, upshot of Hume’s argument might be sim-
when saying that various would-be proofs ply to identify the Uniformity Principle as a
of UP are impossible, refutable, circular or basic principle of this kind, and the sceptical
lack any ‘just foundation’ (THN / 89, flavour of his reasoning – in demonstrating / 89–90, / 91; EHU 4.18 / 35, reason’s incapacity to prove UP – need not
19 / 35–6, 21 / 37–8), denying that human carry over at all into the theory of human
knowledge ‘can afford . . . an argument’ that inference that he draws from it. Nevertheless,
‘supports the understanding’ (EHU 4.17 / 34) the sceptical flavour of the famous argu-
in reasoning from past to future, and conse- ment itself would remain, in denying UP a
quently denying that our factual inferences source of rational support that more opti-
‘are built on solid reasoning’ (THN mistic philosophers might have expected it
/ 90). In both the Treatise (see section 2.2 to enjoy. And although the argument also
above) and Enquiry (see section 1), he later delivers the important positive principle that


9780826443595_Ch03_Final_txt_print.indd 71 2/9/2012 9:38:51 PM


the Uniformity Principle is presupposed by reject some non-Humean notion of reason;

all factual inference (D), even in the Enquiry indeed this style of interpretation became
we have to wait until the final section to see extremely popular in the 1980s. Before then,
this wielded as part of an effective theoretical the general image of Hume was of a highly
defence against the ‘Pyrrhonian’ sceptic.33 destructive sceptic, intending through his
famous argument to maintain that induct-
3.1 DEBATES ABOUT HUMEAN ‘REASON’ ive arguments lack all rational justifica-
tion. Barry Stroud wittily expressed what
What we get much sooner, of course, and in he took to be Hume’s conclusion: ‘As far as
all three works, is Hume’s positive account of the competition for degrees of reasonable-
how our inductive inferences operate through ness is concerned, all possible beliefs about
custom or habit – what he calls in the title of the unobserved are tied for last place.’36
Enquiry Section 5 his ‘sceptical solution’ to the Some of these extreme sceptical interpret-
earlier ‘sceptical doubts’. But as David Owen ations – most influentially those of Antony
observes, it seems odd to suppose that a psy- Flew37 and David Stove38 – took Hume to
chological account of how belief functions be starting from the assumption of deduc-
could in any way ‘solve’ genuine epistemo- tivism, that an inference is rationally justi-
logical doubts; Owen accordingly suggests fied only if it is logically guaranteed.39 But
that the famous argument is itself best under- deductivism sits very uneasily with the (fal-
stood as exclusively concerned with psycho- lible but reasonable) empirical judgements
logical mechanisms, and as having nothing that abound within Hume’s contributions to
to do with ‘the warrant of probable reason- ‘the science of human nature’, for example
ing or the justification of belief’.34 The argu- his discussions of the passions, his Essays
ment’s conclusion, that factual inference is not on politics and economics, and his vari-
founded on reason, may initially seem unam- ous pieces on religion. The ‘wise man’ of
biguously normative, but Owen interprets Enquiry 10.4 / 110, who ‘proportions his
‘reason’ here as signifying what he takes to belief to the [empirical] evidence’, clearly
be a Lockean conception of the mechanism cannot be a deductivist; hence Flew, in dis-
by which human reasoning operates, namely, cussing ‘Of Miracles’, was forced to accuse
through stepwise inference via intermediate Hume of ‘flagrant and embarrassing’ incon-
ideas. Thus he is able to read Hume’s conclu- sistency.40 Even more flagrantly inconsist-
sion that factual inference ‘is not determin’d ent, from this perspective, were the passages
by reason’ as purely descriptive: as denying in which Hume, after his famous argument
that we actually draw factual inferences in the had apparently denied causal, factual infer-
stepwise, mediated manner that Locke suppos- ence a place within the realm of ‘reason’,
es.35 The famous argument accordingly serves then quite explicitly treated it as one of rea-
to reject this Lockean conception of probable son’s central operations, for example:41
reasoning in favour of the more immediate
and instinctive Humean model, thus providing with regard to reason . . . The only con-
a contribution to empirical psychology, rather clusion we can draw from the existence
than an exercise in sceptical epistemology. of one thing to that of another, is by
Owen is by no means the first to read means of the relation of cause and effect
Hume’s famous argument as designed to (THN / 212)


9780826443595_Ch03_Final_txt_print.indd 72 2/9/2012 9:38:52 PM


reason, in a strict and philosophical proposed alternative was to see the argument
sense, can have an influence on our con- as presupposing a perceptual model of reason,
duct . . . by informing us of the existence according to which we draw inferences through
of something which is a proper object the perception of evidential connexions.
of [a passion]; or when it discovers the
This had the virtue of identifying a plausible
connexion of causes and effects, so as to
(and substantial) target of Hume’s argument,
afford us means of exerting any passion.
(THN / 459) namely, Locke’s view – complacently assumed
and stated in his Essay concerning Human
Understanding47 but never worked out in any
But such gross inconsistency was hard to detail – that probable reasoning is founded
credit, and given Hume’s evident commit- on the perception of probable connexions.
ment to inductive moral science (and to its Against this, Hume’s rival model of probable
reasonableness in comparison with ‘super- inference based on custom – and introduced
stition’), it seemed to most later scholars far immediately after his famous argument had
more plausible to interpret his famous argu- refuted the alternative – stood out as a radi-
ment not as genuinely sceptical, but instead cal (and highly sceptical) departure. Moreover
as a way of rejecting or undermining the the ubiquitous hold of the traditional percep-
deductivist notion of reason on which it was tual view of reason, which goes back to the
thought to be based (by revealing its com- ancients and was shared by Hume’s contem-
plete incapacity to underwrite any factual poraries, could help to make sense of his own
inference). From this it would follow that apparent assessment of the famous argument
Hume must use the term ‘reason’ in at least as having a significant sceptical impact. A mere
two distinct senses: one narrowly deductiv- denial that induction has deductive force, or
ist or ‘rationalistic’ notion within the famous yields total certainty, would hardly be worthy
argument, and a broader, more ‘naturalistic’ of notice in the wake of Locke.48 But denying
notion elsewhere, such as in his discussions that induction is founded on perception of any
of the passions and morality. good reason whatever would have vastly more
This anti-deductivist style of interpretation sceptical significance.
was pioneered in 1975 by Tom Beauchamp Garrett’s approach was quite different,
and Thomas Mappes,42 whose work was and in context more radical. He insisted –
quickly followed by numerous variations on against the prevailing orthodoxy – that
the theme.43 However it lost favour after I Hume employs but a single sense of ‘reason’,
and Don Garrett (independently, in 1995 and taking this to be Hume’s name for ‘the gen-
1997)44 pointed out the implausibility, under eral faculty of making inferences or produ-
careful analysis, of reading Hume’s argument cing arguments – just as it was for Locke’.49
as employing a deductivist notion of rea- Leaving aside the (debatable) attribution to
son. Deductivism proves very hard to square Locke,50 the obvious advantage of this inter-
with the argument’s logic,45 and it also seems pretation was precisely its lack of any need to
strange – if Hume’s purpose is to wield the posit an ambiguity in ‘reason’:
argument in order to reject the notion of reason
that it employs – that he should then continue Hume . . . [is] making a specific claim,
to assert its sceptical conclusion in the appar- within cognitive psychology, about the
ently sincere terms we saw earlier.46 My own relation between our tendency to make


9780826443595_Ch03_Final_txt_print.indd 73 2/9/2012 9:38:52 PM


inductive inferences and our inferential/ is . . . that [inductive inferences] are rea-
argumentative faculty: he is arguing that sonings which are not themselves caused
we do not adopt induction on the basis by any piece of reasoning (including, of
of recognizing an argument for its reli- course, themselves). Inductive inferences
ability, for the utterly sufficient reason require that we bridge a gap between
that there is no argument (‘reasoning’ observation and prediction, and for
or ‘process of the understanding’) that someone not already disposed by nature
could have this effect. . . . this does not to bridge that gap, no argument for doing
mean that inductive inferences are not so would be persuasive. Hence, . . .53
themselves instances of argumentation
or reasoning . . . His point is rather that
they are reasonings that are not them- One surprising effect of this change was to
selves produced by any piece of higher bring Garrett’s interpretation rather close to
level reasoning: there is no argument Owen’s, because his detailed analysis of how
that could lead us to accept the conclu- mediated inferences operate made them ipso
sion that inductive reasonings will be facto inferences that are ‘determin’d by rea-
reliable if we did not already accept that
son’. On this account, a demonstrative infer-
conclusion in practice. Hence, in just
ence from A to D, mediated by the intuitive
this sense, they are a class of ‘reasonings’
(inferences or arguments) that ‘reason’ connexions of A to B, B to C, and C to D, will
(the faculty of making inferences or giv- include as part of its processing the intermedi-
ing arguments) does not itself ‘determine’ ate inference connecting B to D.54 This makes
(cause) us to make.51 the latter inference, according to Garrett, a
cause of the overall inference from A to D;
hence that overall inference is ‘determin’d by
Like Owen, Garrett saw Hume’s argu- reason’ in the sense of being caused by another
ment as an exercise in descriptive psych- inference. Thus Garrett agrees with Owen
ology rather than normative epistemology, that Hume’s conclusion involves a denial
delivering a result about the causation of that induction proceeds by stepwise ratiocin-
inductive inference rather than its ‘eviden- ation. All this may seem somewhat artificial,
tiary value’. However, this sat uneasily with and increasingly distant from anything to be
Garrett’s emphasis (as in the quotation found in Hume’s text, but it has the nice fea-
above) on the recognition of higher-level ture of accommodating a genuinely Humean
arguments for the reliability of induction, point – that probable inference is characteris-
and under challenge, he modified his ori- tically immediate and instinctive rather than
ginal reading of Hume’s conclusion to make mediated and reflective – within a framework
it more general. Here are the relevant parts which, unlike Owen’s, avoids any need to treat
of the passage above, edited to reflect the the notion of ‘reason’ that is operative in the
adjusted reading:52 famous argument as one that Hume rejects.
Garrett has consistently urged this last point
Hume . . . [is] making a specific claim, against rival interpretations: that Hume’s
. . . about the underlying causal mechan- famous argument gives little internal clue that
ism that gives rise to inductive inferences: he is employing some special notion of reason
namely, that it is not itself dependent on which he aims to reject. And although both
any reasoning or inference. . . . His point Owen and I sought to mitigate the impact of


9780826443595_Ch03_Final_txt_print.indd 74 2/9/2012 9:38:52 PM


this criticism on our interpretations (by stress- which means ‘Having the power of discov-
ing that we saw Hume’s response to the argu- ering truth immediately without ratiocin-
ment as changing not the scope of his notion ation’. All this seems to fit with Hume’s own
of reason, but rather its presumed method usage: he refers to ‘deductions’ and ‘ratiocin-
of operation),55 the lack of any obvious and ation’ in contexts where stepwise argument
deliberate ambiguity or equivocation on is clearly intended (e.g. THN / 156,
Hume’s part has remained by far the strongest EHU 5.22 / 55, EPM 1.4 / 170; EHU 4.23
weapon in Garrett’s armoury. / 39, 12.17 / 155), and he is happy to refer
Garrett’s interpretation has also seemed to ‘arguments’, ‘inference’ and ‘proof’ that
attractive for a more specious reason, are ‘intuitive’, and hence do not proceed in
namely, the extent to which Hume expresses a stepwise fashion (THN / 173,
his conclusion in terms of the impossibility / 408; EHU 4.21 / 37, 8.22n18 / 94n,
of founding induction on ‘reasoning’ (e.g. LDH 1.187, 91).58 Overall, therefore, the lan-
THN / 90; EHU 4.15 / 32, 4.16 / 34, guage in which Hume expresses his famous
4.23 / 39), ‘proof’ (e.g. THN / 92; conclusion is no argument (sic.) in favour of
Abs. 14 / 651–2), or ‘argument’ (e.g. Abs. 15 Garrett’s interpretation.
/ 652; EHU 4.16–17 / 34–5, 4.21–3 / 38–9, Perhaps the most serious problem for
5.2 / 41). Today we read these terms as sig- Garrett’s interpretation, as for Owen’s, has
nifying complex inference involving inter- been in making sense of the logic of Hume’s
mediate steps, but in Hume’s day they were famous argument. For as we have seen, that
understood rather differently. Johnson’s dic- argument does not in fact put much emphasis
tionary of 1756 tells us that ‘reasoning’ is on a general absence of stepwise processing
derived from ‘reason’, and defines it simply or ratiocination within inductive inference.59
as ‘argument’.56 The first sense of ‘argument’ Instead, it focuses on the very specific step of
is given as ‘A reason alleged for or against any extrapolation from observed to unobserved –
thing’, and Johnson implicitly confirms that that is, the supposition of the Uniformity
he takes this as its primary sense in specify- Principle (UP) – and then it attacks in turn
ing – as one of the non-discursive senses of the props on which that principle ‘may be
‘reason’ – ‘Argument; ground of persuasion; suppos’d to be founded’, showing that none
motive’. Likewise the first sense of ‘proof’ of them can ‘afford any just conclusion of this
is given as ‘Evidence; testimony; convincing nature’ (THN / 90). This move makes
token’, supplemented in later editions by perfect sense on an epistemological interpret-
the clauses ‘convincing argument; means of ation of the argument, because if any essen-
conviction’. The words that Johnson favours tial evidential step in an inductive inference
for stepwise inference are ‘deduction’ and lacks a ‘just foundation’, then the inference
‘ratiocination’,57 as in the specification of as a whole will, apparently, be undermined.60
the two discursive senses of ‘reason’: ‘The On the Garrett/Owen style of psychological
power by which man deduces one propos- interpretation, however, the move looks
ition from another, or proceeds from prem- almost irrelevant – even if it sufficed to show
ises to consequences’, and ‘Ratiocination; that UP is not itself reached through mediated
discursive power’. These both contrast with ratiocination, that would not exclude such
‘intuition’, which is ‘Knowledge not obtained ratiocination from playing some other role
by deduction of reason’, and ‘intuitive’, within inductive inference. This objection can


9780826443595_Ch03_Final_txt_print.indd 75 2/9/2012 9:38:52 PM


be sharpened by posing a dilemma over what has shown that UP has no ‘just’ foundation
role UP itself is supposed to play here.61 If in demonstrative or factual reasoning.65 But
Hume is saying that UP functions as an inter- why should this be thought to exhaust the
mediate step in inductive inference, then it possibilities of relevant reasoning? Hume
looks as though he thinks inductive inference quite often refers to ‘arguments’ or ‘rea-
does involve stepwise ratiocination (via UP soning’ that he considers ‘absurd’, ‘falla-
itself), in which case he is contradicting the cious’, or ‘sophistical’ (e.g. THN /
very conclusion that Owen takes him to be 43, / 247; DNR 9.189–92); on a
drawing from the argument. If, on the other psychological interpretation of the famous
hand, Hume is denying that UP can play any argument, these should be as relevant to his
psychological role within inductive inference theory as the ‘just’ reasonings that he is able
(on the basis that it has no ‘just foundation’), to rule out.66 Suppose, for example, it were
then it is unclear why he should take this to suggested that induction can be founded on
imply anything further about the actual psy- the principle that every change must have a
chological mechanism of inductive inference. cause, and hence the ultimate causal laws
The only apparently plausible answer here is to must be consistent over time. This would
see Hume as placing a conditional constraint bring into play the attempted demonstra-
on how stepwise ‘reason’ could work: ‘If rea- tions of the Causal Maxim that Hume
son determin’d us, it wou’d proceed upon’ refutes at THN–8 / 80–2: even if
UP (THN / 89; emphasis added). But fallacious, they could still be contenders as
this conditional statement appears only in the psychological explanations of our induct-
Treatise presentation, and even there, it is fol- ive assumptions. In short, Hume has done
lowed three paragraphs later by the uncondi- nothing to refute the hypothesis that UP
tional statement that ‘probability is founded may be believed on the basis of an invalid
on the presumption of’ UP (THN / demonstrative argument; hence on the inter-
90; emphasis added).62 Owen largely builds pretations of Owen and Garrett, his famous
his interpretation around the conditional,63 argument is hopelessly incomplete.67
but it is straining credibility to rely so heav- But things are even worse than this, for yet
ily on one statement in the Treatise, when we another strategy that remains open on their
have seen so much evidence in sections 2.1–3 readings was actually used by Hume’s friend
above that the versions in the Abstract and and correspondent Richard Price in the
(especially) the Enquiry are more carefully first chapter of his Review of the Principal
crafted. Questions in Morals. Price claims that the
On a psychological interpretation, more- Causal Maxim is known intuitively, ‘nothing
over, Hume should not be so confident that being more clearly absurd and contradict-
mediated ratiocination for a factual conclu- ory, than the notion of a change without a
sion can proceed only via UP.64 At best, he changer’.68 Then, in a footnote a few pages
can claim that any rationally sufficient rea- later, he explains how this can provide a
soning for such a conclusion must involve basis for inferring future events from past
UP. And likewise, his argument seems com- regularity:
pletely inadequate to show that UP itself
could not be believed on the basis of some The conviction produced by experience
mediated ratiocination. At best, again, he is built on the same principle, with that


9780826443595_Ch03_Final_txt_print.indd 76 2/9/2012 9:38:52 PM


which assures us, that there must be a we do than with why those beliefs are
cause of every event . . . Because we see unjustified.71
intuitively, that there being some reason
or cause of this constancy of event, it must But when articulating Hume’s conclusion,
be derived from causes regularly and con-
Owen apparently moves towards a claim
stantly operating . . . And the more fre-
about the functioning of all individual
quently and uninterruptedly we knew this
had happened, the stronger would be our inductive inferences, and strongly contrasts
expectation of its happening again, . . . 69 this with the alternative view as ascribed to

Hume produced no fewer than seven editions Hume is here denying that such infer-
of his Enquiry after Price published this, and ences can be explained as an activity of
it would be astonishing if he did not know the faculty of reason conceived as func-
tioning by the discovery of intermedi-
of it, given that the Enquiry itself – under its
ate ideas . . . Garrett says that Hume ‘is
original title Philosophical Essays – is men-
denying only that we come to engage
tioned twice within the vicinity of these quo- in this species of reasoning as a result
tations (in footnotes on pages 12 and 41). As of any piece of reasoning about it’.72. . .
interpreted by Owen and Garrett, however, My main objection to Garrett’s inter-
Hume’s famous argument completely fails to pretation is that he treats Hume as ask-
engage with Price’s justification of induction. ing about the cause of our engaging in
That justification starts from what Hume probable reasoning . . . Hume’s question
would no doubt claim to be a fallacious ‘intu- is not what Garrett takes it to be. Hume’s
ition’, but again there is no obvious psycho- question is: how is it that we manage to
logical obstacle to erecting an argument on a make these inferences?73
fallacy: humans do it all the time!
Another related issue concerns the intended As we have seen already, however, Garrett’s
scope of the famous argument: is it supposed interpretation in his 1997 book was
to be proving something about every indi- quickly modified (in his 1998 debate with
vidual factual inference, or about the genesis me), at which point he clarified that, like
of our general practice of factual inference? Owen, he took Hume’s conclusion to be
When developing his interpretation, Owen one that applies to all individual inductive
writes repeatedly as though it were the latter, inferences:
for example:
[From] Cognition and Commitment . . . ,
If the uniformity principle were some- Millican understandably infers that on my
thing we knew or believed, prior to our interpretation ‘it is only the general practice
engaging in probable reasoning,70 we of induction that fails to be determined by
could explain probable reasoning as reason, and each of our particular inductive
being based on reason. . . . [P]rior to our inferences is itself an instance of the oper-
engaging in probable reasoning, we . . . ation of our reason.’ . . . The crucial dis-
neither know nor believe the uniform- tinction for Hume, however, is . . . between
ity principle. . . . Hume’s argument . . . an inference being an instance of reason-
has more to do with the failure of reason ing and the same inference being caused by
to account for why we have the beliefs (another instance of) reasoning.74


9780826443595_Ch03_Final_txt_print.indd 77 2/9/2012 9:38:52 PM


The significance of this point is that the of probable reasoning, it will involve
bulk of Hume’s argument, and especially a crucial step that is not supported by
his conclusion – in both the Treatise and any argument . . . ; that is the point of
the Enquiry – seems to intend a result about Hume’s discovery. Reasoning cannot
cause the crossing of an inductive gap.77
every particular factual inference (cf. sec-
tion 2.4 [H] above). But on the interpret-
ations of Owen and Garrett, the argument However, my objection to which this was a
seems to lead far more plausibly to a result response referred not to the situation where
about factual inference in general. Note, for ‘one piece of probable reasoning is part of
example, that Hume takes many induct- another piece of probable reasoning’, but
ive inferences to be reflective and mediated, rather, where one inductive inference’s con-
especially those that involve ‘inferences from clusion (e.g. ‘a general principle of uniform-
contrary phænomena’ (THN / 133) ity’ as at THN / 104–5) is then given
or the application and balancing of ‘general ‘the role of a premise in further inductive
rules’ (THN–12 / 146–50, 1.3.15 / inference’.78 Garrett’s response thus preserves
173–5). Moreover, inductive inferences from what he takes to be Hume’s conclusion, as
‘only one experiment of a particular effect’ universally applicable to factual inferences,
can – Hume says – be mediated by explicit only by stipulatively treating the inference
reflection on a principle which looks very which was used to establish a proposition
similar to UP (THN / 104–5).75 as itself a part of the further inference which
Owen is well aware of this,76 but he does then takes that proposition as a premise. In
not apparently recognize the threat to his the context of discussing the epistemology
own interpretation, under which such medi- of induction, this might seem reasonable
ated inferences become counterexamples to enough: if the proposition in question has a
Hume’s conclusion, at least if that is read problematic foundation, then those problems
(as Hume’s own words seem to require) as a will be inherited by any further inference
claim about each and every factual inference. built on it. But if Hume’s famous argument
Garrett is also aware that ‘not all probable is to be interpreted as involving the psycho-
inferences are immediate’, but he endeavours logical mechanism of individual inductive
to explain how nevertheless Hume’s conclu- inferences – as Garrett intends – then the
sion can be seen to apply even to those that move looks artificial and ad hoc, smudging
are mediated: over the manifest difference between argu-
ing for some proposition and taking it as
it may well happen, as Millican notes, assumed or established (on the basis of pre-
that one piece of probable reasoning is vious argument).
part of another piece of probable rea-
To sum up so far, we have yet to find an
soning . . . But as Hume states his con-
interpretation that is genuinely satisfying. Any
clusion . . . , it is that ‘in all reasonings
from experience, there is a step taken by extreme sceptical reading leaves Hume’s phil-
the mind which is not supported by any osophy hopelessly inconsistent. The anti-de-
argument or process of the understand- ductivist reading and that of Owen both have
ing’ [EHU 5.2 / 41; emphasis added]. serious difficulty making sense of the logic of
Where a piece of probable reasoning his argument, and also have to rely on the text-
does occur as part of a second piece ually questionable claim that Hume’s notion of


9780826443595_Ch03_Final_txt_print.indd 78 2/9/2012 9:38:53 PM


‘reason’ within the argument is a target rather is looking far more like a discussion of the
than sincere (a problem which also beset my epistemology of our general presumption of
own previous interpretation).79 Garrett’s read- uniformity.
ing has the significant merit of avoiding this
last pitfall, but again has difficulty squaring 3.2 ‘REASON’ AS THE COGNITIVE FACULTY
with the argument’s text and logic, and has
been forced to adapt accordingly over time. One promising route towards a better under-
Initially, Garrett understood Hume’s conclu- standing of Humean ‘reason’ is to look at the
sion as the straightforward claim that ‘we do usage of Hume’s contemporaries in Scotland
not adopt induction on the basis of recogniz- and England, and especially those who –
ing an argument for its reliability’.80 This soon unlike Locke – were enthusiastic about the
changed into the more complex claim that ‘the language of ‘faculties’ and relatively con-
underlying causal mechanism that gives rise to sistent in their usage.84 Francis Hutcheson,
inductive inferences . . . is not itself dependent Professor of Moral Philosophy at Glasgow
on any reasoning or inference’.81 Meanwhile, and correspondent with Hume from 1739,
since the Enquiry argument clearly ranges provides the closest spatio-temporal match
beyond narrow ‘reasoning or inference’, to the Treatise, having published in 1742 no
Garrett suggested that here Hume ‘expands fewer than four works containing an out-
the famous conclusion to rule out any “rea- line of the faculties, at least one of which –
soning or process of the understanding,” Philosophiae Moralis Institutio Compendiaria
thereby eliminating such non-inferential proc- (later translated as A Short Introduction to
esses of the understanding as intuition or the Moral Philosophy) – he sent to Hume.85 This,
perception of a probable connection between like the Synopsis Metaphysicae (Synopsis of
even a single “proof” and a conclusion’.82 But Metaphysics)86 was a Latin teaching text, and
pushing in the opposite direction, his recog- contains a discussion of ‘The parts or pow-
nition that Hume acknowledges the role of ers of the soul’.87 The other two works were
explicit (and sometimes complex) ratiocin- An Essay on the Nature and Conduct of the
ation within some inductive inferences has led Affections and Illustrations on the Moral
to a narrowing of the supposed conclusion, to Sense, published together as a third edition
focus on the very specific logical step which is of both. To the former, Hutheson added a
‘the crossing of an inductive gap’.83 Even after footnote on the faculties,88 and to the latter,
all this, as we have seen, Garrett’s defence of a new paragraph:
the interpretation looks suspiciously ad hoc,
holding that conclusion to be true even of an Writers on these Subjects should
inductive inference which explicitly argues remember the common Divisions of
across the inductive gap using an anteced- the Faculties of the Soul. That there is
1. Reason presenting the natures and
ently established Uniformity Principle, sim-
relations of things, antecedently to any
ply on the basis that some previous inference
Act of Will or Desire: 2. The Will, . . .
was required to establish that principle. But or the disposition of Soul to pursue
by now the interpretation has been diluted what is presented as good, and to shun
beyond recognition, and we seem to have lost Evil. . . . Both these Powers are by the
any focus on the actual psychological mech- Antients included under the Λόγος or
anism of individual inductive inferences – this λογικòν μέρος. Below these they place


9780826443595_Ch03_Final_txt_print.indd 79 2/9/2012 9:38:53 PM


two other powers dependent on the This sceptical doubt, both with respect
Body, the Sensus, and the Appetitus to reason and the senses . . . ’Tis impos-
Sensitivus, in which they place the par- sible upon any system to defend either
ticular Passions: the former answers to our understanding or senses . . . ’
the Understanding, and the latter to (THN / 218; emphasis added)
the Will. But the Will is forgot of late,
and some ascribe to the Intellect, not
only Contemplation or Knowledge, but Consider also the following two footnotes,
Choice, Desire, Prosecuting, Loving.89 the first of which was expanded and moved
to create the second:

It is clear from his alternation between when it [the imagination] is oppos’d

‘Reason’ and ‘the Understanding’ that to the understanding, I understand the
Hutcheson takes these to be one and the same faculty, excluding only our demon-
strative and probable reasonings. (THN
same; indeed his Essay’s new footnote says as / 371n; emphasis added)
much.90 This equivalence is also asserted (or
manifested through the same sort of elegant when I oppose it [the imagination] to
reason, I mean the same faculty, exclud-
variation of terminology that we see above)
ing only our demonstrative and prob-
by various other writers known to Hume, for
able reasonings. (THN /
example Shaftesbury, Butler and Price,91 so it 117–18n; emphasis added)
was evidently commonplace, though writers
in the Scottish common-sense school later
preferred to use ‘reason’ more narrowly, in Again, the switch from ‘the understanding’
much the way that Garrett favours.92 Hume to ‘reason’ looks purely stylistic, perhaps
himself, however, is clearly aligned with prompted by the clumsiness of ‘. . . the under-
the former group, interchanging between standing, I understand . . . ’.
‘reason’ and ‘the understanding’ dozens of There is thus overwhelming textual evi-
times – purely for the sake of stylistic vari- dence that Hume generally treats ‘reason’
ation – just as he does between ‘the fancy’ and ‘the understanding’ as one and the
and ‘the imagination’:93 same. And virtually all major writers of the
period take ‘the understanding’ to refer to
the mind . . . is not determin’d by reason, our principal cognitive faculty, usually draw-
but by certain principles, which associate ing a general division between it and ‘the
together the ideas of these objects, and will’.94 This division between the domains
unite them in the imagination. Had ideas
of the understanding and the will is indeed
no more union in the fancy than objects
essentially the same as the modern distinc-
seem to have to the understanding, . . .
(THN / 92; emphasis added) tion between cognitive and conative mental
functions, a dichotomy whose fundamental
There are no principles either of the
nature is often now expressed in terms of a
understanding or fancy, which lead
us directly to embrace this opinion . . . ‘direction of fit’ between world and mind:
The . . . hypothesis has no primary rec- the understanding aims to conform our
ommendation either to reason or the beliefs to the way the world is, while the will
imagination . . . (THN / 211; aims to change the world to conform to our
my emphasis) desires. Reid characterizes this in terms of a


9780826443595_Ch03_Final_txt_print.indd 80 2/9/2012 9:38:53 PM


distinction between our intellectual (or con- This treats the senses themselves as ‘oper-
templative) and active powers: ations of the understanding’, a tendency
common enough for Price to make a point
We shall . . . take that general division of criticizing it.98 But Hutcheson’s Synopsis
which is the most common, into the pow- of Metaphysics, within two sentences, first
ers of understanding and those of will. implicitly places the senses within the under-
Under the will we comprehend our active standing and then gives them a subordinate
powers, and all that lead to action, or reporting role, which suggests that the former
influence the mind to act; such as, appe-
placement is just a shorthand way of indicat-
tites, passions, affections. The under-
ing that the senses fall within the understand-
standing comprehends our contemplative
powers; by which we perceive objects; by ing’s sphere of influence:
which we conceive or remember them;
by which we analyse or compound them; we might reasonably reduce [the powers
and by which we judge and reason con- of the mind] to two, namely, the faculty
cerning them. . . . The intellectual powers of understanding and the faculty of will-
are commonly divided into simple appre- ing, which are concerned respectively
hension, judgment, and reasoning.95 with knowing things and with render-
ing life happy. The senses report to the
understanding, . . .99
Although Reid is somewhat critical of this
framework, which he takes to be ‘of a very
general reception’, his clear account of it is The Synopsis goes on to give Hutcheson’s
helpful in setting the scene for Hume, whose most detailed account of his faculty frame-
understanding of it – though again critical – work, with Chapter 1 of Part II devoted to a
seems to be very similar.96 categorization of the powers associated with
The earlier quotation from Hutcheson’s the understanding, including external sensa-
Illustrations on the Moral Sense likewise tion (sect. 3), internal senses or consciousness
recognizes this ‘common Division of the (sect. 4), reflexive or subsequent sensations
Faculties of the Soul’ between ‘Reason’ (or ‘the (sect. 5), memory, the power of reasoning and
Understanding’) and ‘The Will’, while suggest- imagination (sect. 6).100 This again suggests a
ing a hierarchical structure in which the senses hierarchical structure, with these various pow-
‘answer to’ the understanding, and the passions ers ‘reporting to’ an overseer faculty – reason
to the will. His Short Introduction to Moral or the understanding proper – which perceives
Philosophy, however, paints a cruder picture and judges the deliverances of the subordin-
which is closer to that outlined by Reid above: ate faculties in order to establish truth. Thus
‘Reason is understood to denote our Power of
The parts or powers of the soul . . . are all finding out true Propositions’.101 Price talks
reducible to two classes, the Understanding in a similar spirit of ‘the power within us that
and the Will. The former contains all the understands; . . . the faculty . . . that discerns
powers which aim at knowledge; the truth, that views, compares, and judges of all
other all our desires. . . . [Of] the several ideas and things’.102 And a similar conceptual
operations of the understanding . . . The linkage between reason or the understanding
first in order are the senses . . . Senses are and the search for truth is common to many
either external, or internal.97 other writers.103


9780826443595_Ch03_Final_txt_print.indd 81 2/9/2012 9:38:53 PM


We are now in a position to appreciate the is entirely unreasonable, must proceed

significance of Hume’s repeated statements from some other faculty than the under-
that align him strongly with this general con- standing. . . . Even after we distinguish
ception of reason as cognition:104 our perceptions from our objects, ’twill
appear presently, that we are still incap-
able of reasoning from the existence of
Reason is the discovery of truth or fal- one to that of the other: So that upon
shood. (THN / 458); the whole our reason neither does, nor is
That faculty, by which we discern it possible it ever shou’d, . . . give us an
Truth and Falshood, . . . (EHU 1.4n – assurance of the continu’d and distinct
1748/1750 editions);105 existence of body. (THN / 193)

Thus the distinct boundaries and offices ’tis a false opinion that any of our . . .
of reason and of taste are easily ascer- perceptions, are identically the same
tained. The former conveys the knowl- after an interruption; and consequently
edge of truth and falsehood: . . . (EPM . . . can never arise from reason, . . .
App. 1.21 / 294); (THN / 209)
we may observe a conjunction or a rela-
reason, in a strict sense, as meaning the
tion of cause and effect betwixt differ-
judgment of truth and falsehood, . . .
ent perceptions, but can never observe
(DIS 5.1 / 24; cf. THN / 417).
it betwixt perceptions and objects. ’Tis
impossible, therefore, that from the exist-
In all these contexts, Hume is stressing that ence or any of the qualities of the former,
reason, since it is purely cognitive, cannot we can ever form any conclusion con-
also be conative: that is, it cannot be contrary cerning the existence of the latter,106 or
ever satisfy our reason in this particular.
to any passion or – by itself – provide any
(THN / 212)
motive to action or the will. This is the crux
of one of Hume’s three most famous argu-
ments concerning the incapacity of reason, It might appear strange that in one of these
which concludes that ‘Since morals . . . have arguments, reason seems to embrace both
an influence on the actions and affections, it truth and falsehood, whereas in the other, it
follows, that they cannot be deriv’d from rea- is normatively connected with truth. But this
son’ (THN / 457). For the purpose of sort of linguistic variation is commonplace,
this argument, it is enough that reason is con- and it is worth noting that an unambiguous
fined to the domain of truth and falsehood, identification of reason with the cognitive
though Hume’s talk of discovery, discern- faculty is consistent with a fairly wide range
ment and knowledge suggests a normative of nuances of meaning. Given such an iden-
bias towards truth rather than falsehood. tification, ‘reason’ might most naturally be
This normative flavour is far more explicit used to refer to the human (or animal) fac-
in another of the three famous arguments, ulty of truth-apprehension, however well
this time concerning the external world (the and by whatever processes it operates (as,
third, of course, concerns induction): for example, when Hume compares the ‘rea-
son’ of people and animals at THN /
The vulgar confound perceptions and 610). But sometimes there might be debate
objects . . . This sentiment, then, as it over these processes, in which case we could


9780826443595_Ch03_Final_txt_print.indd 82 2/9/2012 9:38:54 PM


find ourselves referring to processes that are our natural faculties, then one would expect
commonly taken to be involved in truth- that his arguments about reason’s capabil-
apprehension, even if they turn out not to ities would start from a relatively straightfor-
be truth-conducive (as suggested by Hume’s ward and conventional understanding of our
‘scepticism with regard to reason’ of Treatise cognitive functions. Coming from a Lockean
1.4.1). Alternatively, we might wish to apply background, it is no surprise to find Hume
a stricter criterion under which ‘reason’ recognizing the cognitive faculties of the
would be confined to processes that operate senses, memory, intuition, and demonstrative
successfully to apprehend truth (thus giving and probable reasoning.109 The senses can
the normative flavour of the passages from be either external (i.e. sight, touch, hearing,
Treatise 1.4.2 quoted above).107 A differ- smell, gustatory taste) or internal (i.e. reflec-
ent strict usage is to refer to the faculty of tion) – these provide the impressions from
truth-apprehension acting entirely alone, which our ideas are copied, and those ideas
independently of other faculties such as the are represented to us either through the mem-
senses or memory (this seems to be Hume’s ory or the imagination. It follows that all of
intention at THN / 470).108 Finally, our thinking, except in so far as it confines
there is in the early modern period a com- itself to memory, must involve representation
mon metonymy, under which ‘reason’ is used of ideas in the imagination, which is appar-
to refer to its product, namely true belief as ently to be thought of as something like a
successfully achieved using our rational fac- multi-layered or multi-dimensional canvas
ulty (hence the pairing of ‘truth and reason’ on which sense-copied ideas appear, with dif-
at THN / 415, / 461, and ferent degrees of ‘force and vivacity’ and ‘in a
THN App. 1 / 623; cf. also note 84 above). perpetual flux and movement’ (THN
Notice that acknowledgement of all these / 252).110 Thus our faculties of intuition, dem-
nuances is quite different from supposing an onstration, and probable reasoning must
ambiguity in ‘reason’, because they all arise inevitably act on our imagination, through
naturally from the core meaning, and there such processes as bringing ideas into mind,
need be no suggestion that the word has been dismissing others, or – most importantly
coincidentally assigned two or more distinct given Hume’s analysis of belief (summarized
meanings. With this understood, much of the at the end of section 2.1 above) – changing
evidence that has previously been adduced their degrees of force and vivacity. Even when
for the ambiguity of ‘reason’ is significantly we make judgements about the deliverances
undermined, and it becomes more plausible of our senses and memory, it is their force
to suggest that the term has, for Hume, a sin- and vivacity in the imagination which appar-
gle core meaning, namely what we now call ently constitutes our assent to them.111 Hence
cognition. Hume’s comment, leading into the sceptical
anxieties of the Conclusion of Treatise Book
3.3 REASON AND THE IMAGINATION 1, that: ‘The memory, senses, and understand-
ing are, therefore, all of them founded on the
If reason, for Hume, is just our overall cogni- imagination, or the vivacity of ideas.’ (THN
tive faculty, and if his general epistemological / 265). This comment might be read as
approach is – as set out in section 1 above – suggesting that the imagination is itself active,
to begin by ascribing default authority to but earlier in the same paragraph, Hume


9780826443595_Ch03_Final_txt_print.indd 83 2/9/2012 9:38:54 PM


makes clear that he is talking of principles the opprobrious character of being the
(namely experience and habit) that ‘operate offspring of the imagination. By this
upon the imagination’. The initial framing of expression it appears that the word,
his discussion of induction in the Treatise (as imagination, is commonly us’d in two
different senses; and tho’ nothing be
quoted earlier in section 2.1) may give a differ-
more contrary to true philosophy, than
ent impression: ‘the next question is, Whether
this inaccuracy, yet in the following rea-
experience produces the idea by means of the sonings I have often been oblig’d to fall
understanding or imagination; whether we into it. When I oppose the imagination to
are determin’d by reason to make the tran- the memory, I mean the faculty, by which
sition, or by a certain association and rela- we form our fainter ideas. When I oppose
tion of perceptions?’ (THN / 88–9). it to reason,115 I mean the same faculty,
But Hume’s answer to his own question – excluding only our demonstrative and
repeated numerous times – will be that our probable reasonings. (THN
causal reasoning is determined by custom,112 / 117–18n)
and he never says that it is determined by the
imagination itself. So at least in this context, This note was inserted by means of a ‘cancel’
the imagination is apparently only the virtual leaf, prepared by Hume while the Treatise
stage on which the mind’s various principles – was going through the press, and I believe
either of reason or custom – orchestrate their he saw the need for this on rereading the end
dance of perceptions.113 of THN / 108:116 ‘All this, and every
In other contexts, however, the imagination thing else, which I believe, are nothing but
does appear as an active agent, having the lib- ideas; tho’ by their force and settled order,
erty to transpose and change its ideas (THN arising from custom and the relation of cause / 10, / 10–11, / 85, and effect, they distinguish themselves from / 97; EHU 2.4 / 18, 5.10 / 47–8, 5.12 the other ideas, which are merely the off-
/ 49), to distinguish and separate them (THN spring of the imagination.’ A related pas- / 40, / 54–5, / 79–80, sage is at THN / 225, where Hume / 233), to suggest (THN / addresses the complaint that he has criticized
23–4) or raise them up (THN / 27), ‘the antient philosophers’ for being guided by
and to generate fictions (THN / 16, imaginative fancies, whilst building his own / 200–1, / 205, philosophy on principles of the imagination:
/ 209, / 215,–5 / 220–1, / 224–5,–7 / 253–5).114 The In order to justify myself, I must dis-
distinction between the two classes of oper- tinguish in the imagination betwixt the
ation seems to be explained by the footnote at principles which are permanent, irresist-
ible, and universal; such as the custom-
THN / 117 which we encountered in
ary transition from causes to effects, and
section 3.2 above:
from effects to causes: And the principles,
which are changeable, weak, and irregu-
In general we may observe, that as lar; such as those I have just now taken
our assent to all probable reasonings notice of [e.g. the ‘inclination in human
is founded on the vivacity of ideas, it nature to bestow on external objects
resembles many of those whimsies and the same emotions, which it observes
prejudices, which are rejected under in itself’, as attributed to the ‘antient


9780826443595_Ch03_Final_txt_print.indd 84 2/9/2012 9:38:54 PM


philosophers’]. The former are the foun- our memory’ (EHU 4.3 / 26). He then iden-
dation of all our thoughts and actions, so tifies the crucial step of such inference: the
that upon their removal human nature extrapolation from observed to unobserved
must immediately perish and go to ruin. which is encapsulated in his Uniformity
The latter are neither unavoidable to
Principle. If this is to qualify as founded on
mankind, nor necessary, or so much as
reason, then there must be some cognitive
useful in the conduct of life . . .
operation that grounds it, and which does
so through genuine cognition (rather than
All three passages point to a distinction some fallacy or confusion). In the Treatise
drawn within the class of principles that and Abstract, Hume apparently takes it to be
‘operate on the imagination’ – that is, which obvious that the only plausible contenders
affect our thinking. Some of these are ‘the here are demonstrative reasoning and prob-
foundation of all our thoughts and actions’, able (i.e. moral or factual) reasoning. In the
the ‘permanent, irresistible, and universal’ Enquiry he is more thorough, and rules out
principles that ground ‘our demonstrative also both intuition and sensory knowledge
and probable reasonings’, and are therefore as sources of foundation for the Uniformity
appropriately dignified with the name of rea- Principle. Since memory is taken for granted
son or the understanding. The other princi- in the experiential observations from which
ples are those that we more naturally think of the inference starts, this exhausts all the
as belonging to the imagination itself: those standardly recognized sources of evidence
that ground our free play of ideas, fictions, with which reason might operate. It is there-
whimsies and prejudices. Hence in this nar- fore no coincidence that the four sources
rower sense the imagination is in opposition considered – and rejected – in Hume’s argu-
to reason, though both sets of principles per- ment in the 1748 Enquiry match up exactly
form on the same stage – the imagination in with those itemized in his 1745 Letter from
the broader sense – where all our non-mem- a Gentleman: ‘It is common for Philosophers
ory ideas are represented. It is this broader to distinguish the Kinds of Evidence into
sense which enables Hume to refer, without intuitive, demonstrative, sensible, and moral;
paradox, to ‘the understanding, that is, . . . . . .’ (LFG 22). If reason is understood by
the general and more establish’d properties Hume in the standard contemporary way –
of the imagination’ (THN / 267).117 as the overall cognitive faculty – then we
should indeed expect it to embrace all four
3.4 AN OPERATION OF REASON WHICH IS ‘Kinds of Evidence’.
‘NOT DETERMIN’D BY REASON’ Notice that this way of reading Hume’s
argument has the clear implication that
Equipped with this understanding of the inductive (i.e. probable, moral, factual)
Humean faculties, let us now try to clarify inference is being treated as an operation of
the significance of his famous argument. In reason throughout, which at least strongly
the Treatise, he considers a paradigm causal suggests that it would be a mistake to inter-
inference ‘from the impression to the idea’; pret Hume’s conclusion – that such inference
in the Abstract and Enquiry, he widens this is ‘not determin’d by reason’ – as deposing
to any factual inference ‘beyond the present it from that status. (For if that were indeed
testimony of our senses, or the records of his intention, one might reasonably expect


9780826443595_Ch03_Final_txt_print.indd 85 2/9/2012 9:38:54 PM


such an apparently paradoxical move to A similar theme can be seen in the other of
be far more clearly signalled.) So we need Hume’s most famous arguments that assigns
to understand how Hume, as a result of a vital role to the imagination, on ‘Scepticism
his famous argument, can coherently view with Regard to the Senses’ (THN 1.4.2 /
induction as an operation of reason which is 187–218). Here he takes on the natural and
not ‘determin’d by’ reason.118 naive assumption that external objects – dis-
The obvious answer, given both our inter- tinct from us and continuous over time – are
pretation of reason and the structure of directly and straightforwardly perceived
Hume’s argument, is that he views induction through the senses. To refute this, he shows
as a cognitive process which depends on a that identification of objects over time
non-cognitive sub-process. So he is think- requires a process that goes beyond anything
ing at two levels, with inductive inference we perceive, latching onto patterns of ‘con-
being a manifest operation of our conscious stancy’ and ‘coherence’ in our distinct impres-
reason, causally driven by a subconscious sions, and smoothing over gaps and changes
process that involves the customary enliven- to generate an illusion of continuity. Again
ment of our ideas. This underlying process the process involved is naturally categorized
is of a type which is naturally categorized as as ‘imaginative’, and so Hume describes
‘imaginative’ rather than ‘rational’, because his argument as showing that our ‘assur-
it works through an associative mechanism ance of the continued and distinct existence
which automatically and mindlessly extrapo- of body . . . must be entirely owing to the
lates beyond anything that we have perceived imagination’ (THN / 193). Like
or otherwise detected in the world (whether his argument concerning induction, there-
objective events, or evidence). It is therefore fore, this can be seen as making a significant
in sharp contrast with the underlying pro- contribution to both cognitive science and
cess hypothesized by Locke, who supposed epistemology, by highlighting how the infor-
inductive (i.e. probable) inference to be mational processes that are implicit in the
driven by a perceptual process, namely the temporal identification of physical objects
rational apprehension of objective probable go well beyond anything that is directly per-
connexions. Locke therefore saw induction ceived. Indeed modern-day cognitive science,
as a cognitive process which depends on a through the development of ‘artificial intel-
cognitive sub-process – apparently ‘cognitive ligence’ visual systems, has provided striking
all the way down’ because it is ultimately vindication of Hume, by showing how even
founded on direct perception of evidential the identification of physical objects at a time
connexions. Hume’s argument destroys this requires ‘imagination-like’ processes of edge-
illusion by showing that there is no plausible detection, region identification, shadow and
source for such perception: it cannot derive texture interpretation, and so forth. So far
from examining relations of ideas (because it from being merely passive, visual perception
depends on the experienced world), but nor involves many active – albeit unconscious –
can it derive from experience, either current processes, without which the manifold of our
(because the senses detect no such evidential sensory impressions would be completely
connexions) or remembered (because induct- incomprehensible.
ive extrapolation is the very process whose This interpretation of Hume involves
perceptual basis we are seeking). understanding his talk of faculties as


9780826443595_Ch03_Final_txt_print.indd 86 2/9/2012 9:38:54 PM


descriptive of types of process rather than as stress the liberty of the imagination (EHU
references to parts of the mind, and indeed 2.4 / 18, 5.10 / 47–8, and 5.12 / 48–50). The
this seems anyway to be required in the light contrast is especially striking in the case of
of his general scepticism about any faculty reason, because whereas the Treatise speaks
language that pretends to be more than a of reason itself as a determining cause (e.g.
functional description (THN / 224; THN / 88–9, / 92,
DNR 4.162–3). For Hume, as for Locke, a / 97, / 180), the Enquiry never does
faculty just names a power.119 Nevertheless, so. In the later work, Hume’s preference is
at least in the Treatise, he has an unfortunate to talk instead of reasoning processes (e.g.
tendency to talk of faculties in the way that EHU 4.23 / 39, 5.4 / 42, 9.6 / 108), which
Locke rightly deplored, as ‘so many distinct were never mentioned as such in the Treatise.
Agents’.120 Taking such language literally, his Meanwhile, custom in the Enquiry is said
famous argument paints the absurd picture to act on the imagination (EHU 5.11 / 48,
of reason attempting in vain to make an 9.5 / 106–7) and is never said (or implied)
inductive inference, and needing the imagi- to be itself an operation of the imagination,
nation to step in to lend a hand. But induc- thus avoiding the complications that arise
tion is such a central cognitive process that from trying to place it consistently within the
it ought by definition to be an operation of conventional faculty structure.122 Moreover,
reason (just as remembering is by definition Hume no longer refers to custom as a prin-
an operation of memory); hence if we think ciple of association of ideas (cf. THN
of faculties as distinct agents or areas of the / 97),123 but says instead that it is a process
mind, custom – as the underlying process analogous to the association of ideas, which
that drives induction – should itself be part ‘is of a similar nature, and arises from simi-
of reason. Presumably it is this sort of con- lar causes’ (EHU 5.20 / 53–4). He continues,
sideration that led Hume, in the wake of his however, to draw a contrast between cus-
famous argument, to reassign ‘the general tom and reason (EHU 5.5 / 43, 5.20 / 53–4),
and more establish’d properties of the imagi- thereby retaining the core of his theory that
nation’ – which must surely include custom – inductive inference is determined by a sub-
to reason or the understanding (THN process which is not itself cognitive.
/ 267, cf. section 3.3 above).121 But then we
get into a muddle if we want to hold on to
his conclusion that inductive inference is ‘not
determin’d by reason’, given the frequency 4. CONCLUSION: SCEPTICISM AND
with which he says that inductive inference is RATIONAL FOUNDATIONS
indeed determined by custom. Little wonder,
perhaps, that both Hume and his interpreters After all this, how sceptical is Hume’s pos-
sometimes seem to exhibit confusion of the ition? His famous argument has shown
faculties! that inference from past to future crucially
As so often, the Enquiry brings consid- involves a process of extrapolation that can-
erable improvement, and in a number of not be independently justified by anything
respects. Now the faculties are rarely spo- within our cognitive grasp. This crucial step
ken of as agents in their own right, with the is instead due to a mechanical associative
harmless exception of those passages that process in the mind, whereby past experience


9780826443595_Ch03_Final_txt_print.indd 87 2/9/2012 9:38:54 PM


raises certain ideas about the future and enli- of impressions), he expresses his conclusion
vens them into beliefs. Such a process – given in a way that ignores the obvious and essen-
its automatic, non-reflective nature, and its tial role of the senses:
lack of any rational insight or apprehension
of reality – is naturally classified as ‘imagina- That opinion must be entirely owing to
tive’ rather than ‘cognitive’, and Hume’s the imagination (THN / 193;
faculty language is best interpreted accord- emphasis added).
ingly, as a way of categorizing types of pro-
cess, rather than as a theory of distinct agents Hume seems to be assuming here that even
within our minds. So when he claims that the one imaginative step is sufficient to charac-
imagination plays a crucial role in inductive terize the entire process of which it is a part
inference, he should be understood as saying as one that is determined by the imagination
simply that our process of making inductive (rather than by reason or the senses), just as
inferences itself crucially involves an imagi- one invalid step within a sequential inference
nation-like sub-process. typically renders the entire inference invalid.
As we have seen (in section 3.1 above), Such a focus when speaking of inferential
Hume is well aware that many inductive infer- processes is indeed quite natural, since we
ences also involve reason-like sub-processes, are typically interested in the weakest link in
as for example when we consciously take any chain of support rather than the strong-
into account the ‘rules by which to judge of est. The same applies to other supportive or
causes and effects’ of Treatise 1.3.15 / 173–6, foundational relationships: thus a climber can
or attempt to identify underlying mathemat- properly be described as ‘supported only by a
ical patterns (e.g. EHU 4.13 / 31, 7.29n17 rope’, whether that rope itself is secured to a
/ 77n). But he is clearly far more interested mountain, a building, a heavy vehicle or any
in the crucial imaginative step, even to the other relatively reliable anchor.124 Likewise,
extent of describing it as solely responsible an argument or legal case which crucially
for the inference: depends on some imaginative fabrication,
even if it also depends on numerous points
When the mind . . . passes from the idea that are logically unassailable, can appro-
or impression of one object to the idea priately be said to be ‘founded on fantasy’.
or belief of another, it is not determin’d But if we follow through this line of thought,
by reason, . . . The inference . . . depends
then since inductive inference depends on a
solely on the union of ideas. (THN
sub-process of ‘imaginative’ extrapolation / 92; emphasis added)
which itself has no rational grounding, we
all reasonings are nothing but the effects seem forced to conclude that any proposition
of custom; . . . (THN / 149;
that can be established only by such infer-
emphasis added; cf. EHU 5.5–6 / 43–4)
ence must apparently in turn be disqualified
from counting as founded on reason. Yet as
Similarly with his argument concerning our we have seen in sections 3.1 and 3.4 above,
belief in the continued and distinct existence Hume continues to treat induction as a legit-
of body (which aims to show that it depends imate operation of reason.125 There is, at the
crucially on various associative processes, least, a sceptical tension here: can we really
constructing ‘fictions’ from the passing show suppose that he would consider a process


9780826443595_Ch03_Final_txt_print.indd 88 2/9/2012 9:38:55 PM


genuinely rational which rests on a purely foundation for belief and also a cause. By
mechanical, non-reflective sub-process? contrast, ‘imagination-like’ processes such
To address this worry, suppose that Hume as custom may cause belief, but they cannot
were to take the alternative view, that any at the same time provide a cognitive foun-
rational process must have a rational foun- dation: that is indeed precisely why they do
dation. It would then immediately follow not qualify as processes of reason.127 Hume
that for anything to be founded on reason seems to have embraced this distinction, if
at all, it must be founded on reason ‘all the not perhaps immediately, for his language in
way down’ (i.e. it would have to be solidly the Abstract and Enquiry (though not in the
founded on evidence or principles, which Treatise) precisely fits it. In the Treatise, he
are either immediately apprehended by repeatedly talks about custom (or the prin-
reason, or else themselves solidly founded ciples of the imagination) as providing a
on evidence or principles which are either foundation for inductive inference.128 In the
immediately apprehended by reason or . . . , Abstract and Enquiry, by contrast, he never
etc.). Hume would thus be committed to a does, but there are no fewer than 19 passages
strongly rationalistic notion of reason, the that describe the influence of custom in terms
demands of which would be impossible to that are either explicitly causal, or naturally
fulfil without abandoning the heart of his interpretable as such.129 This strongly sug-
philosophy. At no point would he be able to gests that Hume himself came to recognize a
halt the foundational regress by acknowledg- firm distinction between what in the Enquiry
ing that ultimately the principles of our rea- he calls a foundation, and what he there calls
son can (legitimately) be grounded on basic a determining cause. Thus understood, it is
psychological mechanisms. So the only pos- clear that custom qualifies only as a cause,
sible outcomes would be either rationalism whereas reasoning processes or sources of
or incurable scepticism. Some interpreters information can potentially provide a cogni-
have indeed seen Hume as impelled towards tive foundation.130
radical scepticism by precisely this kind of All this brings the possibility of posing
regressive train of thought.126 But it would coherent but unanswerable questions, such
be completely at odds with his efforts to as that which introduces Hume’s discussion
ground a conception of reason on the con- in Part 2 of Enquiry Section 4: ‘if we still
tingent operations of the human mind, and carry on our sifting humour, and ask, What
flatly incompatible – in the light of his own is the foundation of all conclusions from
investigations – with treating induction as a experience? this implies a new question . . .’
genuine operation of reason. (EHU 4.14 / 32). If custom cannot qualify
As we saw in section 3.4 above, Locke as a foundation, then Hume’s ultimate con-
implicitly follows the path that Hume rejects, clusion that ‘All inferences from experience
by attributing probable reasoning to the per- . . . are effects of custom, not of reasoning’
ception of probable connexions. And indeed (EHU 5.5 / 43) excludes any foundation at
direct perception – conceived of as a process all. For in competing successfully for the
of transparent apprehension – seems to be a causal explanatory role, custom effectively
paradigm of what reason requires if it is to excludes anything else from the foundational
be ‘cognitive all the way down’. Such percep- role (which it is nevertheless unable to fulfil
tion could at once provide both a rational itself). Perhaps, then, there is hidden depth


9780826443595_Ch03_Final_txt_print.indd 89 2/9/2012 9:38:55 PM


in Hume’s declaration of intent: ‘I shall con- inference possible for human beings, despite
tent myself, in this section, with an easy the sceptical impact of his famous argument
task, and shall pretend only to give a nega- which shows that it cannot be founded in any
tive answer to the question here proposed.’ of the ways that previous tradition would
(EHU 4.15 / 32). The upshot is that ‘if we still countenance.131 He reveals how we actu-
carry on our sifting humour’ in the search ally reason inductively, rather than falling
for ultimate foundations, we hit rock bot- back on the aprioristic supposition that this
tom with something that has a cause but no can only be through the rational perception
foundation. And that is the tendency, rooted of evidential connexions. That traditional
in our animal nature, to infer from past to notion is decisively refuted by his sceptical
future, from experienced to not-yet-expe- argument, but his own position is very far
rienced. This is radically different from the from sceptical. On the contrary, as we saw in
kind of perceptual foundation presupposed section 1 above, Hume sees very good reason
by traditional conceptions of reason, differ- to accept our faculty of inductive inference
ent enough to make Hume’s position seem as it is (at least when suitably disciplined by
outrageously sceptical by comparison. But in general rules etc.), and no good reason to
reality it is quite the reverse: he is providing reject it. We have, indeed, no alternative, nor
an account of reason which makes inductive any compelling reason for desiring one.132



Hume’s Own Statement of the Propositions It is allowed on all hands, that there is
Identified in the Structure Diagram no known connexion between the sen-
sible qualities and the secret powers . . .
1. By means of [Cause and Effect] alone
(EHU 4.16 / 33)
can we go beyond the evidence of our
4. . . . every effect is a distinct event from its
memory and senses. (EHU 4.4 / 26)
cause. It could not, therefore, be discov-
2. All reasonings concerning matter of fact
ered in the cause, and . . . the conjunc-
seem to be founded on the relation of
tion of it with the cause must appear . . .
Cause and Effect. (EHU 4.4 / 26)
arbitrary; since there are always many
. . . all arguments concerning existence
other effects, which, to reason, must
are founded on the relation of cause and
seem fully as consistent and natural.
effect . . . (EHU 4.19 / 35)
(EHU 4.11 / 30)
. . . all our evidence for any matter of
5. . . . the knowledge of [cause and effect]
fact, which lies beyond the testimony
is not, in any instance, attained by rea-
of sense or memory, is derived entirely
sonings a priori; but arises entirely from
from the relation of cause and effect . . .
experience . . . (EHU 4.6 / 27)
(EHU 12.22 / 159)
. . . causes and effects are discover-
3. No object ever discovers, by the quali-
able, not by reason, but by experience
ties which appear to the senses, either
. . . (EHU 4.7 / 28)
the causes which produced it, or the
In vain, therefore, should we pretend
effects which will arise from it . . . (EHU
to . . . infer any cause or effect, without
4.6 / 27)


9780826443595_Ch03_Final_txt_print.indd 90 2/9/2012 9:38:55 PM

9780826443595_Ch03_Final_txt_print.indd 91

(1) Only the relation of cause and

effect can take us beyond the
evidence of our memory and senses

(2) All factual inferences to the

unobserved are founded on the (7) All reasonings from experience
relation of cause and effect are founded on the Uniformity
Principle (UP)

(4) Any effect is quite distinct from (6) All factual inferences to the
(8) All factual inferences to the
its cause, and many different effects unobserved are founded on
are equally conceivable experience unobserved are founded on UP No factual inference to the
unobserved is founded on reason

(14) A change in the course of (13) Two kinds of argument are

nature can be distinctly available (for proving UP):
(5) Causal relations cannot be known
conceived, and hence is possible demonstrative and factual
a priori, but can only be discovered
by experience



(15) Future uniformity cannot be (16) If there is a good argument

inferred demonstratively from for UP, it must be a factual
(3) Sensory perception of any object
past uniformity inference
does not reveal either its causes or its
effects, and there is no known
connexion between the sensible
qualities and its ‘secret powers’

(9) UP is not founded on anything (17) Any factual inference to UP (18) There is no good argument
that we learn through the senses would be circular of any kind for UP
about objects’ ‘secret powers’
(10) UP can be founded on reason
only if it is founded on experience (19) UP is not founded on reason
(of uniformity)
(12) UP can be founded on
reason only if it is founded on
argument (via some medium
enabling it to be inferred from
past experience of uniformity)

(11) The inference from past

uniformity to future uniformity is not
2/9/2012 9:38:55 PM

the assistance of observation and experi- to considering experiential arguments

ence. (EHU 4.11 / 30) for it:] As to past Experience, it can be
. . . our knowledge of that relation [of allowed to give direct and certain infor-
cause and effect] is derived entirely from mation of those precise objects only,
experience . . . (EHU 4.19 / 35) and that precise period of time, which
6. . . . nor can our reason, unassisted by fell under its cognizance: but why this
experience, ever draw any inference experience should be extended to future
concerning real existence and matter of times, and to other objects, which for
fact . . . (EHU 4.6 / 27) aught we know, may be only in appear-
In vain, therefore, should we pretend ance similar; this is the main question
to determine any single event . . . with- on which I would insist. (EHU 4.16 /
out the assistance of observation and 33–4)
experience. (EHU 4.11 / 30) 11. The connexion between these proposi-
7. . . . we always presume, when we see tions [I have found that such an object
like sensible qualities, that they have like has always been attended with such an
secret powers, and expect, that effects, effect and I foresee, that other objects,
similar to those which we have experi- which are, in appearance, similar, will
enced, will follow from them . . . (EHU be attended with similar effects] is not
4.16 / 33) intuitive. (EHU 4.16 / 34)
We have said, that . . . all our experi- 12. There is required a medium, which may
mental conclusions proceed upon the enable the mind to draw such an infer-
supposition, that the future will be con- ence, if indeed it be drawn by reasoning
formable to the past . . . (EHU 4.19 / 35) and argument. (EHU 4.16 / 34)
. . . all inferences from experience sup- 13. All reasonings may be divided into two
pose, as their foundation, that the future kinds, namely demonstrative reasoning,
will resemble the past, and that similar or that concerning relations of ideas,
powers will be conjoined with similar and moral reasoning, or that concern-
sensible qualities . . . (EHU 4.21 / 37) ing matter of fact and existence. (EHU
8. [This proposition is implicit in the infer- 4.18 / 35)
ential sequence:] We have said, that all 14. . . . it implies no contradiction, that the
arguments concerning existence are course of nature may change . . . May I
founded on the relation of cause and not clearly and distinctly conceive [such
effect; that our knowledge of that rela- a thing]? (EHU 4.18 / 35)
tion is derived entirely from experience; 15. That there are no demonstrative argu-
and that all our experimental conclu- ments in the case, seems evident . . . (EHU
sions proceed upon the supposition, 4.18 / 35)
that the future will be conformable to . . . whatever is intelligible, and can be
the past. (EHU 4.19 / 35) distinctly conceived, implies no contra-
9. . . . the mind is not led to form such a diction, and can never be proved false by
conclusion concerning [sensible quali- any demonstrative argument or abstract
ties and secret powers’] constant and reasoning a priori . . . (EHU 4.18 / 35)
regular conjunction, by any thing which 16. If we be, therefore, engaged by arguments
it knows of their nature . . . (EHU 4.16 to put trust in past experience, and make
/ 33) it the standard of our future judgment,
10. [This proposition is implicit in Hume’s these arguments must be probable only,
transition from considering ‘a priori’ or such as regard matter of fact and real
evidence for the Uniformity Principle existence . . . (EHU 4.19 / 35)


9780826443595_Ch03_Final_txt_print.indd 92 2/9/2012 9:38:56 PM


17. To endeavour, therefore, the proof [that 25–39, and is outlined in sections 2.1–4 below.
the future will be conformable to the In discussions of induction it is commonly
past] by probable arguments, or argu- referred to as ‘Hume’s famous argument’, a
ments regarding existence, must be evi- convenient shorthand that I shall adopt. Note
also that ‘induction’ is the modern term for the
dently going in a circle, and taking that
topic of his argument; he himself never uses the
for granted, which is the very point in word in this sense.
question. (EHU 4.19 / 35–6) 2
This is the summary of the Section 4 argument
18. . . . it may be requisite . . . to shew, that alluded to earlier. Note, however, that the pre-
none of [the branches of human knowl- vious clause brings in a point from the Section
edge] can afford such an argument . . . 7 discussion of the idea of necessary connexion,
(EHU 4.17 / 35) which does not feature in Section 4 itself.
. . . we have no argument to convince Hume does not reject the Causal Maxim, but
us, that objects, which have, in our experi- says that it ‘must . . . arise from observation and
ence, been frequently conjoined, will like- experience’ (THN / 82), hinting that he
will return to it later (though he never does). For
wise, in other instances, be conjoined in
detailed discussion, see Peter Millican, ‘Hume’s
the same manner . . . (EHU 12.22 / 159) Determinism’, Canadian Journal of Philosophy
19. . . . it is not reasoning which engages us 40 (2010), pp. 611–42, sects II, IV, VI.
to suppose the past resembling the future, 4
Section 1.3.7 will in due course move on to the
and to expect similar effects from causes, third component, ‘the nature and qualities’ of
which are, to appearance, similar . . . the belief-idea.
(EHU 4.23 / 39) Hume continues to mention the imagination’s
. . . nothing leads us to [expect con- power to mix and separate its ideas (e.g.
stant conjunctions to continue] but cus- Abs. 35 / 662, EHU 5.10 / 47–8), but the
tom or a certain instinct of our nature Separability Principle as such is never again
invoked as it had been, very significantly, in the
. . . (EHU 12.22 / 159)
Treatise (e.g. THN / 18–9, /
20. I say then, that, even after we have expe- 36–7, / 54–5), arguably sometimes with
rience of the operations of cause and absurd results (e.g. THN / 222, /
effect, our conclusions from that experi- 233, / 245–6, App. 12 / 634).
ence are not founded on reasoning, or 6
At THN / 77, Hume had stressed
any process of the understanding. (EHU that (single-case) contiguity and succession
4.15 / 32) are insufficient to characterize a cause and
. . . in all reasonings from experience, effect relationship, pointing out that ‘There
there is a step taken by the mind, which is a necessary connexion to be taken into
is not supported by any argument or pro- consideration’. Now at THN / 87, he
reminds us that ‘Contiguity and succession are
cess of the understanding. (EHU 5.2 / 41)
not sufficient to make us pronounce any two
All belief of matter of fact or real objects to be cause and effect’, and he expresses
existence [is due merely to] a species of satisfaction at having unexpectedly ‘discover’d
natural instincts, which no reasoning or a new relation . . . This relation is their con-
process of the thought and understand- stant conjunction.’ The link between the
ing is able, either to produce, or to pre- passages is evident both from the content and
vent. (EHU 5.8 / 46–7) the capitalization.
John Locke, An Essay concerning Human
Understanding, ed. Peter H. Nidditch, (Oxford:
NOTES Clarendon Press, 1975), IV.xv.1, IV.xvii.2.
Humean demonstration corresponds to what
The argument appears in Treatise 1.3.6 / is now called deductive reasoning, in the
86–94, Abstract 8–16 / 649–52, and Enquiry 4 / informal sense of an argument whose premises


9780826443595_Ch03_Final_txt_print.indd 93 2/9/2012 9:38:56 PM


conceptually guarantee the truth of the conclu- relation, which is either intuitively or demon-
sion. For more on this, see Peter Millican, stratively certain’. The Abstract and Enquiry
‘Humes Old and New: Four Fashionable make clear that the circularity is logical.
Falsehoods, and One Unfashionable Truth’, Before drawing this conclusion, Hume adds
Proceedings of the Aristotelian Society, supp. (what I have called) a ‘coda’ to his argument
vol. 81 (2007), pp. 163–99, sect. V. (THN–10 / 90–1), dismissing an
This result comes from Hume’s theory of attempt to get round it by appeal to objects’
relations, at THN–3 / 73–4 (for criti- powers. This attempt is refuted by the simple
cism, see Peter Millican, ‘Hume’s Fork, and observation that induction needs to be presup-
His Theory of Relations’, forthcoming in posed to enable us to draw an inference from
Philosophy and Phenomenological Research). the powers of past objects to the powers of
In brief, THN 1.1.5 / 13–5 enumerates what future objects. For discussion of this coda and
Hume takes to be the seven different kinds its implications, see Peter Millican, ‘Hume’s
of relation, which THN / 69–70 then Sceptical Doubts concerning Induction’, in
divides into two classes. The four relations ‘that Peter Millican (ed.), Reading Hume on Human
depend solely on ideas’ are the sources of strict Understanding (Oxford: Clarendon Press,
‘knowledge’, with resemblance, contrariety 2002), pp. 107–73, sects 9–9.2.
and degrees in quality amenable to intuition He also refers back to it in a footnote at THN
(THN / 70), and proportions in quan- / 163, feeding into his discussion of
tity or number the basis for demonstration. Of the idea of necessary connexion.
the three ‘inconstant’ relations, identity and For discussion of some of the nuances of termi-
relations of time and place are amenable to per- nology for referring to this kind of reasoning,
ception (THN / 73–4), leaving causation see Millican, ‘Hume’s Sceptical Doubts con-
as ‘the only one, that can be trac’d beyond our cerning Induction’, sect. 3.1, which distin-
senses, and informs us of existences and objects, guishes between probable inference, factual
which we do not see or feel’ (THN / inference, factual inference to the unobserved,
74). Hume thus identifies probable with causal and inductive inference. Hume generally takes
reasoning, and the rest of Book 1, Part 3, for granted that all of these coincide.
entitled ‘Of Knowledge and Probability’, is The argument from THN–10 / 90–1 is
accordingly devoted to ‘the idea of causation also very briefly summarized, in the last two
. . . tracing it up to its origin’ (THN / sentences of paragraph 15. For more on this
74–5). Strangely, the word ‘probability’ does not ‘coda’, see note 12 above.
appear at all in this Part before THN Notice that Hume seems entirely happy to
/ 89, except in the title of the Part itself and of take perception and memory for granted here,
Section 1.3.2: ‘Of Probability; and of the Idea of fitting with the strategy described in section 1
Cause and Effect’. above, of allowing default authority to our
Notice that causal inference is categorically faculties. Scepticism regarding the senses is
stated to be founded on that presumption – addressed at THN 1.4.2–4 / 187–231 and
there is no suggestion here of the conditional- EHU 12.6–16 / 151–5, but Hume’s ultimate
ity that we had at THN / 89: ‘If reason attitude to it remains far less clear than his
determin’d us, it wou’d proceed upon that position on induction.
principle . . .’ (emphasis added). Nor is there This notion of a proof plays a significant role
any such conditionality at THN / in Hume’s argument concerning miracles in
91–2, or in either the Abstract (Abs. 13–14 / Section 10 of the Enquiry.
651) or the Enquiry (EHU 4.19 / 35–6, 4.21 / This is taken from Millican, ‘Hume’s Sceptical
36–7, 5.2 / 41–2). Doubts concerning Induction’.
11 19
THN / 90 expresses the circularity in Thus there is no evidence here, as influentially
causal terms: ‘The same principle cannot be claimed by David Stove, Probability and
both the cause and effect of another’, appar- Hume’s Inductive Scepticism (Oxford: Oxford
ently in order to make a joke: ‘and this is, University Press, 1973), p. 50, that Hume’s
perhaps, the only proposition concerning that method of argument shows him to be a


9780826443595_Ch03_Final_txt_print.indd 94 2/9/2012 9:38:57 PM


‘deductivist’, presupposing that only deduc- the Uniformity Principle as implicit rather than
tively valid arguments are legitimate. A similar explicit, a principle we exhibit by our infer-
point, though less obvious, can be made about ential behaviour rather than one we always
the Abstract (‘The mind can always conceive consciously consider. Such an interpretation
any effect to follow from any cause, and nicely squares Hume’s repeated commitment
indeed any event to follow upon another’, to the Principle’s role in all inductive inference
Abs. 11 / 650) and the Treatise (‘When we pass (see note 10) with his clear recognition at THN
from a present impression to the idea of any / 103–4 that we characteristically
object, we might possibly . . . have substituted ‘draw inferences from past experience, without
any other idea in its room’, THN / 87). reflecting on . . . that principle’. See also note
This case of applied mathematics (cf. also 31 below.
THN / 413–4) shows that Hume This suggests that if the inference were intui-
is quite comfortable with demonstrative, tive, it would count as ‘reasoning and argu-
mathematical reasoning being applied to a pos- ment’ notwithstanding the lack of a ‘medium’.
teriori premises. For discussion of this point, Indeed, as we shall see later (section 3.1), in
see Millican, ‘Humes Old and New’, sect. V. Hume’s day the terms ‘reasoning’ and ‘argu-
Hume’s talk of ‘secret powers’ is new in the ment’ did not imply complex ratiocination.
Enquiry, and seems to reflect a more sophis- Hume is fond of elegant variation, frequently
ticated understanding of scientific reasoning using a variety of terms for the same concept.
than is evident in the Treatise and Abstract. ‘Moral reasoning’, ‘reasoning concerning
In the Treatise, science is generally treated matter of fact and real existence’, ‘probable
as involving predictions of discrete types of arguments’ and ‘arguments concerning exist-
event based on previous patterns of conjunc- ence’ are all ways of referring to what we are
tion or difference (as in the ‘rules by which to here calling factual reasoning. See note 14
judge of causes and effects’ of THN 1.3.15 / above.
173–176). The Enquiry, by contrast, evinces That he takes these to be equivalent was
an awareness (e.g. at EHU 4.13 / 31 and EHU made clear by EHU 4.2 / 25–6, where he first
7.29n17 / 77n) that science more typically explained the notion of a matter of fact.
deals with events having continuously vary- For an earlier occurrence of this last implicit
ing characteristics – such as the velocity of a inference, see note 22 above.
billiard ball – whose prediction involves the As in the Treatise (note 12 above) and Abstract
interplay of mathematically determined forces. (note 15 above), Hume rounds off the argu-
For more on this, see Peter Millican, ‘Against ment in the Enquiry with a coda (EHU 4.21
the New Hume’, in Rupert Read and Kenneth / 36–8) in which he refutes the attempt to
A. Richman (eds), The New Hume Debate: circumvent his argument by appeal to objects’
Revised Edition (London: Routledge, 2007), powers. He also adds a parting shot at EHU
pp. 211–52, at pp. 232–3. 4.23 / 39 which emphasises the unlikelihood
Hume obviously means us to infer accord- that peasants, infants or ‘brute beasts’ should
ingly – though he does not explicitly state – form their inductive expectations on the basis
that [8] all factual reasoning, since it has to of ‘any process of argument or ratiocination’.
be founded on experience, presupposes such a Though the point is well made, however, its
resemblance (i.e. the Uniformity Principle). See philosophical significance is less clear, because
also note 27 below. those who take induction to be rationally
In EHU 4.16 / 33–4 itself, Hume oscillates founded need not be committed to supposing
between reference to the activity of inference that animals (etc.) function purely ration-
from observed to unobserved, and to the ally – see Millican, ‘Hume’s Sceptical Doubts
presupposition of resemblance on which such concerning Induction’, sect. 9.3.
inference is based. Indeed it seems that he takes Recall that [12] is the claim that ‘There is
the foundation of the inference to be the same required a medium, which may enable the
as the foundation of the presupposition that it mind to draw such an inference, if indeed
manifests. This supports an interpretation of it be drawn by reasoning and argument.’


9780826443595_Ch03_Final_txt_print.indd 95 2/9/2012 9:38:57 PM


(EHU 4.16 / 34) – that is, because the the existence of body, the metaphysics of
Uniformity Principle cannot be established dir- causation, and the self-annihilation of reason),
ectly through sensory perception or intuition, which the simple assumption of uniformity
if it is to be established by reason at all, then never does. The Enquiry’s response to the
this must be on the basis of some stepwise Pyrrhonian sceptic, starting from a rejection
argument or ratiocination. of extreme antecedent scepticism, would not
The other implicit final stages are also stated be nearly as effective against varieties of con-
explicitly elsewhere: [18] ‘we have no argu- sequent scepticism that bring to light genuine
ment to convince us, that objects, which have, contradictions – rather than simply lack of
in our experience, been frequently conjoined, ultimate grounding – in our faculties, and this
will likewise, in other instances, be conjoined might explain why Hume very much down-
in the same manner’ (EHU 12.22 / 159); [19] plays these more problematic topics in the
‘it is not reasoning which engages us to sup- Enquiry. His attitude to them seems to be that
pose the past resembling the future, and to they are best left alone: for example, meta-
expect similar effects from causes, which are, physical enquiries into the nature of matter
to appearance, similar’ (EHU 4.23 / 39). are likely to lead to contradiction or unin-
See note 23 above for the nature of this telligibility (EHU 12.14–15 / 153–5) unless,
presupposition, which need not be conscious, perhaps, we fall back on a notion of matter so
but is implicitly manifested by the making of empty as to be unexceptionable (EHU 12.16
the inference. So UP need not take any very / 155). Hume’s final recommendation is for a
explicit or determinate form (contrary to the mitigated scepticism that inspires a suitable
impression given by THN / 88–9), and degree of ‘doubt, and caution, and modesty’
is best understood as something like a general (EHU 12.24 / 162), and which also focuses our
principle of evidential relevance between enquiries on ‘such subjects as are best adapted
observed and unobserved, more in line with to the narrow capacity of human understand-
the expression of the Enquiry: ‘we . . . put trust ing’ (EHU 12.25 / 162), notably those where
in past experience, and make it the standard we are able to progress either through math-
of our future judgment’ (EHU 4.19 / 35); we ematical demonstration (EHU 12.27 / 163) or
take ‘the past [as a] rule for the future’ (EHU induction from experience (EHU 12.28–31 /
4.21 / 38). This seems right: in taking such an 163–6).
inference to be better informed than an a priori David Owen, Hume’s Reason (Oxford: Oxford
inference, we are ipso facto presuming that University Press, 1999), p. 136.
what happened in the past provides evidence Ibid., p. 132.
that is positively relevant to what will happen Barry Stroud, Hume (London: Routledge &
in the future. For more on the Uniformity Kegan Paul, 1977), p. 54.
Principle and its presupposition, see Millican, Antony Flew, Hume’s Philosophy of Belief
‘Hume’s Sceptical Doubts concerning (London: Routledge & Kegan Paul, 1961).
Induction’, sect. 3.2 and especially sect. 10.2. Stove, Probability and Hume’s Inductive
Moreover this sequence of argument seems to Scepticism.
be entirely deliberate, because it occurs very Stroud himself (Hume, pp. 56–7) reacted
explicitly twice, first within the main argument against this, suggesting that what he saw as
at EHU 4.16 / 32–4, and then again in the Hume’s extreme scepticism could more plaus-
coda at EHU 4.21 / 37. ibly be attributed to a ‘potentially regressive
In the conclusion of Book 1 of the Treatise, aspect of the notion of reason or justification’
Hume’s attempt to meet the sceptical challenge whereby evidence E can count as a reason
says very little about the issue of induction, for believing P only if one has some reason R
except as part of a general concern regarding for taking E as a reason. If we then ask about
the role of ‘the imagination, or the vivacity the basis for R in turn, and continue in this
of our ideas’ (THN / 265). There the way, we get a regress which can apparently
more pressing problems are those that threaten be terminated only by ‘something we could
inevitable error and contradiction (notably not fail to be reasonable in believing’ (ibid.,


9780826443595_Ch03_Final_txt_print.indd 96 2/9/2012 9:38:58 PM


p. 62), such as an immediate experience or self- deductivist must consider any merely probable
evident truth. Hume’s invoking of UP within argument as evidentially worthless from the
his argument is indeed somewhat in this spirit, start. If Hume were a deductivist, indeed,
but when considering UP’s own foundation, he then he could dismiss induction in a single
seems content to stop with (fallible) sensa- step with his Conceivability Principle. For
tion or memory, not only with the certainty of more detail on all this, see Millican, ‘Hume’s
intuition or demonstration, while the appeal Argument concerning Induction’, pp. 123–4,
to factual inference generates a circle rather 136; ‘Hume’s Sceptical Doubts concerning
than an infinite regress. Nevertheless Stroud’s Induction’, pp. 155–6; Garrett, Cognition and
account is illuminating, in stressing the seduc- Commitment, pp. 86–8).
tive assumption that justification is required at See in particular the passages quoted near the
each step if scepticism is to be resisted. Hume’s end of the first paragraph of section 3 above.
strategy outlined in section 1 above rejects this These and others are cited in this connexion
by shifting the onus onto the sceptic. by Millican, ‘Hume’s Argument concerning
Flew, Hume’s Philosophy of Belief, p. 171. Induction’, pp. 127, 136; ‘Hume’s Sceptical
For some other passages in a similar spirit, see Doubts concerning Induction’, pp. 161–2;
note 125 below. Garrett, Cognition and Commitment,
Tom Beauchamp and Thomas Mappes, ‘Is pp. 85–6.
Hume Really a Sceptic about Induction?’, Locke, Essay, IV.xvii.2.
American Philosophical Quarterly 12 (1975), e.g. Locke, Essay, IV.xv.2.
pp. 119–29. Garrett, Cognition and Commitment, p. 92.
43 50
Barbara Winters, ‘Hume on Reason’, Hume Locke’s usage is somewhat variable, though I
Studies 5 (1979), pp. 20–35 was perhaps most consider perception to be more fundamental to
influential in promoting the idea that Hume’s Lockean reason than inference (see my ‘Hume’s
notion of reason is ambiguous in this way. Argument concerning Induction’, p. 137, or for
Tom Beauchamp and Alexander Rosenberg, more detail, ‘Hume’s Sceptical Doubts concern-
Hume and the Problem of Causation ing Induction’, sect. 2). Note that both of these
(Oxford: Oxford University Press, 1981); are distinct from the intermediate idea charac-
N. Scott Arnold, ‘Hume’s Skepticism about teristic which Owen considers fundamental to
Inductive Inference’, Journal of the History Lockean reason.
of Philosophy 21 (1983), pp. 31–55; Janet Garrett, Cognition and Commitment, pp.
Broughton, ‘Hume’s Skepticism about Causal 91–2.
Inferences’, Pacific Philosophical Quarterly The modified interpretation first appeared in
64 (1983), pp. 3–18; and Annette C. Baier, A Don Garrett, ‘Ideas, Reason and Skepticism:
Progress of Sentiments: Reflections on Hume’s Replies to My Critics’, Hume Studies 24
Treatise (Cambridge MA: Harvard University (1998), pp. 171–94, but his 2002 piece ‘The
Press, 1991) all gave slightly different anti- Meaning of Hume’s Conclusion concerning
deductivist readings, some of the nuances of “Inductive” Inferences’ (in Peter Millican,
which are discussed by Don Garrett, Cognition Reading Hume on Human Understanding,
and Commitment in Hume’s Philosophy pp. 332–4) was based directly on the two
(Oxford: Oxford University Press, 1997), relevant sections of his 1997 book (Cognition
pp. 83–91. and Commitment), reworded accordingly.
44 53
Peter Millican, ‘Hume’s Argument concerning Garrett, ‘The Meaning of Hume’s Conclusion’,
Induction: Structure and Interpretation’, in p. 333.
Stanley Tweyman (ed.), David Hume: Critical Or, presumably, A to C. Garrett’s suggestion
Assessments (London: Routledge, 1995), (‘Ideas, Reason and Skepticism’, pp. 182–3)
vol. 2, pp. 91–144; Garrett, Cognition and is that in attempting to infer from A to D, we
Commitment. first observe that A implies B intuitively (i.e.
Most obviously, the famous argument treats self-evidently), leaving a gap between B and
probable argument as a potential founda- D. We then set out to fill that gap, by noticing
tion for the Uniformity Principle, whereas a that B implies C, and C implies D. We put


9780826443595_Ch03_Final_txt_print.indd 97 2/9/2012 9:38:58 PM


these last two implications together, deducing See for example Owen, Hume’s Reason,
that B implies D (this is Garrett’s intermedi- pp. 9–10, 120–2, 127–30, 141, 148. The books
ate inference). Now from A implies B and B of both Owen and Garrett present only the
implies D, we can deduce that A implies D. Treatise version of the argument, and indeed
See, for example, Peter Millican, ‘Hume on Owen’s analysis hardly mentions the Abstract
Reason and Induction: Epistemology or or Enquiry. Garrett takes relevant quotations
Cognitive Science?’, Hume Studies 24 (1998), from the later works, but states without analysis
pp. 141–60, at pp. 145–7, and David Owen, ‘A that ‘the structure and language of the other ver-
Reply to My Critics’, Hume Studies 26 (2000), sions of the argument are parallel’ to that in the
pp. 323–37, at pp. 329–30. Treatise (Cognition and Commitment, p. 82).
56 64
Samuel Johnson, A Dictionary of the English Descartes’s Meditations, for example, presents
Language (London, 1756). the Ontological Argument for the existence
‘To deduce’ is defined in three clauses: ‘1. of a perfect God, and then appeals to God’s
To draw in a regular connected series. 2. non-deceptive nature to vindicate various
To form a regular chain of consequential factual beliefs about the unobserved, all appar-
propositions. 3. To lay down in regular order.’ ently without any essential reference to the
‘Ratiocination’ is defined in just one clause: Uniformity Principle.
‘The act of reasoning; the act of deducing Or – if we take the Enquiry version – in any
consequences from premises.’ deliverance of the senses or intuition.
58 66
Also, of course, Hume’s own theory of induct- This objection goes back to my ‘Hume on
ive inference implies that it typically does not Reason and Induction’, sect. VII (1998), and is
proceed in a stepwise manner, but essentially also discussed by Garrett, ‘Ideas, Reason and
reduces to conception (see THN / Skepticism’, p. 187; Millican ‘Hume’s Sceptical
96–7n); yet he never hints that terms such as Doubts concerning Induction’, pp. 157–8;
‘argument’, ‘inference’, ‘proof’ or ‘reasoning’ Louis Loeb, ‘Psychology, Epistemology,
are thereby rendered inappropriate to these and Skepticism in Hume’s Argument about
transitions of thought. So it is hard to see how Induction’, Synthese 152 (2006), pp. 321–38,
he could consistently refuse to apply them – on at pp. 328–9; and Abraham Sesshu Roth,
grounds of immediacy – to ‘intuitive inference’. ‘Causation’, in Saul Traiger (ed.), The
Such an emphasis comes later, with the positive Blackwell Guide to Hume’s Treatise (Oxford:
account in terms of instinctive custom (e.g. at Blackwell, 2006), pp. 95–113, at pp. 108–11.
THN / 103–4 and EHU 5.8 / 46–7). The case of faulty factual arguments (e.g. in
At least, this looks like a plausible implication, Hume’s coda at THN / 91 and EHU
just as one invalid step within a mathemat- 4.21 / 36–8) is less clear, because if they pre-
ical proof is enough to render the entire proof suppose UP, then the famous argument – as
invalid. But as we shall see later (section 4), interpreted by Garrett and Owen – can still
things are not quite so straightforward here. get a purchase on them. For critical discus-
For another way of sharpening this sort of sion, see Loeb, ‘Psychology, Epistemology, and
objection, see Millican, ‘Humes Old and New, Skepticism’ (p. 329), who goes on to suggest
sect. VI, which expands on Millican, ‘Sceptical his own explanation of why Hume fails to
Doubts concerning Induction’, pp. 158–60. consider faulty arguments here: ‘The answer
There I focus on the very last step of Hume’s must be that Hume imposes an epistemic
argument, whereby he concludes that because constraint on any causal explanation of induct-
factual inference is founded on UP, and UP is ive inference: the explanation of our making
not founded on reason, it follows that factual inductive inferences must be compatible with
inference is not founded on reason. This step their being justified’ (p. 330). Helen Beebee,
looks very dubious if ‘reason’ here is supposed Hume on Causation (London: Routledge,
to mean stepwise ratiocination (or, indeed, 2006), pp. 55–6, takes a similar line, and both
higher-level argument). are discussed in my ‘Humes Old and New’,
See note 10 above for equivalent passages in pp. 186–8. In brief, I find their approach text-
the Abstract and Enquiry versions. ually unsupported and also in tension with


9780826443595_Ch03_Final_txt_print.indd 98 2/9/2012 9:39:00 PM


the sceptical tone of Hume’s famous argument Garrett, ‘Ideas, Reason and Skepticism’, p. 184.
and of his later references to it. A far simpler Ibid. It seems to be a logical rather than psycho-
solution is to see ‘reason’ as referring to our logical point that some such step must be pre-
cognitive faculty – see section 3.2 below. sent in every inductive inference, given that – as
Richard Price, A Review of the Principal Garrett acknowledges – the ‘supposition of UP’
Questions and Difficulties in Morals (London: that it exhibits can be entirely unconscious.
1758), p. 34. Locke starts his chapter ‘Of Reason’ (Essay,
Ibid., p. 40n. IV.xvii) with the remark that ‘The Word
Owen also adds a note at this point: ‘The Reason in the English Language has different
qualification, “prior to our engaging in Significations: sometimes it is taken for true,
probable reasoning”, is important, because and clear Principles: Sometimes for clear, and
Hume thinks that once we are engaged in fair deductions from those Principles: and
the practice of probable reasoning, we come sometimes for the Cause, and particularly the
to believe the uniformity principle and use it final Cause.’ He then goes on to say that his
in probable reasoning. . . . This requires an chapter concerns yet another signification,
account of how we first engage in probable for that ‘Faculty in Man . . . whereby Man is
reasoning, before the principle is available to supposed to be distinguished from Beasts, and
us.’ (Hume’s Reason, p. 128n30) wherein it is evident he much surpasses them’.
Owen, Hume’s Reason, pp. 128–30. Earlier, at Essay II.xxi.17–20, Locke ridi-
Garrett, Cognition and Commitment, p. 94. cules the language of ‘faculties’ as a source of
Owen, Hume’s Reason, pp. 132–4. philosophical error. For more on his view, see
Garrett, ‘Ideas, Reason and Skepticism’, Millican, ‘Hume’s Sceptical Doubts concerning
pp. 180–1. Induction’, sect. 2, and cf. note 50 above.
75 85
Unlike Garrett and Owen, however, I do not As acknowledged in Hume’s letter of
take the principle in question, ‘that like objects, 10 January 1743 (LDH 1.45, 19).
plac’d in like circumstances, will always Francis Hutcheson, Synopsis of Metaphysics
produce like effects’, to be identical to the (1744), trans. Michael Silverthorne, in Francis
Uniformity Principle. The former concerns the Hutcheson, Logic, Metaphysics, and the
consistency of events within our experience, Natural Sociability of Mankind (Indianapolis:
whereas UP concerns the evidential relevance Liberty Fund, 2006)
of observed to unobserved. Without UP, Francis Hutcheson, A Short Introduction to
experienced consistency (or, indeed, any other Moral Philosophy (1747), ed. Luigi Turco
experienced pattern) could not be extrapolated (Indianapolis: Liberty Fund, 2007), p. 25.
from past to future. Francis Hutcheson, An Essay on the Nature
Owen, Hume’s Reason, pp. 131, 170–1. and Conduct of the Affections. With
Garrett, ‘Ideas, Reason and Skepticism’, p. 184 Illustrations on the Moral Sense, 3rd edn
Millican, ‘Hume on Reason and Induction’, (London: 1742), pp. 30–1.
p. 153. Ibid., pp. 219–20.
79 90
This is the main respect in which my own Note that the quoted paragraph also treats
views have changed over time, largely in ‘the Intellect’ as just another elegant variation
response to Don Garrett’s criticisms. Most on ‘Reason’ and ‘the Understanding’. Hume
other aspects of my previous interpretation does the same, albeit more rarely (DNR 3.153,
remain in place; for example it will become 3.156), though he quite often refers in a simi-
clear in section 4 below that a perceptual lar spirit to the ‘intellectual faculties’ (THN
notion of reason makes a highly plausible / 138, / 437; EHU 5.5n8 /
Humean target, even if we do not suppose that 43–4n, 9.6 / 108; EPM 1.9 / 173, EPM App.
he was employing such a notion himself within 1.11 / 290, 13 / 291, 18 / 293, 3.9 / 307; ‘Of
the famous argument. the Standard of Taste’, E 240–1). Garrett talks
Garrett, Cognition and Commitment, p. 92. of Hume as giving ‘an argument against the
Garrett, ‘The Meaning of Hume’s Conclusion’, intellect’ (Cognition and Commitment, p. 20),
p. 333. but this is misleading unless ‘the intellect’ here


9780826443595_Ch03_Final_txt_print.indd 99 2/9/2012 9:39:00 PM


is understood to mean ‘the intellect conceived Reid, Intellectual Powers, I.vii, pp. 67–8.
of as a faculty of non-sensory ideas’ (a con- Note, for example, the general division
ception that Garrett traces through Descartes, within the Treatise between Book 1 ‘Of the
Spinoza and Leibniz, but is not shared by Understanding’ and Book 2 ‘Of the Passions’
Locke, Berkeley or Hume). (including Part 3 ‘Of the will and direct
See, for example, Anthony Ashley Cooper, passions’), and also the footnote at THN
Lord Shaftesbury, Characteristicks of Men, / 96, where Hume criticizes the
Manners, Opinions, Times, 3 vols, 3rd edn [the ‘universally receiv’d’ threefold ‘division of the
edition purchased by Hume in 1726] (London, acts of understanding’ which Reid describes.
1723), vol. 2, II.ii, p. 118; Joseph Butler, The Hutcheson, A Short Introduction to Moral
Analogy of Religion, Natural and Revealed, Philosophy, pp. 25–6.
to the Constitution and Course of Nature, Price deprecates ‘the division which has
2nd edn, corrected (London, 1736),, been made by some writers, of all the
p. 174; Price, Review, I ii, p. 23. powers of the soul into understanding and
See, for example, Henry Home, Lord Kames, will; the former comprehending under it,
Essays on the Principles of Morality and all the powers of external and internal
Natural Religion, 2nd edn (London, 1758), sensation, as well as those of judging and
essay VII, p. 268n, and James Beattie, An reasoning’. By contrast, he says, ‘I all along
Essay on the Nature and Immutability speak of the understanding, in the most
of Truth; in Opposition to Sophistry and confined and proper sense of it . . . and
Scepticism (Edinburgh, 1770), I.i, pp. 37–8. distinguished from the powers of sensation’
James Oswald, An Appeal to Common Sense (Review, I.ii, p. 20n). Note, however, that
in Behalf of Religion (Edinburgh, 1766), Price implicitly equates the understanding
I.ii.1, p. 80 and Thomas Reid, Essays on the with reason (ibid., p. 23) thus using ‘reason’
Intellectual Powers of Man (Edinburgh, 1785), in a broader sense than those such as Kames
VII.i, p. 671 are likewise keen to insist on a (cf. note 92 above) who exclude intuition
narrow use of ‘reasoning’, distinguished from from its scope.
‘judging’, though they allow both operations Hutcheson, Synopsis of Metaphysics, p. 112.
to be subsumed under ‘reason’. The original Latin of the final clause is ‘Ad
For other relevant passages from the Treatise, Intellectum, referentur Sensus’.
see, for example, THN / 88–9, Sect. 6 ends with a short paragraph on / 149–50, / 180, ‘Natural associations of ideas’, which
/ 186–7, / 193, / 267–8, Hutcheson sees as playing an important role–6 / 413–6,–18 / 462–3, in both imagination and memory; sect. 7
and / 468–9. For passages from the briefly discusses what is pleasing or distress-
Abstract, see Abs. 11 / 650–1, 27 / 657, and ing to the senses, and our consequent judge-
from the Enquiry, EHU 4.0–1 / 25, 5.5 / 43, ments (of good and evil) and passions; then
5.22 / 55, 7.28 / 76 and 9.0–1 / 104. Further sect. 8 discusses habit, and sect. 9 relative
examples may be found in Hume’s other works. ideas and judgements.
94 101
Together with Hume himself (EHU 1.14 / Hutcheson, Illustrations on the Moral Sense,
13–14, 8.22 / 93), see, for example, René I, p. 215.
Descartes, Fourth Meditation, in The Price, Review, pp. 19–20.
Philosophical Writings of Descartes, trans. For example Locke: ‘the understanding
John Cottingham, Robert Stoothoff and . . . is the most elevated Faculty of the Soul
Dugald Murdoch (Cambridge: Cambridge . . . Its searches after Truth, are a sort of . . .
University Press, 1984), vol. 2, pp. 39–40; Hunting’ (Essay, Epistle to the Reader, para-
Locke, Essay,; George Berkeley, graph 1); David Hartley: ‘The Understanding
A Treatise concerning the Principles of is that Faculty, by which we . . . pursue Truth,
Human Knowledge, ed. Jonathan Dancy and assent to, or dissent from, Propositions.’
(Oxford: Oxford University Press, 1998), (Observations on Man, His Frame, His Duty,
I.27. References to Hume’s contemporaries and His Expectations, Bath and London,
Hutcheson, Price and Reid will follow. 1749, vol. 1, Introduction, p. iii).


9780826443595_Ch03_Final_txt_print.indd 100 2/9/2012 9:39:01 PM


Other passages that identify reason with the each (at THN / 417–8 and /
discovery of truth, though usually less expli- 339 respectively). Judgemental ‘taste’ is called
citly, are at THN / 414,–6 a faculty in ‘Of the Standard of Taste’, E
/ 415–6, / 417, / 456–7, 240–1, and spoken of as having ‘a product- / 464n,–7 / 467–70, ive faculty’ in a famous passage at EPM App. / 496, THN App. 1 / 623; EPM 1.7 1.21 / 294.
/ 172, EPM App. 1.6 / 287. All of these occur The model of a canvas is obviously most
in a context where Hume is contrasting appropriate to visual ideas, which indeed
reason (or the understanding) with conative seem to dominate Hume’s thought, although
rather than cognitive notions, thus corrob- ideas may correspond to any of the senses –
orating its identification as the overarching including internal ‘reflection’ – and only the
cognitive faculty. ideas of sight and touch will be spatially
This note (which can be found as a Textual arranged (not necessarily within a single
Variant on p. 177 of my edition of Hume’s space). Note that he takes all of our ideas to
Enquiry concerning Human Understanding, be sense-copied; hence as Garrett observes (cf.
Oxford World’s Classics, 2007) is of particu- note 90 above), Hume denies any separate
lar interest because it credits Hutcheson with faculty that can take a ‘pure and intellectual
distinguishing between ‘the Understanding’ view’ of ‘refin’d and spiritual’ ideas, unsullied
and ‘That Faculty . . . by which we perceive by sensory input (THN / 72).
Vice and Virtue’, although Hutcheson himself At least, this seems to be what Hume is saying
considered the moral sense to be one of the in THN / 265. At THN /
‘reflexive or subsequent sensations’, thus fall- 107–8, he appears instead to take the force
ing within the domain of the understanding. and vivacity of the ‘impressions or ideas of
Price, Review, p. 12n, mentions Hume’s note the memory’ – like that of ‘an immediate
in the course of criticizing Hutcheson. impression’ – as itself constituting assent,
Notice that Hume is implying that if we thus providing a basis for explaining the
could observe a conjunction of cause and assent that derives from causal inference.
effect, then we could form such a ‘conclu- The relationship between memory and the
sion . . . concerning . . . existence’ and ‘satisfy imagination remains somewhat obscure,
our reason in this particular’. So as in the though Hume’s talk of ‘impressions of
famous argument concerning morals (cf. the the memory’ (‘somewhat intermediate
quotation from THN / 459 near betwixt an impression and an idea’, THN
the beginning of section 3.1 above), Hume is / 8) suggests that the memory is
clearly here treating causal, factual inference furnishing ideas that are sufficiently firm
as an operation of reason. and vivid – sufficiently impression-like – to
Such nuances can apply with many words that establish copy-ideas in the imagination:
are associated with some achievement. For ‘The impressions of the memory never
example, a cure that does not work is strictly a change in any considerable degree; and
contradiction in terms, but it is fairly natural to each impression draws along with it a
say, in appropriate circumstances, ‘that cure is precise idea, which takes its place in the
useless and ought to be banned’. imagination, as something solid and real,
An analogy here would be to an accounting certain and invariable.’ (THN / 110).
error within a company, which on a broad If this is right, then all the ideas that are
interpretation could refer to any error in the actually involved in thinking lie within the
accounts (including faulty data from external imagination, and the role of the senses and
sources), but on a narrower interpretation memory is to supply the ‘impressions’ from
would mean an error due to the accountants which those ideas derive.
themselves. See THN / 97, / 107–8,
On the conative side, Hume hardly ever / 128, / 155–6, /
speaks of ‘faculties’, explicitly referring to ‘the 169–70; and for references from the Abstract
will’ and ‘the passions’ as faculties just once and Enquiry see note 129 below.


9780826443595_Ch03_Final_txt_print.indd 101 2/9/2012 9:39:02 PM


I call it a ‘virtual’ stage to reflect Hume’s inference, ‘the imagination in the narrow
comment at THN / 253 that ‘The sense is performing the customary transition’
comparison of the theatre must not mislead (p. 22), thus denying that custom is reassigned
us. They are the successive perceptions only, to reason along with ‘our demonstrative
that constitute the mind’. and probable reasonings’. Hence he sees the
THN / 198 gives a case intermedi- distinction alluded to at THN /
ate between passivity and activity, in which 117–18 as quite distinct from that drawn at
the imagination, having been ‘set into any THN / 225, a position I find rather
train of thinking, is apt to continue even implausible, given that their stated rationale
when its object fails it, and, like a galley is so similar, namely, to distinguish within
put in motion by the oars, carries on its the imagination between the principles that
course without any new impulse’. Note that ‘are the foundation of all our thoughts and
the listed references involving fictions are actions’ and those that give rise to ‘whimsies
confined to those that involve characteristic and prejudices’.
Humean fictions of philosophical interest, Even in the Treatise, Hume never says in so
rather than arbitrary combinations of ideas many words that custom is a process of the
(i.e. ‘mere fictions of the imagination’ as at imagination, though THN / 88–9 and
THN–5 / 85, / 97, / / 97 strongly suggest this.
108, / 219; EHU 5.12–13 / 49–50, In the Enquiry, unlike the Treatise (e.g. THN
6.3 / 57). / 94, / 100–1), Hume is care-
Recall from section 3.2 that the replaced ful to distinguish between the associational
footnote at THN / 371n said ‘the relation of causation (discussed at EHU
understanding’ here instead of ‘reason’. 5.18–19 / 53) and custom (EHU 5.20 / 53–4).
116 124
It is striking, for example, that these are the We would not usually describe the climber as
only two occurrences in Hume’s writings ‘supported only by’ a rock to which the rope
of the phrase ‘offspring of the imagina- is attached unless the rock was considered
tion’. Presumably he was forced to place the potentially less secure than the rope (e.g.
footnote at the end of the section to minimize suppose the attachment is to a spur of rock
type resetting. that is in imminent danger of cracking –
The understanding is also identified with we might well then say that the climber is
the imagination at THN / 104 and ‘supported only by the spur’).
125 / 440. Together with those quoted earlier, relevant
Don Garrett, Cognition and Commitment, passages include: ‘. . . our reason . . . or, more
p. 92, was, I believe, the first to note this properly speaking, . . . those conclusions
possibility, which is crucial if Hume’s use of we form from cause and effect . . .’ (THN
‘reason’ within his argument is to be under- / 231); ‘. . . these emotions extend
stood as sincere rather than a target. But of themselves to the causes and effects of that
course Garrett’s interpretation of ‘determin’d object, as they are pointed out to us by rea-
by reason’ is very different from my own. son and experience . . .’ ( / 414); ‘. . .
See, for example, EHU 1.13–14 / 13–14 for the operations of human understanding . . .
the equation of ‘faculties’ with ‘powers’, and [include] the inferring of matter of fact . . .’
also THN Intro. 4 / xv, / 123. The ( / 463).
same equation is repeatedly found in Locke, Both Barry Stroud (Hume, pp. 60–2) and
e.g. Essay, II xxi 15, 17, 20. John Kenyon (‘Doubts about the Concept
Locke, Essay, II.xxi.20. of Reason’, Proceedings of the Aristotelian
In a recent debate (Don Garrett and Peter Society, supp. vol. 59, 1985, pp. 249–67,
Millican, Reason, Induction and Causation in at pp. 255–7) attribute this to Hume in the
Hume’s Philosophy, IASH Occasional Paper context of his argument concerning induction,
19, Edinburgh: Institute for Advanced Studies but neither justifies the attribution, and there
in the Humanities, University of Edinburgh, is little evidence of it in Hume’s text (cf. note
2011), Don Garrett argues that in inductive 39 above).


9780826443595_Ch03_Final_txt_print.indd 102 2/9/2012 9:39:02 PM


This assumes the internalist perspective which any process of the understanding, signifi-
dominated the early modern period. A mod- cantly more in the Enquiry – EHU 4.4 / 26,
ern Humean might well take an externalist 4.14 / 32 (twice), 4.15 / 32, 4.21 / 37, 9.5 /
approach, but given Hume’s explicit response 106 and 12.29 / 164) – than he does in the
to the sceptic in Enquiry 12 (as described Treatise – THN / 88 and /
in section 1 above), I am not persuaded by 89–90 (twice).
Louis Loeb’s claim that ‘In light of the massive Other aspects of the logic of Hume’s foun-
evidence that Hume is not a skeptic about dational relation are explored in Millican,
induction, he must reject [the] internalist way ‘Hume’s Sceptical Doubts concerning
of thinking.’ (‘Psychology, Epistemology, and Induction’, sect. 10.1.
Skepticism’, p. 333). Kenneth Winkler empha- There is a thematic parallel here with Hume’s
sizes how Enquiry 12 supports a more sceptical account of causation, which is also commonly
reading of Enquiry 4: see ‘Hume’s Inductive thought of as sceptical, but in fact provides
Skepticism’, in Margaret Atherton (ed.), him with a positive basis for applying causal
The Empiricists (Lanham, MD: Rowman & explanation to the human world. For more
Littlefield, 1999), pp. 183–212, at pp. 193–200. on this, see Peter Millican, ‘Hume, Causal
At THN / 117, / 147, Realism, and Causal Science’, Mind 118 / 165, / 225, and / (2009), pp. 647–712.
265. For numerous discussions on the topics of
At Abs. 15 / 652, 16 / 652, 21 / 654 (twice), reason and induction, I am extremely grateful
25 / 656; EHU 5.4–5 / 42–3, 5.5 / 43 (twice), to Louis Loeb, David Owen and especially
5.8 / 46, 5.11 / 48, 5.20 / 54, 5.21 / 54–5, Don Garrett, as well as many other members
6.4 / 58, 7.29 / 76–7, 8.5 / 82, 8.21 / 92, 9.5 of the Hume Society at its various conferences.
/ 106, 9.5n20 / 107, and 12.22 / 159. Note I am also grateful to Henry Merivale, Hsueh
that this contrast cannot be accounted for in Qu and especially Dan O’Brien for comments
terms of Hume’s moving away from the foun- on this paper, and to the Edinburgh Illumni
dational metaphor more generally. On the for providing the David Hume Fellowship at
contrary, he says that induction is ‘founded’ Edinburgh, thus giving me the opportunity to
on the relation of cause and effect, or experi- work in the delightful context of the Institute
ence, or the Uniformity Principle, and that it for Advanced Studies in the Humanities, over-
is not ‘founded’ on reasoning, argument, or lapping with Don Garrett’s tenure there.


9780826443595_Ch03_Final_txt_print.indd 103 2/9/2012 9:39:02 PM

Lorne Falkenstein

Over the course of Treatise 1.3, Hume pre- mechanisms rather than logically sound
sented what he called a ‘system’ of probable judgement, and declares those beliefs to be
reasoning. He then went on, in Treatise 1.4, ultimately unjustifiable. Despite this sceptical
to argue that sceptical objections would result Hume was able to provide for a logic of
leave us entirely incapable of belief were our probable reasoning, grounded on natural, but
natural inclinations not too strong for philo- unjustifiable beliefs. How he did so is still not
sophical conclusions to be able to restrain well understood. That he was able to do so is
our inferences. In the Enquiry concerning one of his great achievements.
Human Understanding he reversed this pro-
cedure, first offering ‘sceptical doubts’ about
the legitimacy of our inferences concerning 1. THE SYSTEM OF THE TREATISE
matters of fact, then a ‘sceptical solution of
these doubts’, and finally a conclusion that All kinds of reasoning consist in noth-
we can only be legitimately sceptical of ing but a comparison, and a discovery of
claims in religion and school metaphysics, those relations, either constant or incon-
not everyday experience or natural science. stant, which two or more objects bear to
Despite the more optimistic tone, the the- each other. (THN / 73)
ory of the Enquiry is built on the same two In the Treatise Hume’s presentation and
principles as the ‘system’ of the Treatise: the defence of his ‘system’ took the form of a tor-
principle of the association of ideas, and the turous journey down the dead-end lanes and
principle of the genesis of belief in the unob- twisted turns of a meditative path of discov-
served as a consequence of association with ery, supplemented by appeals to observations
sensory experience or memories. The common and experiments, worries over contrary evi-
‘system’ of the Treatise and the Enquiry is dence, and the introduction of refinements to
sceptical because it takes our beliefs to be the accommodate recalcitrant data. His project
product of naturally occurring psychological was to inquire into the basis of reasoning,


9780826443595_Ch04_Final_txt_print.indd 104 2/9/2012 9:39:12 PM


particularly probable reasoning. Reasoning 2. THE PATH TO THE FIRST PRINCIPLE

consists in inferring unknown from known
objects by means of a relation between the One of the infelicities of Hume’s presentation
two (THN / 70, / 73). The is this precipitate assertion that the causal
relations on which all reasoning is based are relation is an inconstant relation, and the only
‘philosophical relations’, which are discov- relation on which probable reasoning can rest.
ered by comparing objects with one another While Hume never adequately justified the lat-
(THN / 73). There are also ‘natural ter claim,1 he went on to give an argument for
relations’. A natural relation is not discov- the former. Our ideas of cause and effect could
ered by comparison or appealed to in order not be ideas of any of the observed qualities
to discover or justify a conclusion. Whether of objects because all objects are causes and
we are aware of it or not, it exercises an effects, and there is no quality that all objects
influence on the imagination, impelling us share in common. For any quality we might
to form an idea of an object. This produces pick on, there is some object that is a cause
a kind of instinctive, counterfeit reasoning or an effect even though it does not have that
(THN 1.1.5 / 13–15, cf. 1.1.4 / 10–13). quality. The idea must therefore be the idea
Relations can be divided into two main of a relation (THN / 75) – indeed, of
kinds: ‘inconstant’ relations, which can alter a relation that can change while the objects
even while the compared impressions or ideas remain the same.
remain the same (e.g. relations of contiguity At this point a further infelicity arises.
and distance in space or time); and ‘constant’ Hume simply assumed that causal relations
relations, which cannot change without a are not immediately perceived or intuited
change in the compared objects (e.g. rela- upon comparing objects, as we immediately
tions of resemblance). Demonstrative rea- perceive relations of contiguity or immedi-
soning, yielding certainty, is founded on the ately intuit relations of resemblance, but are
latter, while probable reasoning is founded instead ‘deriv’d’ from some other relation
on the former (THN–2 / 69–70, (THN / 75). It is only much later that / 73). a reason is given for this assumption (THN
Though we discover a number of con- / 86–7, explained more fully at Abs.
stant and inconstant relations by compar- 11–12 / 650–1). If we could perceive or intuit
ing objects, Hume maintained that there is the existence of causal relations, then we
only one relation that can serve as a basis would be able to tell upon first acquaintance
for demonstrative reasoning, the relation of with a pair of objects whether or not they
degrees in quantity, and only one that can are causally related. But we cannot do so.
serve as a basis for probable reasoning, the Granting that causal relations are not
relation of cause and effect. The other con- immediately perceived but are instead derived
stant relations can always be ‘intuited’ with- from some other relations, what might those
out the need for any demonstration and the relations be? The only relations we discover
other inconstant relations can only be ‘per- when we compare those objects we con-
ceived’ and not used as a basis for inference sider to be causes and effects are contiguity
to the unobserved (THN / 70, in space and succession in time. But while
/ 73–4). Thus all probable reasoning reduces these relations may hold between causes
to causal reasoning. and effects,2 they also hold between objects


9780826443595_Ch04_Final_txt_print.indd 105 2/9/2012 9:39:13 PM


that we consider to be only accidentally con- founded on intuition or demonstration and

joined. And, Hume claimed, we think that concluded that it must therefore be based
causes and effects are not just accidentally on experience (THN 1.3.3 / 78–82). Oddly,
conjoined but necessarily connected. he made no attempt to argue that the same
This is tricky. How can we think that there answer must be given to the second question.
is a necessary connection between causes and It is only later (THN / 86–7) that a
effects if we cannot discover any other rela- reason is given for concluding that our infer-
tions between them than contiguity in space ences from cause and effect are not based
and succession in time? Occasional sug- on demonstration. The reason goes back to
gestions to the contrary notwithstanding,3 an observation on the nature of imagina-
Hume did not want to say that we have no tion Hume made at THN / 10. The
idea of necessary connection. A necessary only limitation on the imagination is that its
connection is simply a connection that has ideas come from things that have been pre-
to be present and cannot be broken. A har- viously encountered in sensory experience.
ness is a connection between a horse and Once given ideas, it can separate and rear-
a carriage. If the harness had to run from range them in any way whatsoever. Given an
the horse to the carriage and nothing could object at a place and a time, the imagina-
break it, it would be a necessary connection. tion can conceive any object whatsoever at
The idea of a necessary connection between any of the contiguous places at the earlier
causes and effects is similarly the idea of or later times. Since causes and effects exist
a ‘tye or connexion . . . which binds them at distinct times, any particular cause could
together’ (EHU 4.10 / 29). Hume’s claim at be imagined to be followed by anything and
this stage in his meditations was not that we any particular effect preceded by anything.4
are not thinking of anything when we think But were it intuitively obvious or demon-
of a tie reaching across time and space to stratively provable that this particular cause
bind cause and effect together. It was rather must have that particular effect, any other
that we cannot discover exactly what does alternative would be inconceivable (THN
the job. We only ever see the horse and the / 86–7).5
carriage. The apparatus harnessing the two Postponing this argument for the moment,
together is not apparent. This does pose a Hume (at THN 1.3.4 / 82–4) simply took it
problem: if we cannot discover any harness, for granted that the connection between par-
why do we think it is there – indeed, that it ticular causes and their effects can only be
must be there and cannot be broken? known by experience and proceeded to ask
Hume proceeded to try the patience of what sort of experience does this. In the pro-
his reader further by pretending to have no cess, he presented himself as suddenly dis-
answer to this question and affecting the need covering a new, third relation obtaining in
to look for one by investigating two related cases of causality (THN–3 / 87–8).
questions: (i) why we consider it necessary The relation is not discovered when compar-
that every event have a cause and (ii) why ing individual instances of cause and effect,
we consider this particular cause necessar- but only when comparing multiple instances.
ily to have that particular effect. In response All instances resembling the cause are spa-
to the first, he argued that our belief that tially contiguous with instances resembling
every event must have a cause could not be the effect, and precede them in time.


9780826443595_Ch04_Final_txt_print.indd 106 2/9/2012 9:39:13 PM


Hume continued to express puzzlement At this point in the course of his medita-
over how this new relation of ‘constant con- tions, Hume finally felt prepared to reveal8
junction’ could lead us to conclude that this the first principle of the ‘system’. Though
particular cause must necessarily have that constant conjunction provides us with no
particular effect. We do not discover anything justification for inferring that causes and
in multiple instances that could not be found effects are necessarily connected, it is a
in just one instance, and we do not discover ‘natural’ relation, which impels the imagin-
anything in one instance that would justify ation to call up an associated idea when
the conclusion. Nor could we take the expe- presented with its partner. It therefore
rience of a constant conjunction to establish produces a kind of counterfeit, instinct-
the likelihood of a necessary connection, or ive ‘reasoning’.9 Observing objects of one
the likelihood that the causes contain some sort being customarily followed by objects
quality, unknown to us, that gives them the of another sort trains a habit of thinking
power to bring about their effects.6 Since into the mind. Once developed, the habit
all we perceive are the observable relations induces the imagination to form an idea of
between causes and effects, none of which is an object upon encountering its customar-
a necessary connection, and the observable ily conjoined partner, even in the absence of
qualities of the causes, none of which is a perceiving any tie binding the two together,
power, the most we could infer is that, in the even in the absence of having any reason to
past, objects like the cause have been con- suppose that the future will be like the past,
tiguous to and followed by objects like the and even in the absence of any recollec-
effect, and that, in the past, the set of quali- tion of or reflection upon the past instances
ties characteristic of the causes has included (THN / 103–4). Although Hume
some further, unknown power. But we are did not draw the conclusion until very much
in no position to infer that similar rela- later (THN 1.3.14), this is why, even though
tions must obtain in the future or that simi- we cannot discover any tie that necessarily
lar collections of observed qualities will be binds cause and effect together, we think
accompanied by similar powers in the future that there must be such a thing.
(THN / 90 and / 91). The
new relation of constant conjunction could
only lead us to draw these inferences with the
aid of a further supposition, that what has 3. THE PATH TO THE SECOND
been observed to occur regularly in the past PRINCIPLE
will continue to be observed to occur in the
future. But this principle is not demonstra- After this long journey of discovery, the
bly true, because there is no contradiction in second principle of the ‘system’ was quickly,
conceiving a change in the course of nature though not easily, uncovered. Hume began
(THN / 89). Moreover, it cannot be by noting that causal inference takes place
proven by appeal to past experience, because only when one of the two associated objects
the question at issue is precisely why we is experienced or remembered and the other
should take regularities in past experience to is not. When both objects are experienced or
establish a rule for what will happen in the remembered, there is no occasion to imagine
future (THN / 89–90).7 either. And when neither is either experienced


9780826443595_Ch04_Final_txt_print.indd 107 2/9/2012 9:39:13 PM


or remembered, we feel impelled to imagine only completed at 1.3.7 / 94–8, after hav-
the one upon having occasion to imagine the ing been itself interrupted by 1.3.6 / 86–94.
other, but do not form any belief in the exist- The account begins with an examination of
ence of either (THN / 73–4, / the ‘impressions of the senses and memory’,
87, 1.3.4 / 82–4). which are the apparent source of the belief
Seeking for an explanation for this variation, based on causal inference. The relevant point
Hume noted that the objects of experience and about sense experience had already been
memory are believed to exist (THN / made much earlier in Treatise 1.1 / 1–25. Any
86). This suggests that the belief we get as a object that can be sensed can be imagined, so
consequence of causal association might be that the difference between sense experience
due to some sort of transfer from an experi- and imagination cannot arise from what is
ence or memory to an associated object. When sensed or imagined. Since there is nonethe-
an object is believed to exist or have existed, less a difference, it must be due to something
the relation of constant conjunction induces else. Hume referred to this other factor as
us not merely to conceive an associated object a different ‘manner’ in which the object is
but to form a belief in that object’s existence at conceived (THN / 96), and tried to
a contiguous place and the appropriately earl- describe further this manner of conception
ier or later time (THN / 97).10 by saying that sensing is more ‘forceful’ and
Hume was discontented with this bald ‘vivacious’, imagining ‘fainter’ and ‘lower’
hypothesis and sought for a justification. (THN / 1–2, / 2–3).
At Treatise / 98–9 he attempted to Turning to memory, Hume observed that
account for the origin of belief as a specific anything that can be remembered can likewise
instance of something that can be observed be merely imagined, and concluded that the
to happen more generally: a natural tendency difference between what is received as a fan-
to confuse readily associated objects. Hume tasy and a memory must consist just in the way
claimed that because the natural transition of it ‘feels’ to remember (THN–5 / 85–6).
thought between objects that have been con- Having isolated these differences between
stantly conjoined is ‘so easy’, it goes unnoticed. sensing and remembering, on the one hand,
Consequently, any mental ‘disposition’ that and imagining on the other, Hume inferred
might happen to attend the latter becomes that ‘the belief or assent, which always
attached to its impostor. Where the partner attends the memory and senses is nothing but
is experienced or remembered, the disposi- the vivacity of those perceptions they present;
tions accompanying experience or memory and that alone distinguishes them from the
are confused with the associated idea. Since imagination. To believe is in this case to feel
those dispositions are always bound up with an immediate impression of the senses, or a
a belief in the existence of the experienced or repetition of that impression in the memory’
remembered object, we end up believing in (THN / 86).
the existence of the associated object as well. These reflections on the nature of sense
This account of the origin of belief is com- and memory led Hume to his conclusion. If
plemented with an account of the nature the belief attending causal inference arises
of belief, infelicitously inserted at Treatise from a transfer from a sensed or remembered
1.3.4–5 / 82–6, where it interrupts the thread object to an imagined object, and if the belief
of argument for the first principle, and in sensed or remembered objects is nothing


9780826443595_Ch04_Final_txt_print.indd 108 2/9/2012 9:39:13 PM


but a more vivacious conception of those to understand belief as a specific case of

objects, then the belief attending causal infer- something that happens more generally. His
ence must likewise be just a more vivacious search for these analogies had mixed results.
conception of an object. They will be discussed later.
This is surprising. Rather than find belief in 2. A more pressing problem is the charac-
an unperceived object to be the product of a terization of belief. Four different characteri-
judgement, justified by appeal to a causal rela- zations of belief have emerged from what has
tion to an experienced or remembered object, been said about Hume’s account. At Treatise
Hume found it to be no different from the belief / 98–9, Hume described belief as a
that attends sense experience and memory. It ‘disposition’ of the mind. Over the course of
consists just in a more vivacious conception Treatise / 628–9 (from THN App. 1)
of the object. Perhaps because he sensed that and / 98–9 this disposition is further
this conclusion would not be readily accepted, described as having to do with drawing and
Hume pretended to remain hesitant about it, focusing attention (‘rendering more present’,
offering two reasons for his hesitancy: ‘weighing more in the thought’, ‘having more
1. Accounting for the origin of belief force and influence’, ‘appearing of greater
involves identifying a cause. According to importance’, ‘fixing the attention’); arous-
Hume’s own account of causality, we can- ing passion (‘elevating the spirits’); inspiring
not discover causes merely by inspecting deliberation (‘having a superior influence on
their effects, nor can we be confident that the imagination’); and inclining us to action.
we have identified causes if we have exam- But THN–6 / 95–7 and /
ined only one instance. Either we must find 628–9 also describe belief as a ‘manner’ in
some analogy between the one instance we which an object is conceived. A further char-
have before us and other instances and dis- acterization of belief is found at THN App. 3
cover some regularity in the succession of / 624–5 and in passages from the Appendix
events in the analogous cases, or we must recommended by Hume for insertion in the
show how some combination of more basic, main body of the Treatise ( / 628,
previously established causes could account / 628–9), where belief is described as a
for the effect. The appeal made at Treatise ‘feeling different from the simple conception’ / 98–9 to a general tendency to con- of an object. Finally, and most notoriously,
fuse readily associated objects is a justifica- belief is described on numerous occasions
tion of the latter sort. But Hume went on to in both the Treatise and the Appendix as a
declare that, while he would be satisfied if more forceful and vivacious idea, with the
his reader found this reason compelling, he terms ‘force’ and ‘vivacity’ often being sup-
himself placed his chief confidence in being plemented by a list of others (e.g. ‘solidity’,
able to uncover a justification of the former ‘firmness’, ‘steadiness’) that are not obvi-
sort (THN / 99; App. 3 / 624–5). He ously synonymous, either with one another
wanted to find analogies between the for- or with ‘force’ or ‘vivacity’.
mation of belief as a consequence of experi- The bare fact that Hume described belief
encing or remembering an object that has in these different ways does not pose a prob-
been constantly conjoined with some other lem as long as the different descriptions
object in the past and other operations of can all be integrated.11 But Hume seems to
the mind – something that would allow us have become worried that the frequent and


9780826443595_Ch04_Final_txt_print.indd 109 2/9/2012 9:39:13 PM


prominent description of belief as a more dispositions – and, importantly, evidence that

forceful and vivacious idea had ‘not been so is defeasible – is suggested by THN /
well chosen, as to guard against all mistakes 112–13 and 9.14 / 113–14.12
in the readers’ (THN App. 1 / 623). Perhaps Whatever frustrations Hume may have
this was because he found readers inclined had with his efforts, it is clear that he meant
to take ‘force and vivacity’ to refer to some to reject the view that to believe is to perform
qualitative feature of the object that is con- an act of assenting to a proposition, where
ceived, like brightness or distinctness, rather a proposition involves asserting a relation
than, as he had all along wanted to insist, a between two or more ideas. In particular, to
‘manner’ in which we conceive this object, believe that something exists is not to assent
specifically, a conception with focused atten- to a proposition joining the idea of that thing
tion, aroused passion, and an impetus to to the idea of existence. Hume rejected this
deliberation and action – these being ‘dispo- possibility by arguing that we have no idea
sitions’ of the mind that are ‘felt’ even when of existence distinct from whatever particular
not acted upon and so not made evident thing we conceive to exist. To conceive some-
to others. In the Appendix and insertions thing as existing is no different from conceiv-
to the Treatise proposed in the Appendix ing it (THN–6 / 66–7, / 94–5,
Hume stressed that by ‘force and vivacity’ THN App. 2 / 623–4). He offered the ineligi-
he had meant conception of an object in this bility of this account of belief as a further rea-
‘manner’ – conception attended with these son to accept the alternative that to believe is
dispositions. just to sense or remember or be instinctively
But Hume also confessed that he found ‘a inclined to form a more vivacious idea.
considerable difficulty in the case; and that In a footnote, Hume went so far as to
even when I think I understand the subject describe the division of the acts of under-
perfectly, I am at a loss for terms to express standing into conception, judgement and
my meaning’ (THN / 628). What may reasoning, and the definitions given of these
have bothered him was that appealing to ‘dis- operations as a ‘remarkable error’. These
positions of the mind’ to explain belief does three acts of the understanding ‘all resolve
not sit well with what he was to go on to say themselves into the first, and are nothing but
about the nature of minds and mental acts (in particular ways of conceiving our objects’
THN 1.4.6 / 251–63 and–7 / 244–6). (THN / 96n). This is an overstate-
This pushed him in the direction of taking ment, since Hume did recognize that we
belief to be a feeling (presumably, the feeling do things like comparing objects with one
of having one’s interest aroused, one’s passions another to discover relations between them,
elevated, and one’s inclinations determined), or executing arithmetical demonstrations
and in turn prompted worries about whether in which one thing is inferred from another
the feeling might be separable from the con- by appeal to a relation between the two.
ception (denied over THN, App. 4–8 / 625–7, Indeed, as will be noted below, Hume went
but affirmed at EHU 5.11 / 48). But that so far as to recognize a class of ‘oblique’ or
Hume took the presence of the dispositions to ‘explicit and indirect’ causal inferences that
be what is ultimately constitutive of belief, and are demonstrative in the classic sense. These
the feelings of being so disposed to be merely operations satisfy the definition of judge-
introspective evidence for the presence of the ment as the ‘separating or uniting of different


9780826443595_Ch04_Final_txt_print.indd 110 2/9/2012 9:39:14 PM


ideas’, and of reasoning as the ‘separating or of fact’ or existence. Like the Treatise, they
uniting of different ideas by the interposition leap to the conclusion that this can be done
of others, which show the relation they bear only by means of causal inference. Unlike
to each other’ (THN / 96n). The the Treatise, they do not proceed to analyse
received definitions of judgement and reason- the causal relation in terms of a problematic
ing apply to those operations that are con- notion of necessary connection. The Abstract
stitutive of knowledge in the demonstrative analyses causality in terms of contiguity in
sciences, particularly arithmetic, and know- space, succession in time and a constant con-
ledge of intuitive truths, such as that orange tiguity and succession in like instances, mak-
is more like red than green.13 They even apply ing no mention of necessary connection. The
to many of the judgements and arguments Enquiry offers no analysis of the causal rela-
found in the empirical sciences. But they do tion at all, though there are passing references
not describe all of those operations that are to a ‘supposed’ tie or connection between
constitutive of belief in the empirical sciences cause and effect (EHU 4.4 / 26–7, 4.10 / 29).
or in everyday life. In particular, they do not In both works, necessary connection, which
describe the most fundamental of those oper- had played such a large role in the Treatise,
ations. We form fundamental beliefs neither comes up for discussion only after the two
by discerning relations between ideas nor by parts of the ‘system’ have been presented.
inferring one idea from another by appeal to Rather than investigate the notion of neces-
an intermediate relation. Instead, we form sary connection, Hume directly proceeded to
fundamental beliefs by having lively concep- ask how causal reasoning enables us to infer
tions given to us in sensation and memory, the existence of unperceived objects.
or by being instinctively compelled to form He first claimed that we cannot do this in
lively conceptions as a consequence of asso- advance of experience, by reference to any-
ciation with what is sensed or remembered. thing we can find in those objects we consider
The latter is ‘not only a true species of rea- to be causes and effects. In contrast to the scat-
soning, but the strongest of all others’ (THN tered, sketchy, unconvincing, and ill-placed / 96n). arguments of Treatise,, and his conclusion was now justified by
two different lines of argument. First, Hume
appealed to everyone’s introspection, assisted
4. THE ARGUMENT OF THE by appeal to cases such as that of encoun-
ABSTRACT AND THE ENQUIRY tering an object for the first time or that of
the biblical Adam, newly created with fully
Hume came to be dissatisfied with the ram- functioning, adult cognitive capacities, but no
bling, quasi-meditative path of discovery he experience. Just as Adam would be unable to
had dragged his reader down when present- say what the effect of any given cause would
ing his ‘system’ in the Treatise. The Abstract be prior to experience, even the effect of the
and the Enquiry offer a far more elegant motion of one billiard ball towards another,
presentation of the same theory.14 They so we find ourselves unable to say what the
replace the opening discussion of the foun- effect of a cause will be or what the cause
dations of probable reasoning with the ques- of an effect was in novel cases (Abs. 11 /
tion of how we reason concerning ‘matters 650–1; EHU 4.6–7 / 27–8). If we think we


9780826443595_Ch04_Final_txt_print.indd 111 2/9/2012 9:39:14 PM


do perceive causal powers in more familiar the earlier argument at THN / 86–7,
cases, it is only because we have forgotten this appears to rule out the possibility of there
what it was like to experience these things for being any such thing as a necessary connec-
the first time (EHU 4.8 / 28–9). While we do tion between the two. In contrast, the version
often anticipate how events will turn out in of the argument in the Enquiry omits any
novel situations (scientific experiments being reference to conceivability as a criterion of
the prime example), the demonstrations that metaphysical possibility.
we employ when doing so appeal to funda- Having established that reasoning from
mental causal rules (cohesion, gravitation, causes and effects is based on past experience,
communication of motion by impulse, etc.) the Abstract proceeds to argue that any rea-
that are not intuitively or demonstrably obvi- soning from that experience would have to
ous, which begs the question of how we have depend on the principle that the future will be
obtained the idea of these fundamental causal like the past. But (i) this principle is not dem-
relations (EHU 4.12 / 30–1). onstrably true, since a change in the course
Second, Hume appealed to variations on of nature is conceivable. And (ii) any attempt
the argument of THN / 86–7: to prove that it is most likely true would run
in a circle, since we could only appeal to the
• According to Abs. 11 / 650–1, effects ‘fol- fact that it has been true in the past to argue
low’ from causes. Consequently, given that it will most likely continue to be true in
any cause existing at one time, we can the future. Even were we to take a constant
conceive any other object to exist at the conjunction in past experience to be evidence
following time. But when something is
of the existence of a power in causes to bring
demonstrable, the opposite is contradic-
about their effects, we only perceive the sens-
tory and inconceivable.
• According to EHU 4.8–11 / 28–30, every ible qualities of bodies, and we can have no
effect is a different event from its cause. assurance that like sensible qualities will con-
Consequently, when conceiving the one, tinue to be conjoined with like powers.15
it is not necessary that we also conceive The Enquiry mounts the same argument,
the other. If we do conceive them together, but addresses it to a different question – not
we are conscious that nothing compels the question of why we suppose that the
us to do so, so that the conjunction is future will be like the past, but the question
effectively arbitrary. But this means that of why we suppose that like objects contain
there can be no demonstration of effects like hidden powers (EHU 4.7 / 28). This is
from causes or causes from effects, again
not an innovation, because the same question
because where there is a demonstration
had been raised in the Treatise (at–10
the opposite is inconceivable.
/ 90–1) and the Abstract (at 15 / 652), though
only as an afterthought,16 and reference to a
The version of the argument in the Abstract supposition that the future will be like the
omits reference to the power of the imagin- past does come up in the Enquiry over the
ation to separate different objects, but it obvi- subsequent course of the argument (at 4.19
ously rests on that assumption and it also / 35–6). In addition to giving the usual rea-
makes more explicit appeal to the principle sons for a negative answer Hume also offered
that ‘whatever we conceive is possible, at least a new argument: since peasants, infants and
in a metaphysical sense’ (Abs. 11 / 650). Like animals are able to infer effects from causes,


9780826443595_Ch04_Final_txt_print.indd 112 2/9/2012 9:39:14 PM


either they do not do so by means of any argu- entertained with speculations, which, however
ment or demonstration, or only by means of accurate, may still retain a degree of doubt
the simplest and most obvious of reasons. Yet, and uncertainty (EHU 5.9 / 47)’. ‘Readers of
unless Hume was more obtuse than a peasant a different taste’ are told that the part may be
or child, there are no such reasons.17 neglected without impairing an understand-
Having raised these ‘sceptical doubts ing of subsequent portions of the book.
about the operations of the understanding’ Part 2 of Enquiry 5 (EHU 5.10–22 /
the Abstract and the Enquiry proceed to offer 47–55) is nonetheless important. As Hume
a ‘sceptical solution of these doubts’ – the stressed in the same breath in which he
same, two-part solution that was presented advised ‘readers of a different taste’ to move
in the Treatise. First, our experience of what on, delving into the question of what belief is
has customarily been the case in the past and how it arises will uncover ‘explications
trains habits of thought into us, so that we and analogies that will give satisfaction’. The
naturally expect the same sorts of things to ‘satisfaction’ Hume had in mind is not just
happen in the future. The expectation is not the satisfaction of idle intellectual curios-
rationally justified, but naturally induced. ity, but the satisfaction of objections to the
Again, the Enquiry adds a new and compel- account of belief laid out in the concluding
ling argument for this conclusion: attributing paragraphs of the Enquiry 5, Part 1. A con-
the inference to habit offers the only plaus- cern with uncovering ‘analogies’ between the
ible explanation for how it is that we come account of belief and other operations of the
to draw a conclusion from many experi- mind is a constant of Hume’s thought about
ments that we would not draw after seeing his ‘system’ (cf. THN / 99; App. 9 /
just one. Secondly, we do not just infer causes 627; Abs. 23 / 655), for good reasons that
and effects from one another but believe the have already been alluded to. There is very
absent partner to exist (at the contiguous little positive argument to justify Hume’s
place and the appropriately prior or poster- ‘system’. The meditative path of discovery of
ior time).18 Because this belief does not arise the Treatise offers only rhetorical support,19
from reasoning, but from habit, it is declared and the Abstract’s and Enquiry’s effective
to be due to ‘a species of natural instincts, reconstruction of that path into a critique of
which no reasoning or process of the thought the view that our causal inferences are justi-
and understanding is able, either to produce, fied by appeal to facts and rules at best put
or to prevent’ (EHU 5.8 / 46–7). Hume in a position to claim that causal infer-
The Abstract and the Enquiry go on to ence is not based on reasoning, not to claim
examine what belief is, reaching the same con- that it is based on a habit of association and
clusions as the Treatise and, in the Enquiry, a transfer of belief from an impression or
even stating them by means of an extended memory. The same can be said of a new argu-
quotation from the Treatise (EHU 5.12 / ment, presented only later, to the effect that
48–50, quoting THN / 628–9 with reasoning is too slow and uncertain in its
minor modifications). Interestingly, in the operations to be entrusted with an operation
Enquiry these further details about belief as important for survival as causal inference
are reserved for a distinct part of Enquiry 5, (EHU 5.22 / 55) and of the Enquiry’s appeal
prefaced by a remark dedicating the part to to the abilities of peasants, children and
‘such as love the abstract sciences, and can be animals to draw causal inferences. The one


9780826443595_Ch04_Final_txt_print.indd 113 2/9/2012 9:39:14 PM


positive argument for the theory presented passages are identical in both) – Hume noted
so far is the Enquiry’s appeal to the problem the following ‘analogies’ between causal infer-
of how we draw a conclusion from repeated ence and other operations of the mind: the
experiments that we cannot draw from just picture of an absent friend ‘enlivens’ the idea
one, and that argument offers a justifica- of that friend, as well as the passions that idea
tion of the least satisfying sort: inference to occasions; the ‘mummeries’ of the Roman
the best explanation. Hume hoped that by Catholic religion ‘enliven’ devotion; ‘sensi-
uncovering ‘analogies’ between belief and ble types and images’, which have a greater
other operations of the mind he would be influence on the fancy than any other, ‘con-
able to offer a more compelling, Newtonian vey that influence’ to the ideas they resemble;
argument by induction from the phenomena objects that are placed in the vicinity of other
to a general rule. The general rule would pro- objects ‘transport’ the mind ‘with a superior
vide justification for the two-part system, as vivacity’ to ideas of those other objects (e.g.
a special case, but it would in turn be sup- passing the house next door on my way home
ported by induction from all the analogous gives me an idea of my home that ‘imitates
cases revealed by experience. an immediate impression’). Importantly, in
However, Hume had come to think that the all of these cases the trigger (the picture, the
public had no taste for this sort of investigation, ceremony, the icon, the neighbouring objects)
particularly if drawn out to any great length must be both experienced and ‘naturally
(Abs. Preface, 1–2 / 643). His solution was to related’ to the target (in the cases mentioned,
drastically abbreviate the argument, focus- by relations of resemblance or contiguity); if
ing just on the exposition of analogous cases, the trigger is merely imagined, the target is
and to invite impatient readers to skip ahead. not enlivened; if the trigger is unrelated, the
In the Treatise he went on at much greater idea of the target does not even arise.
length, not only identifying analogous cases, The case is the same with causal infer-
but worrying about contrary evidence, refin- ence, as Hume proceeded to prove by appeal
ing the system to account for it, and appealing to three experiments (THN–11 /
to the system to account for a wide range of 101–3): suppose the natural relation (in this
other phenomena, thus adding a demonstra- case of constant conjunction) is absent (as,
tion of explanatory power to the other reasons for example, when experiencing a cause for
for accepting the system. Because the Enquiry the first time). Then the associated idea does
merely repeats some of what the Treatise had not arise. Now suppose a constant conjunc-
to say on this score,20 I focus on what Hume tion between the trigger and the target has
had to say in the Treatise in what follows. been experienced in the past. Then, solely for
that reason and without the assistance of any
intermediate process of argument or justifi-
cation or appeal to general rules, a vivacious
5. ANALOGIES, EXPERIMENTS, idea of the target arises upon experience of the
RECALCITRANT DATA, AND trigger. Now suppose the trigger is not experi-
REFINEMENTS enced but only imagined. Then the associated
idea does arise, but it has no vivacity.
In both the Treatise (–5 / 99–100) and By induction from all these phenomena,
the Enquiry (5.15–18 / 51–3) – the relevant Hume declared it to be ‘a general maxim


9780826443595_Ch04_Final_txt_print.indd 114 2/9/2012 9:39:14 PM


in the science of human nature, that when Since the head, shoulders, breast and neck
any impression becomes present to us, it not are neither causes nor effects of the legs and
only transports the mind to such ideas as are thighs, the inference here is not causal, even
[naturally]21 related to it, but likewise com- though Hume recognized that it produces
municates to them a share of its force and belief. The same holds of Hume’s description
vivacity’ (THN / 98). of ‘our approach to any object; tho’ it does not
This maxim was no sooner justified than discover itself to our senses’ as leading that
Hume acknowledged a difficulty. He had object to operate on the mind ‘with an influ-
defined belief to be nothing but a more ‘viv- ence that imitates an immediate impression’
acious’ idea. But he had also maintained that (THN / 100). An ‘influence that imi-
belief only arises from causal inference, not tates an immediate impression’ is just a belief.
from the other natural relations of resemblance The case of resemblance poses more of a
and contiguity, notwithstanding that, according problem. Seeing the son of a long-dead friend
to the maxim, they all enliven ideas. The three (EHU 5.19 / 53) does not produce a belief
claims are inconsistent (THN / 107). in the existence of an unperceived object. At
At this point Hume stood on the brink of best, it rouses old memories and enlivens the
momentous discoveries, presented with an associated passions.23 One reason for this is
opportunity to reassess his earlier, ill-consid- that the resemblance relation calls up an idea
ered position that belief in matters of prob- of a resembling object without giving any fur-
ability can only arise from causal inference. ther indication of where that object is placed
He had discovered that basic causal infer- in space and time, leaving us with no inclina-
ences are inferences from the constancy of tion to ascribe it a location in the real world.
temporal succession in resembling cases. But Our causal and geographical inferences, on
we also draw inferences from the constancy the other hand, involve not just association
of spatial arrangement in resembling cases. of objects, but association with places and
Quite apart from forming any background times where the object is to be found. Not
beliefs about the causes of the immobility of surprisingly, therefore, when resemblance is
landmarks, we rely on the constancy of the bound up with relations of time and place, it
position of houses, trees, the pole star and has the same influence as constant succession
other geographical or astronomical objects to in time and constant conjunction in space.
navigate, and when we do so we reason from This is most notably the case with our beliefs
experienced objects to their unperceived sur- about the identity of objects over time, where
roundings, not to their unperceived causes or we suppose a continuum of intermediate
effects.22 Hume himself recognized this with- states to exist unperceived between observed,
out realizing it when he wrote: resembling, earlier and later states.
This is just a sketch of how Hume might
Suppose I see the legs and thighs of a
have gone on to investigate the possibility that
person in motion, while some interpos’d
probabilistic reasoning involving all three of
object conceals the rest of his body. Here
’tis certain, the imagination spreads out the ‘inconstant’ relations of causality, contigu-
the whole figure. I give him a head and ity and identity could be grounded in the three
shoulders, and breast and neck. These ‘natural’ relations of constant succession in
members I conceive and believe him to time, constant conjunction in space and closest
be possessed of. (THN App. 4 / 626) resemblance at contiguous places over time.


9780826443595_Ch04_Final_txt_print.indd 115 2/9/2012 9:39:14 PM


Unfortunately, Hume did not take this These reflections mandate a revision to
path. (THN–23 / 194–9 is perhaps Hume’s maxim, although he never said so.
the most lamentable consequence of that deci- While impressions may transport the mind to
sion.) He did go so far as to declare that we any ideas that are naturally related to them,
take our sense experiences and memories to they only communicate a share of their force
constitute a ‘system’ of ‘realities’, and that we and vivacity sufficient to induce belief to
join a second ‘system’ to it, consisting of the ideas of those objects that have been custom-
unperceived causes or effects of these ‘realities’ arily conjoined with them in the past.
(THN–4 / 108). But he never paused This revised maxim explains an attempt at
to consider that what makes the objects of the ‘confirmation’ of the ‘hypothesis’ that Hume
senses and memory a ‘system’ is that each is made at THN–19 / 115–17. The
related to all the others in virtue of its unique attempt appeals to an example that would
location within a single space and time, and otherwise serve more to falsify than to con-
that what enables the causal relation to aug- firm the hypothesis. If the repetition of a
ment the system is that it directs us where conjunction (making it ‘customary’) plays
to localize unperceived objects in this space a more important role in producing belief
and time – something that constant contigu- than the force and vivacity of the impression,
ity and resemblance insofar as it is bound up and if belief is a mental ‘disposition’ involv-
with identity relations could also do. ing things that can be produced merely by
Instead, Hume maintained that because repetition, such as fixed attention, familiarity
any given object resembles and is spatially and stability of the object, then we should
contiguous to a huge variety of other objects, expect that belief could arise from the mere
the mind senses a certain ‘caprice’ or feeling repetition of an idea even in the absence of
of liberty in making the association with just association with an impression or memory.
one. This feeling of liberty prevents the easy Hume considered this in fact to be the case,
and unnoticed transition from one object to most notably with the beliefs produced by
another that Hume had earlier identified as education, which he considered to provide
essential for the transfer of the mental disposi- outstanding confirmation for the hypothesis
tions characteristic of belief. It also introduces because, as he claimed, education is respon-
new feelings of ‘looseness’ and ‘weakness’ that sible for more than half our opinions and is
are contrary to the feelings of stability and more influential than either abstract reason-
strength characteristic of belief. Moreover, ing or experience (THN / 117). He
any tendency we might have to include objects also instanced the tendency of amputees and
thought of under these conditions in the sys- the bereaved to be unable to accept their loss,
tem of ‘realities’ would produce repeated and of people to consider themselves to be
experiences of having our expectations disap- on intimate terms with personages they have
pointed. As a consequence, we would learn to only read about. It is hard not to wonder
associate objects thought of under these condi- about the aptness of these examples or the
tions with fictions (THN / 109–10).24 soundness of Hume’s implicit view of how
Causal relations are very different. Any given education produces belief, but people do
object is related to just one other object as its have a tendency to believe what they hear
cause and just one other object as its effect, so from those around them simply because
there is no ‘looseness’ to the association.25 everyone is saying it and even though no one


9780826443595_Ch04_Final_txt_print.indd 116 2/9/2012 9:39:15 PM


is in a position to testify to the truth of what demonstrably truth-preserving or probabili-

they are saying (the belief in an afterlife being ty-preserving rules to draw inferences from
an outstanding example). the observed to the unobserved. We do not
Besides this appeal to a confirming experi- discern a relation between causes and effects
ment, Hume justified the hypothesis by appeal by comparing them with one another and
to its explanatory power (THN–15 / then appeal to this relation to draw infer-
110–15). The hypothesis is able to explain ences to the unobserved. Instead, we are
such things as (i) why pilgrimages strengthen instinctively impelled to form ideas of objects
belief; (ii) why it is wrongly supposed that of a sort that have in the past been frequently
the communication of motion by collision observed to be constantly conjoined with
could be anticipated in advance; (iii) why we currently sensed or remembered objects,
have a much more vivid conception of the doing so in the absence of any memory of
vastness of the ocean from vision than from those past occasions or conscious inference
hearing; (iv) why we are credulous, even in from them (THN / 103–4). And we
the face of contrary experience; (v) why we do not judge that these objects must exist
cannot take the infinite rewards and pun- but are instinctively impelled to form a more
ishments of an afterlife seriously, even if we ‘lively’ conception of them – a conception
believe in them; and (vi) why we enjoy reli- that ‘gives them more . . . influence; makes
gious discourses and dramatic performances them appear of greater importance; infixes
that excite the disagreeable passions of fear them in the mind; and renders them the gov-
and terror. In all of these cases the explana- erning principles of all our actions’ (THN
tion is the same. Though the natural relations / 629).
of contiguity and resemblance are not able to These are results that Hume trumpeted
produce belief on their own, when a belief in both the Treatise ( / 103) and the
has once been formed, so that there is no Enquiry (5.8 / 46–7), writing that because
sense of ‘caprice’ in its conception, it will be objects have no discoverable connection with
further enhanced by relations of contiguity one another, we can only draw an inference
(i) and resemblance (ii–iv) holding between from the one to the other with the aid of ‘cus-
the impression and the idea, but also weak- tom operating on the imagination’, and that
ened by the opposite relation of dissimilar- ‘all probable reasoning is nothing but a spe-
ity (v), with the weakening of belief in turn cies of sensation’ (THN / 103), so
accounting for (vi). that when we prefer one argument to another
we do nothing but decide from our feelings
concerning the superiority of their influence,
meaning that belief is ‘a species of natural
6. THE PSYCHOLOGICAL instincts, which no reasoning or process of
FOUNDATIONS OF EPISTEMOLOGICAL the thought and understanding is able, either
NORMATIVITY to produce, or to prevent’ (EHU 5.8 / 46–7).
In the Enquiry, Hume pretended that even
Taken together, the two parts of Hume’s though our beliefs are not drawn from obser-
‘system’ would appear to rule out any role vation in accord with truth-preserving or
for logic in probabilistic inference – any role probability-preserving rules, there is a ‘pre-
for the conscious, deliberate application of established harmony between the course of


9780826443595_Ch04_Final_txt_print.indd 117 2/9/2012 9:39:15 PM


nature and the succession of our ideas’ ensur- than one that is more immediate (THN
ing that beliefs will be produced in us in tan- / 144). And just as there are factors
dem with the way causes and effects succeed that lead us to overlook or discount connec-
upon one another in nature, and providing tions found in the past course of nature, so
‘those who delight in the discovery and con- there are factors that lead us to suppose the
templation of final causes’ with ‘ample sub- existence of connections that were not there.
ject to employ their wonder and admiration’ According to the theory, we are disposed
(EHU 5.21 / 54–5). This is a singular instance to consider similar objects to have similar
of misdirection, inconsistent with the cand- causes or effects. But any given object is a
our that is otherwise characteristic of his compound of many different characteris-
work. It is not just that, on his account, there tics, only one of which may be constantly
could at best be a pre-established harmony conjoined with a cause or effect. Even if
between the past course of nature so far as it we have learned to distinguish the essential
has been observed by us and the succession characteristic from the superfluous ones,
of our ideas. Hume’s account entails that when we encounter an object that resembles
there should not even be that much. a cause or effect only in superfluous ways,
According to Hume’s theory, belief is a the resemblance should lead us to conceive
more vivacious idea resulting from asso- the associated effect or cause, and the ease
ciation with an impression or memory. It of the association together with the vivac-
therefore depends on the original vivacity of ity of the encountered object should induce
the impression or memory and the strength a kind of bigoted or prejudicial belief,
of the associative link between the impres- which holds sway despite our recognition
sion or memory and the idea. If either of of the superfluity of the resemblance (THN
these is weakened, the belief will be as well. / 146–7 and / 147–8).
But, as Hume observed in the Treatise, an Nor are these the only such cases. Hume
impression is more vivacious than a memory, noted that education ‘not only approaches
and a recent memory more vivacious than an in its influence, but even on many occa-
older one. A recent observation of a conjunc- sions prevails over that [belief] which arises
tion between types of events also produces from the constant and inseparable union of
a stronger disposition to associate those causes and effects’ (THN / 116). He
events than an earlier one. As the Hume of further noted that we do not regulate our-
the Treatise went on to admit, these factors selves entirely by experience of the governing
entail that the course of our ideas should principles of human nature when deciding
not be in harmony with the past course of whether to believe testimony, but instead
nature. Instead, it should be disproportion- ‘have a remarkable propensity to believe
ately influenced by the most recent observa- whatever is reported, even concerning appari-
tions (THN–2 / 143–4). tions, enchantments, and prodigies, however
This is not all. Hume further observed contrary to daily experience and observa-
that the strength of association is also tion’ (THN / 113). This is in part
affected by the ease with which it is made, due to the influence of the resemblance rela-
so that a causal inference that needs to be tion between ideas (supposed to exist in the
drawn by appeal to a number of intermedi- minds of others on the basis of their words)
ate causes should be less strongly believed and facts, which strengthens the associative


9780826443595_Ch04_Final_txt_print.indd 118 2/9/2012 9:39:15 PM


relation beyond what is warranted by expe- a universal and conspicuous ‘weakness of

rience of their constant conjunction. But it human nature’ (THN / 112).
is also due to the fact that ideas that arouse This raises two problems. If all beliefs are
passions are reciprocally enlivened by those ultimately unjustifiable and all are founded
passions (THN / 120). on the same operation of a transmission of
This looseness of fit between the course of vivacity across associative links, how could
our ideas and the past course of nature is not any of us have come to think that some of
necessarily a bad feature of Hume’s ‘system’. them are better than others? And how could
As a matter of fact, people’s beliefs are more some of us (e.g. a gambler who places bets
strongly influenced by recent experiences; in accord with a calculation of the probabil-
people are less inclined to lend credence to ity of outcomes, ignoring the results of recent
the conclusions of complex arguments; and games) not only think that certain beliefs are
people are disposed to bigotry, blind adher- better than others but manage to form their
ence to received opinions, and credulity. The own beliefs accordingly? Hume had solu-
fact that the course of our ideas fails to track tions for both problems.
the past course of nature in just these ways is
further confirmation that Hume’s theory has
correctly captured the psychological mecha-
nisms responsible for human belief formation 7. EPISTEMOLOGICAL NORMS
and is a further instance of its explanatory
power in accounting for those inferences we In all cases we transfer our experience to
are in fact psychologically compelled to draw. instances, of which we have no experi-
But this is still not an entirely happy result. ence, either expressly or tacitly, either dir-
While many of us form blinkered, prejudicial, ectly or indirectly. (THN / 105)
obstinate and credulous beliefs, we do not all
do so. At least, we do not all do so all of the According to the account that has so far
time. At the very least, we do not all think been presented of Hume’s ‘system’, causal
we should do so, even if, as a matter of fact, inference is an unconscious (‘tacit’), instinct-
we find ourselves irresistibly compelled to do ive (‘direct’) operation resulting from, not car-
so anyway. As Hume himself remarked, infer- ried out in cognizance of, past experience. On
ences skewed by recent experience are not the first few occasions of observing objects of
‘receiv’d by philosophy as solid and legiti- one sort to be followed by objects of another
mate’ because ‘philosophers’ do not think sort, we are unimpressed, and not disposed
that the same event or the same conjunction to draw any inference when encountering
between events should provide less evidence a objects of either sort in the future. But as
month from now than it provides today (THN we make more and more observations of the / 143). Furthermore, education is conjunction of the two objects, we develop
not ‘recogniz’d by philosophers’ because it a habit to think of the one when presented
‘is an artificial and not a natural cause’, and with the other. As the number of observa-
because ‘its maxims are frequently contrary tions increases, the habit strengthens, and as
to reason, and even to themselves in differ- the habit strengthens, more and more of the
ent times and places’ (THN / 117). mental dispositions characteristic of belief
And credulity is, as Hume himself admits, and attendant upon experience and memory


9780826443595_Ch04_Final_txt_print.indd 119 2/9/2012 9:39:15 PM


come to attend the associated idea. Belief, is not entirely uniform, we seldom rely on a
therefore, is something that comes in degrees, gut reaction but instead deliberately recall
varying from conjecture to certainty in pro- the past experiments, count up the number
portion to the strength of the habit and hence of confirming and contrary instances, and
to the number of past observations up to the form beliefs that are stronger or weaker
point where a sufficient number of observa- in accord with a mathematical calculation
tions have been made to produce a habit that of probabilities.27 These are extraordinary
mimics experience in its effects (THN 1.3.12. claims that at first sight seem incompatible
2 / 130–1). Importantly, we do not recall the with the ‘system’. A habit cannot be formed
past observations or appeal to them to justify after just one experience, and a mathematical
our belief. The past observations have made result is based on the perception of a relation
us develop a habit and that habit alone pro- between ideas that have no vivacity, and so
duces the belief (THN / 103–4). should not produce belief.28
But there are twists to this simple story. Hume’s account of how we manage to do
One twist arises from the fact that past expe- these things lays the foundations for episte-
rience is not always uniform.26 Sometimes, mological norms and an account of action in
objects of one sort are not always followed accord with those norms.
by objects of another sort. When that hap-
pens, the contrary experiences weaken the
habit. Over time, we end up with a habit that
would be strong or weak in proportion to the 8. GENERAL RULES
number of confirming instances in the total
number of trials, but for the influence of the Hume claimed that while we do not as a
‘unphilosophical’ factors mentioned above matter of fact recall any past experiences
(THN / 132–3). when drawing inferences concerning con-
The belief we get from inconstant experi- junctions of causes and effects that have
ence is still ‘tacit’ and ‘direct’. But in Treatise been constantly observed since infancy (e.g. / 130–1, Hume declared that there stones fall, fire burns, water suffocates),
are other kinds of causal inference. The kind we do ‘assist the custom and transition of
based on the gradual development of a habit ideas’ by recalling past experiences when
over the course of a uniform past experi- encountering more rare or unusual objects
ence is not, he claimed, to be found in any- (THN / 104). There is nothing
one ‘who is arriv’d at the age of maturity’ about the system that would suggest we are
(THN / 131), and the kind based prevented from doing so. On the contrary,
on inconstant experience is one that ‘we have similar objects can jog the memory as well
but few instances of in our probable reason- as the imagination, particularly in unusual
ings’ (THN / 133). Mature adults cases. And we can be motivated to recall
are supposed to draw causal inferences after past instances by passions such as curios-
just one experience, supposing it has been ity, love of fame, fear and hope. Our causal
obtained in circumstances where ‘all foreign inferences could, therefore, be sometimes
and superfluous circumstances’ have been ‘express’ rather than ‘tacit’.
removed. And Hume maintained that when Just as nothing prevents us from expressly
the conjunction between causes and effects recalling and reviewing past experiences, so


9780826443595_Ch04_Final_txt_print.indd 120 2/9/2012 9:39:15 PM


nothing prevents us from drawing conclu- effects. Instead, we need to be sure that we
sions from those past experiences ‘indirectly’, have eliminated ‘all foreign and superflu-
by explicit appeal to causal rules, learned ous circumstances’ as Hume put it (THN
from past experience. The chief such rule / 104). We first consider it to be
is the general one that like objects, placed at least possible that where two resembling
in like circumstances, will have like effects objects have different effects, all the simi-
(THN / 173). This rule is not jus- larities between them must be foreign and
tified by past experience – no causal rule is. superfluous and the different effects must
But past experience does lead us to form it arise from some hidden respect in which the
and believe it. It is ‘merely habitual’ as Hume causes differ. As it turns out, our experience
put it (THN / 104–5). Once formed, of regularly finding these hidden features
it can be expressly appealed to in order to upon a more exact scrutiny habituates us
justify causal inference from as little as a sin- to that belief (THN / 132). A simi-
gle past experience (THN / 105). lar course of experience habituates us to the
This accounts for why people who have belief that where strikingly different objects
reached the age of maturity will draw causal have the same effect, all the differences
inferences after just one experience rather between them are foreign and superfluous
than needing to be trained by a number of and the effect must arise from some circum-
experiences. stance common to the two. We are further
‘Express and indirect’ causal inferences habituated to believe that any increase or
can be based on other rules besides the diminution in the effect must be due to a
rule that similar objects placed in similar compound cause, and that any cause that
circumstances will have similar effects. As persists over time before being followed
has already been noted, any striking resem- by its effect cannot be the sole cause of
blance between an unfamiliar object and a that effect (THN–12 / 149–50,
familiar one will lead us to suppose that the–10 / 174–5). These ‘rules by which
unfamiliar object has the same causes or to judge of causes and effects’ justify other
effects as the familiar one, even if there are causal inferences. Taken together, they con-
some dissimilarities between the objects and stitute a logic of causal inference (THN
the attendant circumstances are not quite the / 175) – a system of rules that can
same. This is the foundation of prejudice. As be expressly appealed to in drawing indirect
we grow older, we encounter cases in which or ‘oblique’ causal inferences.
our prejudices are disconfirmed and reflec- This is one part of the answer to the ques-
tion on these cases leads us to appreciate tion of how Hume could provide for a logic
the importance of distinguishing between of causal inference.29 Once we have come to
superficial characteristics, which are often form and accept the rules, inferences that are
but not always present in causal conjunc- in accord with them will be approved of as
tions, and essential ones, which are always wise or justified, whereas those that violate
present (THN–12 / 149–50). We them will be condemned as foolish. This will
begin to think that it is not good enough to be the case even if the person making the
suppose that any object that bears any strik- normative assessment is personally unable
ing resemblance to objects we have experi- to follow the rules due to the influence of
enced before will have the same causes or ‘unphilosophical’ factors.


9780826443595_Ch04_Final_txt_print.indd 121 2/9/2012 9:39:16 PM


9. PROBABILITIES or resources to search for it, we have no

recourse but to consider those circumstances
Hume noted that while past experience may that tend to accompany it to be signs of its
have habituated us to the thought that same likely existence, and so to take apparently
objects have same effects, we do not always similar objects to have similar effects, albeit
find this to be the case. Contrary experience with a diminished degree of certainty pro-
does not, however, lead us to diminish our portioned to our experience of the degree of
confidence in the general rule because we regularity in the connection between occur-
have been further habituated to believe that rences of the superfluous circumstances and
where same objects have different effects the effect (THN / 132–3). These
they will be found upon more exact scrutiny inferences are ‘probable’ in the strict sense
to differ in some previously unnoticed way. of being something less than ‘entirely free
We may not always have the opportunity, of doubt and uncertainty’ (THN
the curiosity, the leisure, or the resources to / 124). Importantly, they are not ‘tacit and
search for this hidden circumstance. But this direct’. They do not result from a habit that
just means that, in the absence of a percep- has been strengthened or weakened by past
tion of a previously hidden cause, our habits experience – but also by a host of ‘unphilo-
will lead us to believe that it must exist. The sophical’ factors. They are instead ‘oblique’
case here is similar to what Hume ought to or explicit and indirect. They result from a
have said concerning the spatial contiguity survey of past instances and a calculation of
of causes and effects. If we fail to observe a the proportion of confirming experiments in
succession of cause and effect because clos- the total number of trials.
ing our eyes or turning our heads leads us to It has already been noted why we might
fail to observe the contiguous regions, then be impelled to remember past experiences
we do not think we have observed a fail- and want to survey them. Hume’s remain-
ure of the expected succession to occur. On ing challenge in accounting for our ‘explicit
the contrary, this is exactly the sort of case and indirect’ reasoning concerning matters
in which the habit kicks in and leads us to of strict probability was not to come up
form a belief in the unperceived existence of with a mathematical theory of the calcula-
the cause or effect at the contiguous location. tion of probability. It was to explain how we
The only difference between the two cases come to proportion belief in accord with any
is that this time Hume did not fumble, as he merely mathematical calculation.
did at Treatise / 197–8, but declared The basic facts that need to be accounted
that, ‘from the observation of several parallel for are obvious: in cases where there has
instances, philosophers form a maxim, that never been an exception to a succession, we
the connexion betwixt all causes and effects should form a belief that is ‘entirely free of
is equally necessary, and that its seeming doubt and uncertainty’; in cases where it is
uncertainty in some instances proceeds from no more likely that an event will occur than
the secret opposition of contrary causes’ not, we should have no belief, and in the
(THN / 132). intermediate cases we should have a pro-
But, Hume went on to observe, even though portionally strong or weak belief. So, where
we may believe there is a hidden cause, in the there is what we might call a ‘fifty-fifty
absence of opportunities, curiosity, leisure chance’, we should have no belief; where


9780826443595_Ch04_Final_txt_print.indd 122 2/9/2012 9:39:16 PM


there is a ‘one hundred per cent chance’ we the generic effect will occur, but indifferent
should have certain belief; and where there over which form it will take. For example,
is a ‘seventy-five percent chance’ we should the belief that the die will fall with ‘side one’
have a belief that is half way between indif- facing up would be one-sixth the strength of
ference and certainty. In the last case, import- the belief that it will fall with some side or
antly, we do not form a certain belief in the other facing up – but for the fact that we can-
proposition that the event has a probability not have any degree of belief that ‘side one’
of fifty (or seventy-five) percent;30 we form a will fall face up without disbelieving that
less vivacious conception of the event – one any of the other sides will fall face up. Since,
that is half-way between a certain belief and however, all those possibilities are considered
an indifferently entertained idea as measured to be equally likely, they cancel one another
by the feeling that attends it and the strength out, leaving us with certainty that some side
of the characteristic mental dispositions. We or other will face up, but indifference about
should get a belief of this strength regardless which one.
of what our views are on how to calculate Hume’s account of how we form beliefs
probabilities mathematically, or whether we about the effects of inconstant causes grows
employ any mathematics at all beyond a bare out of a refinement of the account of belief
survey of instances. in chances. If some of the chance alternatives
In explaining how the belief arises Hume resemble one another, e.g. four of the six
distinguished between two cases, that of sides have the same figure on them, a sur-
belief in ‘chances’, and that of belief in vey of the alternatives causes their portions
‘inconstant causes’. Chance arises when a of the original belief to combine. In that case
cause has an effect that can indifferently take the alternatives do not perfectly cancel one
one or another of a finite number of alterna- another out, e.g. the possibility that the die
tive forms, e.g. the toss of a die causes it to will land with the more familiar figure facing
fall with one or other of its six sides facing up receives four of the six portions, only two
up. Causes are ‘inconstant’ when either the of which suffice to cancel the rival possibili-
indifference condition or the finitude condi- ties, leaving us with a residual belief that the
tion is not met. more familiar figure will face up. This resid-
In the first case we have been strongly ual belief is two-sixths of the way between
habituated to associate the generic effect with indifference and conviction.31
the cause. So, for example, upon witnessing a Belief in the outcome of inconstant causes
die being tossed we form a strong belief that is like this, except that in this case we do not
it will fall with one side facing up. But now start off with a strong belief in the occurrence
suppose we ask ourselves exactly which side of a generic effect. Hume instead supposed
will face up. Because the generic effect has a that we develop the habit of expecting the
number of equally possible, mutually incom- future to be like the past. The strong belief
patible forms the strong belief in the generic in this uniformity principle plays the same
effect becomes divided, with an equal por- role as the strong belief in the occurrence of
tion going to each alternative. But, because a generic effect. If in the past one egg in every
the alternatives are all mutually incompat- crate has been rotten, we will transfer that
ible, the divided beliefs cancel one another past experience to the future in the sense that
out. We are left with a very strong belief that any subsequent impression or memory of a


9780826443595_Ch04_Final_txt_print.indd 123 2/9/2012 9:39:16 PM


crate of eggs will produce a strong belief in possibilities from more frequent ones reap-
the existence of eleven sound eggs and one pears at EHU 10.4 / 110–11.
rotten one. On picking out any one egg from Whatever ambivalence Hume might or
the crate and forming an idea of what we might not have had about his account, he
will smell upon cracking it open, that strong suggested in the Treatise ( / 7–8) and
belief is divided into twelve packets, eleven stated outright in the Abstract (4 / 646–7) that
of which resemble one another in being he had offered the only adequate explanation
attached to ideas of sound eggs, and one of probable belief that had ever been given.
of which is attached to the idea of a rotten There could be no demonstrative account
egg. Since the possibilities are incompatible, of probable belief because it is in principle
we cannot believe them both to result from impossible to demonstrate that the event that
cracking the egg. Since one of them comes is most often observed will occur (in that case,
up so much more often in a mental survey it would not be merely probable). There can
of alternatives, we cannot be perfectly indif- be no probable account of probable belief
ferent, either. The odd possibility is cancelled because those who claim that a survey of
out by one of the opposing packets leaving past results can at least make us certain about
us with a belief that this egg will prove to which event is most likely to occur are actually
be sound that is ten-twelfths as strong as the doing no more than uttering the trivial claim
original belief that there will be eleven sound that the event that has come up most often
eggs in the crate. in a survey of past results is the event that
There are oddities about Hume’s presen- has come up most often in a past survey of
tation of this account in both the Treatise results, not giving us a reason to believe, with
and the Enquiry. The account offered by the any degree of conviction, that this event will
Treatise is repeated four times over (first at occur on any future occasion.32 This is further
THN–12 / 133–5, a second time at proof, in Hume’s eyes, that belief must be a–18 / 135–7, again at more vivacious manner of conception rather
/ 137–8, and a fourth time at–2 than the product of a judgement.
/ 138–40). Whether this is because Hume
was unsure of himself or particularly proud
of his result is unclear. In Enquiry 6 / 56–9
he retreated from the attempt to provide 10. ‘PHILOSOPHICAL’ BELIEF
a calculus of the strength of belief, at first
attributing probabilistic belief to ‘an inex- By all that has been said the reader will
plicable contrivance of nature’ but then say- easily perceive, that the philosophy con-
ing that his account of belief as a firmer and tained in this book is very sceptical, and
stronger conception of an object allows us to tends to give us a notion of the imperfec-
explain matters a bit further by saying that tions and narrow limits of human under-
standing. (Abs. 27 / 657)
where ‘a great number of views . . . concur
in one event’ they ‘fortify’ the conception of
that event in the imagination and so pro- Hume had no sooner presented his account
duce belief. This reticence notwithstanding, of how we arrive at weaker or stronger
the full theory of belief as proportioned in beliefs about matters of probability than he
accord with the subtraction of less frequent remarked that recent experiments weigh more


9780826443595_Ch04_Final_txt_print.indd 124 2/9/2012 9:39:16 PM


than earlier ones in our assessments of the philosopher’s commitments have no other
probability of outcomes, whether we realize foundation than (at best, and not always)
it or not. The freshness of an experiment ‘has the past course of their own experience – a
a considerable influence on the understand- foundation that the vulgar can appeal to as
ing, and secretly changes the authority of the well. The difference between the vulgar and
same argument, according to the different the philosophers is that they have had differ-
times, in which it is propos’d to us’ (THN ent experiences. Each judges in accord with / 143). He also said that this is an what their own experiences have made them
effect that ‘has not had the good fortune’ to and neither is in a position to appeal to their
be ‘receiv’d by philosophers, and allow’d to experiences as a higher authority.
be [a] reasonable foundation of belief and Nor is this the only impediment to ‘philo-
opinion’ (THN / 143). As already sophical’ belief. Just because ‘philosophers’
noted, this is just one of many ‘unphilosoph- approve of the ‘rules by which to judge of
ical’ influences on belief, influences that we causes and effects’, it does not follow that they
are as a matter of fact susceptible to, for rea- will always form their beliefs in accord with
sons that Hume’s ‘system’ succeeds admir- those rules. Very few of us have been habitu-
ably well at explaining. ated to think (1) that whenever a supposed
We might conjecture that ‘philosophers’ cause fails of its usual effect a hidden cause
are led to condemn these influences as a con- will be discovered upon more exact scrutiny,
sequence of having been habituated to accept (2) that the bare passage of time does not des-
certain rules, such as the rule that explicit troy the authority of an experiment, or (3) that
and indirect probable inferences are more beliefs that are based on recent, lurid anec-
often correct than tacit and indirect ones. dotes will more often prove to be false than
But not everyone is habituated to accept the those that are based on an impartial survey of
same rules. Hume noted that ‘the vulgar’ are cases. The reason most of us accept these and
not habituated to accept that same causes other such rules is not personal experience in
must have same effects or that when a cause the laboratory, but uncritical acceptance of
fails of its usual effect a more exact scrutiny the testimony of others, education or the influ-
will uncover some previously hidden circum- ence of a few, recent, notable errors. That is
stance that is the true cause. Instead, their perhaps not a bad thing. Ironically, if anything
experiences have habituated them to accept enables ‘philosophers’ to draw their infer-
that causes are not perfectly regular in their ences just in accord with the rules by which
operations, even though nothing impedes to judge of causes and effects, it is not having
them (THN / 132). Someone with been habituated to accept them but having
that outlook will be less inclined to recog- been educated to accept them.33 Education is
nize a distinction between superfluous and among the most illegitimate but also the most
essential components of causes (hence more powerful of the factors influencing our belief.
inclined to prejudice), and less inclined to Personal past experience conveys a degree of
accept that the course of nature cannot vivacity to associated ideas that is vulnerable
change (hence more inclined to form their to being artificially enhanced or diminished
beliefs on the basis of recent evidence). And by resemblance, contiguity, passion or the
a philosopher is in no position to convince passage of time, whereas education is largely
them of the error of their ways since the impervious to those influences.


9780826443595_Ch04_Final_txt_print.indd 125 2/9/2012 9:39:16 PM


Hume’s position is not entirely sceptical. an experience of the force of sceptical argu-
Although he never said so in quite so many ments.36 He thought that someone who has
words (but see THN / 271–2), he once been convinced of the weakness and
would probably have agreed that a survey fallibility of our powers of knowledge will
would show that ‘philosophical’ beliefs have be permanently changed by that experience.
more often turned out to be correct than Forever afterwards, they will be doubtful
‘vulgar’ ones. The rules followed by ‘phi- about all their beliefs and hesitant about
losophers’ in arriving at their beliefs, there- forming them. This doubt and hesitancy will
fore, have a title to be considered the logic naturally dispose them to distrust testimony
of probable reasoning rather than just an and education and refrain from peremptory
anthropological description of the epistemo- (tacit, direct) judgement (EHU 5.1 / 40–1,
logical aspirations of a certain social class.34 12.24–6 / 161–3; DNR 1.133–4). It will also
But the rules have this title only for those extend to philosophical beliefs, as Hume
already habituated to accept that the future made clear in the last words of Treatise 1.4.7
will be like the past, and to accept that expli- / 274.
cit and indirect probable inferences, based
on an impartial survey of past cases, are to . . . we are apt not only to forget our scep-
be preferred to tacit and indirect ones. Those ticism, but even our modesty too; and
not habituated (or educated) to accept these make use of such terms as these, ’tis evi-
dent, ’tis certain, ’tis undeniable; which a
rules could justly reject any appeal to a sur-
due deference to the public ought, per-
vey of past instances as question-begging.
haps, to prevent . . . . such expressions
And even those who do accept the presup- were extorted from me . . . , and imply no
positions of the argument will not always be dogmatical spirit, nor conceited idea of
able to resist the other factors inducing them my own judgement, which are sentiments
to form beliefs. ‘Unphilosophical’ belief will that I am sensible can become no body,
persist, even among those who know better. and a sceptic still less than any other.
This explains the characteristic pessimism
displayed in Hume’s Essays, Natural History But a sceptical disposition will at least
of Religion, and History of England.35 Under put philosophical beliefs on an equal footing
certain social conditions the arts and sciences with ‘unphilosophical’ ones, where they have
will flourish and philosophical learning will a greater chance of winning assent after due
triumph over vulgar superstition. But once consideration.
entrenched in a society, there is no guaran- Seen in this light, Hume’s account of the
tee that the arts and sciences will progress. ‘unphilosophical’ influences we are subjected
A change in circumstances, beyond anyone’s to and of the ultimate lack of any founda-
ability to control or even predict, altering the tion even for ‘philosophical’ belief constitutes
course of people’s experiences, can change the best sceptical lesson, and thereby the
their inferential practices and the intellectual best lesson in logic, that anyone could have.
culture that had developed can be supplanted In the Treatise Hume overplayed this result.
by the crudest barbarism. Realizing the salutary effects that a sceptical
Hume did think that there is one means disposition could have, he set out to blast
by which the influence of ‘unphilosoph- his readers with the most extreme – but also
ical’ factors on belief might be mitigated: the most strained – sceptical arguments he


9780826443595_Ch04_Final_txt_print.indd 126 2/9/2012 9:39:16 PM


could invent. Not content to raise sceptical the idea of what, specifically, connects causes
doubts about causal inference in Treatise 1.3 necessarily to their effects.
Combined with either an identification of
/ 69–179, Hume went on to raise sceptical
objects with perceptions of objects (as at THN
arguments against the existence of an external 1.4.2 / 187–218 or EHU 12 / 149–165) or an
world, against the existence of persisting sub- appeal to a conceivability criterion of meta-
stances, and even against the validity of dem- physical possibility (as at THN / 233),
onstration and the existence of a self. It is as this argument entails the independence of all
objects occupying distinct locations in space
if he thought that his reader needed first to be
and/or time, and so rules out the existence
driven into a deep sceptical crisis in order to of necessary connections between causes and
be adequately prepared to undertake a prop- effects. The extent of Hume’s commitment
erly scientific study of the foundations of mor- to the metaphysical impossibility of neces-
als in the passions, as taken up in Treatise 2 sary connections between causes and effects
remains controversial. For discussion, see
and 3.37 But he seems to have quickly realized
H. Beebee, Hume on Causation (London:
that the excessive sceptical arguments had the Routledge, 2006) and her contribution to this
opposite effect. By contesting received opin- volume, or the papers collected in R. Read and
ions too forcefully, he had only led readers to K.A. Richman (eds), The New Hume Debate
reject the Treatise as a whole. The Enquiry (London: Routledge, 2000).
For more on Hume’s view of intuition and
takes a different tack. Though it mentions rea-
demonstration, see D. Owen, Hume’s Reasons
sons for scepticism both about the existence (Oxford: Oxford University Press, 1999),
of an external world and about the validity chap. 5.
of probable reasoning, it also discounts them At THN / 90–1 Hume stressed that in
both, presenting them merely as a means to allowing for the possibility of hidden powers
he was making a concession to his opponent
inducing a properly scientific attitude. Hume
for the sake of argument, not admitting that the
seems to have realized that there is no better possibility is a real one.
sceptical argument than the presentation of 7
The reader who has been keeping track of the
the ‘system’ with its consequences. The ‘sys- references will note that I have juggled Hume’s
tem’, which diagnoses the problem, also cures order of presentation. This avoids the repetition
of the same arguments under different headings
it by means of that very diagnosis.
that mars the exposition of the Treatise.
Here what I have charitably described as a
‘meditative path of discovery’ takes on a rhet-
orical dimension. Prepare your reader to accept
NOTES what you have to say by first inducing a deep
sense of puzzlement. Then offer what you have
For a critique of such reasons as Hume gave to say as a solution to the puzzle and trust to
(at THN / 73–4), see L. Falkenstein and the reader’s sense of relief to induce acceptance
D. Welton, ‘Humean Contiguity’, History of of your solution, even in the absence of sup-
Philosophy Quarterly 18 (2001), pp. 279–96. porting argument.
Hume expressed some doubts about whether Like necessary connection, the nature and role
spatial contiguity and succession are always of reasoning in causal inference is controver-
necessary for causality, but considered the ques- sial. For discussion, see P. Millican, D. Owen
tion moot (THN / 75 and / 76). and D. Garrett in the symposia on Garrett’s
THN / 162 might be read as suggest- Cognition and Commitment (Oxford: Oxford
ing that we have no idea of necessary connec- University Press, 1997) printed in Hume Studies
tion. But Hume’s concern there was not to deny 24 (1998), pp. 141–59, 171–94 and Philosophy
that we have the idea of a connection, or even and Phenomenological Research 62 (2001),
of a necessary connection, but just that we have pp. 191–6, 205–15.


9780826443595_Ch04_Final_txt_print.indd 127 2/9/2012 9:39:16 PM


I here amplify on Hume’s actual statement. no vivacity. My belief that I have correctly
When discussing belief, Hume did not mention intuited this relation does.
the spatial and temporal contiguity condi- This view of the relation between the works
tions he previously identified as involved in is controversial. For an opposed view see
the causal relation. He spoke just of belief in P. Millican, ‘The Context, Aims and Structure
the existence of an object, not of belief in its of Hume’s First Enquiry’, in P. Millican (ed.),
existence at a prior or subsequent time and at Reading Hume on Human Understanding,
a contiguous place. This omission has momen- pp. 27–65, esp. pp. 40–8.
tous consequences, generating pseudo-prob- Unlike the Treatise, the Abstract contains no
lems, most notably at THN / 197–8. explicit pronouncement to the effect that in
An integrated account of the four features is speaking of ‘[t]he powers, by which bodies
presented at THN / 628–9. operate’, Hume was indulging common but
THN / 630–1 grapples with a fur- false ways of thinking (cf. THN–10 /
ther problem that might have exercised Hume: 90–1). In contrast, the parallel discussion in
works of fiction focus attention and arouse the Enquiry contains a qualification, this time
passions without prompting belief. For further occurring in a footnote, to the effect that the
discussion of the problems with and prospects talk of hidden powers is ‘loose and popular’
for including mental dispositions within the and that a ‘more accurate explication’ of the
larger framework of Hume’s theory, see J. notion would further buttress the conclusion
Bricke, Hume’s Philosophy of Mind (Princeton: to be drawn here.
Princeton University Press, 1980), chap. 3; As noted earlier, the exposition of the Treatise
S. Everson, ‘The Difference between Feeling is made more elegant by bringing this after-
and Thinking’, Mind 97 (1988), pp. 401–13; thought forward.
L. Loeb, Stability and Justification in Hume’s Compare THN 1.3.16 / 176–9, which appeals
Treatise (Oxford: Oxford University Press, to the abilities of animals as a further reason to
2002), pp. 60–100; and J. Smalligan Marušić, accept the account of belief.
‘Does Hume Hold a Dispositional Account As in the Treatise Hume continued, in both the
of Belief?’, Canadian Journal of Philosophy Abstract and the Enquiry, to omit this impor-
40 (2010), pp. 155–83. For further discussion tant detail.
of Hume’s ambivalence about his account See note 8 above.
of belief see M. Bell, ‘Belief and Instinct in The interested reader is invited to consult
Hume’s First Enquiry’, in P. Millican (ed.), THN-C: lxv–lxvii, which sets out the extent
Reading Hume on Human Understanding of Hume’s quotations from the Treatise in
(Oxford: Clarendon Press, 2002), pp. 175–85. Enquiry 5, Part 2 (EHU 5.10–22 / 47–55) and
Hume was later to argue that intuition and his modifications to those passages.
demonstration reduce to probability (THN This is obviously intended.
1.4.1 / 180–7). But even then his claim was Objects can of course move around. But
not that we do not intuit relations between just as we do not assume that two objects
ideas or demonstrate truths in mathematics are relatively immovable upon having once
by appeal to a chain of intuitions. It was seen them alongside one another, so we do
that because we sometimes have the wrong not assume that they are causally related
intuitions, our assurance of the results of upon once having seen them in succession.
a demonstration has to be informed by And just as the bare experience of a constant
considerations of how likely it is that we are conjunction in time suffices to impel us to
mistaken. Intuition and demonstration do not associate them independently of any further
‘reduce’ to probability in the sense of turning justification by appeal to secret powers
out to be nothing other than more vivacious producing a necessary connection, so the
conceptions of an idea. They ‘reduce’ to bare experience of a constant conjunction in
probability in the sense of presupposing space suffices to impel us to associate them
second-order beliefs about the reliability independently of any further justification
of our intuitive judgements. My intuitive by appeal to the causes of their mobility or
judgement that eight plus seven is fifteen has immobility.


9780826443595_Ch04_Final_txt_print.indd 128 2/9/2012 9:39:17 PM


The effect of resemblance in raising religious cause and effect are present, the case is one of
devotion is of this sort. The icons and cere- perception rather than causal inference (THN
monies enliven an antecedent belief in the past / 73). The cases where one turns
existence of people and events, the exception one’s head or closes one’s eyes are precisely
being that in this case the antecedent belief is the occasions on which causal inference is
grounded on testimony rather than memory. called for. Had Hume been right at THN
Belief is only enhanced by the experience of the /197–8, there would be no such
icon or ceremony, not created. thing as causal inference. There would only
This is the first appearance of the important be perceptions of regularities in the succession
notion of correction by appeal to general rules. of causes and effects and perceptions of the
This attempt to distinguish causality from con- failure of those regularities to occur, without
tiguity and identity is a failure. When one adds any attendant instinct to form a belief on
a specification of direction and distance to the latter occasions. The ‘system’ rules that
contiguity relations, and of temporal distance possibility out without any need to invoke
to identity relations, they become as restrictive the elaborate mechanism proposed at THN
as causal relations. Any given object is causally–43 / 199–210 to provide for belief in
related to a huge number of others as well, if the continued existence of objects when not
we do not consider whether the objects lie in perceived.
the direction of cause or effect, or distinguish These pronouncements seem inconsistent with
between proximate and remote causes and THN–14 / 103–5, which declares
effects. that ‘in all the most establish’d and uniform
When I speak here of a lack of uniformity in conjunctions of causes and effects, such as
past experience, I mean a verified lack of uni- those of gravity, impulse, solidity, &c. the mind
formity, where careful scrutiny of the contigu- never carries its view expressly to consider any
ous regions is unable to uncover any evidence past experience’, illustrating the claim with
of the existence of the inferred object, not an the point that a traveller who runs into a river
unverified lack of uniformity, where one fails does not consult past experience when forming
to observe a cause or effect simply because the belief that walking out on the water will
one failed to look for it or (as in the case of be followed by sinking and suffocating. The
historical inference) was in no position to two passages can be reconciled if Hume’s point
observe it. Hume at one point grossly over- in THN / 131 and / 133 is
stated the extent of the lack of uniformity in taken to concern just those causal inferences
our experience, pretending that the turning of involving new, rare and unusual objects and
the head or closing of the eyes could prevent the formation of new causal laws. Our every-
us from considering a succession between day inferences concerning familiar objects can
two types of objects to be perfectly constant be considered to proceed in accordance with
(THN / 197–8). This is an artifact of habits learned in infancy.
a mistake lamented in a number of previous For Hume, mathematical results are matters
notes: Hume’s persistent neglect of the point of knowledge, not belief, where the knowledge
that, according to his own theory, causes and arises from not being able to conceive things in
effects are not merely conceived to exist, but any other way (THN / 124, cf.
to exist at a certain place at a certain time. If / 95). One can have knowledge without belief
I am habituated vividly to imagine a cause or because belief involves attention, elevation of
effect in one place, and I consider myself to passion and inclination to action, and merely
have turned my head to look at another place appreciating the impossibility of conceiving
or to have closed my eyes and not be looking things any other way need have none of those
at all, I am not going to suppose that I have results.
experienced a failure of my expectations. As For earlier versions of the account given here
Hume himself pointed out elsewhere, when see F. Wilson, Hume’s Defence of Causal
we reason from causes to effects or effects to Inference (Toronto: University of Toronto
causes, it is always the case that the object Press, 1997), chap. 2, esp. pp. 123–40 and
we reason to is unperceived. Where both W.E. Morris, ‘Belief, Probability, Normativity’,


9780826443595_Ch04_Final_txt_print.indd 129 2/9/2012 9:39:17 PM


in Saul Traiger (ed.), The Blackwell Guide to For notable instances see E 135–7; E 528–9;
Hume’s Treatise (Malden: Blackwell, 2006), H 5.67; and NHR as a whole.
pp. 77–94, esp. pp. 85–91. For more on this, see D.F. Norton, ‘How
Hume denied this at THN / 127. a Sceptic May Live Scepticism’, in J.J.
If we calculate probabilities in the common way, MacIntosh and H.A. Meynell (eds), Faith,
this one-third belief would correspond to a two- Scepticism, and Personal Identity (Calgary:
thirds probability. The point to keep in mind is The University of Calgary Press, 1994),
that Hume was not concerned about accounting pp. 119–39, esp. pp. 128–32.
for the mathematical probability of an outcome, This idea has been pursued in more detail by
but for the strength of belief in that outcome. Loeb, Stability and Justification in Hume’s
For further commentary on this argument Treatise, pp. 6, 36, 215–29. However, Loeb
see C. Howson, Hume’s Problem is inclined to attribute Hume’s overstated
(Oxford: Clarendon Press, 2000), p. 14 scepticism to a ‘somewhat perverse’ (ibid., p. 16)
and chap. 4. aspect of his ‘temperament’ (ibid., p. 229) rather
For another ironic twist, see THN / than an attempt to cause the reader to become
149–50. more reflective and hesitant about all his or her
In this I follow Beebee, Hume on Causation, beliefs.
pp. 71–4.


9780826443595_Ch04_Final_txt_print.indd 130 2/9/2012 9:39:17 PM

Helen Beebee

1. INTRODUCTION each interpretation. Roughly speaking, these

interpretations take Hume to be, respectively,
It is difficult to say anything about Hume’s a regularity theorist, a non-cognitivist, and a
views on causation and necessary connection realist about causation. Finally, in section 5,
without making claims that are hotly dis- I sum up the current state of play, which, as
puted amongst interpreters of Hume’s work. I see it, is something of a stand-off between
Some interpreters take Hume to be a causal the projectivist and sceptical realist interpret-
realist, while others hold that he is a regular- ations, and say something about the specific
ity theorist. Some take him to hold that ‘caus- problems that each interpretation needs to
ation’ is an irretrievably defective notion that overcome if it is to prevail over its rival.
could not possibly apply to any worldly phe-
nomena, and some take him to be a non-cog-
nitivist about our causal talk and thought.
Some take him to hold that there is such a 2. HUME’S BASIC ACCOUNT OF
thing as objective necessary connection, while CAUSATION
others take him to be a subjectivist about
necessity. And so on. In this chapter, I shall For Hume, causal thinking lies right at the
strike a path through these and other inter- heart of our conception of the world: all ‘rea-
pretative controversies as follows. I begin sonings concerning matters of fact and exist-
in section 2 by sketching, in what I hope is ence seem to be founded on the relation of
a reasonably interpretatively neutral way, Cause and Effect. By means of that relation
the bare bones of Hume’s account of caus- alone we can go beyond the evidence of our
ation, and in section 3 I discuss his famous memory and senses’ (EHU 4.3 / 26). In other
‘two definitions’. In section 4, I outline the words, our access to external reality, beyond
three main classes of interpretative position – the ‘evidence’ of current experience and
what I shall call the traditional, sceptical memory, entirely depends on reasoning from
realist and projectivist interpretations – and causes to effects (and vice versa). When I form
briefly examine the main items of evidence the belief that my dinner will not poison me,
that are normally marshalled for and against or that what I am reading in the newspaper


9780826443595_Ch05_Final_txt_print.indd 131 2/9/2012 9:34:59 PM


is true, or that the kettle I turned on a few is simply a mental mechanism that, given
minutes ago will have boiled by now, I do so relevant past experience, conveys the mind
on the basis of such reasoning. from the impression of a C (one billiard ball
This being so, the primary task that Hume striking another, for example) to a belief that
sets himself is to uncover the nature of this an E will follow (the second ball moving).
mysterious relation – or rather, to discover And it turns out, in ‘Of the idea of necessary
what our idea of causation consists in, since it connexion’ (THN 1.3.14 / 155–72; EHU
is ‘impossible to reason justly, without under- 7.26–30 / 73–9), that it is the operation of
standing perfectly the idea concerning which this very mechanism that furnishes us with
we reason; and ’tis impossible to understand the impression, and so the idea, of neces-
any idea, without tracing it up to its origin, sary connection. Before Hume can establish
and examining that primary impression from this latter claim, however, he needs to show
which it arises’ (THN / 74–5). He that we have no sensory impression of nec-
quickly discovers that ‘whatever objects are essary connection. Before briefly rehearsing
consider’d as causes and effects are contigu- his argument, it is worth saying something
ous’ (THN / 75) and are also such that about the importance of the issue for Hume.
the cause is temporally prior to its effect. But One of Hume’s main aims is to provide a
those two conditions clearly do not exhaust ‘science of man’: an account of how the mind
our idea of causation, for ‘there is necessary works that is based on a clear-headed, ‘exper-
connexion to be taken into consideration’ imental’ investigation. A plausible account of
(THN / 77). his primary target is given by Edward Craig,1
Hume’s search for the impression-source who sees Hume’s major adversaries as those
of the idea of necessary connection is a long who uphold what he calls the ‘Image of God’
one, for it turns out that this crucial com- doctrine (we are made in God’s image) and a
ponent of ‘the idea concerning which we corresponding epistemological doctrine: the
reason’ has its source in that very reasoning ‘Insight Ideal’. According to the Insight Ideal,
itself. So Hume needs to discover the nature the nature of reality – or at least some of
of our reasoning from causes to effects (what it – is in principle accessible to reason; thus
I shall call ‘causal reasoning’) before he can philosophers before Hume had subscribed
locate the impression-source of the idea of to the self-evident or a priori status of the
necessary connection. claim that every event has a cause, or had
Hume’s investigation into causal reason- claimed that the essence of objects can be
ing – what is traditionally described as his known by what Descartes calls ‘purely men-
discussion of the ‘problem of induction’ – tal scrutiny’.2 Hume, by contrast, sets out to
yields two results that are significant for undermine systematically the claim that any
present purposes. First, causal reasoning aspect of the nature of reality is knowable a
proceeds on the basis of past observed regu- priori, and, moreover, to show that no ‘mat-
larity: on observing a C, we infer that an E ter of fact’ can be inferred a priori from any
will follow just when we have experienced other distinct matter of fact.
sufficiently many Cs followed by Es. Second, Hume takes himself to have established
that reasoning proceeds not by considera- this claim in his discussion of causal reason-
tion of any argument, but as a matter of ing; but it is a claim to which he returns in
‘custom or habit’ (EHU 5.5 / 43). There the negative phase of his discussion of the


9780826443595_Ch05_Final_txt_print.indd 132 2/9/2012 9:35:00 PM


idea of necessary connection (EHU 7.1–25 / entail that there is no observable connection
60–73; THN–18 / 155–65), where or ‘tie’ between cause and effect (necessity1).
he argues that the idea does not have a sen- It might easily be that we can observe such
sory impression-source. Hume assumes, in a connection – that is, we can observe the
this discussion, that such an impression- causal relation – despite the fact that we can-
source for the idea of necessary connection not observe any power, in the cause itself,
would have to be such that it delivers cer- that produces certainty about the effect.
tainty that the effect will follow: ‘were the It seems that Hume does indeed run these
power or energy of any cause discoverable two distinct notions together; however, by
by the mind, we could foresee the effect, his own lights it is not clear how much of
even without experience [of past constant a problem this is. Hume’s central concern,
conjunction]; and might at first, pronounce remember, is with inference from one mat-
with certainty concerning it, by the mere ter of fact to another – that is, from causes
dint of thought and reasoning’ (EHU 7.6 / to effects (and vice versa). The observability
63). With this assumption (to which I return of necessity1 – of a mere causal ‘tie’ between
below) in place, it is an easy matter to estab- causes and effects – would make no substan-
lish Hume’s negative conclusion, since, as we tive difference to the account that Hume has
already know from his discussion of causal already offered of such reasoning, since such
reasoning, observation of a particular event a tie could only be observed by, as it were,
never delivers such certainty: we can always observing c-causing-e as a package deal,
imagine the cause happening without its and could therefore not serve as the basis of
effect, and so it is always epistemically pos- causal reasoning. For how would we then
sible that the effect will not occur. reason, when confronted with a C? The best
The assumption just mentioned has caused we could do is reason that, since Cs have
much puzzlement amongst commentators. always been observed to cause Es in our past
Hume is apparently arguing for a phenom- experience, the currently observed C will
enological claim – that on first observing likewise cause an E. But this inference cannot
them, ‘[a]ll events seem entirely loose and sep- be a priori, since it is still epistemically possi-
arate. One event follows another; but we can ble that the former is true and the latter false.
never observe any tie between them’ (EHU So we need to postulate a different mental
7.26 / 74) – and yet his argument for this mechanism that supplies the inference, and
claim proceeds by way of pointing out that that mechanism would turn out (according
we cannot ‘pronounce with certainty’ that to Hume’s own argument) to be custom or
the effect will follow, on observing the cause. habit. Thus the only difference between the
But how is the former claim supposed to fol- account just canvassed and Hume’s own
low from the latter? In particular, is Hume account of causal reasoning is that accord-
not confusing two distinct notions? On the ing to the former account the impression of
one hand, we have the claim that there is no causation is supplied by sensation, whereas
observable power, within the cause itself (e.g. according to Hume’s account it has another
the striking of one billiard ball by another), source.
such that observing it would deliver certainty That source, of course, turns out to be the
that the effect will follow. (J.L. Mackie calls very inference from causes to effects itself:
such a power ‘necessity2’.3) But this does not the ‘transition arising from the accustom’d


9780826443595_Ch05_Final_txt_print.indd 133 2/9/2012 9:35:00 PM


union’ (THN / 165), that is, the mind that considers them’ (THN
habit that takes the mind from an impression / 168). This suggests that it is a mistake to
of the cause, together with experience of past think that events in the world really are nec-
constant conjunction, to the belief that the essarily connected to one another.
effect will follow. As Hume puts it, 2. He does not appear to suggest that
the idea of causation can be stripped of the
when one particular species of event has component idea of necessary connection: he
always . . . been conjoined with another, does not appear to respond to the mistake
we make no longer any scruple of fore- just identified by advocating a revisionary
telling one upon the appearance of the
account of the concept of cause.
other, and of employing that reasoning,
3. He appears to think that our causal
which can alone assure us of any matter
of fact or existence. We then call the one thought and talk is truth-apt. (Certainly it
object, Cause; the other, Effect. (EHU is subject to normative constraints: in the
7.27 / 74–5) Treatise the very next section lists ‘rules by
which to judge of causes and effects’ (THN
Hence the ‘connexion . . . which we feel in the 1.3.15 / 173–6).)
mind, this customary transition of the imagi- The inconsistency is easy to see: by (2), we
nation from one object to its usual attendant, really do deploy the idea of necessary con-
is the sentiment or impression from which nection when we engage in causal talk and
we form the idea of power or necessary con- thought; and, by (3), that causal talk is in
nexion’ (EHU 7.28 / 75). general entirely legitimate (not least because
Hume thus finally achieves what he set it is subject to normative constraints; it is
out – nearly a hundred pages earlier, in the hard to see how this could be so if all such
Treatise (THN / 74) – to achieve: the talk was irredeemably defective). And yet, by
impression-source for the idea of necessary (1), that talk is irredeemably defective. There
connection. Unfortunately, however, it is far can be nothing in the world that answers to
from clear what consequences Hume takes the idea of necessary connection, since that
his discovery to have for our causal talk and idea derives solely from the ‘determination
thought, for he appears to endorse three posi- of the mind’. The three broad interpretative
tions that are mutually inconsistent: rivals discussed in section 4 below resolve
1. He seems to think that we are apt to the inconsistency in different ways. Roughly
project the impression – and the idea – of speaking, the traditional interpretation denies
necessary connection into the world: ‘the (2),4 and both the projectivist and sceptical
mind has a great propensity to spread itself realist interpretations finesse (1): they hold
on external objects, and to conjoin with Hume to the claim that there is a mistake
them any internal impressions, which they in the offing, but deny that the mistake in
occasion’ (THN / 167). And he question is that of thinking of events in the
seems to suggest that this projection is a mis- world as causally or necessarily connected.
take: ‘we are led astray by a false philosophy’ Finally – and this is what distinguishes these
when ‘we transfer the determination of the two interpretations – the projectivist inter-
thought to external objects, and suppose any pretation finesses (3) as well: Hume does
real intelligible connexion betwixt them; that endorse our causal talk and thought, but
being a quality, which can only belong to the that talk and thought is to be understood in


9780826443595_Ch05_Final_txt_print.indd 134 2/9/2012 9:35:01 PM


non-cognitivist terms. For the sceptical real- This interpretation removes the need to
ist, by contrast, our causal talk and thought think of the impression of necessary connec-
is straightforwardly referential: there is tion as a ‘feeling of expectation’ or similar;
something in the world that answers to our how things seem, when the impression arises,
idea of causation (or at least we believe that is, precisely, necessarily connected, just as it is
there is, and there is nothing defective about by virtue of the impression of red that things
that belief). seem red to us. It also helps to explain why
One more piece of the already difficult Hume offers a non-phenomenological argu-
puzzle needs to be put on the table, namely ment for the claim that there is no sensory
Hume’s famous two definitions of causa- impression of necessary connection. After all,
tion; this is the topic of the next section. if all events really did seem entirely loose and
Before leaving Hume’s discussion of the separate to us, Hume would not need such
origin of the idea of necessary connection, an argument; unbiased phenomenological
however, it is worth noting what is, in my reflection would do the job just by itself (and
view, an often misunderstood feature of his would thereby establish the stronger claim
account. It is routinely taken for granted that we have no impression of either neces-
that in Hume’s view, all events ‘seem entirely sity1 or necessity2).7
loose and separate’. That is, phenomenologi-
cally speaking, our experience is merely as
of one event following another, even once
the habit of inference has been established, 3. THE TWO DEFINITIONS
and so even once the impression of necessary
connection is present. This has forced some Here are Hume’s two definitions of cau-
interpreters to cast the impression of neces- sation – or, more precisely, definitions of
sary connection as, for example, simply a ‘a cause’ – as they appear in the Treatise,
‘peculiar feeling’5 or a ‘feeling of helplessness towards the end of his discussion of the idea
or inevitability’(where the inevitability is the of necessary connection:
inevitability of one’s expectation, and not the
inevitability of the effect itself).6 (D1) ‘An object precedent and contigu-
Hume is not, in fact, committed to this ous to another, and where all the objects
resembling the former are plac’d in like
view. What he says is that ‘all events seem
relations of precedency and contiguity to
entirely loose and separate’; he does so in
those objects, that resemble the latter’.
the context of ‘single instances of the oper- (THN / 170)
ation of bodies’ (EHU 7.26 / 73) – that is,
(D2) ‘[A]n object precedent and con-
when we first observe a pair of contiguous
tiguous to another, and so united with
events, and hence before the habit of associa-
it, that the idea of the one determines
tion has arisen. He does not explicitly state a the mind to form the idea of the other,
view about how things seem once the habit and the impression of the one to form
of association has arisen, but it is plausible a more lively idea of [that is, a belief in]
to suppose that he takes the impression of the other’. (ibid.)
necessary connection to affect, precisely, how
things seem – that is, to affect the nature of Unfortunately, rather than clarifying the situ-
visual experience itself. ation, the two definitions are problematic in


9780826443595_Ch05_Final_txt_print.indd 135 2/9/2012 9:35:01 PM


their own right. The first thing to note is that of ‘idealized spectator’, that is, an observer
they are not even extensionally equivalent. who observes representative samples of all
The conditions in (D2) can easily be satis- kinds of constantly conjoined events.
fied without (D1) holding, by someone hav- A third solution – similar to Garrett’s
ing observed an unrepresentative sample of ‘subjective’ reading and to Robinson’s pro-
‘objects’ (or events), so that the two kinds of posal concerning the second definition – is
event are constantly conjoined in their expe- to read both definitions not as ‘definitions’ in
rience but not universally. And (D1) can be the standard contemporary sense at all, but
satisfied without (D2) holding, for instance in as saying how it is we come to believe that
the case of constant conjunction between two one thing is a cause of another. As Edward
kinds of events that nobody has observed, so Craig puts it, the definitions characterize the
that nobody’s mind is determined to move ‘circumstances under which belief in a causal
from the idea of one to the idea of the other. connection arises, one concentrating on the
A standard solution to this problem8 has outward situation, the other on the state of
been to claim that Hume only intends (D1) the believer’s mind that those outward facts
as a genuine definition of causation, while induce’.11
(D2)’s aim is different: to explain the con- My own view is that none of these solu-
ditions under which we do, in fact, come tions are satisfactory, because they all ignore
to make causal claims, for example. This Hume’s preceding remark in the Treatise that
move seems somewhat ad hoc, however, ‘two definitions of this relation may be given
given Hume’s claim that the two definitions of this relation, which are only different, by
present ‘a different view of the same object’ their presenting a different view of the same
(THN / 170). object, and making us consider it either as a
A second solution, offered by Don Garrett,9 philosophical or as a natural relation; either as
is to distinguish a ‘subjective’ from an ‘abso- a comparison of two ideas, or as an associa-
lute’ reading of each definition. Roughly, a tion betwixt them’ (THN / 169–70).
subjective reading of (D1) would take ‘all Hume’s distinction between philosophical
objects’ to mean all objects observed so far and natural relations is a distinction between
by a particular person. This would then be two kinds of mental operation. Roughly, the
coextensive with (D2) read subjectively, former is the conscious ‘placing’ of two ideas
with ‘the mind’ understood as referring to under a relation (hence ‘plac’d’ in (D1)), and
the same person, since an ‘object’ that meets the latter is the unconscious ‘transition’ of the
(D1) now will be such that the mind of the mind from one idea to another. For example,
person in question is indeed determined by resemblance – which is the other relation
the idea of the object to form the idea of its that is both natural and philosophical – can
effect. As Garrett puts it, read subjectively, operate in two distinct ways, as when I con-
the two definitions tell us when an object sider whether a painting of a particular scene
‘functions psychologically’ as a cause.10 An resembles the image of a particular remem-
absolute reading of (D1) takes ‘all objects’ bered scene I have in my mind and come to
to be unrestricted, so that (D1) appeals to judge that it does (philosophical relation), or
universal constant conjunction. This is coex- when I see a picture of the Queen and my
tensive with (D2) read absolutely, with ‘the mind is automatically drawn to the idea of
mind’ now understood to refer to some sort the Queen herself (natural relation). Similarly


9780826443595_Ch05_Final_txt_print.indd 136 2/9/2012 9:35:01 PM


for causation: I can ‘place’ two ideas under in the most basic cases, we come to make
the relation of causation, and will (or should) causal judgements – judgements of the form
do so precisely when the conditions speci- ‘c caused e’, or ‘c will cause e’ – on the basis
fied in (D1) are met (contiguity, precedence of the temporal priority and contiguity rela-
and observed constant conjunction), thereby tions that hold between these two events, and
coming to form the judgement that one event the observed constant conjunction between
is the cause of the other. And I ‘naturally’ events of the kinds that c and e instantiate.
judge one event to be the cause of another In such cases we infer that e will occur on the
when I have acquired the relevant habit of basis of having observed c, and we also think
association: my mind is drawn from the idea of c as a cause of e, in a way that, somehow or
(or impression) of the first event to the idea of other, involves deploying the idea of necessary
(or belief in) the second, and, again, I thereby connection. Second, the impression-source of
come to form the judgement that the first that idea is the inference just described: the
event is the cause of the second.12 impression of necessary connection is not
Note that neither of the last two of the a sensory impression but an impression of
four interpretative positions just outlined reflection. Finally, there is – or can be – some-
delivers any verdict about the meaning of thing awry in our deployment of the idea of
‘cause’, since both deny that the definitions necessary connection. Perhaps the easiest way
are definitions in anything like the standard of seeing the difference between the three
contemporary sense. Instead, the definitions main interpretative positions is to consider
tell us something about how it is we come to how they interpret the second and third of
make causal judgements. As we shall see, the the claims just described. I shall thus start
availability of these interpretative options my brief account of each of the positions by
with respect to the two definitions under- describing their attitudes to those claims.
mines a key component of the motivation
for the traditional interpretation of Hume on THE TRADITIONAL INTERPRETATION
causation, according to which he is a naive
regularity theorist. We are not required to According to what I am calling the ‘tradi-
read the first definition as a definition of tional’ interpretation, Hume is a naive regu-
the meaning of ‘cause’, and so we are not larity theorist about causation: c causes e if
required (at least not required by the two and only if c is contiguous with and prior to e,
definitions) to hold Hume to the view that and events similar to c (the Cs) are constantly
causation consists in contiguity, precedence conjoined with events similar to e (the Es).13
and constant conjunction. In other words, ‘c causes e’ just means ‘all
Cs are followed by Es’. What is awry in our
deployment of the idea of necessary connec-
tion, in this view, is, precisely, that we deploy
4. INTERPRETATIONS: it at all: given that the source of that idea is
TRADITIONAL, SCEPTICAL REALIST the transition of the mind, it cannot possi-
AND PROJECTIVIST bly represent any mind-independent feature
of the world. Thus (and this is an issue on
It is uncontroversial that Hume endorses which versions of the traditional interpreta-
the following three theses. First, at least tion differ) either (i) necessary connection is


9780826443595_Ch05_Final_txt_print.indd 137 2/9/2012 9:35:01 PM


not, in fact, part of the meaning of ‘cause’ as SCEPTICAL REALIST INTERPRETATIONS

we actually deploy the idea ‘cause’, or (ii) it
is part of the ordinary meaning of ‘cause’ but In stark contrast to the traditional interpreta-
Hume is in effect offering a revisionary con- tion, the sceptical realist interpretation takes
ceptual analysis of ‘cause’ (shorn of the trou- Hume to hold that our causal talk refers –
blesome concept of necessary connection), successfully – to real, mind-independent
or (iii) it is part of the ordinary meaning of causal connections in nature, or what I shall
‘cause’ and Hume offers no such revision: call ‘causal powers’.14 Exactly what role the
‘cause’ is irredeemably defective. idea of necessary connection plays here is
The major problem with the traditional something that sceptical realist interpreters
interpretation is an almost total lack of evi- differ on. In particular, John Wright holds
dence in its favour. The major source of evi- that the idea of necessary connection refers
dence has standardly been thought to be the to genuine mind-independent necessary con-
‘two definitions’ – or rather, the first defin- nections, so that if only we could penetrate
ition. However, as we saw in section 3, the into the true, underlying nature of causes,
interpretation of the two definitions relied on we would be able to discern those (causal)
here (according to which the first definition powers by which causes really do abso-
really is Hume’s attempt to offer a concep- lutely necessitate their effects: he ‘retained
tual analysis of ‘cause’) has recently come an ideal of knowledge of true causes which
under attack. In addition, the accounts given was derived from the Cartesians’,15 so that
of the role of the idea of necessary connec- the ‘true manner of conceiving a particular
tion offered by different versions of the trad- power in a particular body’ (of which we are
itional interpretation fit badly with the text. in fact incapable) would involve being ‘able
Against alternative (i) above, Hume does not to pronounce, from a simple view of the one,
suggest (setting aside the first definition) that that it must be follow’d or preceded by the
the idea of necessary connection is no part of other’ (THN / 161).16 Thus the mis-
the actual meaning of ‘cause’; nor, contra (ii), take we are apt to make when it comes to the
does he suggest that the concept of causation idea of necessary connection is the mistake of
needs to be revised. He does, of course, sug- holding that it derives from an impression of
gest that something is apt to go awry when sensation, and thus holding that we really do
we deploy the idea of necessary connection; perceive necessity. Hence our idea of neces-
but he does not suggest that the appropriate sary connection is defective, in that – being
response is to stop deploying it; indeed, it is derived from an impression of reflection – it
hard to see how Hume could even consider cannot adequately represent real necessity in
this to be a genuine psychological possibil- nature; nonetheless, it succeeds in referring to
ity. Finally, against (iii), which I take to be real necessity. In other words, Hume agrees
Barry Stroud’s position, Hume clearly and with his opponents when it comes to the
persistently endorses a wide range of causal metaphysics of the Image of God doctrine, as
claims, and indeed provides ‘rules by which far as causation is concerned, but he disagrees
to judge of causes and effects’ (THN 1.3.15 with them over the Insight Ideal: we lack the
/ 173–6). This is extremely difficult to square God-like ability to penetrate into the essences
with the claim that he takes all causal talk to of things in a way that would reveal their true,
be equally and irredeemably false. effect-guaranteeing, underlying nature to us.


9780826443595_Ch05_Final_txt_print.indd 138 2/9/2012 9:35:02 PM


According to Galen Strawson, by contrast, except what is drawn from something extra-
Hume takes all necessity (whether causal or neous and foreign to it’ because we ‘have no
logical) to be purely subjective: our idea of idea of this connexion, nor even any distinct
necessity does not track any real necessity in notion what it is we desire to know, when we
nature, but is ‘just a feeling we have about endeavour at a conception of it’ (EHU 7.29
certain things – about 2 + 2 = 4, and about / 76–7).
what this billiard ball does to that one, and It is possible to reinterpret these claims in
about the sum of the angles of a triangle’.17 a way that does not commit Hume to belief
Nonetheless, our idea of causation suc- in causal powers. In particular, one might
ceeds in referring to more in the world than take the claims about nature’s ‘secret’ pow-
mere regularities: it refers to (to use Hume’s ers to be mere suppositions for the sake of
own expression) that upon which the ‘regu- the argument.19 Or one might point out that,
lar course and succession of objects totally since Hume explicitly takes ‘power’ to be syn-
depends’ (EHU 5.22 / 55); or, in Strawson’s onymous with ‘cause’ (THN / 157),
words, ‘whatever it is about the universe (or and since he doubtless thinks that there are
matter) which is that in virtue of which it is additional, not-yet-known regularities under-
regular’.18 Again, a central thought here is lying the observed behaviour of objects, talk
that our idea of causation is inadequate to of secret powers presents no problem for the
what it represents: we can have a ‘relative’ other interpretations.20 But there is nothing
idea of it, but not what Strawson calls a ‘pos- in the text of the first Enquiry itself to moti-
itively contentful’ idea of it. Thus Wright’s vate such a reinterpretation (or so defend-
Hume conceives of the referent of ‘cause’ ers of sceptical realism maintain): the most
to be objective necessity – the feature of the natural interpretation of the first Enquiry,
cause that absolutely guarantees that the taken on its own terms, reads Hume’s claims
effect will occur – whereas Strawson’s Hume at face value, as expressions of belief in, but
conceives of the referent of ‘cause’ to be a ignorance concerning the nature of, causal
regularity-guaranteeing feature of nature. powers.
The starting point for the claim that there One advantage of Wright’s version of
is serious textual evidence to justify (some the sceptical realist interpretation over the
version of) the sceptical realist interpreta- traditional interpretation is that it makes
tion is to take the first Enquiry rather than sense of the tension noted in section 2:
the Treatise as expressing Hume’s consid- Hume endorses our (necessary-connection-
ered view about causation. This is because it involving) causal thought and talk, and yet
is in the first Enquiry that Hume refers, on he apparently thinks there is something
several occasions, to the ‘powers and forces, wrong with the idea of necessary connection.
by which [the course of nature] is governed’ Wright’s account resolves the tension by iden-
that are ‘wholly unknown to us’ (EHU 5.21 / tifying what is ‘wrong’ with the idea of nec-
54), the ‘powers and principles on which the essary connection as our tendency to think
influence of these objects entirely depends’ that necessary connections are perceivable,
that nature ‘conceals from us’ (EHU 4.16 and our corresponding tendency to think
/ 33), and so on. It is also the place where that we have thus penetrated into the essence
he says that our idea of causation is ‘imper- of bodies: ‘[t]he vulgar mistake an associa-
fect’, and that it admits of no ‘just definition, tional connection for a genuinely perceived


9780826443595_Ch05_Final_txt_print.indd 139 2/9/2012 9:35:02 PM


rational connection’.21 This tendency does endorses our necessary-connection-involving

not, according to Wright’s account, affect the causal talk and thought, the same can be said
meaning of our causal talk, however. Our of the traditional interpretation. My own
habits of inference give rise to the belief that view is that Strawson’s suggestion does not
there are necessary connections in nature, adequately capture the thought that the idea
and that belief is (for all we know) true; the of necessary connection really is part of the
mistake we tend to make is the mistake of meaning of ‘cause’: Hume really does seem
thinking that our idea of necessary connec- to think that our causal thought involves the
tion is fully adequate to what it represents. claim that causes and effects are necessarily
When we ‘make the terms of power and effi- connected, and not merely that they happen
cacy signify something, of which we have a to give rise to a certain kind of ‘feeling’.
clear idea, and which is incompatible with A second advantage that the sceptical realist
those objects, to which we apply it, obscu- interpretation has been claimed by Strawson
rity and error begin then to take place’ (THN to have over the traditional interpretation is / 168). This is because we are in that the latter saddles Hume with the pre-
effect claiming that there is some feature of posterous (according to Strawson) claim that
the external world that is adequately repre- there is nothing more to the world than mere
sented by an idea whose impression-source regularity (‘one of the most baroque meta-
is a mere transition in the mind, when in fact physical suggestions ever put forward’).23 The
there can be no such feature. But this mis- traditional interpretation need make no such
take plays no role in the meaning of ‘cause’: claim, however. According to the traditional
our idea of necessary connection really does interpretation, Hume’s claim is only that
refer to real necessity, even if we are apt to be our thoughts cannot successfully reach out
mistaken about the nature of what it is we to any mind-independent relations between
thereby represent. causes and effects, aside from priority and
It is unclear whether Strawson’s interpre- contiguity. This is not to assert positively that
tation also has this advantage over the tra- such relations do not, or cannot, exist – only
ditional interpretation, given his claim that that we cannot succeed in referring to them
Hume takes all necessity to be subjective. in our causal thought and talk (if we try, we
Strawson says that ‘the E-intelligible [that is, ‘lapse into obscurity and error’). We can (as
positively contentful] meaning of the term Strawson says) form a ‘relative’ idea of such
“causation” can only encompass certain relations – we can consider the possibility
aspects of the experience Causation [that is, that they exist (without being able to form
causal power] gives rise to . . . [including] adequate ideas of what they might be like),
the feeling of determination’.22 So according but to do so would be, at best, to indulge in
to Strawson’s account, the idea of necessary idle metaphysical speculation. From a seman-
connection is only a part of the ‘meaning’ of tic point of view, then, the crux of the diffe-
‘causation’ in the sense that it is the idea of rence between the traditional interpretation
an experience that causation ‘gives rise to’. and Strawson’s version of sceptical realism is
But of course such a claim could equally that for Strawson, the ‘relative’ idea we form
be made by a defender of the traditional is a relative idea of real causal powers, which
interpretation; so if Strawson’s claim here are what our ordinary causal talk refers to.
makes adequate sense of the fact that Hume According to the traditional interpretation,


9780826443595_Ch05_Final_txt_print.indd 140 2/9/2012 9:35:03 PM


by contrast, we can form a relative idea of and so on), we are expressing a sentiment or
some possible not-further-specifiable relation moral attitude towards that action or person,
between causes and effects, but that idea is rather than attributing to them some mind-
not the idea of causation: the idea of caus- independent moral property. Moreover (this
ation is exhausted by contiguity, priority and further move is admittedly more controver-
constant conjunction. sial), we do not merely express the relevant
sentiment; we project it onto the object of
THE PROJECTIVIST INTERPRETATION24 our experience or judgement, so that the
painting we are observing looks beautiful or
The projectivist interpretation in some sense the murder will seem vicious thanks to the
represents a middle ground between the scep- projection of the relevant sentiment, and we
tical realist and traditional interpretations correspondingly judge them to be so thanks
(indeed Angela Coventry calls it the ‘interme- to the projection of the relevant idea. Thus
diate interpretation’).25 It shares with scepti- Michael Smith notes that when Hume says
cal realism the thought that our causal talk that you ‘never can find’ the viciousness of
and thought does more than merely register a murder ‘till you turn your reflexion into
the existence of regularities in nature, and your own breast, and find a sentiment of dis-
with Wright’s version of sceptical realism approbation’ (THN / 468–9), he ‘is
the thought that this involves the legitimate precisely trying to focus our attention away
deployment of the idea of necessary connec- from where it is naturally focused when we
tion. However, the projectivist interpretation judge a wilful murder to be wrong: that is,
shares with the traditional interpretation a away from the murder itself, and onto an
broadly meaning-empiricist interpretation of otherwise quite unnoticed “calm passion” he
Hume, according to which experience places supposes to arise in us’.26
strict limits on what can be represented, via Part of the point here, according to the
our ideas, in our thought and talk. In particu- projectivist line on morality, is that Hume
lar, the two interpretations agree that noth- is not merely making a straightforward phe-
ing in reality answers to our idea of necessary nomenological claim; it is not supposed to
connection: causation ‘in the objects’, as it is be just obvious to us that there is nothing in
sometimes put, amounts to no more than con- the murder itself that constitutes its vicious-
tiguity, priority and constant conjunction. ness. On the contrary, he only gets to this
The projectivist interpretation squares the claim after quite a lot of argument. So – the
apparent circle by adopting a non-cognitiv- thought is – it seems to us that our moral and
ist approach – on the one hand, the idea of aesthetic judgements are responses to genu-
necessary connection is deployed legitimately ine features of external objects, and their
in our causal talk and thought, but on the so seeming is due to the projection of the
other, no aspect of reality answers to this relevant sentiment. Our sentiment-derived
idea. A standard interpretation of Hume’s judgements are, however, subject to norma-
ethical (and indeed aesthetic) writing takes tive constraints – so, for example, there are
him to be a ‘sentimentalist’ about moral (and ‘rules of art’ which deliver a true standard
aesthetic) claims: when we make an evalu- of taste and sentiment’,27 even though ‘no
ative claim about an action or a person (that sentiment represents what is really in the
they are good or bad, brave or cowardly, object’.28 Hence moral and aesthetic claims


9780826443595_Ch05_Final_txt_print.indd 141 2/9/2012 9:35:03 PM


can legitimately be regarded as correct or fact A1s, and that A2s are often not followed
incorrect (there is a ‘true standard’ for them by Bs. This would give me good grounds to
to meet) on the basis of those normative revise my initial judgement that As cause Bs
constraints.29 in favour of the judgement that A1s cause Bs
In the case of causation, the analogue of but A2s do not, since the ‘difference in the
‘sentiment’ is, of course, the impression – effects of two resembling objects must pro-
and hence the idea – of necessary connection. ceed from that particular, in which they dif-
According to the projectivist view, we are apt fer’ (THN / 174; rule 6). And this
to mistake the impression of necessary con- would be so even if the habit of inference that
nection for a sensory impression (this is a has been established naturally inclines me, on
point of agreement with Wright), and we are next observing an A, to expect a B, and thus
apt to do this because we project the impres- to judge that they are causally connected
sion onto the external objects that trigger it. if a B does indeed follow; I know that that
(My own proposal here is that the ‘impres- judgement is hostage to information I do not
sion’ of necessary connection is in fact, for possess, namely whether or not the observed
Hume, not a self-standing ‘feeling’ at all, but A is an A1 or an A2. In other words, the rel-
merely the modification of visual experience evant rule acts as a normative constraint on
that occurs once our habit of expectation is the causal judgement I am naturally inclined
formed; see section 2 above.) And when we to make.
‘describe’ what is going on – when we ‘call Direct textual evidence for the projectiv-
the one object, Cause; the other, Effect’ (EHU ist interpretation is admittedly rather thin.
7.27 / 74–5) – we similarly project the idea of Indirect support comes largely from similari-
necessary connection onto those objects. Our ties between Hume’s treatment of causation
causal talk and thought, then, is not descrip- on the one hand, and moral and aesthetic
tive: it does not attribute a mind-independent judgements on the other. In each case, we
relation to causes and effects, but projects have the thought that our judgement (causal,
our idea of necessity onto them.30 moral or aesthetic) ‘adds’ something to mind-
As with the moral and aesthetic cases, independent matters of fact. In the causal
however, this is not to say that Hume is a case, Hume says that the mind does this via
subjectivist about causation, for there are its ‘propensity to spread itself on external
norms that govern the appropriateness of objects’ (THN / 167); in the moral
our causal claims to their objects, for exam- and aesthetic cases, he postulates ‘a produc-
ple in his ‘rules by which to judge of causes tive faculty, and gilding or staining all natu-
and effects’ (THN 1.3.15 / 173–6).31 More ral objects with the colours, borrowed from
generally, Hume is certainly in a position to internal sentiment, raises in a manner a new
regard our natural, instinctive causal judge- creation’ (EPM App. 1.21 / 294).
ments as eminently revisable in the light of A further piece of evidence – and at the
the evidence.32 Thus, for example, As might same time a response to the charge that
all have been followed by Bs in my experi- Hume shows no serious positive inclina-
ence, so that I judge that the As are causes of tion towards a non-cognitivist account of
Bs. But I might then find out that there are causation – comes from his selective use of
two distinct kinds of A (call them A1s and the terms ‘matter of fact’ and ‘belief’. Hume
A2s), that all the As I have observed are in never talks about causal ‘beliefs’ or considers


9780826443595_Ch05_Final_txt_print.indd 142 2/9/2012 9:35:03 PM


causal claims to fall within the class of ‘mat- is largely independent of beliefs about the
ters of fact’. We make causal judgements, but causal structure of reality. (For example ‘cau-
these would appear not to have the status of sation’ does not even appear in the index
belief for Hume, nor would they appear to of Colin Howson’s 2000 book, Hume’s
be judgements about matters of fact – or at Problem: Induction and the Justification of
least, Hume never claims that they are.33 In Belief [Oxford: Oxford University Press].)
other words, Hume appears to restrict ‘mat- Hume is, of course, the philosopher from
ter of fact’ – and correspondingly, ‘belief’ – to whom we are supposed to have learned that
the relata of causation. From the perspective causal thinking is both suspect and dispensa-
of the traditional and sceptical realist inter- ble. The irony is that this is not Hume’s view
pretations, this is rather puzzling. For in both at all: for Hume, causal thinking is central to
interpretations, there are perfectly good facts our understanding of, and beliefs about, the
about causation to be had, and so our causal nature of reality. But how are we to under-
‘judgements’ should count as beliefs, every bit stand what ‘causal thinking’ amounts to for
as much as the existence of a moving billiard Hume? As we have seen, the range of inter-
ball is a matter of fact and our expectation pretative possibilities is very wide indeed.
that the billiard ball will move is a belief. My own view is that the traditional inter-
From the perspective of the projectiv- pretation, in all its forms, is untenable: Hume
ist interpretation, by contrast, there is no holds that causal thinking amounts to more
real anomaly here. Causal reasoning, Hume than belief in regularities, for it involves the
says, is reasoning from one matter of fact to idea of necessary connection. Moreover, that
another. If causal judgements are projections idea is entirely legitimate, when correctly
of the idea to which this reasoning gives rise, understood, and it does serious philosophi-
then those judgements are not beliefs about cal work for him.37 But this leaves both pro-
matters of fact, any more than moral and jectivism and at least one version of sceptical
aesthetic judgements are beliefs about mat- realism still in the running, and deciding
ters of fact: they are projections of our habits between these possibilities is a difficult task.
of thought onto matters of fact, and so do Part of the difficulty lies in the differences
not constitute beliefs about matters of fact.34 between the Treatise and the first Enquiry.
Reading the first Enquiry as a reworking of
the Treatise, with no substantial change in
the philosophical views presented, inclines
5. PROBLEMS AND PROSPECTS one towards projectivism, correspondingly
encouraging one to reinterpret Hume’s talk
For quite a large part of the twentieth cen- of secret powers and the like in the Enquiry
tury, most analytic philosophers steadfastly so that they do not express a commitment to
avoided appealing to the concept of cause: the existence of real causal powers. Reading
a ‘horrid little word’, according to Peter van the first Enquiry on its own, by contrast, with
Inwagen,35 and ‘a truly obscure’ concept, no preconceptions carried over from reading
according to John Earman.36 In particu- the Treatise, on the grounds that Hume took
lar, discussion of the problem of induction the first Enquiry to best express his consid-
overwhelmingly proceeded as though infer- ered philosophical view, inclines one towards
ence from the observed to the unobserved sceptical realism. But of course which strategy


9780826443595_Ch05_Final_txt_print.indd 143 2/9/2012 9:35:03 PM


one should adopt is largely a question of his- (Cambridge: Cambridge University Press,
torical fact that further attention to the texts 1996), p. 21.
J.L. Mackie, The Cement of the Universe
themselves will not resolve.38
(Oxford: Oxford University Press, 1974),
There is, however, at least one reason to pp. 12–13.
be sceptical about sceptical realism. Sceptical 4
In fact, Barry Stroud’s interpretation (which
realism (or at least Wright’s and Kail’s ver- I class as a version of the traditional inter-
sions thereof) inevitably saddles Hume with pretation) does not deny (2); in Stroud’s view,
we simply do, inevitably, ‘come to believe,
a deeply puzzling view. On the one hand, all
mistakenly, that there are objective necessary
sides agree that Hume rejects what Craig connections between events’ (B. Stroud, Hume
calls the ‘Insight Ideal’: we cannot pen- (London: Routledge, 1977), p. 87). So Stroud
etrate into the ‘essence’ of bodies in such apparently embraces, rather than attempts to
a way as to reveal any features that would resolve, the tension described above.
Stroud, Hume, p. 86.
license a priori inferences from causes to 6
H. Noonan, Hume on Knowledge (London:
effects. Nonetheless, Wright’s and Kail’s ver- Routledge, 1999), p. 142.
sions of sceptical realism attribute to Hume 7
For more on this, see my Hume on Causation
one aspect of the ‘Image of God’ doctrine, (Abingdon: Routledge, 2006), sect. 4.3.
the metaphysical position that underlies See, for example, J.A. Robinson, ‘Hume’s
Two Definitions of “Cause”’, Philosophical
the Insight Ideal, for they both attribute to
Quarterly 12 (1962), pp. 162–71; A. Coventry,
Hume ‘an ideal of knowledge of true causes Hume: A Guide for the Perplexed (London:
which was derived from the Cartesians’.39 In Continuum, 2007), p. 110; G. Dicker, Hume’s
other words, ‘true causes’ are such that we Epistemology and Metaphysics (London:
would, if only we could penetrate into their Routledge, 1998), p. 115; Noonan, Hume on
Knowledge, p. 151.
nature, be able to infer effects from causes a 9
D. Garrett, Cognition and Commitment in
priori. And the question is: why would Hume Hume’s Philosophy (New York: Oxford
commit himself to such a metaphysical posi- University Press, 1997), chap. 5.
tion? After all, the only motivation for such a Garrett, Cognition and Commitment, p. 108.
view is the thought that the nature of reality Craig, The Mind of God, p. 108; see also
Garrett, Cognition and Commitment, p. 106.
must be such that God himself (and so we, if 12
For more on this, see my Hume on Causation,
we are sufficiently God-like) can infer effects pp. 94–107.
from causes a priori. And this is a motivation 13
Interpretations that come under this gen-
that Hume himself clearly lacks: he has no eral heading include T.L. Beauchamp and
reason whatsoever to want to cling on to a A. Rosenberg, Hume and the Problem of
Causation (New York: Oxford University
picture of the nature of reality that derives
Press, 1981); Stroud, Hume; Mackie, The
from views about God’s epistemic access to Cement of the Universe; Garrett, Cognition
that reality. and Commitment. There are important differ-
ences between the interpretations offered by
these authors, however, on the troublesome
issue of what to do about the idea of neces-
NOTES sary connection; see my Hume on Causation,
chap. 5, for discussion.
E. Craig, The Mind of God and the Works of The major contributions to the sceptical
Man (Oxford: Clarendon Press, 1987), chap. 1. realist literature have been N. Kemp Smith,
R. Descartes, Meditations on First Philosophy The Philosophy of David Hume (London:
(1641), trans. and ed. J. Cottingham Macmillan, 1941); J.P. Wright, The Sceptical


9780826443595_Ch05_Final_txt_print.indd 144 2/9/2012 9:35:03 PM


Realism of David Hume (Manchester: Continuum, 2006). For a general discussion of

Manchester University Press, 1983); G. the notion of ‘projection’ see Kail, Projection
Strawson, The Secret Connexion: Causation, and Realism, chap. 1, and for a discussion
Realism, and David Hume (New York: of Hume’s (allegedly problematic) attitude
Oxford University Press, 1989); S. Buckle, towards ‘projection’, see B. Stroud, ‘“Gilding
Hume’s Enlightenment Tract (Oxford: Oxford or Staining” the World with “Sentiments” and
University Press, 2001); and P.J.E. Kail, “Phantasms”’, in R. Read and K. Richman
Projection and Realism in Hume’s Philosophy (eds), The New Hume Debate, pp. 16–30.
(Oxford: Oxford University Press, 2007). Coventry, Hume’s Theory of Causation.
15 26
Wright, The Sceptical Realism of David Hume, M. Smith, ‘Objectivity and Moral Realism:
p. 147. Kail takes the same line, holding that On the Significance of the Phenomenology
the ‘causal necessity of which we may be igno- of Moral Experience’, in J. Haldane and
rant can be understood as that which, were we C. Wright (eds), Reality, Representation, and
to be acquainted with it, would yield a priori Projection (New York: Oxford University
inference and render it impossible to conceive Press, 1993), p. 246.
cause without effect’ (Kail, Projection and D. Hume, ‘Of the Standard of Taste’, E 230.
Realism, p. 103). Ibid.
16 29
Quoted in Wright, The Sceptical Realism of See Coventry, Hume’s Theory of Causation,
David Hume, p. 140. pp. 117–33, for further discussion.
17 30
Strawson, The Secret Connexion, p. 157. See, however, P.J.E. Kail, ‘Book Review: Helen
Ibid., p. 126. Beebee, Hume on Causation’, Mind 117
See A.J. Jacobson, ‘From Cognitive Science (2008), pp. 453–4, for a worry about the use
to a Post-Cartesian Text: What Did Hume of the notion of projection here.
Really Say?’, in R. Read and K. Richman (eds), See Coventry, Hume’s Theory of Causation,
The New Hume Debate (London: Routledge, pp. 133–7.
2000), pp. 156–66. See my Hume on Causation, pp. 160–7.
20 33
See K. Winkler, ‘The New Hume’, in R. Read The same is generally true of the moral and
and K. Richman (eds), The New Hume Debate, aesthetic cases, although, as Coventry notes,
pp. 52–87, and S. Blackburn, ‘Hume and Thick in ‘Of the Standard of Taste’, Hume twice
Connexions’, in R. Read and K. Richman appears to count matters for which there is a
(eds), The New Hume Debate, pp. 100–12. ‘standard’ (as in ‘standard of taste’) as ‘mat-
See my Hume on Causation, pp. 180–93 for ters of fact’; see Coventry, Hume’s Theory of
further discussion of both options, and Kail, Causation, p. 119, and Hume, E 230, 242.
Projection and Realism, pp. 118–21, for more See my Hume on Causation, pp. 152–4, for
critical assessment. more discussion in the context of Hume’s
Wright, The Sceptical Realism of David Hume, distinction between reason and taste.
p. 95. P. Van Inwagen, An Essay on Free Will
Strawson, The Secret Connexion, pp. 132–3. (Oxford: Clarendon Press, 1983), p. 65.
23 36
Ibid., p. 87. J. Earman, A Primer on Determinism
The projectivist interpretation is suggested by (Dordrecht: D. Reidel, 1986), p. 5.
some remarks of Simon Blackburn’s; see, for See my Hume on Causation, sect. 6.5.
example, S. Blackburn, Spreading the Word For a defence of the latter strategy, see G.
(Oxford: Oxford University Press, 1984), Strawson, ‘David Hume: Objects and Power’,
pp. 210–12 and ‘Hume and Thick Connexions’, in R. Read and K. Richman (eds), The New
pp. 107–11. It is given a fuller treatment Hume Debate, pp. 31–51. For brief discussion,
in my Hume on Causation, chap. 6, and in see my Hume on Causation, pp. 221–5.
A. Coventry, Hume’s Theory of Causation: Wright, The Sceptical Realism of David Hume,
A Quasi-Realist Interpretation (London: p. 147.


9780826443595_Ch05_Final_txt_print.indd 145 2/9/2012 9:35:03 PM

Alan Bailey

1. HUME’S RADICAL SCEPTICISM supposes that the senses continue to oper-

ate, even after they have ceas’d all manner of
The title of the concluding part of Book One of operation’ (THN / 188). And he fol-
Hume’s Treatise is ‘Of the sceptical and other lows this with a repudiation of the supposi-
systems of philosophy’, and Hume wastes no tion that the senses are capable of persuading
time in setting before the reader two scepti- us that some of the things of which we are
cal arguments that attack, when taken in tan- immediately aware in perception have the
dem, the justified status of the overwhelming ability to exist independently of perception
majority of our beliefs. In the opening section, or are representations of entities of another
‘Of scepticism with regard to reason’, the kind that do possess such a power (see THN
argument outlined by Hume is presented as–14 / 189–93).
targeting all beliefs that are the product of any It seems, therefore, that if we take Hume
kind of inference. According to this argument, as endorsing the arguments that he has set
both demonstrative and probabilistic infer- out in such detail in these key sections of
ences are ultimately unsuccessful in giving the Treatise, then we must interpret him as
their conclusions any positive degree of justi- accepting that no claim about mind-inde-
fication (see THN–6 / 180–3). pendent objects and no claim that requires
The attempt to found beliefs about mind- support from any form of inference is ever
independent physical objects on the senses any better justified than the logical contra-
rather than some form of inference then dictory of that claim. And as Hume appears
comes under attack in the very next sec- to hold that the inner and outer senses are
tion of the Treatise. In ‘Of scepticism with unable, without some form of inferential
regard to the senses’, Hume rapidly dismisses supplementation, to tell us anything about
the suggestion that the senses are capable events that are not happening to us at the
by themselves of revealing that some of present moment, this implies that Hume
the objects seemingly encountered by us in holds that only beliefs about very sim-
perception continue to exist even after they ple necessary truths that do not need to be
have ceased to appear to the senses: any such grasped through a process of inference and
supposition is ‘a contradiction in terms, and beliefs about the content of our present ideas


9780826443595_Ch06_Final_txt_print.indd 146 2/9/2012 9:35:12 PM


and impressions are potentially capable of positive view of sceptical argumentation.

being rationally justified beliefs. Immediately after Hume has set out the argu-
It will be convenient from this point ment that no process of inference succeeds in
onwards to refer to these few beliefs that conferring justification on its conclusions, he
Hume is seemingly prepared to accept as jus- specifically addresses the question of whether
tifiable as H-minimal beliefs. Now it seems he sincerely gives his assent to this argument
evident that an inquirer who has engaged and whether he is ‘really one of those scep-
in substantial epistemological reflection but tics, who hold that all is uncertain, and that
nevertheless fails to accept that any beliefs our judgement is not in any thing possest of
other than H-minimal beliefs are rationally any measures of truth and falshood’ (THN
justified beliefs is someone who has embraced / 183). And his answer is one that
a radically sceptical posture. It appears, could easily be interpreted as a repudiation of
therefore, that we can legitimately reject the the argument that he has just constructed:
conclusion that Hume is himself a radical
sceptic only if we can uncover substantial I shou’d reply, that this question is
grounds for supposing that Hume does not entirely superfluous, and that neither I,
genuinely endorse the arguments that he has nor any other person was ever sincerely
and constantly of that opinion. Nature,
chosen to elaborate in ‘Of scepticism with
by an absolute and uncontroulable neces-
regard to reason’ and ‘Of scepticism with
sity has determin’d us to judge as well
regard to the senses’. There must be a pre- as to breathe and feel; nor can we any
sumption, however, that an author who puts more forbear viewing certain objects in
lengthy arguments before the reader without a stronger and fuller light, upon account
explicitly distancing himself from those argu- of their customary connexion with a pre-
ments does endorse their conclusions. And it sent impression, than we can hinder our-
is difficult to see Hume’s comments in the selves from thinking as long as we are
paragraph that brings these two sections to awake, or seeing the surrounding bodies,
an end as anything other than an attempt to when we turn our eyes towards them in
make it clear that he stands firmly behind his broad sun-shine. (ibid.)
negative epistemological arguments:
Commentators who doubt the propri-
’Tis impossible upon any system to ety of classifying Hume as a radical sceptic
defend either our understanding or our also tend to seize on his explanation of his
senses; and we but expose them farther purposes in presenting his master argument
when we endeavour to justify them
against inferential justification. He indicates
in that manner. As the sceptical doubt
that this argument has been included in the
arises naturally from a profound and
intense reflection on those subjects, it Treatise as a way of adding credibility to his
always encreases, the farther we carry preferred account of belief as something that
our reflections, whether in opposition or arises when ideas acquire additional force
conformity to it. (THN / 218) and vivacity as a result of their associative
links with impressions:
It does have to be conceded that there are
other passages in the Treatise that initially My intention then in displaying so care-
seem to indicate that Hume has a much less fully the arguments of that fantastic sect


9780826443595_Ch06_Final_txt_print.indd 147 2/9/2012 9:35:12 PM


[the sceptics], is only to make the reader H-minimal beliefs ever possess any positive
sensible of the truth of my hypothesis, degree of rational justification. But when we
that all our reasonings concerning causes look more closely at his explicit reservations
and effects are deriv’d from nothing but about sceptical argumentation, we discover
custom; and that belief is more properly
that he seems intent on distancing himself
an act of the sensitive, than of the cogi-
only from the supposition that such argumen-
tative part of our natures. (THN
/ 183) tation is capable of radically reshaping our
non-epistemic beliefs. He is keen to empha-
size that no amount of reflection on sceptical
In similar fashion, Hume’s discussion arguments can permanently dislodge every-
of our belief in the existence of mind-inde- day beliefs about the existence and proper-
pendent physical objects sees him surround- ties of such things as trees, chairs and books.
ing his presentation of the relevant sceptical This does not mean, however, that he holds
arguments with observations that generate that sceptical arguments permit us to regard
significant perplexity about his personal atti- such beliefs as rationally justified.
tude towards such argumentation. Thus we As we have seen, Hume is sometimes pre-
find that the paragraph that opens ‘Of scep- pared to insinuate that it would be appro-
ticism with regard to the senses’ includes the priate to regard him as sincerely assenting to
following declamation: the argument that no beliefs based on infer-
ence are ever rationally justified only if that
We may well ask, What causes induce us argument were to cause him to abandon all
to believe in the existence of body? but inferential beliefs. Thus we find that Hume
’tis in vain to ask, Whether there be body presents our supposed inability to ‘forbear
or not? That is a point, which we must viewing certain objects in a stronger and
take for granted in all our reasonings. fuller light, upon account of their custom-
(THN / 187) ary connexion with a present impression’
as sufficient of itself to guarantee that it is
It might be suggested, accordingly, that if we impossible for anyone to give their sincere
must indeed hold as a presupposition of all assent to this argument (THN / 183).
our reasoning that mind-independent bodies It would normally be supposed, though, that
exist, this precludes anyone from sincerely a distinction can be drawn between sincere
endorsing sceptical arguments on this topic. assent to the conclusion that a particular
If we are genuinely committed to the conclu- belief is not a rationally justified belief and
sion that mind-independent bodies exist, how the actual abandonment of that belief. Many
can we simultaneously attach any weight to people suffering from acrophobia, for exam-
arguments that purport to call that conclu- ple, sincerely accept that at least some of
sion into question? their fears about being in high places are not
It is apparent, therefore, that the manner rationally justified fears. Nevertheless this
in which Hume expounds his philosophical does not prevent them from continuing to be
position in the Treatise does place some dif- afflicted by those fears. It seems, accordingly,
ficulties in the way of the conclusion that that Hume’s insistence that neither he nor
this work incorporates sceptical arguments anyone else is capable of eschewing all infer-
that have genuinely persuaded him that only ential beliefs in response to the argument set


9780826443595_Ch06_Final_txt_print.indd 148 2/9/2012 9:35:12 PM


out in ‘Of scepticism with regard to reason’ is betwixt a false reason and none at all’ (THN
entirely compatible with the supposition that / 268), we have compelling grounds
he sincerely assents to that argument in the for asserting that this argument’s conclusion
sense that he endorses as true its conclusions about the extremely limited availability of
about the limitations on our ability to arrive rationally justified belief does indeed enjoy
at epistemically justified beliefs. Hume’s full support.
This latter interpretation of Hume’s pos- This same pattern is also discernible
ition gains additional credibility from other within Hume’s discussion of the status of
remarks that he makes about the argument our beliefs about the existence of mind-in-
against reason: these remarks are very dif- dependent physical objects. It is indeed true
ficult to explain if we are not prepared to that he emphatically repudiates the suppos-
regard him as holding that only someone ition that any of his arguments could bring
who has already embraced the view that about more than a fleeting alienation from
scarcely any of our beliefs are rationally jus- those beliefs (see THN / 218).
tified is in a position to deny, in good faith, Nevertheless it seems plain that he regards
that this argument provides rationally com- the sceptical attack on their supposed
pelling support for just that assessment of rational credentials as completely unanswer-
our epistemic situation. Hume boldly claims, able. Instead of diagnosing this sceptical
for instance, that the argument against rea- critique as relying on some fallacious infer-
son proves that the principles that lead us ential step or a premise that is either false or
to correct our judgements on any subject readily deniable, Hume affirms that sceptical
through ‘the consideration of our genius and worries about the justification of the suppos-
capacity, and of the situation of our mind, ition that mind-independent physical objects
when we examin’d that subject’ would, if we exist are irrefutable: ‘This sceptical doubt,
persisted in following them, reduce all puta- both with regard to reason and the senses, is
tive evidence to nothing and ‘utterly subvert a malady, which can never be radically cur’d,
all belief and opinion’ (THN / 184). but must return upon us every moment, how-
He also explicitly bases his affirmation that ever we may chace it away, and sometimes
we can safely conclude that ‘reasoning and may seem entirely free from it’ (ibid.). From
belief is some sensation or peculiar manner Hume’s perspective, no process of reflect-
of conception, which ’tis impossible for mere ive thought can place that supposition on a
ideas and reflections to destroy’ (ibid.) on his justified basis, and the intense examination
confidence that no one will be able to find of sceptical arguments, even if undertaken
any error in the arguments contained in ‘Of with the intention of refuting them, simply
scepticism with regard to reason’. It seems makes their irrefutability more salient and
clear, therefore, that when we consider the potentially strengthens their psychological
foregoing pieces of corroborative evidence impact to such an extent that they can fleet-
in combination with Hume’s comments in ingly alienate us from our common-sense
the final section of Book One of the Treatise belief in physical objects. Hume accordingly
that he has shown, by means of the argu- rejects the project of constructing a philo-
ment against reason, that the understanding sophical answer to scepticism in favour of
entirely undermines itself when it operates delineating a way of arriving at a psycho-
alone and that we are left with no choice ‘but logical accommodation with scepticism that


9780826443595_Ch06_Final_txt_print.indd 149 2/9/2012 9:35:13 PM


serves to minimize its potentially disrup- in almost every instance and attempts to
tive effects. As philosophical inquiry cannot reassure us that our natural instincts are suf-
locate any deficiencies in the arguments of ficiently powerful to ensure that it can never
the sceptics that are not shared by all other ‘undermine the reasonings of common life,
forms of reasoning, Hume suggests that the and carry its doubts so far as to destroy all
best way of responding to these epistemo- action, as well as speculation’ (EHU 5.1–2 /
logical difficulties is to relax our intellectual 41). Moreover, the piece of abstract reason-
efforts: ‘Carelessness and in-attention alone ing Hume displays as an example of the kind
can afford us any remedy’ (ibid.). of argument held in check by these natural
It seems, therefore, that the most plaus- instincts is nothing other than a summary of
ible interpretation of the explicit discussions the argument offered in Section 4 (See EHU
of sceptical argumentation in the Treatise 5.2 / 41). And even more striking is the man-
is that Hume combines the conviction that ner in which Hume shapes the discussion of
these arguments make it impossible for him, scepticism in the final section of the Enquiry.
or indeed any inquirer who has fully under- In Part 2 of Section 12 he uses the terms
stood their implications, to maintain in ‘Pyrrhonism’ and ‘excessive scepticism’ to
good faith that beliefs other than H-minimal refer to a form of scepticism that attempts
beliefs are ever rationally justified with to ‘destroy reason by argument and rati-
the judgement that arriving at this conclu- ocination’ (EHU 12.17 / 155).1 According
sion has little psychological impact on our to him, this form of scepticism purports ‘to
everyday, non-epistemic beliefs. However, find objections, both to our abstract reason-
the credibility of this assessment of Hume’s ings, and to those which regard matter of
stance also depends on the extent to which it fact and existence’ (EHU 12.17 / 156); and
fits with his presentation of his views in the when he comes to describe the most force-
Enquiry concerning Human Understanding. ful of these sceptical objections to reasonings
And we find, in fact, that this work displays concerning matters of fact, the argument
numerous elements that prevent us from he ascribes to the Pyrrhonist is essentially a
regarding it as less sceptical than Book One reprise of Hume’s own negative claims about
of the Treatise. the grounding of causal inferences.
Several arguments deployed in the Treatise In addition to the foregoing greater will-
are explicitly acknowledged by Hume as scep- ingness to identify his own arguments as
tical arguments only when they appear in the sceptical arguments, further evidence within
Enquiry concerning Human Understanding. the Enquiry of Hume’s commitment to some
Consider, for example, the revisions he makes form of radical scepticism is furnished by
to the presentation of his analysis of causal his discussion of the relationship between
reasoning. Not only does he give the initial Pyrrhonean and Academic scepticism. In the
stages of this analysis the forthright title final section of the Enquiry Hume seems to
‘Sceptical Doubts concerning the Operations imply that the only significant criticism that
of the Understanding’, but we also find that can be levelled against the arguments used
the immediately following section begins by Pyrrhonean sceptics is that they lack the
with a discussion of ‘the academical or power to dislodge our non-epistemic beliefs
sceptical philosophy’ in which he says that (see EHU 12.23 / 159–60). And he indicates
such a philosophy is harmless and innocent that the causal resultant of natural instinct


9780826443595_Ch06_Final_txt_print.indd 150 2/9/2012 9:35:13 PM


and Pyrrhonean scruples about rational jus- to hold that Hume does not embrace the
tification is an enduring and advantageous sceptical view that very few of our beliefs are
disposition to form one’s non-epistemic ever rationally justified.
beliefs in a cautious and undogmatic man-
ner. According to Hume, the most conveni-
ent way of acquiring this disposition is ‘to be
once thoroughly convinced of the force of 2. THE TENSION BETWEEN
the pyrrhonian doubt, and of the impossi- SCEPTICISM AND ABSTRUSE INQUIRY
bility, that any thing, but the strong power of
natural instinct, could free us from it’ (EHU As we have just seen, discounting Hume’s
12.25 / 162). Thus mitigated scepticism or commitment to a radically sceptical perspec-
the Academical philosophy, which is strongly tive in respect of the limited availability of
recommended to the reader in both Sections rationally justified belief does not appear to
5 and 12 of the Enquiry, is presented not be a viable interpretative strategy. However,
as the outcome of recognizing flaws in the if we accept that reflection on sceptical argu-
arguments of radical sceptics but rather as ments has led Hume to the conclusion that
the outcome of our natural belief-forming only H-minimal beliefs can ever qualify as
mechanisms causally interacting with the rationally justified beliefs, there does seem to
realization that sceptical arguments can be a real danger that Hume’s overall intellec-
only be repudiated in good faith by some- tual posture ceases to be a coherent one.
one who already accepts that no beliefs other The subtitle of the Treatise is Being an
than H-minimal beliefs are ever rationally Attempt to Introduce the Experimental
justified.2 Method of Reasoning into Moral Subjects.
Ultimately, then, the considerations in And in the introduction to the Treatise, Hume
favour of interpreting Hume as committed presents this work as primarily an attempt to
to a radically sceptical stance in both the explain the best way of making progress in
Treatise and the Enquiry concerning Human the science of man.
Understanding seem very strong indeed. Hume maintains that the science and the
Prominent arguments in Book One of the philosophy of his era are replete with systems
Treatise terminate in the conclusion that and theories that display such undesirable
no beliefs other than H-minimal beliefs are qualities as ‘principles taken upon trust, con-
rationally justified; and once we have given sequences lamely deduc’d from them, want
due weight to the distinction between seeing of coherence in the parts, and of evidence in
a belief as devoid of rational justification and the whole’ (THN Intro. 1 / xiii). These glar-
actually abandoning that belief, there seems ing deficiencies ‘seem to have drawn disgrace
to be little explicit basis in the text for sup- upon philosophy itself’ (ibid.), and they have
posing that Hume does not endorse that con- also given rise to what Hume describes as
clusion. When, therefore, we find a similarly ‘that common prejudice against metaphysical
radical array of sceptical arguments deployed reasonings of all kinds, even amongst those,
in the Enquiry alongside seemingly robust who profess themselves scholars, and have a
affirmations of their strength and irrefutabil- just value for every other part of literature’
ity (see EHU 12.14 / 153–4, 12.22 / 159, and (THN Intro. 2 / xiv). As he explains, this
12.25 / 162), it surely ceases to be plausible prejudice against metaphysical reasonings is


9780826443595_Ch06_Final_txt_print.indd 151 2/9/2012 9:35:13 PM


not simply a prejudice against a particular rise to those theories. This methodology is,
branch of inquiry but has come to constitute in turn, something that can beneficially be
a general aversion to ‘every kind of argument, applied to the study of human nature and in
which is in any way abstruse, and requires particular to the study of how human beings
some attention to be comprehended’ (ibid.). make inferences and form judgements. And
Hume’s declared ambition is to alleviate with the assistance of the ensuing improved
this prejudice by helping to put science and account of human psychology, we will be
philosophy on a sound basis. His proposed able to develop a still better methodology of
way of doing this involves applying the inquiry for use not only within the natural sci-
methods of the natural sciences, as exempli- ences themselves but also within other intel-
fied within the most compelling theories that lectual disciplines that lack acknowledged
have arisen within that realm of inquiry, to achievements on a par with the theories of
the study of human nature. Hume holds that Newton, Galileo and Copernicus.
once we have a better understanding of the Significantly, Hume is adamant that the
mental activities of human beings, this will progress in the natural sciences that has
shed fresh light on a wide variety of intellec- finally given us some theories that are both
tual disciplines: intellectually satisfying and stable in the face
of examination is the product of the experi-
’Tis evident, that all the sciences have mental method of investigation. He specific-
a relation, greater or less, to human ally places his project in the Treatise in the
nature; and that however wide any of context of an intellectual movement, exempli-
them may seem to run from it, they still
fied in his eyes by ‘some late philosophers in
return back by one passage or another.
England, who have begun to put the science
Even Mathematics, Natural Philosophy,3
and Natural Religion, are in some meas- of man on a new footing, and have engag’d
ure dependent on the science of man; the attention, and excited the curiosity of the
since they lie under the cognizance of public’ (THN Intro. 7 / xvii), that is attempt-
men, and are judg’d of by their powers ing to transfer the application of experi-
and faculties. (THN Intro. 4 / xv) mental philosophy from natural to moral
subjects. Hume maintains that this attempt
Of course natural philosophy is the arena in forces us to acknowledge that the science of
which Hume has found examples of convin- man can only be built on a foundation ‘laid
cing theories built on a credible investigative on experience and observation’ (THN Intro.
methodology, an investigative methodology 7 / xvi). If we construct the science of man
he has deliberately extended to the moral sci- in this manner, then our conclusions about
ences. His strategy, therefore, involves making human psychology will assist us in refining
use of a virtuous cycle of mutual improve- and improving our investigative techniques
ment. Such exemplary theories within the across a broad range of subjects. But if we
natural sciences as ‘newton’s explication of abandon the guidance of experience, then we
the wonderful phenomenon of the rainbow’ will be led to conclusions that ‘ought, at first,
(DNR 1.136) and ‘the copernican system’ to be rejected as presumptuous and chimer-
(ibid. 138) serve to validate the experimen- ical’ (THN Intro. 8 / xvii).
tal and experience-based investigative meth- These forthright scientific and philosoph-
odology that Hume regards as having given ical aspirations, however, raise the question


9780826443595_Ch06_Final_txt_print.indd 152 2/9/2012 9:35:13 PM


of how they cohere with Hume’s radical scep- will corroborate whatever has been said
ticism. Given their prominence at the start of concerning the understanding and the
the Treatise, how does it come about that passions. (THN / 455)
Book One seems to terminate in a reaffirm-
ation of sceptical arguments indicating that We also need to keep in mind that if Hume
only beliefs with H-minimal content are ever had seen the sceptical crisis of Book One as
rationally justified beliefs? No science can an unwelcome or unexpected interruption in
be constructed that consists exclusively of the unfolding of his overall line of argument
beliefs with such attenuated content. Should in the Treatise, then he had ample oppor-
we conclude, therefore, that Hume’s positive tunity to omit it from the version of the text
project has catastrophically imploded in the sent for publication or to include additional
course of writing Book One and that he has guidance to the reader about how to place
unintentionally found himself immersed in this element of his discussion within his posi-
a sceptical quagmire from which there is no tive aims. His letters confirm that he had no
principled escape? qualms about deleting some of the more obvi-
The supposition that Hume sees the out- ously irreligious material from the text before
come of Book One of the Treatise as incon- finally placing it in the hands of his publisher
sistent with or even tangential to his initial (see, in particular, LDH 1.25, 6). Thus the
ambitions to put the study of human nature prominent role ultimately allocated in the
on a sounder footing is difficult to reconcile Treatise to the presentation of sceptical lines
with the untroubled opening of Book Two. of thought presumably reflects Hume’s con-
In Book Two he presents an account of the sidered judgement that the sceptical aspects
operations of the human passions that seems of his thinking and the attempt to build the
perfectly in accord with his professed inten- science of human nature on a sound founda-
tions to make progress in the moral sciences, tion form part of one coherent and mutually
and there is no indication whatsoever that supportive intellectual enterprise.
he regards this account as undermined or in Once again Hume’s revised presentation of
any way rendered problematic by the scepti- his views in the Enquiry concerning Human
cal reflections developed in the final stages of Understanding provides important corrobor-
Book One. Similarly, his discussion of moral- ation of what is implied by the text of the
ity in Book Three shows no signs of being Treatise. In Section 1 of the Enquiry Hume
infected by destabilizing sceptical worries. develops a defence of ‘profound reasonings,
It is true that in the opening paragraph he or what is commonly called metaphysics’
ventures some observations about the lim- and abstract philosophy (EHU 1.7 / 9) that
ited persuasive force of abstruse reasoning, presents it as providing us with, amongst
but the overwhelming tenor of his discussion other benefits, an accurate delineation of ‘the
is one of quiet confidence in the satisfactory powers and faculties of human nature’ (EHU
nature of his arguments and conclusions: 1.13 / 13). Hume even suggests that the ana-
logy with the development of Newtonian
I am not, however, without hopes, that mechanics means that we can legitimately
the present system of philosophy will hope that such profound and abstruse rea-
acquire new force as it advances; and sonings will ultimately serve to ‘discover, at
that our reasonings concerning morals least in some degree, the secret springs and


9780826443595_Ch06_Final_txt_print.indd 153 2/9/2012 9:35:13 PM


principles, by which the human mind is actu- while it remains in its full force and vig-
ated in its operations’ (EHU 1.15 / 14). We our. We need only ask such a sceptic,
are forced to conclude, therefore, that Hume’s What his meaning is? And what he pro-
intention to assist in the construction of a sci- poses by all these curious researches? He
is immediately at a loss, and knows not
ence of human nature remained intact at the
what to answer. (EHU 12.23 / 159–60)
time of writing the Enquiry despite the effort
lavished in the Treatise on the elaboration of
sceptical lines of argument. Yet the Enquiry Hume explains this objection in terms of the
also sees Hume deploying another battery of inability of sceptical argumentation to prod-
sceptical arguments that are scarcely any less uce stable changes in people’s beliefs and
radical than those found in the Treatise. In behaviour analogous to those that can arise
some ways, indeed, the Enquiry even seems from exposure to the Copernican system of
to place a greater emphasis than the Treatise astronomy or the moral exhortations of the
on the force and tenaciousness of sceptical Stoics and Epicureans. According to Hume,
argumentation. It appears, therefore, that the the radical or Pyrrhonean sceptic ‘cannot
Enquiry concerning Human Understanding expect, that his philosophy will have any
replicates, without any obvious sign of constant influence on the mind: Or if it had,
unease or tension on Hume’s part, the same that its influence would be beneficial to soci-
blend of philosophico-scientific ambitions ety’ (EHU 12.23 / 160). It seems, therefore,
and sceptical conclusions about the limited that Hume is suggesting that we are entitled
availability of rationally justified beliefs that to dismiss reflection on sceptical reasoning as
can be discerned in the Treatise. pointless or, at best, a frivolous and esoteric
Once we accept, however, that Hume form of amusement unless there is some indi-
regards radical scepticism and the attempt cation that it can alter a person’s set of beliefs
to construct a science of man as complemen- in stable and useful ways.
tary activities, it is natural to wonder what Somewhat ironically Hume’s own attempt
kind of science can exist in a context where to emphasize the transient nature of the dox-
all the core beliefs of this putative science are astic changes induced by sceptical arguments
viewed by its own adherents as totally devoid actually terminates by drawing attention to
of rational justification. Moreover, if the one stable change to which these arguments
most profound and advanced forms of scien- can give rise. He insists that despite the initial
tific thinking are as compatible with radical impact of the sceptic’s arguments, it is evident
scepticism about justification as Hume seems that ‘the first and most trivial event in life
to think, what does this say about the value will put to flight all his doubts and scruples’
of engaging in sceptical reflection? (ibid.). Thus he describes the sceptic as awak-
In the final section of the Enquiry concern- ening from a dream and laughing at himself
ing Human Understanding, Hume presents for being swayed by these abstruse episte-
what he plainly regards as a potentially dev- mological objections. Nevertheless this reim-
astating criticism of radical scepticism: mersion in quotidian life does not entirely
negate the influence of sceptical reasoning.
For here is the chief and most confound- The sceptic may indeed be led to describe
ing objection to excessive scepticism, that his arguments as nothing more than amusing
no durable good can ever result from it; puzzles, but Hume also represents the sceptic


9780826443595_Ch06_Final_txt_print.indd 154 2/9/2012 9:35:14 PM


as retaining a distinctive perspective on the follow such reasoning. Some rather unusual
rationality of his and our beliefs. Even when people might find that the study of sceptical
the sceptical fugue that can be generated argumentation serves them as a mildly enter-
by prolonged reflection on epistemological taining pastime, but Cleanthes undoubtedly
arguments passes, the sceptic retains a robust speaks for most of us when he expresses
sense of an aspect of the human predicament a preference in the Dialogues concern-
that escapes the attention of more superficial ing Natural Religion for less demanding
inquirers: and cerebral forms of amusement: ‘But for
my part, whenever I find myself disposed
the whimsical condition of mankind, who to mirth and amusement, I shall certainly
must act and reason and believe; though choose my entertainment of a less perplex-
they are not able, by their most diligent ing and abstruse nature. A comedy, a novel,
enquiry, to satisfy themselves concerning or at most a history, seems a more natural
the foundation of these operations, or recreation than such metaphysical subtilties
to remove the objections, which may be and abstractions.’ (DNR 1.137)
raised against them. (ibid.) It might be suggested at this point that the
sceptic can mount an effective response to
Given that Hume draws our attention Hume’s objection by drawing a distinction
to this enduring consequence of the careful between the non-epistemic beliefs that are
study of sceptical reasonings, why does he an integral part of everyday life and the non-
also make the seemingly contradictory claim epistemic beliefs generated by such activities
that radical scepticism fails to exercise any as philosophy and theoretical science. Even if
constant influence on the mind? The most these everyday beliefs cannot be disturbed or
plausible answer here is that he is thinking changed by sceptical argumentation, perhaps
in terms of influence over our everyday non- the arguments deployed by radical sceptics
epistemic beliefs. The radical sceptic might can reshape our philosophical and scien-
well acquire a settled disposition to deny that tific commitments. Suspension of judgement
any beliefs other than H-minimal beliefs are at the everyday level is certainly not some-
ever rationally justified beliefs, but this does thing that flows from reflection on sceptical
not lead to suspension of belief on the exist- reasonings, but that does not guarantee that
ence of cattle, chairs and chimneys. Indeed the more abstract and sophisticated non-
the radical sceptic’s expectations about what epistemic beliefs found within philosophy
will happen, unlike his beliefs about what and the sciences are similarly invulnerable to
we are justified in thinking will happen, sceptical attack.
tend to be remarkably similar to our own. This indeed is Philo’s initial response to
Thus the objection to radical scepticism that Cleanthes’ charge that his professed scepti-
Hume emphasizes in the Enquiry concern- cism is an insincere affectation. Cleanthes
ing Human Understanding seems to amount challenges Philo to exhibit his scepticism in
to the contention that if the argumentative his behaviour:
case assembled so laboriously by sceptical
thinkers has no or almost no impact on Whether your scepticism be as absolute
our non-epistemic beliefs, then we have no and sincere as you pretend, we shall
motive for expending the effort required to learn bye and bye, when the company


9780826443595_Ch06_Final_txt_print.indd 155 2/9/2012 9:35:14 PM


breaks up: We shall then see, whether This second component of Philo’s defence
you go out at the door or the window; of scepticism could be construed as overlook-
and whether you really doubt, if your ing the possibility that refined arguments
body has gravity, or can be injured that run wide of common life might some-
by its fall; according to popular opin-
times be arguments that should be allowed
ion, derived from our fallacious senses
to influence us. However, it would be more
and more fallacious experience. (DNR
1.132) charitable to interpret Philo as attempting
to draw a distinction between complicated
and intricate arguments that are neverthe-
In this crude form Cleanthes’ attack is easily less rooted in the same underlying principles
parried by Philo. Philo simply appeals to the as those that generate stable conviction in
involuntary nature of these everyday beliefs. ordinary life and even more ambitious and
The sceptic sincerely and genuinely holds speculative arguments that disregard those
that even straightforward and seemingly common-sense principles. Thus we find that
uncontroversial everyday beliefs are entirely Philo is happy to endorse a form of theoriz-
devoid of rational justification. However, ing in both the natural and moral sciences
this conviction fails to dislodge these beliefs that keeps closely to patterns of inference
because they are held in place by psycho- and evidence assessment that exert a major
logical forces that are causally impervious to influence on our quotidian beliefs:
any influence that can be exerted by sceptical
arguments: ‘To whatever length any one may He considers besides, that every one,
push his speculative principles of scepticism, even in common life, is constrained to
he must act, I own, and live, and converse have more or less of this philosophy; that
from our earliest infancy we make con-
like other men; and for this conduct he is
tinual advances in forming more general
not obliged to give any other reason than the
principles of conduct and reasoning; that
absolute necessity he lies under of so doing.’ the larger experience we acquire, and the
(DNR 1.134) stronger reason we are endowed with,
Philo also wishes, however, to leave room we always render our principles the
for radical scepticism to affect at least some more general and comprehensive; and
of our non-epistemic beliefs in a potentially that what we call philosophy is nothing
beneficial way. Consequently he suggests that but a more regular and methodical oper-
abstruse theoretical beliefs can be subverted ation of the same kind. (DNR 1.134)
by sceptical arguments:
It seems, therefore, that Philo’s response to
But it is evident, whenever our argu- the question of the practical value of reflec-
ments . . . run wide of common life, that tion on sceptical arguments is to claim that
the most refined scepticism comes to be
such reflection is useful because it serves to
upon a footing with them, and is able
subvert and sweep away speculative beliefs
to oppose and counterbalance them.
The one has no more weight than the that are not rooted in the core psychological
other. The mind must remain in suspense mechanisms that generate our enduring every-
between them; and it is that very sus- day beliefs. At the same time this reflection is
pense or balance, which is the triumph not dangerous because those everyday beliefs
of scepticism. (DNR 1.135–6) and even the more sophisticated beliefs that


9780826443595_Ch06_Final_txt_print.indd 156 2/9/2012 9:35:14 PM


arise from the methodical and systematic appears to leave him potentially vulnerable to
application of common-sense principles of the charge that these views turn out to have
inference are founded in permanent features no practical consequences whatsoever. If the
of human nature that cannot be overturned radical sceptic responds at the level of his
by any form of abstract reasoning. non-epistemic beliefs to what he would call
Is it really the case, however, that we can ‘putative’ or ‘alleged’ evidence in exactly the
satisfactorily combine the view that radical same manner that the non-sceptic responds
scepticism is intellectually safe with the sup- to what he views as real evidence, then it
position that it is genuinely useful? In the seems that even if we disregard the issue of
Dialogues Cleanthes returns to the attack by sincerity, there is nothing of any substantial
alleging that self-professed sceptics invari- value to be gained from the mental exertion
ably seem to conduct even their theoretical required to sustain a sceptical stance. On the
inquiries in the same manner as people who other hand, if the radical sceptic genuinely
do not regard themselves as sceptics: finds, as Philo suggests, that his mental pos-
ture is modified so that refined and abstruse
But I observe . . . with regard to you, reasoning loses much of its usual persuasive-
philo, and all speculative sceptics, that ness, how can a radical sceptic be convinced
your doctrine and practice are as much by the refined and abstruse reasoning that
at variance in the most abstruse points
underpins Hume’s science of human nature?
of theory as in the conduct of common
life. Wherever evidence discovers itself,
you adhere to it, notwithstanding your
pretended scepticism; and I can observe,
too, some of your sect to be as decisive 3. SCEPTICISM AND THE CURBING
as those who make greater professions of OF THE RESTLESS IMAGINATION
certainty and assurance. (DNR 1.136)
It is clear that a satisfactory explanation of
It is now possible to appreciate the full how Hume sees scepticism and the science of
force of Cleanthes’ attack on radical scepti- human nature as mutually reinforcing com-
cism. Initially it appeared that he was con- ponents of a genuinely illuminating world-
centrating on the issue of whether a person view needs to do more than merely point to
can sincerely embrace the thesis that no his commitment to the supposition that many
beliefs other than H-minimal beliefs are ever of our beliefs are the inevitable causal prod-
rationally justified. And Philo’s appeals to ucts of psychological mechanisms that can-
psychological necessitation seemed to pro- not be disrupted by abstract reasoning. That
vide a satisfactory explanation of why the commitment accounts for Philo’s insouciance
sceptic shares so many non-epistemic beliefs in the face of Cleanthes’ misguided conten-
with people who would vehemently repudi- tion that a true sceptic would have no motive
ate radical scepticism. However, Philo’s ensu- for leaving via the door and stairs rather than
ing attempt to locate an area of inquiry in a high window. But it does not explain how
which the radical sceptic’s distinctive views radical scepticism can give rise to real changes
about the limited availability of rationally amongst a person’s non-epistemic beliefs
justified beliefs generate a plausibly benefi- while leaving refined and abstruse reason-
cial reordering of his non-epistemic beliefs ing with enough persuasiveness to support


9780826443595_Ch06_Final_txt_print.indd 157 2/9/2012 9:35:14 PM


an edifice of theory with the complexity with the psychological foundations provided
manifested by the science of man that Hume by experience and observation can resume:
is attempting to develop in the Treatise. In
order to find Hume’s explicit account of how If the reader finds himself in the same
this is possible, we must turn our attention to easy disposition, let him follow me in my
the final section of Book One. future speculations. If not, let him fol-
Hume appears content in this section to low his inclination, and wait the returns
present sceptical arguments as serving to of application and good humour. (THN
expose our lack of rational justification for / 273)
our beliefs. Indeed, he even admits that these
arguments can fleetingly alienate us from our Hume describes the conduct of a man who
most stable beliefs about contingent matters engages in philosophical inquiry when the
of objective fact. He also emphasizes that inclination spontaneously arises within him,
‘the intense view’ of the ‘manifold contradic- or in ‘this careless manner’, as ‘more truly
tions and imperfections in human reason’ sceptical than that of one, who feeling in him-
has effects that cannot be dispelled through self an inclination to it, is yet so over-whelm’d
the efforts of reason itself (THN / with doubts and scruples, as totally to reject
268–9). Nature, rather than reason, ultim- it’ (ibid.).4 The point here is that resisting
ately brings such doxastic alienation to an one’s natural propensities is something that
end, but the initial outcome of this fortuitous requires a rationale. Sceptical arguments,
cure is a state of ‘spleen and indolence’ (THN however, subvert standard epistemic reasons: / 270) that is not conducive to any it is no longer possible for a sceptic, even of
further philosophical or scientific inquiry. As a mitigated or Academic variety, to think of
time passes, however, and the inquirer grows himself as striving against his inclinations in
‘tir’d with amusement and company’ (THN order to construct a set of beliefs in accord- / 270), an inclination to return to ance with the demands of reason. Someone
intellectually taxing investigations spon- who strenuously strives to shed his beliefs on
taneously arises again, and new attempts the grounds that this is what reason demands
at constructing intellectual systems begin. of an inquirer who no longer views his beliefs
Hume is adamant, however, that this process as rationally justified has failed to appreciate
operates in accordance with a natural cycle. that this alleged constraint of reason is itself
Until such time as a psychological propensity something that sceptical arguments expose as
to find some principles and claims immedi- devoid of rational justification.
ately appealing and convincing is activated, Now a suspensive sceptic might possibly
and the mind also recovers its enthusiasm for claim that he finds himself convinced that
intricate reasoning falling outside the bound- only widespread suspension of belief can pro-
aries of common life, it is simply not psycho- tect him against erroneous beliefs that are not
logically possible to arrive at any fresh beliefs balanced by equally valuable true beliefs, and
in either the moral or the natural sciences in that this constitutes a motive for suspending
the face of the influence exerted by remem- judgement. Hume’s reply to this contention
bered sceptical arguments. But once these would presumably be that the most plausible
mechanisms are in play once more, the con- psychological theories available to us imply
struction of a science of man in accordance that sceptical arguments, as opposed perhaps


9780826443595_Ch06_Final_txt_print.indd 158 2/9/2012 9:35:14 PM


to insanity, cannot subvert our belief-forming can there be for denying that it might also
mechanisms so completely. Moreover, he be successful in uncovering the truth in the
can also point out that there is no recorded moral sciences?
instance of anyone arriving at such a state, The other direction from which pressure
rather than a state of verbally asserting that falls on a suspensive sceptic who professes
he is suspending judgement on a radical scale, to suspend judgement on the results of all
as a result of exposure to any kind of philo- abstruse reasoning is his practice of forming
sophical argument, and that the behaviour of numerous opinions within the arena of eve-
this self-professed suspensive sceptic makes it ryday life. According to Hume, ‘philosophi-
clear that he too has numerous beliefs. Indeed cal decisions are nothing but the reflections
it is difficult to see how anyone, no matter of common life, methodized and corrected’
how carried away by intense philosophical (EHU 12.25 / 162). As long as we remain
reflection, could deny in good faith that he connected in this way to the reasoning we
generally believes that if wholesale suspen- employ and find persuasive outside of our
sion of belief were to be instantiated, then ‘all philosophical investigations, there is no
human life must perish’ (EHU 12.23 / 160). sharp division to be drawn between every-
More plausibly, perhaps, someone who day and philosophical reasoning, and hence
purports to be unwilling to engage in phil- there is, it might be thought, no principled
osophico-scientific theorizing and system- motive for refusing to participate in philo-
building might argue that he simply finds the sophical inquiry. After all, if one appeals
results of abstruse and complicated reason- once more to the existence of unanswerable
ing, at least about matters of fact, to be so Pyrrhonean arguments, or indeed Humean
full of ‘errors and delusions’ (THN Intro. 3 / genetic arguments for scepticism based on
xiv) that he sees no point in continuing with alleged discoveries about the nature of the
an enterprise that has such a disappointing psychological mechanisms responsible for
outcome. Even this position, however, seems our beliefs,5 it is an evident fact, and one that
an untenable one. Cleanthes’ comments in cannot be denied in good faith by any philo-
the Dialogues are particularly apposite here: sophical sceptic, that these arguments tell
‘In reality, would not a man be ridiculous, equally against the justification of the every-
who pretended to reject newton’s explica- day beliefs that all sceptics have in profusion
tion of the wonderful phenomenon of the and are equally ineffectual at a psychological
rainbow, because that explication gives a level against these everyday beliefs and some
minute anatomy of the rays of light; a sub- scientific theories. As Philo contends in his
ject, forsooth, too refined for human com- own defence of scepticism: ‘To philosophise
prehension.’ (DNR 1.136) on such subjects [natural and moral subjects]
Very few people are prepared in practice is nothing essentially different from reason-
to repudiate all the results of refined and ing on common life; and we may only expect
abstruse reasoning, especially where this greater stability, if not greater truth, from
yields verifiable predictions of novel phenom- our philosophy, on account of its exacter
ena or generates new technologies that assist and more scrupulous method of proceeding.’
in the manipulation of the world around us. (DNR 1.134)
And if such reasoning is perforce embraced Now it might be suggested that this argu-
in the case of the natural world, what basis ment from the close relationship between


9780826443595_Ch06_Final_txt_print.indd 159 2/9/2012 9:35:15 PM


philosophical reasoning and everyday rea- relatively complicated arguments in the natu-
soning overlooks Hume’s own explanation of ral sciences do patently command the assent
the inability of radical sceptical arguments to of even the most determined sceptics. As rea-
dislodge our everyday non-epistemic beliefs. soning in the natural and moral sciences is
Hume accepts that these arguments are plau- ‘nothing but a more regular and methodical
sibly seen as continuations of patterns of operation of the same kind’ (DNR 1.134) as
reasoning that we embrace and respond to reasoning in common life, this affords a strong
in the course of daily life (see, in particular, presumption in favour of its being very simi-
THN–9 / 183–5). However, despite lar in its persuasiveness to the reasoning that
this intimate relationship with common-sense shapes our beliefs in everyday contexts. Now
reasoning, they fail to have much impact on the suspensive sceptic rightly injects a note of
our non-epistemic beliefs. Hume explains caution here. The relationship between eve-
this in terms of the obstacles the imagination ryday reasoning and arguments for radical
encounters when it attempts to participate scepticism conspicuously fails to ensure that
in complicated and subtle reasoning: ‘Where sceptical arguments have a similar influence
the mind reaches not its objects with easiness over our non-epistemic beliefs. But when we
and facility, the same principles have not the discover that complicated arguments in the
same effect as in a more natural conception natural sciences do sometimes lead to sci-
of the ideas; nor does the imagination feel a entific theories that even the most hardened
sensation, which holds any proportion with sceptic would find it ridiculous to reject, then
that which arises from its common judg- the presumption under discussion here must
ments and opinions’ (THN / 185). surely carry a great deal of weight.
An important feature of this explanation, The only recourse left open to the suspen-
however, is that complication and subtlety sive sceptic would be to challenge the assimi-
are matters of degree. From Hume’s perspec- lation of arguments in the human sciences to
tive, arguments gradually lose their persua- arguments in the natural sciences rather than
siveness as their intricacy and complexity radical Pyrrhonean arguments. Arguments
increase: this loss of persuasiveness does in the natural sciences and radical scepti-
not stem from a change in the fundamental cal arguments both manifest a high level of
nature of the argumentation presented to abstraction and complexity. However, some
us. Thus the sceptic who maintains that it is arguments of this kind within the field of the
appropriate for him to suspend judgement natural sciences generate stable non-epistemic
on all conclusions reached by abstruse and beliefs, yet abstract and complicated argu-
refined reasoning might be tempted to dis- ments of a radically sceptical kind appear to
miss the significance of the close connections lack this capacity. What, then, explains this
between everyday reasoning and the reason- difference? Until we understand why the out-
ing used in philosophy and science. After all, comes are so dissimilar, we might be mistaken
sceptical reasoning and everyday reason- in treating arguments in the human sciences
ing also have much in common, but Hume as potentially performing in the manner of
concedes that they affect our non-epistemic their counterparts in the natural sciences.
beliefs very differently. The explanation that Hume seems to
The problem with this response is that it endorse is that arguments in the natural sci-
fails to take into account the fact that some ences gain persuasiveness and psychological


9780826443595_Ch06_Final_txt_print.indd 160 2/9/2012 9:35:15 PM


force from their strong links with experience insufficiently rigorous. Repeated experiences
and observation: of that kind might tend to reinforce our com-
mitment to the standards of self-criticism
It is only experience, which teaches us exploited by radical scepticism. But even this
the nature and bounds of cause and might, in turn, be subverted by a recognition
effect, and enables us to infer the exist- that such self-criticism would also suggest
ence of one object from that of another.
that we are being excessively rash in taking
Such is the foundation of moral rea-
our apparent observations of falsifying phe-
soning, which forms the greater part of
human knowledge, and is the source of nomena at face value. Essentially, therefore,
all human action and behaviour. (EHU the Humean view is that compelling scientific
12.29 / 164) arguments and inferences gain persuasiveness
from being directly linked to the enlivening
Sceptical arguments, in contrast, are rela- effects of sense perception, whereas sceptical
tively detached from experience, and they arguments have little, if any, scope to exploit
represent instead attempts at exploring the this source of conviction. Thus the contest
implications of various abstract epistemic between good scientific arguments and even
norms. Hume says little about the origins the most powerful sceptical arguments is, at
of these norms, possibly because he believes the non-epistemic level, a distinctly unequal
that all his readers will readily acknowledge contest:
both the psychological pull exerted by these
norms and the key prescriptions they lay These [sceptical] principles may flour-
down. For our current purposes, however, the ish and triumph in the schools; where it
key point is that sceptical arguments seem to is, indeed, difficult, if not impossible, to
refute them. But as soon as they leave
rely entirely on such norms’ antecedent psy-
the shade, and by the presence of the real
chological force. And as the argumentative
objects, which actuate our passions and
links with this original source of vivacity and sensations, are put in opposition to the
persuasiveness become increasingly compli- more powerful principles of our nature,
cated and intricate, the conclusions become they vanish like smoke, and leave the most
less and less convincing. In the case of argu- determined sceptic in the same condition
ments in the natural sciences, however, their as other mortals. (EHU 12.21 / 159)
persuasiveness can be reinforced by adding
additional lively impressions to our stock of When it comes to epistemic appraisals,
supporting evidence or by securing obser- in contrast, the struggle is one that scep-
vations that confirm the predictions of the tical arguments can win. It is one thing to be
theories built upon this argumentation or the steadfastly convinced that it is true that p and
effectiveness of the technology arising from quite another to be convinced that one’s belief
those theories. that p meets the abstract epistemic standards
The closest that sceptical arguments can required for it to qualify as a rationally justi-
come to exploiting such links with lively fied belief. The first level of conviction can be
impressions is when we take ourselves to reached simply with the assistance of direct
be encountering impressions that falsify perception or a causal inference based on a
beliefs we have formed on the basis of rea- lively impression and associative links created
soning that sceptical arguments condemn as by past patterns within one’s impressions.


9780826443595_Ch06_Final_txt_print.indd 161 2/9/2012 9:35:15 PM


However, the second level of conviction loose associative links. Theories built upon
requires the belief to be assessed against a set solid causal reasoning are not similarly vul-
of epistemic norms that are not themselves nerable; so they survive the encounter with
directly rooted in perception. And as positive even the most powerful sceptical arguments
assessments are themselves the products of a without any substantial damage to their
relatively artificial and strained mental pos- credibility. The only upshot in this latter case
ture, they can be readily counterbalanced by is a modest prompting towards a seemly
sceptical arguments. degree of intellectual modesty and a willing-
The vulnerability of our epistemic apprais- ness to consult the opinions of other people
als derives from the fact that the realm of such (see EHU 12.24 / 161–2). Thus the applica-
appraisals, as opposed to the realm of beliefs tion of radically sceptical arguments, when
about mere matters of physical and psycho- their power and irrefutability is genuinely
logical fact, is itself a relatively rarefied arena internalized, is an extremely efficacious way
of thought that has only an indirect connec- of orienting the mind towards theories sup-
tion with what we are told by observation ported by exemplary causal reasoning and
and experience. Because it is not a natural away from specious theories supported only
and instinctive turn of thought, conviction is by weaker principles of mental association.
not as strong and as stable as it often is con- The outcome for a person who has what
cerning straightforward matters of fact: Hume regards as the appropriate attitude
towards sceptical arguments is that some
No wonder, then, the conviction, which theory-building continues even after attain-
arises from a subtile reasoning, dimin- ing a sceptical apotheosis. Causal arguments
ishes in proportion to the efforts, which that possess substantial experiential backing
the imagination makes to enter into
are almost completely impervious to scep-
the reasoning, and to conceive it in all
tical arguments at the level of psychological
its parts. Belief, being a lively concep-
tion, can never be entire, where it is not persuasiveness, and they can therefore sup-
founded on something lively and easy. port conclusions in the natural and human
(THN / 186) sciences of a recondite and abstruse kind.
Engaging in serious intellectual inquiry
It is not just our epistemic beliefs, how- does, however, leave the inquirer susceptible
ever, that are vulnerable to sceptical argu- to sudden shifts in his attitude towards his
ments. Non-epistemic judgements that arise apparent discoveries. According to Hume,
from over-lively imaginations that respond such inquiry will inevitably see us yield-
to weak associations of ideas, rather than ing from time to time ‘to that propensity,
the strong associations that underpin causal which inclines us to be positive and certain
inference, are also sufficiently weakly rooted in particular points, according to the light,
in our natural belief-forming mechanisms in which we survey them in any particular
that they too can be swept aside by radical instant’ (THN / 273). Hume sees
sceptical arguments. And it is this vulnerabil- this as an inevitable psychological conse-
ity that allows sceptical argumentation to quence of giving our full attention to specific
perform its valuable function of eliminating pieces of evidence and reasoning. While no
all those metaphysical positions that are built countervailing considerations are before the
upon excessively exuberant imaginations and mind, we are pulled along by the immediate


9780826443595_Ch06_Final_txt_print.indd 162 2/9/2012 9:35:15 PM


plausibility of the narrow and unbalanced radical scepticism to claim that this is a self-
range of thoughts present to the mind at that refuting conclusion. Now it is true that it is,
particular moment. No direct effort of will by its own standards, not a rationally justi-
can suppress these momentary lapses into fied conclusion. But this does not entail that
exuberant dogmatism. The only way of elim- it is a false conclusion, nor does this give us
inating them would be to ‘forbear all exam- any grounds for supposing that some con-
ination and enquiry’ (THN / 274). trary conclusion is better justified. Moreover,
However, the consequences of doing this at if it is urged at this point that we are under
the level of everyday enquiry would be, as we an intellectual obligation to believe that we
noted before, the destruction of all human ought to eschew beliefs when we cannot
life (see EHU 12.23 / 160). And although the regard them as rationally justified, this sim-
outcome of eschewing all enquiry of a sci- ply raises the issue of the status of this sus-
entific and philosophical kind would be less pensive norm. And a little reflection seems
disastrous, it would still be absurd for those to establish that it is this norm, at least as it
with an inclination towards such researches is supposed to hold sway in anyone’s mind,
to engage in the struggle with their natural that is, unlike radical scepticism, genuinely
impulses that this would involve. self-subverting.
The arguments of radical sceptics have the Radical scepticism as a set of opinions
psychological power to force some of us to about the epistemic status of beliefs is fully
abandon our initial belief in the existence internally coherent unless it incorporates the
of certain types of reasons for actions and suspensive norm currently under discussion,
opinions. However, they do not generate any and there is no obvious basis for concluding
new reasons to guide people’s behaviour and that it should embrace this norm. Indeed, it
judgements. They subvert the framework appears that there are compelling grounds
of epistemological rationality, but they put for supposing that it is not a coherent norm
nothing new in its place. Moreover, it would to endorse if one is a radical sceptic. Radical
be a serious mistake to suppose that this scepticism can be true and we can believe it
subversion fails to encompass the supposed to be true without transgressing any doxastic
normative requirement to eschew beliefs that norms internal to itself even if it does imply,
one cannot sincerely view as rationally jus- as a thesis, that it is not a position that pos-
tified. Hume’s science of man critiques that sesses any positive justification. But if radical
alleged norm at a variety of levels. Thus he scepticism across the range of topics envis-
takes great delight, for example, in repeat- aged by Hume is true and we believe it to be
edly exposing our psychological inability to true, then the suspensive norm undermines
put this alleged norm into practice. But the itself. Let us begin by trying to assume that
key observation to make in respect of the this norm is true. Then it instructs us to eject
implications of radical scepticism is that this it from our minds, as the suspensive norm
norm is a self-subverting norm. comes to be its own target given its status as
Suppose that one has been driven, through a belief that we cannot regard as justified. On
the contemplation of sceptical arguments, the other hand, if we assume that it is a false
to the conclusion that no beliefs other than norm, then we clearly do not wish to allow
H-minimal beliefs are ever rationally justi- it any influence over us. It is, therefore, not a
fied. It is a commonplace of discussions of norm we should allow to shape our doxastic


9780826443595_Ch06_Final_txt_print.indd 163 2/9/2012 9:35:15 PM


preferences, because it is either a false norm conclusion: ‘A true sceptic will be diffident
or a norm that tells us, as radical sceptics, to of his philosophical doubts, as well as of his
disregard it. philosophical conviction; and will never ref-
Only a superficial sceptic, accordingly, use any innocent satisfaction, which offers
would combine radical scepticism about all itself, upon account of either of them.’ (THN
beliefs with more than H-minimal content / 273)
with the judgement that this places him under One’s willingness to accept that particu-
some obligation to suspend judgement on all lar beliefs cannot be rationally justified is,
such matters. A more profound sceptic rec- moreover, entirely compatible with firmly
ognizes that thoroughgoing scepticism about regarding those beliefs as true and also hav-
rational justification actually allows for a ing the meta-belief that these beliefs are the
level of doxastic promiscuity that is entirely products of psychological mechanisms that
compatible with possessing as rich a set of mostly latch on to the truth in nearby pos-
beliefs as would be possessed by any non- sible worlds. Viewing beliefs as incapable of
sceptical inquirer who is guided by observa- being supported by any form of reasoning or
tion and experiment rather than the fanciful putative evidence that does not exhibit circu-
products of an overheated imagination. larity or dependence on arbitrary and unde-
We have seen, therefore, that radical scep- fended assumptions is something that needs
ticism of a near-global kind does not place to be carefully distinguished from a stance of
its adherents under an obligation to prac- regarding those beliefs as arising from psy-
tise near-universal suspension of judgement. chological mechanisms that would generate
It should be seen instead as merely expos- false beliefs in many possible worlds close to
ing the hollowness of the supposed epi- the one within which we take ourselves to
stemic reasons that non-sceptics regard as exist.
constraining the beliefs they ought to hold. The first stance is one the Humean sceptic
All other motives for holding beliefs remain adopts in respect of almost all beliefs. The
potentially intact, although the balance of second stance, however, is potentially one
psychological forces that generate and sus- with much more radical implications at a
tain our beliefs is subtly rearranged. Beliefs doxastic level. This latter stance is associated
founded on loose, unstable principles of with either genuine suspension of judgement
association prove to be causally incapable on non-epistemic matters or, in those cases
of withstanding prolonged sceptical scru- where suspension of judgement fails to super-
tiny, and the problems exposed within our vene, a state of alienation from one’s beliefs
epistemic practices by sceptical argumenta- that mimics the attitude that a self-aware
tion have the effect of reducing intellectual kleptomaniac or sufferer from a phobia takes
arrogance and dogmatism. It follows that if towards the beliefs caused by his condition.
one is so constituted that one takes pleas- Widespread suspension of judgement or dox-
ure in complicated intellectual investiga- astic alienation is not a phenomenon associ-
tions, it is not the case that scepticism, even ated with the Humean sceptic, and it is the
when accepted as true, provides any reason foregoing distinction that explains how this
why one should puritanically refrain from lack of disruption manages to co-exist with
such activities. In the words Hume uses as extremely wide-ranging scepticism about the
Book One of the Treatise approaches its availability of rational justification. It also


9780826443595_Ch06_Final_txt_print.indd 164 2/9/2012 9:35:15 PM


serves to explain how the Humean sceptic It is one thing, however, to defend a habit of
can possess a high degree of confidence in inquiry from criticisms that we should aban-
the abilities of human beings to uncover the don this practice and to point out that such
truth about even recondite matters of inquiry. inquiry will inevitably lead us into momen-
The Humean sceptic views himself as fortu- tary spasms of complete conviction, and quite
nately located in an actual world that is one another thing to maintain that such moments
of those possible worlds in which many key provide us with beliefs that can be assembled
mechanisms of belief formation latch on to into a system that is beyond all criticism and
the truth and also do this in the vast major- challenge. Hume confesses, therefore, that
ity of neighbouring possible worlds. Thus although he has sometimes made use in the
he does not see his beliefs as the product of Treatise of ‘such terms as these, ’tis evident,
psychological mechanisms that generally fail ’tis certain, ’tis undeniable; which a due
to track the truth, and he is consequently deference to the public ought, perhaps, to
neither alienated from those beliefs nor is he prevent’ (THN / 274), these terms
inclined to abandon them. At the same time are to be understood merely as spontaneous
he does accept that their truth is still, in some expressions of momentary flashes of convic-
respects, a matter of unearned good fortune tion that fail to retain their full vigour when
in the sense that it is impossible to construct placed in a broader context of countervailing
any argument to defend the supposition considerations and general worries about the
that the actual world has the properties he adequacy of human reason.
ascribes to it that is not patently arbitrary or Hume, accordingly, ends Book One of the
question begging. Thus he does regard all his Treatise with an observation that once again
beliefs, other than his H-minimal beliefs, as alerts us to Hume’s wish to fuse philosoph-
lacking rational justification.6 ical scepticism and the experimental method
We arrive, therefore, at the combination into one integrated and mutually supportive
of thoughts that gives rise to the distinct- methodology for investigating both the nat-
ive character of Humean scepticism. The ural world and human psychology. Hume
Humean sceptic is confident that he has declares that these emphatic declarations
investigative techniques at his disposal that of certainty and conviction ‘were extorted
can reliably yield more true beliefs than from me by the present view of the object,
false beliefs. These are the techniques of ‘the and imply no dogmatical spirit, nor conceited
experimental method of reasoning’, which idea of my own judgement, which are senti-
is mentioned on the title pages of all three ments that I am sensible can become no body,
books of the Treatise. He also accepts, how- and a sceptic still less than any other’ (ibid.).
ever, that any argument intended to show As Hume is referring here to characteristics of
that this is the case must be probatively his own psychological constitution and deny-
defective. Ultimately all that underpins this ing that he possesses a ‘dogmatical spirit’ or
confidence is animal instinct and the posi- ‘conceited idea of my own judgement’, there
tive psychological reinforcement we receive is surely no good motive for Hume to say that
when our beliefs steer us away from pain or he is aware that these are dispositions that are
towards pleasure and also permit us to avoid not appropriate for a sceptic unless Hume is,
encountering too many situations where we at this point, thinking of himself as a scep-
are driven to revise our previous beliefs. tic.7 Thus Book One of the Treatise begins


9780826443595_Ch06_Final_txt_print.indd 165 2/9/2012 9:35:16 PM


and ends with statements that are revelatory mitigated form of scepticism seems to have
of the two central pillars of his overall philo- been Robert Fogelin’s paper ‘The Tendency