Consciousness, Time, and Scepticism in Hume's Thought 103267783X, 9781032677835

Citation preview

“A carefully crafted and detailed examination of some problems in Hume’s philosophy that have not received the attention they deserve. Falkenstein has provided an exacting analysis of the various relations between time, space, and the mind in Hume’s philosophy, a gift to Hume scholarship.” Wade Robison, Rochester Institute of Technology, USA “This book is an original, deep, and unfailingly valuable interpretation and interrogation of central aspects of Hume’s epistemology, metaphysics, and philosophy of mind. It merits and rewards close study.” Don Garrett, New York University, USA

Consciousness, Time, and Scepticism in Hume’s Thought

David Hume’s philosophical work presents the reader with a perplexing mix of constructive accounts of empirically guided belief and destructive sceptical arguments against all belief. This book reconciles this conflict by showing that Hume intended his scepticism to be remedial. It immunizes us against the influence of “unphilosophical” causes of belief, determining us to proportion our beliefs to the evidence. In making this case, this book develops Humean positions on topics Hume did not discuss in detail but that are of interest to contemporary philosophers: consciousness and the unity of consciousness, temporal experience, visual spatial perception, the experience of colour and other qualia, objective experience, and spatially extended minds. It also challenges currently accepted interpretations of Hume’s views on the finite divisibility of space and time, vacuum, the duration of unchanging objects, and identity over time. It deals with criticisms of Hume that were raised by his contemporaries, notably by Thomas Reid, draws attention to earlier seventeenth‑ and eighteenth‑century work that has bearing on the interpretation of Hume’s thought, and compares Hume’s achievements with those of later nineteenth‑century psychologists and philosophers. Consciousness, Time, and Scepticism in Hume’s Thought will appeal to scholars and advanced students interested in Hume, history of philosophy, and early modern theories of perception, time, and consciousness. Lorne Falkenstein is Emeritus Professor of Philosophy at Western Univer‑ sity, Canada. He is a co‑author of Logic Works: A Rigorous Introduction to Formal Logic (Routledge: 2022), co‑editor of the Broadview editions of Hume’s Enquires, Dissertation, and Natural History (2011–13), and has written many articles on early modern philosophy.

Routledge Studies in Eighteenth‑Century Philosophy

Human Dignity and the Kingdom of Ends Kantian Perspectives and Practical Applications Edited by Jan‑Willem van der Rijt and Adam Cureton System and Freedom in Kant and Fichte Edited by Giovanni Pietro Basile and Ansgar Lyssy Perspectives on Kant’s Opus postumum Edited by Giovanni Pietro Basile and Ansgar Lyssy Adam Smith and Modernity 1723–2023 Edited by Alberto Burgio Metaphysics as a Science in Classical German Philosophy Edited by Robb Dunphy and Toby Lovat Kant on Freedom and Human Nature Edited by Luigi Filieri and Sofie Møller Condillac and His Reception On the Origin and Nature of Human Abilities Edited by Delphine Antoine‑Mahut and Anik Waldow Consciousness, Time, and Scepticism in Hume’s Thought Lorne Falkenstein

For more information about this series, please visit: https://www.routledge.com/Routledge‑ Studies‑in‑Eighteenth‑Century‑Philosophy/book‑series/SE0391

Consciousness, Time, and Scepticism in Hume’s Thought

Lorne Falkenstein

NEW YORK AND LONDON

First published 2024 by Routledge 605 Third Avenue, New York, NY 10158 and by Routledge 4 Park Square, Milton Park, Abingdon, Oxon, OX14 4RN Routledge is an imprint of the Taylor & Francis Group, an informa business © 2024 Lorne Falkenstein The right of Lorne Falkenstein to be identified as author of this work has been asserted in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988. With the exception of the Introduction and Chapter 1, no part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. The Introduction and Chapter 1 of this book are freely available as downloadable Open Access PDFs at http://www.taylorfrancis.com under a Creative Commons Attribution-Non Commercial-No Derivatives CC-BY-NC-ND 4.0 license. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. ISBN: 978‑1‑032‑67783‑5 (hbk) ISBN: 978‑1‑032‑67789‑7 (pbk) ISBN: 978‑1‑032‑67788‑0 (ebk) DOI: 10.4324/9781032677880 Typeset in Sabon by codeMantra

Contents

Acknowledgementsix Abbreviations and short titles for primary sources xi

Introduction: Hume’s remedy for unphilosophical belief

1

1 Impressions: Colour, consciousness, temporal experience

19

2 Finite divisibility; manners of disposition; points

68

3 Time and our experience of time

117

4 Identity

146

5 The conception and perception of a vacuum

165

6 Belief: Normativity; objects

205

7 Causes of the belief in bodies

250

8 Reasons for scepticism about the external existence of bodies

278



303

Conclusion: Hume’s remedy

Index313

Acknowledgements

I am grateful to the University of Western Ontario for continuing to ­support my research in retirement. I am especially grateful to the staff at ­Western’s Weldon Library for their prompt attention to many digitiza‑ tion and other material requests. Great thanks also to Andrew Wecken‑ mann and R ­ osaleah Stammler at Routledge, and to Assunta Petrone at CodeMantra for their support and their assistance in bringing this work so quickly and smoothly to press. Over my years at Western, I have been inspired by many colleagues interested in the history of the enlightenment, the history of mathematics, the history of medicine, and the philosophy of mind: Roger Emerson, John Bell, David Bellhouse, Robert E. Butts, R ­ obert ­DiSalle, Corey Dyck, Benjamin Hill, Garth Kidd, Thomas M. ­Lennon, Doug Long, John Metcalfe, Robert Muehlmann, Paul Potter, Robert Stain‑ ton, Jean‑Pierre Schachter, John Thorp, and Christopher Viger. My work on Hume has benefitted from conversations and correspondence with Laura Berchielli, Graham Clay, Alastair Crosby, Maité Cruz, Dante Dauksz, Giovanni Grandi, James Harris, Hannes Ole Matthiessen, Nick Nash, Katharina Paxman, Dario Perinetti, ­Margaret Schabas, Scott Stapl‑ eford, Mikko Tolonen, James Van Cleve, David Welton, and Fred W ­ ilson. This work has been significantly improved in light of suggestions and criti‑ cisms offered by Dante Dauksz, Brigitte Sassen, and two anonymous re‑ viewers for the press. I am especially grateful to Donald Baxter for drawing attention to shortcomings with foundational parts of an earlier version of the manuscript. We have both devoted careers to putting Hume’s views on space and time in a more positive light. I regret that I have not been able to find a way to see these matters quite the way he does. His taking up as disa‑ greeable a task as studying and offering helpful comments on work criti‑ cal of his own is an instance of extraordinary ­generosity and dedication. I expect he will continue to disagree with much of what is said here, but it is much better for what he has already said. As I write this, I remember those whose support, material, exemplary, and professional, has made my work possible: Rochus Falkenstein (1926–2001); Rosemarie Falkenstein (1928–2003); Robert E. Butts (1928–1997).

Abbreviations and short titles for primary sources

Works by Hume References to Hume’s work by commentators who use other editions have been converted to references to the following editions. Where commenta‑ tors make dual references, only those to the following editions are retained.

Ab

Abstract

An abstract of a book lately published; entit‑ uled, A Treatise of Human Nature, &c. Cited by paragraph number in Treatise, 403–17.

Ad

“Advertisement”

Advertisement to the 1777 and remaindered earlier editions of the second volume of ETSS. Cited from Enquiry, 1.

Ax

Appendix

Appendix to Treatise. Cited by paragraph num‑ ber in Treatise, 396–401.

D

Dialogues

Dialogues Concerning Natural Religion, edited by Dorothy Coleman. Cambridge: Cambridge University Press, 2007. Cited by part and para‑ graph numbers.

DP

Passions

“A Dissertation on the Passions” in A Disserta‑ tion on the Passions; The Natural History of Religion: A Critical Edition, edited by Tom L. Beauchamp. Oxford: Clarendon Press, 2007. Cited by section and paragraph numbers. Edi‑ torial material is cited by page number follow‑ ing a comma.

xii  Abbreviations and short titles for primary sources E

Essays

Essays, Moral, Political, and Literary: A Criti‑ cal Edition, edited by Tom L. Beauchamp and Mark A. Box. 2 vols. Oxford: Clarendon Press, 2021. Work by Hume is cited by this edition’s specified short title for the individual essay and by paragraph number. Editorial material is cited by page number. (Page numbers are continuous across the two volumes.)

EHU

Enquiry

An Enquiry Concerning Human Understanding: A Critical Edition, edited by Tom L. Beauchamp. Oxford: Clarendon Press, 2000. Cited by sec‑ tion and paragraph number. Editorial material is cited by page number following a comma.

EPM

Morals

An Enquiry Concerning the Principles of Mor‑ als: A Critical Edition, edited by Tom L. Beau‑ champ. Oxford: Clarendon Press, 1998. Cited by section or appendix (EPMAx) and para‑ graph number. Essays and Treatises on Several Subjects. A collec‑ tion comprised of E, EHU, DP, EPM, and NHR issued in multiple editions with varying contents between 1756 and 1777. For full publication de‑ tails, see E1, xxiv–xxvii, 404–36, and 446–95.

ETSS

HE 1754

L

History

The History of England, edited by William B. Todd. 6 vols. Indianapolis: Liberty Fund, 1983. Cited by chapter number and, after a comma, page number of the volume containing that chapter. Where the occasion warrants, refer‑ ences are to the first (1754) edition of the first volume to be published (Edinburgh: Hamilton, Balfour, and Neill).

“Immortality”

“Of the Immortality of the Soul.” Cited by page number in Miller.

Introduction

Introduction to the Treatise. Cited by para‑ graph number in Treatise, 3–6.

Letters

Letters of David Hume, edited by J. Y. T. Grieg. 2 vols. Oxford: Clarendon Press, 1932. Cited

Abbreviations and short titles for primary sources   xiii by volume and letter number with page num‑ ber following a comma from the compilation by Mark C. Rooks for InteLex Corporation’s Past Masters Database, https://www.nlx.com. LG

A Letter from a Gentleman to His Friend in Ed‑ inburgh. Cited by paragraph number in Trea‑ tise, 419–31. New Letters

New Letters of David Hume, edited by Ray‑ mond Klibansky and Ernest C. Mossner. Oxford: Clarendon Press, 1954. Cited by ­ volume and letter number with page number following a comma from the compilation by Mark C. Rooks for InteLex Corporation’s Past Masters Database, https://www.nlx.com.

Miller

Essays Moral, Political, and Literary, edited by Eugene F. Miller. Indianapolis: Liberty Fund, 1987.

MOL

My Own Life

My Own Life. Cited by page number in Miller, xxxi–xli.

NHR

Natural History

The Natural History of Religion, in A Dissertation on the Passions; The Natural History of Religion: A Critical Edition, edited by Tom L. Beauchamp. Oxford: Clarendon Press, 2007. Cited by section and paragraph numbers. Editorial material is cited by page number following a comma.

“Suicide”

“Of Suicide” cited by page number in Miller.

Treatise

A Treatise of Human Nature: A Critical Edi‑ tion, edited by David Fate Norton and Mary J. Norton. 2 vols. Oxford: Clarendon Press, 2007. Volume 1, containing Hume’s text, is cited as T or Treatise, followed by book, part, section, and paragraph numbers, or just a page number when editorial content is being cited. Volume 2 is cited as T2 with page numbers following a comma. Where the occasion warrants, cita‑ tions are from the original editions of 1739 and 1740 (London: John Noon and London: Thomas Longman) using the imprints appear‑ ing in Eighteenth Century Collections Online.

NL

T T2 1739 1740

xiv  Abbreviations and short titles for primary sources Works by other authors

NIT

Port Royal

Arnauld, Antoine and Nicole, Pierre. 1996. Logic or the Art of Thinking, translated by Jill Vance Buroker. Cambridge: Cambridge Uni‑ versity Press.

Metaphysics

Aristotle. 1995. Metaphysics. In Aristotle Se‑ lections, translated by Terence Irwin and Gail Fine. Indianapolis: Hackett.

Confessions

Augustine of Hippo. 1997. Confessions. In Basic Issues in Medieval Philosophy, edited by Richard N. Bosley and Martin Tweedale. ­Peterborough: Broadview, 147–55.

Barrow 1685

Barrow, Isaac.  1685. Lectiones Mathematicæ XIII. London: George Wells.

Barrow 1734

Barrow, Isaac. 1734. The Usefulness of Math‑ ematical Learning Explained and Demon‑ strated, translated by John Kirkby. London: Stephen Austen.

Dictionnaire

Bayle, Pierre. 1991. Historical and Critical Dictionary: Selections, edited and translated by Richard H. Popkin. Indianapolis: Hackett.

Essay

Beattie, James. 1770. An Essay on the Na‑ ture and Immutability of Truth. Edinburgh: A. Kincaid and J. Bell. Citations note any dis‑ crepancies between this 1st edition and the 5th edition of 1774 (Edinburgh: William Creech), the last to have appeared in Hume’s lifetime. Berkeley, George. 1710. An Essay Towards a New Theory of Vision. 2nd ed. Cited from Ayers by paragraph number.

NTV

Principles

Berkeley, George. 1734. A Treatise Concern‑ ing the Principles of Human Knowledge. 2nd ed. Cited from Ayers by paragraph number.

Abbreviations and short titles for primary sources   xv

AT VII

Ayers

Berkeley, George. 1995. Philosophical Works Including the Works on Vision, edited by Michael R. Ayers. Everyman Library London: J M Dent.

Cyclopedia

Chambers, Ephraim. 1728. Cyclopedia. Lon‑ don: James and John Knapton and others. Cited by article.

Clarke‑Collins

Clarke, Samuel. 2011. The Correspondence of Samuel Clarke and Anthony Collins, 1707–08, edited by William L. Uzgalis. Peterborough: Broadview.

Demonstration

Clarke, Samuel. 1998. A Demonstration of the Being and Attributes of God, edited by Ezio Vailati. Cambridge: Cambridge University Press.

Traité

Condillac, Étienne Bonnot, Abbé de. 1947. Traité des sensations. In Oeuvres Philosophiques de Condillac, edited by Georges Le Roy. Vol. 1. Paris: Presses Universitaires de France.

Meditations

Descartes, René. 1984. Meditations on First Philosophy. In The Philosophical Writings of Descartes, edited by John Cottingham, Rob‑ ert Stoothoff and Dugald Murdoch, Vol. 2. Cambridge: Cambridge University Press. Cited by the pagination of volume 7 of the Adam and Tannery revised edition of Oeuvres de Descartes. Paris: Vrin / C.N.R.S., 1964–76.

Principles

Descartes, René. 1985. Principles of Philosophy. In The Philosophical Writings of Descartes, edited by John Cottingham, Robert Stoothoff and Dugald Murdoch, Vol. 1. Cam‑ bridge: Cambridge University Press. Cited by part and section number.

xvi  Abbreviations and short titles for primary sources Dioptrique

Descartes, René. 2000. Discourse on Method, Optics, Geometry, and Meteorology, trans‑ lated by Paul J. Olscamp. Rev. ed. Indianapo‑ lis: Hackett. Cited by the pagination of volume 6 of the Adam and Tannery revised edition of Oeuvres de Descartes. Paris: Vrin / C.N.R.S., 1964–76.

Euclid 1703

Euclid.  1703. Quæ supersunt omnia, trans‑ lated by David Gregory. Oxford. Commonly catalogued as Euclidis Elementorum libri XV. Accessit XVI. de solidorum regularium comparatione.

Euclid 1705

Euclid.  1705. The English Euclid, translated by Edmund Scarburgh. Oxford.

Euclid 1714

Euclid.  1714. The Elements of Euclid, trans‑ lated by Andrew Tacquet. 3rd ed. London: J. Roberts.

Euclid 1726c

Euclid.  1726. The Elements of Euclid Explain’d, English translation of the French translation of F. Claud. Francis Milliet de Chales (Claude‑François Milliet Dechales). 7th ed. London: James and John Knapton and others.

Euclid 1726h

Euclid. 1726. The Six First, Together with the Eleventh and Twelfth Books of Euclid’s Ele‑ ments, translated by Henry Hill London: Wil‑ liam Pearson.

URC

Feder 1787

Feder, Johann Georg Heinrich. 1787. Ueber Raum und Caussalität: zur Prüfung der Kan‑ tischen Philosophie. Göttingen: Dietrich.

PO

Physiological Optics

Helmholtz, Hermann von. 1962. Helmholtz’s Treatise on Physiological Optics, translated by James P. C. Southall. 3 vols. New York: Dover. Cited by volume, section, and page number.

AT VI

Abbreviations and short titles for primary sources   xvii

A/B

Elements

Hobbes, Thomas. 1992. Elements of Philoso‑ phy. In The English Works of Thomas Hobbes, edited by Mark C. Rooks. Charlottesville: In‑ teLex Corporation, 1992. https://www.nlx.com. Cited by part, chapter, and paragraph number.

Human Nature

Hobbes, Thomas. 1994. Human Nature and De Corpore Politico, edited by J. C. A. Gaskin. Oxford: Oxford University Press.

Principles

James, William. 1981. Principles of Psychol‑ ogy. 2 vols. Cambridge: Harvard University Press. Cited by volume and chapter number, in the pagination of the Harvard edition from the electronic edition Charlottesville: InteLex Corporation, 2008. https://www.nlx.com.

Kritik

Kant, Immanuel. 1781 and 1787. Kritik der reinen Vernunft. Riga: Johann Friedrich Hart‑ knoch. My own translations cited by the pagi‑ nation of the 1781 (A) and 1787 (B) editions.

Keill 1720

Keill, John. 1720. An Introduction to Natural Philosophy. London: William and John Innys.

Leibniz‑Clarke

Leibniz, Gottfried Wilhelm. 1717. A Collec‑ tion of Papers, Which Passed Between the Late Learned Mr. Leibnitz, and Dr. Clarke, in the Years 1715 and 1716, translated by Sam‑ uel Clarke. London: James Knapton. Cited by letter and paragraph number.

Essay

Locke, John. 1975. An Essay Concerning Hu‑ man Understanding, edited by Peter H. Nid‑ ditch. Oxford: Oxford University Press. Cited by book, chapter, and section number.

Search

Malebranche, Nicolas. 1997. The Search af‑ ter Truth, translated by Thomas M. Lennon and Paul J. Olscamp. Cambridge: Cambridge University Press. Referenced by book, part, chapter, and section number with page num‑ ber following a colon.

xviii  Abbreviations and short titles for primary sources

EIP

M

Dioptrica Nova

Molyneux, William. 1709. Dioptrica Nova. 2nd ed. London: Benjamin Tooke.

Principia

Newton, Isaac. 1729. The Mathematical Prin‑ ciples of Natural Philosophy, translated by Andrew Motte. 2 vols. London: Benjamin Motte. Cited by volume and page number.

Opticks

Newton, Isaac. 1730. Opticks. 4th ed. L ­ ondon: William Innys.

Pardies 1734

Pardies, Ignace Gaston. 1734. Short, But Yet Plain Elements of Geometry, translated by John Harris. 7th ed. London: D. Midwinter.

Review

Price, Richard. 1758. A Review of the Principal Questions and Difficulties in Morals. London: A. Millar.

“Abstract”

Reid, Thomas “[Abstract of the Inquiry].” In Inquiry, 257–62.

Inquiry

Reid, Thomas. 1997. An Inquiry into the Hu‑ man Mind, edited by Derek R. Brookes. Ed‑ inburgh: Edinburgh University Press. Cited by chapter and section number with page number following a comma.

Intellectual Powers

Reid, Thomas. 2002. Essays on the Intellectual Powers of Man, edited by Derek R. Brookes. Edinburgh: Edinburgh University Press. Cited by essay and chapter number with page num‑ ber following a comma. Sextus Empiricus. 1997. Adversus Mathema‑ ticos. In Hellenistic Philosophy, 2nd edition, translated by Brad Inwood and L. P. Gerson, 302–97. Indianapolis: Hackett.

Abbreviations and short titles for primary sources   xix Sextus Empiricus. 1997. Outlines of Pyrrhon‑ ism. In Hellenistic Philosophy, 2nd edition, translated by Brad Inwood and L. P. Gerson, 302–97. Indianapolis: Hackett.

PH

Opticks

Smith, Robert. 1738. A Compleat System of Opticks. 2 vols. Cambridge: Cornelius Crownfield. Cited by book, chapter, and sec‑ tion number.

Introduction Hume’s remedy for unphilosophical belief

Reacting to David Hume’s first, anonymously published book, Thomas Reid wrote, if [the mind] is indeed what the Treatise of human nature makes it, I find I have been only in an inchanted castle, imposed upon by spectres and apparitions. (Inquiry 1.6, 22)1 Hume would have approved of this assessment. Here is why. I.1  Epistemic determinism The themes taken up in Consciousness, Time, and Scepticism in Hume’s Thought are unified by their contribution to a remedy Hume proposed for a problem with belief. The problem is a practical problem. The remedy is likewise a practical remedy. Hume wrote that belief is the necessary result of placing the mind in such circumstances. It is an operation of the soul, when we are so situated, as unavoidable as to feel the passion of love, when we receive benefits; or hatred, when we meet with injuries. (EHU 5.8) Chief among these circumstances is what he called “custom.” For Hume, “custom” often refers to what is commonly experienced to be the case. But he also recognized a second kind of custom: what is commonly opined to be the case.2 Only the first kind of custom is “recogniz’d by philosophers” (T 1.3.9.19). Hume also identified other circumstances that can influence or necessitate belief (EHU 10.16–19; T 1.3.9–10; T 1.3.13). Many of these circumstances “have not had the good fortune” to be recognized by philosophers. Some are “disclaimed” by them (T 1.3.13.1; T 1.3.13.2). DOI: 10.4324/9781032677880-1 This chapter has been made available under a CC BY NC ND 4.0 license.

2  Introduction The distanced tone of these pronouncements is striking. “This is what philos‑ ophers have declared,” he said. He held back from endorsing their verdicts.3 Hume famously argued that there is no justification for believing that what has customarily happened up to now will continue to happen that way (EHU 4; T 1.3.6.4–11). This poses a problem. What entitles philoso‑ phers to approve of beliefs that are proportioned to common experience and disclaim those based on other factors? This is a problem that con‑ cerned Hume (EHU 1.11–12; T 1.4.7); it is at the root of the dispute be‑ tween commentators who take Hume’s primary message to be destructive and sceptical and those who take it to be constructive and naturalistic; and it has inspired as many critical observations, and as many interpretations as there are critics and defenders of his epistemology.4 As a determinist about belief, Hume approached this problem from a different angle than those who accept what he at one place called a “fan‑ tastical system of liberty” (T 2.3.1.15). Even were there a justification for basing belief on common experience, there would be no practical point to appealing to it unless our awareness of that justification were a cir‑ cumstance that could determine us to behave accordingly.5 On Hume’s account, philosophers have not in fact been determined to proportion be‑ lief to the evidence by reason or justifications. They have been placed in circumstances that have determined them to discover and accept a general rule to that effect.6 There is nothing they can say to those who have not been similarly placed that will convince them to follow their example or endorse their beliefs. They can only be apprehensive that those others will gain some measure of social control. Worse, when philosophers come to apply their own rule to particular cases, they often find that they are them‑ selves unable to resist the force of the special circumstances of those cases. Those special circumstances necessitate them to form beliefs that are dis‑ proportioned or even contrary to the evidence. Considering this discovery, philosophers find themselves naturally determined to look for and place themselves and others in some further circumstance that might more ef‑ fectively determine them to abide by the general rule. There is some indication that Hume thought that being impressed with the force of sceptical arguments is such a determinant. The indication is strongest in the Enquiry (12.24) and the Dialogues (1.8), but it is not ab‑ sent from the Treatise (1.4.7.11), which puts on more of a show of dismay over sceptical results, even while it is more desperate to establish them. Hume was fond of observing that no sceptical argument is strong enough to overcome the natural circumstances inducing us to form beliefs (EHU 12.23; T 1.4.7.9–10; and elsewhere). But he also maintained that some sceptical arguments naturally induce an enduring distrust (Morals 8.8 calls it “diffidence”) of our cognitive powers, a consequent degree of hesitation over our beliefs, and a greater readiness to abandon them as new

Section I.1  3 circumstances arise. For someone impressed by the force of sceptical argu‑ ments, only those beliefs repeatedly inculcated by prevailing circumstances may be able to overcome this diffidence. Those are the beliefs produced by repeated experience.7 Beliefs determined by circumstances that are more varied, temporary, and conflicting have less of an influence and that influ‑ ence does not last as long. As it turns out, the beliefs determined by the cir‑ cumstances recognized by philosophers are the ones those impressed by the force of sceptical arguments are compelled to accept, whereas the beliefs disclaimed by philosophers are the ones those impressed by the force of sceptical arguments hesitate over accepting. In Hume’s hands, scepticism is not destructive or problematic. It offers a remedy for “unphilosophical” belief. This answer to the question of how to reconcile the naturalist and scepti‑ cal tendencies in Hume’s thought has its roots in the work of commenta‑ tors who have recognized the role of “unphilosophical” factors in Hume’s psychology of belief,8 and of commentators who have underscored the Pyrrhonian elements in Hume’s thought.9 Rather than recognize opposed naturalistic and sceptical trends in Hume’s thought, it takes a practical problem to be exposed by Hume’s account of the causes of belief. It takes sceptical arguments to offer a practical remedy to this problem. It thereby weaves Hume’s naturalism and his scepticism into a consistent whole un‑ der the auspices of his determinism.10 Philosophers are professionally disposed to value considerations of war‑ rant over those of motivation. Philosophers disposed to defend Hume can be further disposed to look to his writings for a justification for proportion‑ ing belief to the evidence. This is a legitimate project. Hume did express concern over the proper guide for belief (EHU 1.10–11; T 1.4.7).11 But his struggles with this topic are compatible with recommending scepticism as an effective antidote to precipitate, obstinate, and intolerant belief (EHU 5.1 and 12.24), particularly for those not determined to respect reasoning. Treatise 1.4.7.11 suggests that sceptical arguments might serve as an anti‑ dote to beliefs that are not well supported by the evidence, there offering Hume hope that he might be able to pursue the research of the following two books of the Treatise in a way that is both scientific and sceptical, and the one because it is the other.12 Hume had many goals. One of them was “to fix some general rules, by which we may know when [objects] really are [causes or effects to each other]” (T 1.3.15.2). More generally, he was concerned to determine what we should endorse as a proper guide to belief. But it does no good to dis‑ cover the “rules by which to judge of causes and effects” (T 1.3.15) if we prove to be incapable of following them, even though we know and accept them. In addition to being a philosopher, determined to identify which beliefs are optimal, Hume was a historian, who was under no illusions

4  Introduction about the popular efficacy of any “logic” he or other philosophers might recommend. To oppose the torrent of scholastic religion by such feeble maxims as these, that it is impossible for the same thing to be and not to be, that the whole is greater than a part, that two and three make five; is pre‑ tending to stop the ocean with a bull‑rush. Will you set up profane reason against sacred mystery? No punishment is great enough for your impiety. And the same fires, which were kindled for heretics, will serve also for the destruction of philosophers. (NHR 11.5) Hume offered the outstanding early modern account of the motives for belief, surpassing even that offered by his fellow British determinist, Hob‑ bes.13 That achievement merits scrutiny. This book examines the features of Hume’s thought that give rise to his remedial proposal, assesses what his sceptical arguments do to real‑ ize it, and considers whether he overplayed those arguments to achieve his goal.14 It takes for granted that Hume was a determinist about belief. Granting that much, it studies the principles on which Hume’s sceptical arguments are founded (Chapters 1–5), his account of the causes of belief and how they give rise to the situation that calls for remedy (Chapter 6), and whether his sceptical arguments can perform the remedial task he as‑ signed to them (Chapters 7 and 8). I.2  Approach, method, and objectives In pursuing these topics, Consciousness, Time and Scepticism in Hume’s Thought (hereafter short‑titled Hume’s Remedy) draws on things Hume said to develop positions on topics he did not discuss: consciousness, ex‑ perience of the very recent past, and objectivity. It also criticizes him for saying what he did about the finite divisibility of space and time, the endur‑ ance of unchanging objects, the conceivability of a vacuum, identity, and the causes of the belief in body. Section I.3 discusses how these topics are connected. Hume’s Remedy gives pride of place to what Hume wrote in his au‑ thorized publications, generally giving them priority in discussion, and identifying divergences between them and the anonymous works.15 This having been said, Hume’s Remedy lets the chips fall where they may on the contentious issue of the relation between the authorized works and the Treatise. As a work in the history of philosophy, Hume’s Remedy aims to get the historical facts right. That means being open to the likelihood that the

Section I.2  5 historical subject, being a human being, will have made mistakes, taken wrong turns, and overlooked opportunities, including opportunities to fur‑ ther develop ideas first floated in earlier work and then abandoned. The last of these possibilities is not to be discounted. In Hume’s case, there are reasons for considering how his thought failed to evolve. Hume was not much given to changing his mind about things. He made it a principle never to reply to critics (MOL, xxxvi) and seems to have taken this so far as not to revise his works in any way that might be construed as making a reply to an increasingly large body of them.16 But this is just part of the story. Hume seems to have been constitutionally ill‑disposed to revising his earlier work. His response to difficulties was not to make changes but to make cuts, narrowing down on what he was most confident he had gotten right and excising the rest. While he frequently re‑edited, significant revisions or insertions tend to appear only within the first two subsequent editions. After that, his alterations are almost exclusively sty‑ listic, with one exception. He could be persuaded to shut up about things. There continue to be deletions and extractions to appendices or footnotes. He may never have given up on much of the deleted material. He may sim‑ ply have decided that his work would make more of an impact if it were omitted or downplayed. Because he was largely disinclined to rethink his original views, there are inconsistencies that were never eliminated, mis‑ takes that went uncorrected, and missed opportunities: ways in which he might have further developed his ideas that he did not pursue. Part of the job of the historian, having recognized shortcomings in the work and seri‑ ous objections raised by the more astute contemporary critics, is to assess the magnitude of those shortcomings. That can only be done by consid‑ ering what the historical figure would have been able to do to repair the work, given the resources available at the time. This means drawing only on examples, information, and ways of thinking that would have been cur‑ rent at the time. Even then, any repair that is proposed must be proposed as such, not attributed to the historical figure as if it were their own idea. Getting the history right means recognizing failures of clarity, cogency, and development, as well as mistakes, and not presenting improvements made on the figure’s behalf as if they were that figure’s own thoughts. Proposing ways in which the philosophy might have been better de‑ veloped at the time amounts to treating the philosophy, as well as the philosopher, as a historical subject, and considering how it might have developed had surrounding events taken a different course. The question is what Hume would have been able to say in reply to Reid or Kant, or what use he might have been able to make of theses put forward by Condillac, not what we would now be able to say on his behalf. The question is also what Hume did in fact think, for better or worse, on the topics that come up for examination. This second question continues

6  Introduction to preoccupy the main part of Hume’s Remedy. On some topics (the con‑ ception of a vacuum and the endurance of unchanging objects) Hume’s Remedy reaches verdicts opposed to the current scholarly consensus. I.3 Overview Hume offered many sceptical arguments. Some contribute to the prob‑ lem that his remedial proposal addresses; others figure in the remedy. The foundational arguments concern causal inference and necessary connec‑ tions. They establish that empirically guided inference cannot be rationally justified. The remedial arguments are offered over the course of Enquiry 12 (correspondingly, at Treatise 1.4.2.44–9 and 1.4.4.6–14).17 Scepti‑ cism about causal inference reappears at this second stage (at EHU 12.22, though not in T 1.4), but the remedial sceptical arguments Hume had prin‑ cipally in mind are those that establish doubt about our knowledge of an external world.18 Whereas scepticism about causal inference applies to all experience of regularities, whether in the motion of billiard balls or in the sequence of pain, passion, and volition, external world scepticism rests on a veil of per‑ ception argument, premised on a distinction between “bodies” (T 1.4.2.1) and “images presented by the senses” (EHU 12.8) or perceptions (T 1.2.6.7–9 and 1.4.2.2 at the end). The most serious challenge to Hume’s case for external world scepticism arises from his cavalier attitude to the nature of the “objects” as he was disposed to ambiguously call them, of our experience.19 Enquiry 2.1–2 opens by talking about impressions as exemplified by the pain of heat, the pleasure of moderate warmth, and a fit of anger. These are private sensory states. Publicly observable objects, instanced by mon‑ sters, golden mountains, and virtuous horses make a brief appearance in connection with ideas formed in imagination (EHU 2.4). Otherwise, En‑ quiry 2 and 3 only consider ideas as derived from impressions, and discuss their association. But in Enquiry 4 reference to impressions and ideas is abruptly abandoned, not to reappear except briefly, in Enquiry 5.2 when accounting for belief, and in Enquiry 7 when accounting for the idea of necessary connection. The bulk of Hume’s famous account of causal infer‑ ence appeals to regularity in the succession of species of external objects, like billiard balls, bread, wine, fire logs, and swords. It seems as if these objects are as directly perceived as impressions of pain or anger. But then, when Hume turned to talk about external world scepticism in Enquiry 12, “images presented by the senses” make a sudden appearance, posing a challenge to our acquaintance with external objects. Like the Enquiry, the Treatise fumbles with objectivity. Unlike the ­Enquiry, it draws a distinction between simple and complex perceptions (T 1.1.1.2). But then it proceeds to ignore the question of what limits the

Section I.3  7 complexity of perceptions. According to the Treatise, an apple is a com‑ plex impression, as if it were obvious why we parse the sensible points constitutive of a visual or tactile field as outlining that “image” rather than some other. And, like the Enquiry, for all its opening talk of impressions, ideas, and perceptions, whenever the Treatise turns to talk about relations (causal relations in particular) the language of impressions, ideas, and per‑ ceptions is abruptly abandoned in favour of a preponderant use of the term “object” exemplified by publicly observable external objects. Like a complex perception of an apple, an image captures how some‑ thing appears when displayed against a contrasting background. It is something that has been cut out, presenting a unity of form within a larger context. A visual perception, consisting just of coloured points disposed in space out to the edges of the visual field (like a piece of abstract art), is not an image. It is something more primitive. It is like a page of dots in a child’s game book, which reveals one figure when the dots are connected according to the instructions, but something else should they be connected in a different way. The dots are given where they are in space. But the child decides to connect them one way rather than another. Connecting the dots brings a figure/background distinction out of the whole field of spatially disposed points. It takes some work to account for how a visual experience, consist‑ ing of various spatially and temporally disposed coloured points, takes on the character of an image of a table or an apple. It then takes more work to identify images, which are temporary and perspective‑dependent, with multi‑faceted, mobile objects that change in regular ways over time. Kant referred to these operations as unifying a sensory “manifold” in the conception of an object. In speaking of images as being presented by the senses, the Enquiry takes the first of these operations, and so the second, for granted. Even on the supposition that images just are objects, Hume is exposed to the classic Kantian objection that he was only able to offer an empirical account of causal inference (as based on observed regularities in the succession of species of object) by taking the achievement of recogniz‑ ing objects for granted, neglecting the essential role of a priori concepts, such as those of substance and cause, in this operation.20 Hume’s Remedy argues that Hume had the resources to avoid the Kan‑ tian objection. His views on association by contiguity in space and on identity over time are key to that resolution. However the resolution calls for a reassessment of his sceptical arguments. Foundational work on how we recognize objects among the “bloom‑ ing buzzing confusion” of sensory experiences had already begun in the period before Hume. It was partly inspired by geometrical optics, which suggests that the information transmitted by the eye underdetermines what we see. More radically, it was inspired by Cartesian dualism, extended to idealism by Berkeley. The doctrine that minds are unextended scuttles

8  Introduction the view that we perceive the spatial properties and relations of things by experiencing extended, mental images, leaving Cartesians and Berkeleians attempting to account for how the mind gathers information about spatial properties and relations from a succession of purely qualitative sensations. Descartes assumed the mind must somehow be able to contemplate images imprinted in the brain or “corporeal imagination,” and Berkeley only got as far as accounting for visual depth perception, visual perception of objec‑ tive magnitude, single vision, and erect vision, without managing to reduce localization on the two‑dimensional field of view (let alone tactile experi‑ ence) to something more primitive. The goal of completing the project was bequeathed to their disciples. Berkeley’s claim that an “intelligence” capable of seeing but not feeling would not be able to understand the first principles of geometry (NTV 153–9) was a further step in that direction, as well as an early instance of a turn to consider the limits of what each sense tells us on its own, and how one might inform the others. Reid wrote his Inquiry as a series of investigations into what we are able to learn from each of our senses, asking what a being capable only of having tactile sen‑ sations could know about space (Inquiry 5.6) and what beings confined to a two‑dimensional world would think about the objects around them (Inquiry 6.9). Condillac argued that even though colours are disposed in space, touch needs to educate vision to see them that way (Traité 1.11.8, 3.3.1–13), and is the only sense that can acquaint us with an external world (Traité 2.4–5). In doing his own work on the elements of the science of human nature, Hume had a resource to draw on that these other authors did not. He had no commitment to dualism (“Immortality”; T 1.4.5), and so felt no pressure to recognize a problem of how the mind localizes pains or colour sensations in space. He could take it to be a fact, revealed by experience, that visual and tactile sensations are originally given as spatially disposed. If that means that mental states are spatially disposed, so much the worse for the dualists. The thesis that sensory experience exhibits spatial and temporal struc‑ ture antecedent to any operations of the imagination or understanding is one that Hume shared with Kant. Kant ruminated that it reflects subjec‑ tively necessary conditions under which it is possible for us to have sensa‑ tions; Hume just took it to be a further feature of our experience. Those early modern philosophers who were convinced that minds are not in space were forced to either maintain that the spatial structure of experience is constructed rather than intuited or invoke innate ideas. Kant had the op‑ posite problem. He needed to look for special reasons to deny the “tran‑ scendental reality” of space and time.21 Hume’s approach makes it possible to offer a strictly empirical account of objectivity (Sections 6.10–12).

Section I.3  9 Taking the spatial and temporal arrangement of coloured and tangible points to be intuited (or “perceived”) is a crucial first step to avoiding Kantian synthesis under a priori concepts. Hume devoted parts of Treatise 1.2 to discussing our experience of space and time. It was there that he introduced the notion that space and time are experienced as “manners of disposition” of unextended impressions and ideas. He did not return to that topic in his authorized publications. Recognition of the spatial complexity of our most primitive experiences is nonetheless as much a feature of the Enquiry as the Treatise. While there is no dedicated discussion of spatial “manners of disposition” in the En‑ quiry, they are still there, from the shades of blue of Enquiry 2.8, which are said to be experienced as descending along a line in space in order of their phenomenally experienced intensity, to the assertion that we as‑ sociate ideas depending on how those ideas are spatially and temporally disposed (EHU 3.3 and 5.17), to the talk of “images” that are “presented by the senses.” Recognizing space and time as manners in which impressions are dis‑ posed means recognizing that what the author of the Treatise called “complex impressions” are not the product of any cognitive operation performed on simple impressions.22 Section 1.6 shows that Hume thought that complex sense impressions are “original,” that is, “such as without any antecedent perception arise in the soul” (T 2.1.1.1). The simple parts into which they might be divided are only rarely given in isolation, as simple impressions. In the case of visual or tactile points, they never are. A point is only experienced as such insofar as it is hemmed in on all sides by surrounding points. Taking complex impressions to be originally given as extended over space implies that minds are spatially extended. Hume embraced this con‑ sequence. His most astute contemporary critic, Thomas Reid, was deeply troubled by it. Encountering it in the Treatise led Reid to reject his early acceptance of Berkeley’s immaterialism. As Reid understood it, Hume had shown that Berkeley’s principles lead to a consequence far more troubling than the rejection of an external world: that “the mind either is no sub‑ stance, or that it is an extended and divisible substance, because the ideas of extension cannot be in a subject which is indivisible and unextended.”23 In the early 1760s, some 14 years after publishing the first edition of the En‑ quiry, Hume corresponded with Reid about this topic (Inquiry, 255–65). He did not take the occasion to object that Reid had misinterpreted him. Section 1.2 argues that Reid understood Hume correctly. Sections 1.3–5 defend Reid’s interpretation of Hume against objections raised by Reid himself and by James Beattie. In doing this work, Sections 1.2–5 pre‑ sent Hume and Reid as champions of two starkly opposed views of the

10  Introduction representation of the spatial properties and relations of sensations and the objects of experience. On the Humean account, to have an idea of extension is to have an extended idea (T 1.4.5.15). On Reid’s account, to perceive an extended object is to perform an act of conception that acquaints us with that object, even though the act itself is not extended and does not other‑ wise resemble that object (Inquiry 6.20, 168). According to Sections 1.3 and 1.4, the dispute between the two comes down to rival intuitions about the nature of our experience of colour.24 In light of this dispute, it is noteworthy that, in an attempt to argue that we have no conception of a vacuum, Hume conceded to Reid that it is possible to have sensations of colour that are nowhere in space (T 1.2.5.11–12).25 This turn of events requires anyone wishing to defend Hume’s position on representation to engage what he said about vacuum. Section 2.6 and Chapter 5 show that Hume had only one argument for denying the con‑ ceivability of a vacuum, an argument from the non‑entity of unqualified points (T 1.2.3.12–17). Section 2.6 shows that this argument is flawed. Hume also sought to provide an alternative explanation of what we expe‑ rience when we “falsely imagine” we are perceiving a vacuum. Chapter 5 argues that this explanation tacitly invokes the very conception it seeks to replace. Fortunately for those wishing to defend Hume’s position on representation, he was unable to escape the intuition that we can only ex‑ perience two lone coloured points as spatially contiguous or separated by a space where there is nothing visible or tangible. Section 1.9 initiates a parallel discussion of Hume’s views on time. Those views entail that, just as the idea of extension is an extended idea, so the idea of time is an idea that takes time to occur. But, like everyone else be‑ fore Einstein, Hume was a “presentist.” He thought that what is past no longer exists, that the future does not yet exist, and that what does exist is confined to a simple and indivisible moment (T 2.3.7.5, 1.2.2.4, 1.2.3.8).26 This moment is perpetually perishing and perpetually renewed. Such an account would appear to confine all knowledge of the past to the experi‑ ence of presently occurring traces or images. But, barring innate ideas, no collection of presently existing images can give us the idea of succession. Hume accepted that we do have that idea, but he did not consider how that could be possible (Costa 1990, n9). Section 1.9  addresses this difficulty with work that is critically constructive. It presents Hume as someone who was capable of missing opportunities and making mistakes, but it also looks for a way he could have done better, by appealing to other things he had to say. The thesis that impressions and ideas are extended and take time to oc‑ cur does not just have implications for theories of mind and accounts of temporal experience. It creates problems for accounts of consciousness.

Section I.3  11 If impressions and ideas are spatially distributed and take time to occur, and impressions and ideas are conscious states, consciousness is likewise distributed over space and composed of distinguishable temporal parts. William James famously charged that this is impossible, observing that “A succession of feelings, in and of itself, is not a feeling of succession” (Prin‑ ciples 1.15, 591) and that when twelve people standing in a row each think one word of a twelve‑word sentence “nowhere will there be a conscious‑ ness of the whole sentence” (Principles 1.8, 162). Section 1.6 refers to this as the “comprehension problem.” Versions of this problem were raised in Hume’s time by Bayle (Dictionnaire, 130) and Clarke (Clarke‑Collins, 47), and were taken to prove that consciousness, and so the mind must be indivisible and so unextended. There is a further problem, referred to in Section 1.7 as the “limitation problem.” Supposing that impressions and ideas extend over space and time, what fixes their bounds? Hume naïvely thought that the bounds of complex impressions and ideas are fixed by the surfaces of objects like apples and tables. But this opens him to the Kantian objection. What de‑ termines us to draw the bounds as we do rather than in some other way? Section 1.8 argues that Hume had the resources to address these prob‑ lems. One implication of that work is that Hume could have modified his account of complex impressions to provide an account of conscious‑ ness. Another is that, though he seems not to have fully realized it, he cannot have been a psychological atomist.27 The thesis that space and time are originally perceived (not subsequently constructed) entails that complex impressions are not just aggregates of parts but wholes that display parts as spatially and temporally related, where the relations are not determined by anything in the parts considered individually. The parts need not be qualitatively different. They can be differentiated just by how they are disposed within the whole. Their manner of disposi‑ tion is additional information, not to be found in the parts considered individually or in their bare collectivity (T 1.3.1.1). It is instead given in what Hume called the “perception” (immediate experience) of a whole (T 1.3.2.2; Section 2.4.1). Kant thought that a determination of the conditions of the possibility of experience of sensory images and external objects could do something to answer sceptical arguments. If Hume could have offered an account of objectivity, would it have undermined his sceptical arguments and the suc‑ cess of his remedial proposal? In the Enquiry, Hume appealed to a veil of perception argument under‑ written by a Berkeleian argument from the phenomenal status of sensible qualities. He opposed these arguments to a blind and powerful, natural instinct or prepossession to suppose that there is an external world. The author of the Treatise appealed to the same arguments but also sought to

12  Introduction establish that the belief in an external world is based on trivial qualities of the fancy conducted by false suppositions. Chief among the false sup‑ positions is the supposition that interrupted but resembling objects can be identical. Chapter 7 argues that the considerations that would have enabled Hume to explain how we come to recognize objects undermine his claim that the belief in body rests on trivial qualities of the fancy conducted by false suppositions.28 Chapters 6 and 7 also argue that the spatiality of visual and tactile experience makes it possible to justify identity attributions to interrupted objects. Hume was led to think otherwise by a flawed identity theory, hobbled by the supposition that unchanging objects do not endure. Chapter 4 presents an alternative “free Humean identity theory” (one free of the supposition that there are no monotonous successions). Chapter 3 shows that Hume had no good reason for rejecting the endurance of un‑ changing objects.29 Section 3.4 argues that the rejection is inconsistent with his presentism.30 However, Chapter 8 also argues that the sceptical arguments of the En‑ quiry, along with their prototypes in Treatise 1.4.2.45 and 1.4.4.6–14, retain their power.31 “Images presented by the senses” may lead us to rec‑ ognize objects, but these objects are not external objects. They are natu‑ rally taken for external objects, but they are demonstrably distinct from them. As Reid complained, if Hume is right, we find ourselves in an en‑ chanted castle, imposed upon by spectres and apparitions. These remedial sceptical arguments stand as “proof against proof.” They do not disprove a belief based only on disreputable foundations. They op‑ pose a belief that is as well warranted as any can be. They leave us readier to question and abandon our beliefs, but cannot overcome our disposition to be guided by the evidence. Notes 1 See “Abbreviations and Short Titles for Primary Sources” (pp. xi-xviii above) for a key to references to works first published prior to the twentieth century. 2 DP 2.33; T 1.3.9.16–19. Treatise 1.3.9.16 speaks of “other kinds of custom” and remarks that “custom … may operate upon the mind … after two several ways.” The second of these ways, there labelled “education” is said to give rise to opinions that “take such deep root, that ’tis impossible for us, by all the powers of reason and experience, to eradicate them; and this habit not only approaches in its influence, but even on many occasions prevails over that which arises from the constant and inseparable union of causes and effects” (T 1.3.9.17). 3 He did not always do so (Qu 2016, 59). This is only to be expected. He will himself have been determined by circumstances to believe certain things to be improbable, to view those who accept them as ridiculous and contemptible, and to express those sentiments (Falkenstein 1997; Garrett 1997, 157–9). See

Notes  13 the Natural History from start to finish. This did not preclude him from rec‑ ognizing that those sentiments were “extorted” from him (T 1.4.7.15) or from recognizing the force of sceptical arguments against his own practices (EHU 12.22) and of common opinions contrary to his own (T 1.3.14.24, especially at the end). 4 Qu (2019) provides a useful survey of interpretations in the more recent lit‑ erature. The suggestion made in this book, that Hume’s scepticism was reme‑ dial, is not mentioned. The opposed sceptical and naturalistic interpretations of Hume were forged by Reid, Inquiry and Intellectual Powers, and Kemp Smith (1905, 1941). 5 This is not to be discounted. As Hume noted (DP 2.33), our opinions of our‑ selves are determined by the opinions that others have of us, making the desire to earn their admiration and approval one of the driving forces of all our be‑ haviour. This can extend to doing what earns us a reputation for wisdom and good judgment. It just depends on whether we find ourselves in the company of those who admire others naturally endowed with an ability to proportion belief to the evidence, or of people who place a premium on some other quality, such as perseverance in original opinions, loyalty to a party line, or implicit faith in the words of demagogues or priests. 6 EHU 8.13 quoting T 1.3.12.5, EHU 9.5n, “Rise and Progress” (E1, 102). The circumstances enumerated in these passages include character traits, cogni‑ tive abilities, and environmental factors: curiosity and the opportunity and resources to pursue it, along with experience and the abilities involved in gath‑ ering and codifying information from what it offers and drawing logical infer‑ ences from it. The role of general rules and of curiosity in Hume’s thought has been exhaustively studied by Wilson (2008). 7 Tenets promulgated by education and popular opinion can also be repeat‑ edly encountered, but wherever opinion bears on divergent interests it can be counted upon to be factious (interest being a determinant of belief by way of arousing passions [T 1.3.10.4]). Opinion is also frequently contrary to experi‑ ence, and in these cases is often refuted by subsequent events. Seeing is believ‑ ing, and can produce a disposition to distrust the words of others when they lead to disappointment. 8 They include MacNabb (1951, 95–100), Norton (1994), Falkenstein (1997), Loeb (2002, 101–38), and Fogelin, (2009, 29–38). 9 Notably Popkin (1980, esp. 127–32), and Baxter (2008, 9–14). 10 The most complete earlier expression of this approach is Falkenstein (1997). Fogelin (2009) intimates a similar position. Williams (2004, 290–2) comes close but is voluntarist (273, 274). Loeb emphasizes the importance of Hume’s recognition of a distinction between philosophical and unphilosophical beliefs (2013 and the other work referenced there) but seeks to draw epistemological conclusions rather than recognize the remedial role of sceptical arguments in a context where belief is determined by circumstances. Ainslie (2015, 8) consid‑ ers our ability to overcome sceptical arguments to be a crucial challenge for an adequate theory of mind to address. 11 Enquiry 1.12 declares that “Accurate and just reasoning,” rather than scep‑ ticism, “is the only catholic remedy [to abstruse philosophy and metaphysi‑ cal jargon mixed up with popular superstition], fitted for all persons and all dispositions[.]” This is the only case where “remedy” is used as a noun in the Enquiry. In contrast, sceptical arguments are only recommended as something that “may be … useful” (EHU 12.24). However, the preponderance of the

14  Introduction evidence indicates that Enquiry 1.12 ought to be regarded as the kind of hyper‑ bole that is only to be expected in the introduction to a popular essay involving a good deal of accurate and just reasoning. The declaration that accurate and just reasoning is fitted for all persons and all dispositions is contradicted by Enquiry 1.3–6, which allows that not everyone is temperamentally suited to abstract and profound reasoning. Natural History 6.1 goes further, speaking of, “the ignorance and stupidity of the people, and their incurable prejudices.” Enquiry 9.5n enumerates factors preventing many people from having a good capacity for refined reasoning. Hume may have deliberately concealed his pur‑ poses until after displaying his strongest sceptical arguments, like a physician concealing the surgical instruments from the eyes of the patient. This possibility is further considered in Conclusion, Section 2. 12 The concluding two sentences of this paragraph, often referred to as the “ti‑ tle principle” (Garrett 1997, 234), are a follow‑up to this observation, not a self‑standing principle. Hume’s determinism about belief entails that where rea‑ son is lively and joined to some propensity it will be believed, perhaps more in virtue of the “propensity” than the reasoning. The “ought” is more predictive than normative, though predictive of a result Hume felt determined to endorse, that the reasoning that “ought” to be assented to is reasoning that is lively enough to overcome the force of sceptical defeaters. 13 Hume did not offer a dedicated theory of the causes of unphilosophical belief in his authorized works, but he repeatedly and extensively invoked the factors originally identified in the Treatise in his later work. Notable examples are Enquiry 10.4 and 10.16–18 and almost the whole of the Natural History. See Conclusion, Section 2 on Hume’s philosophical psychology in the History. 14 Loeb (2002, 177–207), argues that Hume overplayed his sceptical arguments, but to the detriment of his positive theses rather than because he was pursuing a remedial project. 15 As indicated by his 1751 letter to Gilbert Eliot of Minto (L1 #73, 158) and “Advertisement,” Hume became increasingly dissatisfied with his anonymous Treatise, to the point of advising correspondents not to read it and condemning critics for focusing on it. This topic has been raised to prominence by Mil‑ lican (2002) and is further pursued by Merivale (2018). Taylor (2015, 120–9 and 148–52) identifies various ways in which Morals is superior to Treatise 2–3. In attending to how the authorized and anonymous works differ, I follow Qu (2020) though not always to the same conclusions. We draw on different evidence. 16 There is some evidence to the contrary. The appendix to the Treatise may be in part a response to his friends’ early reactions to its first two books. He revised his Essays in light of comments on the first edition sent to him or his publisher by an anonymous critic (E, 411–12). He declared himself to be diffident of the quality of his work and anxious to receive criticisms and make corrections in a letter of 14 March 1740 to Hutcheson (L1 #16, 38; see also #10, 29 to Pierre Desmaizeaux). But, as he said at a point in the History that I can no longer locate, the character of a King or Queen is only appropriately delineated at the end of the reign because it is only then that we are in a good position to determine, from a lifetime of action, what the governing passions and other settled character traits were. The young author of the Treatise, seeking to make a name for himself, win an audience, and impress important and established contemporaries will have been driven by circumstances not necessarily reflec‑ tive of his character. The mature Hume may have taken himself to be engaged

Notes  15 in a battle with religious bigots to whom he could afford to give no quarter. Or he may have felt more secure and confident in his position and shown his true disposition accordingly. Either way, he was not much given to changing his mind about things. 17 Only the Enquiry excuses sceptical arguments by appealing to their utility for various ends. The Treatise is not so forthright. Though it puts on more of a show of dismay over sceptical results, it seems more desperate to multiply and amplify them, even at the cost of implausibility (especially T 1.4.1), complexity (especially T 1.4.2.23–43), and inconsistency with principles announced earlier (T 1.4.2.3–14 and 20–23). The author of the Treatise seems to have been con‑ cerned to establish that he had been forced by the facts into recognizing scepti‑ cal results, contrary to his intentions and to his own consternation (T 1.4.2.56 and 1.4.7). The mature Hume was more comfortable reporting on select scepti‑ cal arguments and conclusions, which he distanced himself from, but excused (EHU 5.1–2) and declared to have multiple utilities (EHU 12.24–26). 18 Treatise 1.4.1 and Appendix 10–21 mention further arguments, but they are not retained in the authorized publications. (Hume’s position at the close of Treatise 1.4.1 is also not an effective version of scepticism, both because the sceptical argument is questionable [Millican 2018] and because he considered it impossible for us to carry the argument through to its conclusion and so impossible for it to produce the amazement and confusion that have a remedial effect. The Appendix is for its part more a confession of confusion and failure than a sceptical argument.) Enquiry 12.18–20 mentions yet other arguments but appends a note that they might be answered along lines proposed in Trea‑ tise 1.2.1 and 1.2.4. 19 Concerns with what Hume meant by “object” and related concerns with his account of the relation between impressions, perceptions, and objects are wide‑ spread. The classic study of this topic is Grene (1994). 20 See Beck (1979, 76–8) and Price (1940, 7–9) for classic statements of this Kantian objection to Hume. Landy (2015) offers a more recent and extensive Kantian critique of Hume. Rocknak (2013) attributes a quasi‑Kantian position to Hume. Wilson (1997, ch1) contests the thesis that Hume was in any way Kantian. 21 As detailed in Falkenstein (1991), the problem was already drawn to his atten‑ tion in the early 1770s when Moses Mendelssohn and Johann Lambert wrote to him to object that even an idealist must accept that mental states are really successive in time. 22 Hume endorsed an argument given by Malézieu, premised on the assertion that “existence in itself belongs only to unity, and is never applicable to number, but on account of the unites, of which the number is compos’d” (T 1.2.2.3). How‑ ever, that argument concerns what might be called the ontogenesis of wholes and parts whereas the point here concerns the psychogenesis of complex im‑ pressions. The shortcomings of Hume’s endorsement of the Malézieu argument are discussed in Sections 1.6 and 2.5. 23 Reid Inquiry 7, 217. Reid alluded to his early acceptance of Berkeley’s imma‑ terialism at Intellectual Powers 2.10, 142. The “other consequences” referred to there as inducing him to change his mind must have been those described at Inquiry 7, 217. 24 Falkenstein (2000) argues for preferring Hume’s side in this dispute on account of its ability to deal with colour. Hume’s position is also defended by Garrett (2006) independently of a contrast with Reid.

16  Introduction 25 Reid did not pick up on this in any of his published work and, so far as I know, neither has anyone else. Hume’s position on the conception of a vacuum has only received sustained attention from scholars of this generation. (The out‑ standing earlier treatment is Broad 1961, 171–6.) Recent work understandably seeks to show why this hitherto neglected topic merits attention. Were it not for such careful and compelling sympathetic assessments as those offered by Costa (1990), Baxter (2009, 2016), Boehm (2012), Kervick (2016), and Cot‑ trell (2019a, 2019b) the present, more critical assessment would have fallen into a vacuum of its own for lack of suitable opposition. 26 Baxter (2008, 22–3) disagrees. His position is considered in Section 1.9. 27 I take psychological atomism to be the thesis that there are no complex impres‑ sions, only complex ideas. All relations are the products of acts of association or comparison performed on previously given simples. Psychological atomism is attractive to those who think that no mental states are spatially disposed and that all things must be recreated from moment to moment. It was also attrac‑ tive to Bayle and Leibniz whose work on the paradoxes of the continuum led them to declare space and time to be only ideal. Hume denied psychological atomism at Treatise 1.3.2.2, but did not always follow through on that denial, the most serious lapse being at Treatise 1.2.3.16. 28 Loeb (2002, 177–207) argues for a related thesis. 29 This charge is also defended by Falkenstein (2017). 30 This work challenges a currently received account of Hume’s views on dura‑ tion, most fully presented by Baxter (2008, ch2, ch3). 31 In this, I do not agree with Loeb (2002, 208–11).

Bibliography Ainslie, Donald. 2015. Hume’s True Scepticism. Oxford: Oxford University Press. Baxter, Donald L. M. 2008. Hume’s Difficulty: Time and Identity in the Treatise. London: Routledge. Baxter, Donald L. M. 2009. “Hume’s Theory of Space and Time in its Skeptical Con‑ text.” In The Cambridge Companion to Hume, 2nd edition, edited by David Fate Norton and Jacqueline Taylor, 105–46. New York: Cambridge University Press. Baxter, Donald L. M. 2016. “Hume on Space and Time.” In The Oxford Hand‑ book of Hume, edited by Paul Russell, 173–90. Oxford: Oxford University Press. Beck, Lewis White. 1979. “A Prussian Hume and a Scottish Kant?” In McGill Hume Studies, edited by David Fate Norton, Nicholas Capaldi, and Wade L. Robison, 63–78. San Diego: Austin Hill Press. Boehm, Miren. 2012. “Filling the Gaps in Hume’s Vacuums.” Hume Studies 38: 79–99. Broad C. D. 1961. “Hume’s Doctrine of Space.” Proceedings of the British Acad‑ emy 47: 61–76. Costa, Michael J. 1990. “Hume, Strict Identity, and Time’s Vacuum.” Hume ­Studies 16: 1–16. Cottrell, Jonathan. 2019a. “Hume on Space and Time: A Limited Defense.” In The Humean Mind, edited by Angela Coventry and Alex Sager, 83–95. Oxon: Routledge.

Bibliography  17 Cottrell, Jonathan. 2019b. “Hume’s Answer to Bayle on the Vacuum.” Archiv für Geschichte der Philosophie 101: 205–36. Falkenstein, Lorne. 1991. “Kant, Mendelssohn, Lambert, and the Subjectivity of Time.” Journal of the History of Philosophy 29: 227–51. Falkenstein, Lorne. 1997. “Naturalism, Normativity, and Scepticism in Hume’s Account of Belief.” Hume Studies 23: 29–72. Falkenstein, Lorne. 2000. “Reid’s Account of Localization.” Philosophy and Phe‑ nomenological Research 61: 305–28. Falkenstein, Lorne. 2017. “Hume on Temporal Experience.” In The Routledge Handbook of Philosophy of Temporal Experience, edited by Ian Phillips, 42–52. Oxon: Routledge. Fogelin, Robert J. 2009. Hume’s Skeptical Crisis. A Textual Study. Oxford: Oxford University Press. Garrett, Don. 1997. Cognition and Commitment in Hume’s Philosophy. New York: Oxford. Garrett, Don. 2006. “Hume’s Naturalistic Theory of Representation.” Synthese 152: 301–19. Grene, Marjorie G. 1994. “The Objects of Hume’s Treatise.” Hume Studies 20: 163–77. Kemp Smith, Norman. 1905. “The Naturalism of Hume (I).” [and (II)] Mind 14: 149–73 [and 339–47]. Kemp Smith, Norman. 1941. The Philosophy of David Hume. London: Macmillan. Kervick, Dan. 2016. “Hume’s Perceptual Relationism.” Hume Studies 42: 61–87. Loeb, Louis E. 2002. Stability and Justification in Hume’s Treatise. Oxford: Ox‑ ford University Press. Loeb, Louis E. 2013. “Epistemological Commitment in Hume’s Treatise.” In Ox‑ ford Studies in Early Modern Philosophy VI, edited by Daniel Garber and Don‑ ald Rutherford, 309–48. Oxford: Oxford University Press. MacNabb, D. G. C. 1951. David Hume: His Theory of Knowledge and Morality. Reprinted in facsimile. Oxon: Routledge: 2019. Merivale, Amyas. 2018. Hume on Art, Emotion, and Superstition: A Critical Study of the Four Dissertations. Boca Raton: Routledge. Millican, Peter. 2002. “The Context, Aims, and Structure of Hume’s First En‑ quiry.” In Reading Hume on Human Understanding, edited by Peter Millican, 27–65. Oxford: Oxford University Press. Millican, Peter. 2018. “Hume’s Pivotal Argument, and His Supposed Obligation of Reason.” Hume Studies 44: 167–208. Norton, David Fate. 1994. “How a Sceptic May Live Scepticism.” In Faith, Scepti‑ cism and Personal Identity, A Festschrift for Terence Penelhum, edited by J. J. MacIntosh and H. A. Meynell, 119–39. Calgary: University of Calgary Press. Popkin, Richard H. 1980. “David Hume: His Pyrrhonism and His Critique of Pyr‑ rhonism.” In The High Road to Pyrrhonism, edited by Richard A. Watson and James E. Force, 103–32. San Diego: Austin Hill Press. Price, H. H. 1940. Hume’s Theory of the External World. Oxford: Clarendon Press.

18  Introduction Qu, Hsueh M. 2016. “Hume’s Doxastic Voluntarism.” Mind 126: 53–92. Qu, Hsueh M. 2019. “Hume’s Epistemology: The State of the Question.” Southern Journal of Philosophy 57: 301–23. Qu, Hsueh M. 2020. Hume’s Epistemological Evolution. New York: Oxford ­University Press. Rocknak, Stefanie. 2013. Imagined Causes: Hume’s Conception of Objects. Dordrecht: Springer. Taylor, Jacqueline A. 2015. Reflecting Subjects: Passion Sympathy and Society in Hume’s Philosophy. Oxford: Oxford University Press. Williams, Michael. 2004. “The Unity of Hume’s Philosophical Project.” Hume Studies 30: 265–96. Wilson, Fred. 1997. Hume’s Defence of Causal Inference. Toronto: University of Toronto Press. Wilson, Fred. 2008. The External World and Our Knowledge of It. Hume’s Criti‑ cal Realism, an Exposition and a Defence. Toronto: University of Toronto Press.

1 Impressions Colour, consciousness, temporal experience

This chapter examines Hume’s account of impressions and its consequences. Impressions of reflection are passions. Impressions of sensation are “sensible qualities”: phenomenally experienced colours, smells, tastes, sounds, and tactile feelings that are inconceivable to those who lack the associated sense organ. Impressions are in the mind in the sense of being locally conjoined with the mind, not in the sense of being its intentional objects or its states or modifications. Visual and tactile impressions are also disposed in space. Reid accused Hume of taking minds to be either extended and divisible substances or no substance (nothing to speak of). Hume never took exception to this interpretation. This chapter defends it against objections that were raised at the time and considers what it implies for the nature of consciousness and temporal experience. Hume had no theory of consciousness, but the direction of his thought committed him to taking it to be extended without detriment to its unity and to extend into the very recent past, again without detriment to its unity. The latter view is inconsistent with presentist commitments Hume will have shared with everyone else. That tension was never noticed or resolved. 1.1

The nature of impressions

Anyone wishing to study Hume’s enquiries into human understanding must first develop a working notion of what he took impressions to be. Hume thought so himself. The Enquiry gets down to business at 2.1–3, and the Treatise begins at 1.1.1.1 by explaining what impressions are. Both works describe them as one of two main kinds of “perceptions of the mind.” In the Enquiry, “perceptions of the mind” are instanced by feeling the pain of excessive heat or the pleasure of moderate warmth, being in a fit of anger, and being in love. Impressions are further identified using the terms “sentiment,” “sensation,” “emotion,” “passion,” and “affection.” Later, impressions are instanced by the relish of wine and colours in all their hues and shades. The colours that are spoken of in this section are DOI: 10.4324/9781032677880-2 This chapter has been made available under a CC BY NC ND 4.0 license.

20  Impressions likely not wavelengths of light or any other object that the physical ­science of Hume’s day recognized as part of the external world. They are the sort of colours that are inconceivable to those with no sense of sight. The per‑ son who notices that a shade of blue is missing is not said to base that assessment on a measurement of the wavelengths of light reflected by the different colour patches. Hume most likely thought that the person’s judge‑ ment is based on differences between immediately apparent, phenomenal colour qualities, and used sensible quality terms, like “blue,” to refer to such qualities. In the first book of the Treatise, impressions are identified in the same way. The terms “sensations, passions, and emotions” are used to designate species of impression, though in the Treatise, these descriptors are qualified with “as they first make their appearance.” Instances of initially appearing sensations, passions, and emotions are the visual and tactile perceptions experienced while reading a book, the pleasure or uneasiness aroused by what is read (T 1.1.1.1), the colour red as it “strikes our eyes in sun‑shine” (as opposed to how it is thought of in the dark),1 heat, cold, hunger, thirst, pleasure, pain (T 1.1.2.1), and colours, sounds, and tastes considered as qualities that cannot be conceived by someone who lacks the use of the associated sense organ (T 1.1.1.10, 1.1.6.1). The Enquiry’s discussion of what it means to see blue is recopied from T 1.1.1.10. The Treatise also remarks that impressions “arise in the soul originally, from unknown causes,” the examination of which is set aside as belong‑ ing “more to anatomists and natural philosophers” (T 1.1.2.1, see also 1.3.5.2, 2.1.1.2, and EHU 12.11 on “perceptions of the mind” in general). It is stressed that the term “impression” is not meant to describe “the man‑ ner in which our lively impressions are produc’d in the soul” (T 1.1.1.1n; compare EHU 2.3). Nonetheless, Treatise 2.1.1.1 describes “original im‑ pressions” as arising “from the constitution of the body, from the animal spirits, or from the application of objects to the external organs.”2 This passage stresses that “original impressions” arise “without any anteced‑ ent perception,” but passages that suggest that impressions are disposed outside of us and merely conveyed into us by the senses nonetheless occur frequently in both the Enquiry and the Treatise.3 The common description, in both the Enquiry and the Treatise, of im‑ pressions as one kind of “perceptions of the mind” or “soul,” the declara‑ tion that those impressions associated with specific senses are inconceivable to those who lack the use of those senses, the phenomenology of the case of the missing shade of blue, and the examples of pleasure, pain, relish, sentiment, feeling, emotion, passion, and affection all indicate that impres‑ sions are found in sentient beings.4 They are broadly of two sorts: first, “impressions of the senses, and all bodily pains and pleasures” (T 2.1.1.1), and second, passions or emotions.

Section 1.1  21 It is a further question whether impressions of the senses are only to be found in sentient beings (T 1.4.2.12–13). Except for bodily pains and pleasures, the distinction between impressions of the senses and qualities of insentient objects is not evident (EHU 12.15; T 1.4.4.3). It needs to be demonstrated, and Hume found only one of the arguments used to demon‑ strate it to be satisfactory (T 1.4.4.3). That demonstration only establishes a matter of fact. In principle, impressions of the senses could exist outside of sentient creatures (T 1.4.2.38–40), and most people, who are unaware of the only good reason for thinking otherwise, consider them to do so. Much of what Hume said over the first 11 sections of the Enquiry, the first three parts of the first book of the Treatise, and the second and third books of the Treatise is unapologetically written from a naïve or repre‑ sentational realist perspective. If impressions are not literally “conveyed” into the mind from somewhere outside of it, they are still effects of the operation of external objects on bodily sense organs (as, for example, at T 1.2.1.3, 1.2.3.7, and 2.1.1.1). Some way into the Treatise Hume remarked that ’tis universally allow’d by philosophers, and is besides pretty obvious of itself, that nothing is ever really present with the mind but its percep‑ tions or impressions and ideas, and that external objects become known to us only by those perceptions they occasion. (T 1.2.6.7) He went on to infer that “’tis impossible for us so much as to conceive or form an idea of any thing specifically different from ideas and impressions” (T 1.2.6.8). And he concluded that “The farthest we can go towards a conception of external objects, when suppos’d specifically different from our perceptions, is to form a relative idea of them” (T 1.2.6.9). Without further argument, none of these comments rules out coming to know ex‑ ternal objects that are specifically like our perceptions (or just are our per‑ ceptions), as representational and naïve realists maintain (Garrett 2006, 305). Neither do they entail that result. Hume reported further arguments for rejecting knowledge of external objects specifically like our perceptions towards the end of the Enquiry (12.9, 12.15). He not only reported but endorsed those arguments towards the end of Treatise 1.4 (1.4.2.44–5, 1.4.4.6–14). He may have waited so long because the question is genuinely difficult and could not be answered without a great deal of prior study (Chapter 8).5 Hume’s descriptions of impressions as a kind of thing that “arises in the soul” (T 1.1.2.1) and as “perceptions of minds” (EHU 2.3; T 1.1.1.1) raise the question of how impressions are related to minds or souls. Are they in minds as constituent parts, are they in minds as states or modifications

22  Impressions of their manner of being, or are they distinct objects that minds perceive? The first of these alternatives rests on the notion that the relation between impressions and minds is one Hume elsewhere described as “local con‑ junction” (T 1.4.5.8), the second appeals to the relation of inherence of modifications in a substance, and the third would today be considered to involve a relation to an intentional object. There is another possibility, which might be considered a variant of the first one. Insofar as Hume con‑ sidered it in principle possible that sensations could be qualities of insenti‑ ent bodies, he would have considered it equally plausible that they could be qualities of the bodies or brains of sentient creatures. He said as much about bodily pleasures and pains, writing that they “arise originally in the soul, or in the body, which‑ever you please to call it” (T 2.1.1.2). It is a question which, if any, of these accounts should be attributed to Hume.6 The evidence favours the first or its variant. This is indicated by a rare interaction between Hume and Thomas Reid. 1.2

Reid’s objection

At some point before July 1762, Reid asked their mutual friend, Hugh Blair, to pass on a draft of the second to fifth chapters of his Inquiry into the Human Mind to Hume. This occasioned a brief exchange, consisting of a letter from Hume to Blair, an abstract Reid made of his Inquiry appar‑ ently intended for Hume, and a following pair of letters, one from Hume to Reid and a reply from Reid to Hume.7 Judging from the extant Inquiry and Reid’s “Abstract,” Hume will have read that he had been wrong to identify all the “objects of human thought” with impressions and ideas (Inquiry, 257). According to Reid, we have thoughts that are not copies of impressions but are instead “of” other objects. On Reid’s account, stimulation of the sensory organs has two principal effects, not just one. It produces sensations, which are states of feeling with a qualitative character, like the feelings of pleasure and pain, or the relish of a wine. These feelings exist only when they are felt, and they are modi‑ fications of the state of being of some sentient creature. Except for being identified as modifications of the mind, they are the analogues of Hume’s impressions of sensation. But, for Reid, sensory stimulation does not just lead us to experience sensations. It occasions beliefs. The beliefs take the form of belief in the current existence of some object and so involve some lean conception of that object. The object might be the sensation itself. (I cannot have a sensation of smell without believing it to now exist, In‑ quiry 2.3, 27.) But it need not be. There are three objects that any sensory experience leads us to believe in: the sensation itself, a mind that feels or enjoys or otherwise undergoes that sensation (Inquiry 2.6, 32), and a cause of the sensation (Inquiry 2.8, 38–9). The last of these objects might

Section 1.2  23 be originally conceived as something that exists and occasions or is other‑ wise responsible for the mind’s experience of the sensation but is otherwise unknown. This is the case, Reid thought, with the original conceptions of the causes of the sensory experiences had by smell, taste, and hearing, and the tactile sensations of heat and cold. But in the case of tactile sensations of pressure, we do not just believe that there must be some cause of the sensation; we believe that it is composed of parts that are resistant to rela‑ tive motion and that compose a figure of a certain size. We learn by touch that our immediate environment contains many such objects, disposed at various locations, or set in various states of motion, and further discover, by approaching to or retreating from them, that they are responsible for our sensations of smell, taste, and sound. Judging from the extant Inquiry 2–5, the draft Reid prepared for Hume will not have said much more about conception than that it is the simple apprehension of an object (Inquiry 2.4, 29). But if Hume read the first (1764) or any of the subsequent editions of the Inquiry he would have seen Reid draw a painstaking distinction between sensation and the perception of external objects. Sensations are acts of feeling and there is no distinc‑ tion to be drawn between the feeling and the thing felt. “[T]he distinction between the act and the object is not real but grammatical” (Inquiry 6.20, 168). “[F]eeling a pain signifies no more than being pained” (Inquiry 6.20, 168). Perception, in contrast, involves the conception or thought of an object distinct from the act of conceiving or thinking. Perceiving is an act that, like the act of feeling, exists only in the mind and only when felt. But the object need not be in the mind. It can exist when the act does not, and the act can exist when the object does not (though in that case the act is not properly called “perception”). Moreover, the object is nothing like the act. A tree “is made up of a trunk, branches, and leaves; but the act of the mind by which it is perceived hath neither trunk, branches, nor leaves.… [I]t is too simple to admit of an analysis” (Inquiry 6.20, 168). Most importantly, the act of conceiving is not an act of forming a mental image or picture of the object (EIP 4.1, 299–301); it is an act whereby the conceiver becomes acquainted with the object itself. Even though Reid likely did not make these points in so many words in the draft Inquiry, Hume will almost certainly have found them strikingly illustrated by a case likely included in the draft Inquiry 2.3. (The point is critical of Hume’s own accounts of belief and memory, and Reid will have wanted to draw that to Hume’s attention.) Reid there maintained that to remember an impression is not to now form a somewhat less vivacious idea that faintly copies that impression, as the author of the Treatise (T 1.1.3.1) had maintained. It is to perform an act of conceiving or thinking that ac‑ quaints the mind with that very past and no longer existent impression. Using a case of seeing and smelling as an illustration, Reid maintained that

24  Impressions remembering yesterday’s visual object and olfactory sensation is not an act of now forming a less vivacious idea of that object and that sensation. It is an act that has yesterday’s object and yesterday’s sensation as its objects, even though they no longer exist. Suppose that once, and only once, I smelled a tuberose in a certain room where it grew in a pot, and gave a very grateful perfume. Next day I re‑ late what I saw and smelled. When I attend as carefully as I can to what passes in my mind in this case, it appears evident, that the very thing I saw yesterday, and the fragrance I smelled, are now the immediate ob‑ jects of my mind when I remember it. (Inquiry 2.3, 28) Though made about acts of remembering, the case is a clear illustration of three of Reid’s claims about acts of perceiving and the acts of conception they involve: the act is not an image of the object, but an act of coming to be directly acquainted with the object; the act is nothing like the object (not being an image of it); and the object need not exist even though the act does and even though the act acquaints us with the object. In the case of remembering a smell, the object is a sensation, or, in Hume’s language, impression, believed to have once existed. But the same analysis holds for the acts of remembering or perceiving external objects conceived as the causes of sensations. The acts of thinking or conceiving or believing in the past or present existence of the object are nothing like the object believed to exist. The thing thought about is something other than the mind itself and its own sensory states. And the thing can exist apart from being thought of and it can be thought to have properties that are not properties of the act whereby it is conceived. In the “Abstract” (Inquiry, 258–60), Reid stressed that he had attempted to justify these assertions by means of a painstaking introspective investi‑ gation of the nature of the sensations delivered by smell, taste, hearing, and touch. (He also considered vision and did so at greater length than all the other senses combined, but that material was not included in the draft sent to Hume or discussed in the “Abstract.”) This investigation had revealed, he maintained, that we have conceptions of our sensations, and concep‑ tions of “corresponding” qualities of objects that cause those sensations, and that the two are always totally dissimilar (“Abstract,” Inquiry, 260). In both the “Abstract” (Inquiry, 260) and Inquiry 5.7, 70 Reid at‑ tempted to drive this point home by appealing to a particular case, which he described as an experimentum crucis on which he was willing to hang everything. The case concerns the character of our tactile sensations as compared to what we believe about tangible objects. Reid maintained that we have tactile sensations that pass unnoticed. They have not been given a

Section 1.3  25 name in any language. Using a little freedom, we might call them sensations of pressure, or the feelings that we experience when parts of our bodies are pressed upon. (Inquiry 5.2, 56 refers to the sensations received when press‑ ing your head against a pillar, prior to the point where they become pain‑ ful.) Reid maintained that we cannot experience these tactile sensations without simultaneously believing them to be caused by an extended body of a certain shape and degree of hardness. But the conceptions involved in this belief are nothing like the pressure sensations. The sensations are feelings and, as such, states of being of an immaterial substance. They are nowhere in space. The conceptions are conceptions of parts disposed rela‑ tive to one another in space and resistant to motion relative to one another. Supposing Hume had something close to Chapters 2–5 of the extant Inquiry before him in draft, he will likely have seen Reid develop this ac‑ count through a detailed study of the workings of the senses of smell, taste, hearing, and the feelings of heat and cold, where the conceptions are just the conceptions of some otherwise unknown cause of a currently oc‑ current sensation, and of oneself as an object that is in that sensory state. He will then have seen Reid offer an account of tactile perception, where the conceptions are conceptions of a solid object of a determinate shape and size, and the sensations are feelings of pressure that pass largely un‑ noticed and unnamed. Hume will have further seen Reid suggest that we can learn to identify the objects discovered through tactile perception with those responsible for our sensations of smell, taste, sound, heat, and cold, and can learn something about the means by which they affect our sense organs. He will have seen Reid marshal this account to offer a response to external‑world scepticism. And he will have found these points reiterated and emphasized in the “Abstract.” 1.3

Hume’s reception of Reid’s objection

Hume made it a policy never to reply to his critics (MOL, xxxvi). There are no changes in any of the six subsequent lifetime editions of his Essays and Treatises on Several Subjects (which contains the Enquiry) that might be construed as a reply to Reid’s observations regarding sensation and the acts of conception involved in perception. Hume did add an advertisement to the last edition and remaindered copies of the second last in which he desired his critics to level their objections against the works appearing in that collection rather than the Treatise, which he had never acknowledged as his own work. But the Enquiry follows the Treatise in restricting the perceptions of the mind to impressions and ideas and in insisting that ideas can do no more than copy impressions. And the Enquiry and the Treatise are equally cavalier about the relation between impressions and objects (Sections I.3 and 6.12).

26  Impressions Hume was only slightly more engaging in a brief letter he wrote to Reid by way of Blair on 4 July 1762. The letter is as important for what it does not say as for what it does say. Hume did not protest that Reid had misin‑ terpreted him, nor did he suggest that he agreed with what Reid had said about perception and conception. Instead, his letter offers three substan‑ tive criticisms of what Reid had said and protests about “one particular Insinuation” (possibly regarding his character or his motives) that is not any further described. One of Hume’s criticisms was that Reid’s doctrine “leads us back to innate ideas.” This objection was off the mark, though understandable.8 His remaining two objections are better directed. Secondly. The Author supposes, that the Vulgar do not believe the sen‑ sible Qualities of Heat, Smell, Sound, & probably Colour9 to be really in the Bodies, but only their Causes or something capable of producing them in the Mind. But this is imagining the Vulgar to be Philosophers & Corpuscularians from their Infancy.… Philosophy scarce ever advances a greater Paradox in the Eyes of the People, than when it affirms that Snow is neither cold nor white: Fire hot nor red. Thirdly. It surpriz’d me to find the Author affirm, that our Idea of Extension is nothing like the Objects of Touch. He certainly knows, that People born blind have very compleat Ideas of Extension; & some of them have even been great Geometers. Touch alone gives us an idea of three Dimensions. (Letter to Hugh Blair of 4 July 1762 in Inquiry, 256) Reid may have taken these objections as indications that Hume had failed to get the point, in the first case about ambiguity in words used to refer to smells, tastes, sounds, heat, and cold, leading to a misunderstanding be‑ tween philosophers and ordinary people, in the second about the difference between the sensations of touch experienced by the blind and the objects they conceive or perceive on the occasion of experiencing those sensations. Hume’s letter prompted Reid to send Blair the “Abstract,” apparently to make what he had said about the difference between sensation and percep‑ tion more explicit. But there is more to Hume’s remaining objections. Reid maintained that the words used to refer to smells, tastes, sounds, and heat and cold are sometimes used to refer to the sensations we experience when our organs are affected and other times to the qualities in bodies that are responsible for causing us to feel those sensations (Inquiry 2.9, 43).10 “Heat” might refer to a painful feeling or to something that causes that feeling. But while the terms are ambiguous, Reid maintained that no one ever confuses the qualities felt in sensation with qualities inherent in external objects. Ordi‑ nary people do not conceive anything like their sensations of cold to be in

Section 1.3  27 snow. There is at most a misunderstanding. When modern philosophers deny that there is cold in snow, they mean to say that there is nothing in snow that resembles our feelings of cold. Ordinary people hear what the philosophers say and take them to be denying that there is anything in snow that causes their feeling of cold. Philosophers then wrongly assume that ordinary people are objecting because they project their feelings onto external objects. But no such thing is the case.11 It is different with hardness and softness, shapes like round or oblong, or textures like rough and smooth. In these cases, the words are not ambigu‑ ous. They are only ever used, by philosophers and ordinary people alike, to refer to qualities conceived to inhere in the objects that cause our tactile sensations. The tactile sensations aroused in us by these qualities of bodies are nameless and largely unnoticed, except when they become painful, and then they are described as pains.12 Going by these doctrines, Hume was right to observe that Reid main‑ tained that ordinary people are, from infancy, perfectly well versed in the distinction between their personal sensory states and the qualities in exter‑ nal objects responsible for producing those states. Hume would also have been right to observe that Reid considered our idea (conception) of exten‑ sion to be nothing like the “Objects of Touch” provided that, by “Objects of Touch” Hume meant what he can only have been expected to mean: the impressions or feelings we experience by touch. The published Inquiry denies that our tactile sensations are anything like extension (for instance, Inquiry 5.5, 63  lines 24–30) and the draft will likely have contained a similar statement. Of course, Hume meant to offer these remarks as criticisms. There is more at stake here than a dispute over how ordinary people think and speak. There is an underlying difference over what the most painstaking and careful examination (Reid, “Abstract,” Inquiry, 258) reveals about the nature of our sensory experience. Reid was a Cartesian dualist (Inquiry 7, 217; EIP 2.4, 89). He considered minds to be unextended. He accord‑ ingly maintained that our sensory states and acts of feeling and conceiving are nowhere in space (unlike the objects that are conceived). For Reid, the pain of a rotting tooth is not in the tooth and the ache of a gouty toe is not in the toe (Inquiry 6.12, 125; 6.21, 175). These feelings are states of mind (modifications of its manner of being) and as such have no spatial location relative to one another or to anything else. We are simply so constituted that a disorder in a tooth causes us both to feel pain and to conceive a body part to be disordered in some way and likewise for the case of a gouty toe. The body part is perceived to have its own location relative to other body parts, but the nature of its disorder is not immediately perceived. The pain is immediately perceived but it has no location relative to other tactile sensations.

28  Impressions Hume did not think in these terms, perhaps because he considered the thesis that minds are not extended to be open to question (T 1.4.5) and because his own carefully considered sensory experience did not square with Reid’s. For Hume, the opinion that external objects are permeated with the qualities of our sensations (figs with sweetness, fire with heat, and so on) does not just capture how ordinary people think. It captures something that is close to being correct in the case of tactile and visual im‑ pressions. Tactile and visual impressions are not just fictitiously projected onto external objects. They are themselves spatially disposed. The table that the author of Treatise 1.2.3.4–5 saw before him is one he described as a composition of purple points that is alone sufficient to supply the idea of extension. In Hume’s language, “purple,” like “blue,” is a name for a phenomenal feature of visual sensory experience that is inconceivable to the blind. It designates what Reid called a sensation. These sensations supply the idea of extension in virtue, Hume said, of being “dispos’d in a certain manner” (T 1.2.3.4), that is, disposed outside of and alongside one another over two or three spatial dimensions.13 The notion that impressions are disposed in space does not just emerge in part 2, section 3 of the first book of the Treatise, and then disappear from Hume’s authorized work. It is a central and abiding feature of his thought, albeit one that is so central and so pervasive that its presence can pass unnoticed. While Hume maintained that passions and impressions like those of hunger, thirst, desire, and will are not spatially disposed, and did not even consider impressions of smell, taste, and sound to be spatially disposed, the very passages where he made these claims are passages where he made an exception, both for impressions of vision and touch and for the thoughts or “ideas” that copy them. This is the case right from the start. The various shades of blue that have been experienced by the subject of Enquiry 2.8 and Treatise 1.1.1.10 are supposed to be presented as “de‑ scending gradually from the deepest to the lightest.” What is the manner of this descent? If they were presented over time from deepest to lightest, it would not be obvious that the subject could notice a gap. It is hard to remember the precise shade of a colour, even after just a moment. If the shades of blue were presented all at once, but “scattered” rather than “in a line” so to speak, or were merely all simultaneously “felt” without being situated anywhere in space, so that no shade was any more adjacent to any given shade than any other, it is once again highly unlikely that the subject would notice a gap (not to mention that Hume’s stipulation that the shades be presented as “descending” would not be met). Hume’s famous thought experiment only makes sense on the supposition that the visual sensations are disposed along a line in order of brightness or saturation. It is such a natural assumption that we make it without noticing that we are doing so. Without it, the experiment would have no plausibility. That readers do

Section 1.3  29 consider the experiment to have some plausibility shows they do conceive colours (the sort of colours that are inconceivable to those with no sense of vision) to be spatially disposed. It shows just how central the tenet that visual impressions are spatially disposed was to Hume’s (indeed, every‑ one’s) thought and personal experience. Hume would have thought the same way about many pain sensations. He would have thought that I feel a toothache in my tooth, as opposed to in my toe or in my mind, and that my pain sensation in my tooth is dis‑ posed relative to other simultaneously occurring tactile sensations. This is what led him to object that Reid’s claim, that our “Idea of Extension” is nothing like the “Objects of Touch,” is surprising. Properly understanding this objection requires a paradigm shift. On Reid’s account, “the Objects of Touch” are the objects conceived on the occasion of sensory stimula‑ tion, not our tactile sensations, which for Reid are very difficult to notice or attend to, so much so that they have no names in any language (Inquiry 5.2, 55–7; 5.5, 63–4; 5.8, 73  lines 31–8). Reid must have been wrong about at least this much: They do have a name in at least one language: Hume’s language. Hume can be expected to have been speaking to Reid in his own language. In Hume’s language, “Objects of Touch” means “our tactile sensations” (or what is copied from them) which are originally ex‑ perienced as set outside of and alongside one another at locations in space. From his paradigm, Hume reacted to Reid’s account of the origin of the idea of extension by objecting that someone who only has sensations of touch can nonetheless form a very complete notion of extension based on the fact that those sensations are experienced as spatially disposed. Such a person can in fact form a more complete notion of extension than someone who only has visual sensations because tactile sensations are disposed over three dimensions, whereas for Hume (and Reid) visual perceptions are dis‑ posed over only two.14 On Hume’s account, we do not need to have the conception of extension suggested to us or be innately so constituted as to form conceptions of extended objects on the occasion of sensory stimula‑ tion. We experience extension as the manner in which things internal to us, our tactile sensations, are disposed. Reid may have had just as much difficulty making the shift to Hume’s paradigm as Hume had making the shift to Reid’s. Inquiry 5.6 offers a thought experiment that aims to prove that it is impossible to acquire any notion of extension, figure, motion, or space from sensation alone. The experiment asks us to consider someone with no sense of vision who has “lost all the experience and habits and notions … got by touch,” who does not have “the least conception of the existence, figure, dimensions, or extension, either of [their] own body, or of any other; but to have all [this] knowledge of external things to acquire anew, by means of sensation, and the power of reason” (65). Reid maintained this could not be done by

30  Impressions appealing to what the subject could infer from the following experiences: being pricked by a pin, being bruised by a blunt object, being touched by an object that covers a larger or lesser part of the body, experiencing a body move along the hands or face, making an instinctive effort to move that is unsuccessful, and making a successful instinctive effort to move. The cases beg the question at issue between Hume and Reid. In each case, the subject experiences only an undifferentiated sensation or a suc‑ cession of such sensations. But Hume’s claim was that different tactile sensations can be simultaneously experienced as disposed in space. That opposed answers to the question in dispute should be merely asserted by the parties to the dispute is characteristic of a situation where paradigms conflict, and where the parties disagree over what each considers to be intuitively obvious: whether minds are extended and whether sensations are disposed in space.15 At the time of their correspondence, when Hume would only have read the second to fifth chapters of Reid’s Inquiry in draft, he could not have known just how close he was to putting his finger on their principal disa‑ greement. It is only in the sixth chapter that Reid turned to consider vision. In that chapter, Reid maintained that no one, not even painters, gardeners, interior decorators, tailors, dyers, weavers, or those involved in prepar‑ ing and mixing paints and cosmetics ever uses colour terms like “red” or “blue” or “coloured” to refer to the qualities of the sensations experienced by those with a sense of vision. Instead, these terms are only ever used to refer to the unknown causes of those sensations, and so to something that can just as well be conceived by the blind as the sighted (Inquiry 6.2, 79–80; Inquiry 6.4–5; EIP 2.18, 212). “Blue,” for instance, refers to a quality of an object that is not seen or known otherwise than as “whatever it is that causes me to have visual sensation x,” where visual sensation x has no shape, no location, and no name in any language.16 Someone with no sense of vision need only know the number of kinds of visual sensations experienced by the sighted, and the occasions under which those of each kind are experienced, to know as much about “colour” in Reid’s sense of that term, as the sighted. Hume would not have accepted this account of the meaning of colour terms. After encountering it in the published Inquiry, he might simply have viewed it as a tactic employed to evade a plain fact of visual experience: that qualities that, according to “the modern philosophy,” have no exter‑ nal existence, and are only sensations experienced by sentient creatures, are nonetheless disposed at locations on a visual field. Had he chosen to respond to Reid’s account of visual perception he would almost certainly have made at least two points: First, that, on pain of being tongue‑tied, he meant colour terms like “dark shade of blue” to describe the quality of visual sensations, and second, that he meant to say that these qualities

Section 1.4  31 are immediately experienced as disposed over two spatial dimensions to constitute “colour” patches (understanding “colour” in his language, of course) of various shapes, sizes, and positions on an extended visual field. Accepting that visual experience teaches us that much, there is no need to invoke acts of conception to account for ideas of visible position, figure, or magnitude. The ideas copy the manner in which visual sensations are disposed. The same holds for the conception of hardness, which Reid considered to figure in an experimentum crucis for deciding between his account and Hume’s (Inquiry 5.7, 70; “Abstract” in Inquiry 260; Falkenstein, 2002). On Reid’s account, the conception of hardness is the conception of a body that consists of parts that resist motion relative to one another. That con‑ ception arises in us when experiencing tactile sensations of pressure. The pressure sensations are nothing like the conception they occasion. Hume would have accepted that no one pressure sensation conveys the idea of parts that resist motion relative to one another. He made a similar point about solidity at Treatise 1.4.4.13–14. But he further observed that we can get the idea of solidity (and by implication that of hardness) from compounds of visual (and so, presumably, tactile) sensations. The sensa‑ tions need only be disposed in such a fashion as to constitute an extended patch that preserves its figure through motion and collision with other patches, or that preserves its figure notwithstanding the occurrence of vo‑ litions that are normally observed to be followed by the compression or separation of colour patches. The only difficulty arises when these patches are conceived as colourless external bodies. That produces a vicious circle (Treatise 1.4.4.8–9). But there is no circularity if it is allowed that the idea of extension is supplied by impressions of spatially disposed coloured points. Here again, the idea is copied from the manner in which sensations are disposed. There is no call to invoke acts of conception suggested by sensations or occasioned by sensory stimulation.17 1.4

Hume and Reid on extended minds

In both Enquiry 2 and Treatise 1.1.1, Hume described impressions as “per‑ ceptions of the mind.” If calling them “perceptions of the mind” means that they are either components of minds or mental states or modifications of the manner of being of a mind or soul then, since visual and tactile impressions are spatially disposed, minds must be extended over the space their visual and tactile impressions occupy. Hume rejected the notion that impressions inhere in a mental substance (Treatise 1.4.5.2–6) but allowed that it is at least intelligible that they might be “locally conjoined” or disposed alongside one another to de‑ fine a region like the visual field or a tactile field identical with the space

32  Impressions occupied by an animal body.18 That would imply that any entity that holds or contains them must be extended and, in that sense, material. However, not all impressions are spatially disposed. Smells, tastes, and passions are nowhere in space and so cannot be “locally conjoined” with anything ex‑ tended. Any mind they are attributed to could not be material. This raises the question of what it would mean for them to belong to or inhere in such a mind, beyond considering them to be modifications of its manner of be‑ ing (an option Hume continued to grapple with at Treatise 1.4.5.26–7). These conflicting results led Hume to pronounce the question of whether the soul is a material or an immaterial substance to be “absolutely unintel‑ ligible” (T 1.4.5.33). Hume’s position has a consequence he may not himself have fully ap‑ preciated, though it fits with his intriguing reference to “the soul, or … the body, which‑ever you prefer to call it” at Treatise 2.1.1.2. Tactile sen‑ sations are disposed to constitute a tactile field that has the shape of an animal body. Allowing for a limited range of limb motions, this field pre‑ serves its shape and size over time and is solid (resistant to compression). We might therefore say that it is a body. In experiencing a tactile field, we are directly perceiving our own bodies. Hence, Hume’s remark that “Bodily pains and pleasures … arise originally in the soul, or in the body, which‑ever you please to call it.” Pains and pleasures may be the sort of things that only exist in sentient creatures, but that does not mean that they are not spatially disposed. They “arise” at various locations in our bodies, and they make us directly aware of the dimensions and condition of our bodies. As far as our experience is concerned, they constitute our bodies. In addition to experiencing a tactile field, we experience a visual field. At first, these two fields appear as discrete and incommensurable spaces, with qualia of different sorts appearing and disappearing over their extensions. But we learn to correlate colour patches appearing on the visual field with the tactile sensations comprising personal body parts, and we learn to per‑ ceive visual depth. These two achievements lead us to experience our bod‑ ies as extending over and moving through a largely empty ambient visual space occupied by scattered coloured bodies set at some distance from us. Hume’s account of space as a manner of disposition of sensible qualities naturally leads to a form of direct realism. Hume went on to offer sceptical challenges to the belief in an external world,19 and he was perceived to be a sceptic by Reid, though in the In‑ quiry only in the person of “the author of the Treatise.”20 But it was not that author’s external‑world scepticism that most disturbed Reid. It was his thesis that visual and tactile sensations are spatially disposed. Hume’s treatment of this topic had a profound influence on Reid. It may have been what led him to abandon his early endorsement of “the whole of

Section 1.4  33 BERKELEY’s [immaterialist] system” and develop his account of concep‑ tion (EIP 2.10, 142, compare Inquiry 7, 217 and “Abstract” in Inquiry, 257). He remarked that Hume had invoked the doctrine that ideas of sensi‑ ble things are “in the mind” to prove that the mind “either is no substance, or is an extended and divisible substance; because the ideas of extension cannot be in a subject which is indivisible and unextended” (Inquiry 7, 217). He went on to endorse the principal premise, that an idea of exten‑ sion cannot be in an unextended and indivisible subject. But he proceeded to argue modus tollens to Hume’s modus ponens: since the mind “is an unextended and indivisible substance … there cannot be in it any thing that resembles extension.” This leaves the question of how we can be aware of extension and ex‑ tended bodies. Reid’s answer was that careful introspection makes it clear that stimulation of the sense organs does not just occasion changes in the mind’s sensory states or manner of being. It also occasions the mind to per‑ form acts of conceiving an object that is in no way like the act of concep‑ tion or any other occurrent state of mind. How acts of conception manage to acquaint us with their objects is not explained. It is just asserted that we know that we conceive properties of spatially extended objects, we know what our sensations are like, we know that they are nothing like the properties of spatially extended objects, and we know that we observe a constant conjunction between the two. What accounts for this regularity is a matter for further inquiry, and should that inquiry prove unsuccess‑ ful, that would not deny the “law of nature” it instances, which is evident enough to introspection (“Abstract” in Inquiry, 260–1). But there is a troubling detail. Our experience of colour does not fit well with what Reid said about feelings of pleasure and pain and sensory states experienced when smelling, tasting, hearing, and feeling heat and cold. For Hume, for critics of Reid like John Fearn, and even for students of Reid such as Dugald Stewart, introspection suggests that sensations of colour are extended and located on a visual field. Reid could have maintained that sensations of colour, meaning the phe‑ nomenally experienced tints, tones, and shades that we sense and that are inconceivable to those who have no sense of vision, are objects of concep‑ tion, like the other things that, on his account, are ascribed a position, figure, or magnitude. But that was a cost he was unwilling to pay. Admit‑ ting the point for colour would establish a precedent for doing the same for bodily pains like toothaches. (Not to mention that there is introspective evidence for considering bodily pains to be spatially located, even if it may not be as clear, for some, as it is that visual sensations have location on a visual field.) After that, Reid would have had to confront the fact that the things we originally experience to be in space, colours and pains, are things that the science of his day did not recognize to exist in physical space

34  Impressions outside the bodies of sentient creatures. The things we perceive by vision and touch turn out to be things located in some private space. This opens the door to the sort of sceptical objections Hume had floated in Enquiry 12.15 and Treatise 1.4.4, which arise from a disconnect between what the best current science teaches about the nature of reality and the things en‑ countered in sense perception. Rather than countenance this result, Reid imposed a counterinterpreta‑ tion on the language in which it is stated. Colour terminology is only ever used to name the qualities in external objects that occasion nameless col‑ our sensations in us. We call the object “blue” or “scarlet” in virtue of the sort of sensation it occasions, but the terms “blue” and “scarlet” are only ever used to refer to the unperceived cause in the object, not to the qual‑ ity of the visual sensation it causes (Inquiry 6.4, 86–7). Anyone who uses colour terminology to refer to the qualities of our visual sensations is using language idiosyncratically without giving due warning (Inquiry 6.5, 90). Anyone who thinks that the visual sensation itself has position, figure, or magnitude is confused and needs to introspect more carefully on the nature of their visual experience. Hume might have been content to leave it to the reader to perform the introspection and decide who had gotten the facts right. 1.5

The Beattie–Feder objections

James Beattie’s An Essay on the Nature and Immutability of Truth con‑ tains some natural, obvious, and serious objections to Hume’s thesis that all ideas are copied from impressions (NIT, 248–56).21 Among other things, the objections attack the notion that mental states could be spatially ex‑ tended. Beattie’s Essay was published in the last decade of Hume’s life and went through five editions before his death. It is not known which, if any, Hume read, and there is no record of any reply to Beattie’s criticisms.22 They are nonetheless worth considering. Beattie was not alone in raising them. J. G. H. Feder raised a similar objection to Kant’s claim that space is a form of intuition,23 and the Beattie–Feder objections continue to be echoed today.24 Hume had the resources to defend his views against them, even though he chose not to do so. Beattie wrote that it is “inconceivable and impossible” that every idea should be a copy and resemblance of the impression whence it is derived; … that the idea of extension should be extended, and that of solidity solid; — that a thought of the mind should be endued with all, or any, of the qualities of matter. (NIT, 251)

Section 1.5  35 In justifying his criticism, he cited Treatise 1.1.1.1 and 1.4.2.39–40 as evi‑ dence that Hume took the terms, “impression” and “object,” to “amount to the same thing.”25 He proceeded to conclude that for Hume, the idea of a line the shortest that sense can perceive, must be equal in length to the line itself; …and the idea of a line a hundred times as long, must be a hundred times as long as the former idea.… And so it clearly follows … that the idea of an inch is really an inch long; and that of a mile, a mile long. In a word, every idea of any particular extension is equal in length to the extended object. The same reasoning holds good in regard to the other dimensions of breadth and thickness. All ideas, therefore, of solid objects, are [1774: must be] (according to Hume’s [1774: Mr. Hume’s] philosophy) equal in magnitude and solidity to the objects themselves. (NIT, 252–3) Beattie neglected to consider a point Hume discussed in some detail, that there are two different ways of measuring the size of the objects/impres‑ sions of vision (Treatise 1.2.4.19–22).26 The size of objects might be meas‑ ured “relatively” by juxtaposition or superposition with other objects, typically those adopted as standard units of measurement. Berkely called this “tangible magnitude” because we move our own body to bring stand‑ ard measures into tangible juxtaposition with or superposition over the object being measured. When we do this, we discover by experience that as the visual distance between ourselves and objects increases, these rela‑ tive measures continue to yield the same results. A table that measures to a yard in height continues to rise to the one‑yard mark on an adjacent yard‑ stick regardless of how near or far we are from it. Even though the visual image of the table shrinks as we retreat from it, that of the yardstick does as well, preserving the result. But the size of visible objects might also be measured “absolutely,” by the number of minimally visible points (or “pixels,” as we now call them) the image uses. When we retreat from an object, the image shrinks, which is to say that it turns into a different image that uses fewer pixels and that consequently loses some of the detail that was conveyed by the original image. Any image we form of an object is necessarily going to lack a great deal of fine detail (the sort of detail revealed by holding a magnifying glass over the object). The more detailed the impression or image, the less of it is going to fit onto the visual field or onto a corresponding image field. In one sense, therefore, visual impressions and the ideas that copy them can never get larger than the visual field, that is, they can never take up more than the finite number of minimally visible points or pixels on the

36  Impressions field, and analogously for ideas and the imaginative powers of the mind. But in another sense, arbitrarily large objects can be seen, provided they are sufficiently distant and luminescent enough and their details are sufficiently confused. At 2.5 million light years, the Andromeda Galaxy appears as a minimally visible point of light. That a spatially extended idea equal to the visible appearance of a whole house should be present in a mind “lodged in a body of no extraordinary dimensions, … contained in a room ten feet square and ten feet high” (NIT, 253) is no more extraordinary than that an image of Andromeda should appear on a 1920 × 1080 pixel monitor lodged in the same room. That a spatially extended idea equal to “the real tangible magnitude of [a] whole house” should be “lodged in a hu‑ man body” is impossible, but only because the tactile image field is not large enough to contain all of it, not because tactile impressions are not extended. No one has ever felt or had a tactile idea of a whole house all at once unless it is a doll house. Beattie had other objections. He maintained that thoughts could not be endowed with any of the qualities of matter. To this end, he asked some rhetorical questions and made some assertions that he took to need no argument. Is there any warmth in [the] idea of heat? There must, according to Mr Hume’s doctrine; only the warmth of the idea is not quite so strong as that of the impression. (NIT, 249) Do you acknowledge that to be true of the idea of eating, which is cer‑ tainly true of the impression of it, that it alleviates hunger, fills the belly, and contributes to the support of human life? (NIT, 250) [T]hat … the idea of red should be a red idea; the idea of a roaring lion a roaring idea; the idea of an ass, a hairy, long‑eared, sluggish idea, pa‑ tient of labour, and much addicted to thistles; that the idea of extension should be extended, and that of solidity solid; … is, in my judgment, inconceivable and impossible. (NIT, 251) I therefore leave it to the reader to determine, whether, if this doctrine of solid and extended ideas be true, it will not follow, … that I must be greatly hurt, if not dashed to pieces, if I am so imprudent, as to form only the idea of a bomb bursting under my feet. (NIT, 255)

Section 1.5  37 Some of these cases pose more of a problem than others. Hume might have addressed some by appealing to the coherence of taking impressions and ideas to be spatially disposed. Provided ideas can be spatially disposed they can be hairy, since being hairy reduces to having a certain shape; and solid, since being solid reduces to preserving size while hit on all sides. In other cases, the ideas refer to how coloured or tactile shapes, like that of an ass move over time.27 Hume would also have been unperturbed by Beattie’s question‑begging suggestion that an idea of red is not a red idea. Beattie was not helped by his identification of impressions with objects, his claim that thoughts could not be endowed with any of the qualities of matter, and his claim that ideas are best considered to be lodged in minds (NIT, 253). Hume elsewhere observed that It is universally allowed by modern enquirers, that all the sensible quali‑ ties of objects, such as hard, soft, hot, cold, white, black, &c. are merely secondary, and exist not in the objects themselves, but are perceptions of the mind, without any external archetype or model, which they represent. (EHU 12.15; similarly, T 1.4.4.3) Granting that the sensible quality of redness is not a property of “matter” or any external object, it is nonetheless evidently present on our visual fields, where it also has extension and location. It therefore exists some‑ where, and by Beattie’s own admission, the place where it is most “con‑ veniently deposited” (NIT 253) is in the mind, which must therefore be extended to accommodate it, and coloured insofar as it contains a colour. Beattie brought up other cases that Hume could have treated as cases where we discover that causal relations fail to obtain. Ideas of the various tastes involved in having a meal are not followed by the impression of sati‑ ety or the idea of an explosion by impressions of pain. “Every one of him‑ self will readily perceive the difference betwixt feeling and thinking,” Hume maintained, adding, “The common degrees of these are easily distinguish’d” (T 1.1.1.1). We quickly learn that the occurrence of impressions is regularly followed by consequences that do not regularly follow the occurrence of the readily distinguishable ideas that copy and resemble them. However, Beattie’s appeals to heat, cold, and pain are not so readily dealt with. Granting that the idea of red is a red idea, it follows by parity of cases that the idea of heat ought to be a hot idea, only less vivaciously so. But that does not seem to be the case. Even granting that heat is only a sensible quality and not a property of external objects, there is an apparent failure of the general rule that ideas copy impressions.

38  Impressions Hume might have dealt with this problem by noting three things. First, the cases really do point to an exception to a rule that otherwise holds, and the exception can be generalized to all tactile sensations as compared to colours, sounds, smells, and tastes. Most of us have no trouble closing our eyes and visualizing red, even to the point where it almost seems we are experiencing a red after‑image. Again, for many of us to imagine the taste of cinnamon or the smell of apple pie is to come close to having the very impressions. And many of us have had the annoying experience of getting a tune stuck in our heads. These earworms can sometimes play so loudly that we cannot concen‑ trate on what we are trying to think about. (The idea of a roaring lion can indeed be a roaring idea.) But with tactile sensations, it is not so easy. In the case of touch, however, there is an exception to the exception. We can form tactile ideas that almost feel like tactile impressions by sympathy or by causal inference. Were I present at any of the more terrible operations of surgery, ’tis certain, that even before it begun, the preparation of the instruments, the laying of the bandages in order, the heating of the irons, with all the signs of anxiety and concern in the patient and assistants, wou’d have a great effect upon my mind, and excite the strongest sentiments of pity and terror. (T 3.3.1.7) The sound of a dental drill is unbearable for some people, arousing ideas of tooth penetration that are almost as painful as the unanesthetized experience. Hume could have offered a third observation by way of explaining these phenomena. Our other senses can be interrupted. But we are constantly aware of our own bodies and of a stream of sensations coming from many of their parts. We are not going to avoid feeling ourselves freezing to death by having warm ideas or avoid believing ourselves to be perfectly healthy by imagining injurious events (Rocknak 2013, 23; 2019, 89) because the constant stream of much more vivacious, incompatible impressions can‑ cels the limited vivacity of the ideas of warmth or injury, in accord with Hume’s account of probabilistic belief (T 1.2.3.12). It is difficult to have ideas that compete with contrary impressions. It takes exceptional circum‑ stances, drawing on sympathy or arousing strong passions to enliven our tactile ideas to the point where this can happen.28 1.6

The comprehension problem

Hume faced difficulties as serious as Reid’s difficulty with colour percep‑ tion, though they are not among those raised by Beattie. One can be called the comprehension problem, another the limitation problem.

Section 1.6  39 Hume would have encountered the comprehension problem in Clarke (Clarke‑Collins, 47) and Bayle, among other sources (Lennon and Stain‑ ton 2008, Mijuskovic 1974; Kant A 351, called it the “Achilles” of all the dialectical arguments in the pure doctrine of the soul). Consider the shape of the four parts of the world on a globe.… [I]f this globe were capable of knowing the shapes with which it has been deco‑ rated, it would contain nothing that could say, “I know all Europe, all France, the whole city of Amsterdam, the whole Vistula”; each part of the globe could only know the portion of the shape that fell to it; and since that part would be so small as not to represent any place entirely, … [its] acts of knowing would be very different from those that we experience; for they make us know an entire object, an entire tree, an entire horse, and so on[.] (Bayle Dictionnaire, 130) Bayle took this example to prove that “the subject that is affected by the entire image of these objects is not at all divisible into several parts” and consequently that we are not “corporeal or material or a composite of sev‑ eral beings.” As William James was later to put it, referencing Brentano’s Psychologie vom empirischen Standpunkt, Take a sentence of a dozen words, and take twelve men and tell to each one word. Then stand the men in a row or jam them in a bunch, and let each think of his word as intently as he will; nowhere will there be a consciousness of the whole sentence. (Principles 1.6, 162) In the Abstract, Hume claimed that “if any thing can entitle the author to so glorious a name as that of an inventor, ’tis the use he makes of the principle of the association of ideas” (Ab 35). But association cannot do this job. Association presupposes prior experience. We first experi‑ ence one object alongside another (EHU 3.3; T 1.1.4.2). Only after that can we begin to associate one with the other. Rather than explain “uni‑ fied consciousness,” association presupposes it. The subject who first experiences one object alongside another must experience both objects in that relation, on pain of introducing James’s problem case, where two adjacent subjects each experience its own object and neither is af‑ fected by what the other perceives or knows how it stands in relation to the other. In his own discussion of this problem, James criticized those he called the “associationists” for maintaining that entities, whether psychic or ma‑ terial, can sum themselves together.

40  Impressions Each remains, in the sum, what it always was; and the sum itself exists only for a bystander who happens to overlook the units and apprehend the sum as such; or else it exists in the shape of some other effect on an entity external to the sum itself. Let it not be objected that H2 and O combine themselves into “water,” and thenceforward exhibit new properties. They do not. The “water” is just the old atoms in the new position H–O–H; their “new properties” are just their combined effects, when in this position, upon eternal media, such as our sense organs and the various reagents on which water may exert its properties[.] (Principles 1.6, 161) Hume would have found this point being made by Clarke in response to Collins (Clarke‑Collins, 57), though without James’s further refutation of nineteenth century appeals to “mental chemistry” (a variant on even that point can be found in Bayle, Dictionnaire, 70–1). James’s own solution, like Clarke’s before him, was to invoke effects on a soul. The associationists say the mind is constituted by a multiplicity of dis‑ tinct “ideas” associated into a unity. There is, they say, an idea of a, and also an idea of b. Therefore, they say, there is an idea of a + b, or of a and b together. Which is … a palpable untruth. Idea of a, plus idea of b, are not identical with idea of (a + b). It is one, they are two; in it, what knows a also knows b; in them, what knows a is expressly posited as not knowing b.… [T]he two separate ideas can never by any logic be made to figure as one and the same thing as the “associated” idea. … [S]ince we do, as a matter of fact, have the “compounded” idea, and do know a and b together, [the spiritualists] adopt a farther hypoth‑ esis to explain that fact. The separate ideas exist … but affect a third entity, the soul. This has the “compounded” idea, if you please so to call it; and the compounded idea is an altogether new psychic fact to which the compounded ideas stand in relation, not of constituents, but of occasions of production. (Principles 1.6, 163) In offering this solution, James granted that we do “as a matter of fact, have the ‘compounded’ idea.” But he left something out. The compounded idea is more than just ideas of a and b “together.” As suggested by Bayle’s reference to the “entire image” of an object, it consists of a and b disposed in spatial relations to one another and to their surroundings, raising the question of what it means for the compounded idea to be produced in a soul. Since the compound idea is spatially extended, we might think that the soul it is produced in would have to be spatially extended. Bayle’s question of how this spatially extended soul has a unified consciousness of

Section 1.6  41 the extended images with which it imprinted has not been addressed. The problem is greater than Bayle made it out to be. It is as much a problem for the “spiritualists” as the “associationists,” because if we do not think that the soul is spatially extended, we are confronted with the problem of explaining what it means for a spatially extended image to be “produced” in it. Hume at one point wrote, all my hopes vanish, when I come to explain the principles, that unite our successive perceptions in our thought or consciousness. (Ax 20) This might be read as meaning that he gave up on trying to explain how we obtain a unified consciousness of spatially distributed information. How‑ ever, the passage only mentions the union of successive perceptions, not the union of spatially adjacent perceptions.29 And his despair was confined to explaining the principles responsible for the union, not the experience of the union. Hume may not have recognized the comprehension problem as his own. The problem concerns the relation between thought and thinking beings, and Hume was not particularly interested in how thoughts are related to thinking beings. He was more interested in the nature of thoughts or impressions themselves. On that score, he maintained that while some thoughts (passions, tastes, and sounds) are neither extended nor disposed in space, others, our visual and tactile impressions, are spatially extended (T 1.4.5.9). The most vulgar philosophy informs us, that no external object can make itself known to the mind immediately, and without the interposi‑ tion of an image or perception. That table, which just now appears to me, is only a perception, and all its qualities are qualities of a percep‑ tion. Now the most obvious of its qualities is extension. The perception consists of parts. These parts are so situated, as to afford us the notion of distance and contiguity; of length, breadth, and thickness. The ter‑ mination of these three dimensions is what we call figure. This figure is moveable, separable, and divisible. Mobility and separability are the distinguishing properties of extended objects. (T 1.4.5.15) In making this observation, Hume was dismissing anything we might say about how material souls could have inseparable and indivisible thoughts, or about how immaterial souls could be locally conjoined with extended images. As noted earlier (p. 32) he went on to condemn the “question

42  Impressions concerning the substance of the soul” as “absolutely unintelligible” (T 1.4.5.33), and proceeded to consider just the impressions, not the souls in which the impressions conjecturally “inhere.” And he took it to be “ob‑ vious” that some of those impressions are extended. Opponents such as Bayle, Clarke, and James might object that an ex‑ tended impression is nothing more than a collection of simple impressions disposed alongside one another in space. And, as James might have put it, a collection of spatially disposed impressions is not the same thing as the impression of a space, leaving our consciousness of Bayle’s “entire images” of things like trees and horses unaccounted for. Hume was aware of this objection (T 1.4.5.7). He was also opposed to the premise on which it rests. Whatever we might think about the ontogenesis of objects, Humean sense impressions arise “without any antecedent perception” (T 2.1.1.1). This is as true of complex impressions as simple ones. Complex impressions would not be called “impressions” if they had to be compounded from previously given simple impressions. In them, the whole precedes the parts. Being ex‑ tended makes spatially complex impressions divisible into parts, but it does not divide them into parts. The visual field is divided by holding a hand over an eye or otherwise blocking the field of view, the tangible by putting on oven mitts or taking anaesthetic. Otherwise, however variegated they may be, visual and tactile complex impressions are undivided wholes. Parts may be separated from or distinguished within them, but are originally given in them. Moreover, parts are originally given in them as disposed in a certain order. The whole is not just the sum of the parts; it is the parts in certain re‑ lations to one another. Bayle and Clarke and their dualist followers got this backwards. Of course if we start off being a dust cloud of consciousnesses of simple impressions, a whole consciousness, that is, a complex impression, will be difficult to account for. But that is not where Hume started. 1.7

The limitation problem

The solution that has just been offered for the comprehension problem raises an opposite question. What limits the size of impressions? What led Hume to identify an apple as a complex impression (T 1.1.1.2) as opposed to the bowl of fruit that contains the apple, or something yet larger? Where visual and tactile impressions are concerned, there is an empiri‑ cally given answer. Visual and tactile impressions are disposed out to cer‑ tain bounds and stop there. This makes the limits of the visual and tactile fields empirically given limits for visual and tactile impressions. On this ac‑ count, things like apples and tables are not complex impressions. They are portions of complex impressions that are readily distinguished by a sharp contrast in quality between figure and background.

Section 1.7  43 This can be extended to visual and tactile ideas, which are either pro‑ jected onto the visual or tactile fields (as when Macbeth sees a dagger before him, or people fantasize30) or onto an image field surrounding the sensory fields (as when imagining what lies to one side or another). Ac‑ cepting the finite capacities of the mind (T 1.2.1.2) the image fields must be finitely pixilated and so bounded. But for reasons discussed in Section 1.5 that would not prevent them from displaying arbitrarily large or remote images, with due loss of detail. Hume said things that suggest that complex impressions are yet more inclusive. When instancing a complex impression, he included the smell and taste of an apple along with its colour (T 1.1.1.2). Plausibly, he would have included its firmness and coolness as well. This may have been inad‑ equately discussed, but it cannot have been incorrect. All these sensible qualities do figure in what we think of as an apple. At Treatise 1.4.5.11–12, Hume maintained that tastes are not in space and that their ideas are only associated with coloured shapes by the imagination. The visual and tangi‑ ble fields might likewise be said to be associated. Appeals to association resurrect the comprehension problem. We speak loosely of impressions of all sorts being “given.” But given to what? Where visual and tactile impressions are concerned, the troubling ques‑ tion, “given to what?” can be rephrased as “given where?” and answered with “on the visual or tactile sensory fields,” which are bounded, endur‑ ing, and mobile regions of a larger image field. But when complex im‑ pressions are taken to include sensations that are purportedly nowhere in space and visual and tactile spaces are considered to be discrete, it is hard to answer the question, “given to what?” without invoking an external observer or mind “before” which the impressions appear, or a substance in which they inhere. Hume at one point compared the mind to a theatre, where several per‑ ceptions successively make their appearance (T 1.4.6.4), and at other places to collections, heaps, or bundles (T 1.4.2.39–40, 1.4.6.4). But theatres, heaps, and bundles have spatial boundaries that separate their contents from what lies beyond those boundaries, and the members of collections are determined by appeal to some admission criterion. If “the greatest part of beings do and must exist after this manner [nowhere in space]” (T  1.4.5.9–11), the mind they constitute cannot be appropriately com‑ pared to a theatre, heap, or bundle. Our perceptions might satisfy some other criterion for admission to a collection, but if I am only first consti‑ tuted by the creation of the collection, that criterion cannot be that they appear to me. Failing an admission criterion, the members of collections might be arbitrarily selected. But in that case, something outside the col‑ lection makes the arbitrary selection of what to include in the collection.

44  Impressions A vicious circle would again arise if the thing that made the selection were constituted by the collection.31 Hume further described the collections, heaps, or bundles as consisting of related perceptions (T 1.4.6.6). But it is one thing to inspect the con‑ tents of a given collection, heap, or bundle of impressions and discover various relations among its members (or to be such a collection, heap, or bundle that comes to contain ideas of its simplicity and identity over time in virtue of the existence of such relations between its components); it is quite another for the denizens of a universe of separately and indepen‑ dently existing impressions to be gathered into collections, heaps, or bun‑ dles by their own relations.32 Each impression in a universe of separately and independently existing impressions could have many relations to any of the others and be collected with others in many different ways by those relations. All the scarlet impressions make a collection based on their re‑ semblance. All the impressions occupying the centre points of visual fields make a collection, and so on. Everything resembles everything else in some way or another. This makes the powerset of the denizens of a universe of impressions the set of all the collections that are formed based on relations between them. No one way of collecting impressions is better or more a mind than any other. Hume devoted Section 1.4.6 of the Treatise to explaining “What then gives us so great a propension to ascribe an identity to these successive perceptions” (T 1.4.6.5) but that account takes for granted that certain perceptions are gathered in a collection to the exclusion of others and only explains why we think of that collection of distinct things as a single thing that endures over time. It does not identify the principle that originally fixes the contents of the collection. Donald Ainslie has proposed that “Hume’s explanandum is not the real unity of the bundle of perceptions: there is none. Instead, his explanandum is our tendency to believe that the mind is unified when it is under observa‑ tion” (2015, 206). But this supposes that there is some bundle or collection that is under observation. What determines the contents of the bundle? Ainslie later writes that he does not disagree that Hume presupposes that perceptions come pre‑bundled. He just maintains that this is a brute fact, though one that will not satisfy “bundling interpreters who think Hume owes us an account of this fact” (2015, 264 and n.42). But the concern is not that the fact is brute, but that it is unintelligible. Once it is denied that there is a mind in which perceptions inhere, or a body within which they are locally conjoined, we are left without any‑ thing to appeal to by way of defining what the word “bundle” means.33 After allowing that perceptions come pre‑bundled, Ainslie writes, citing T 1.4.6.18,

Section 1.7  45 When I introspectively enter into myself I do not observe all the percep‑ tions in the world; I observe only “my” perceptions. If “we cou’d see clearly into the breast of another,” we [would] “observe that succession of perceptions, which constitutes his mind or thinking principle[.]”. (2015, 264 citing T 1.4.6.18; also 219, 138) But this is a metaphor. The “breast” of another (or my own “self” that I “en‑ ter” into) is either a spatially extended location in which all the perceptions are placed, like a room with a door through which I can enter to view its contents, or it is a mind in which they inhere, and Hume denied that either metaphor can be taken literally, leaving it unclear what is being looked at or entered into when it is claimed that we look into our breasts or enter into ourselves. It would need to be explained “where” we are looking, and why looking “somewhere else” would not count as seeing more of ourselves. It would need to be explained what “self” is being introspectively “entered” into if not a substantial spiritual container. Again, Ainslie writes, “If an omniscient God looked into the breasts of all sentient creatures simultane‑ ously, he would see ‘islands’ of perception‑bundles, not an undifferentiated mass of all the perceptions in the universe” (2015, 264). But an island is a piece of land surrounded by water. For perceptions to be contained on an island is for them to be disposed in a bounded space separated from other spaces. We need an account of what God sees that does not appeal to spa‑ tial containment or inherence in distinctly existing spirits. In the absence of those recourses, the only thing we have to go on is relations between the impressions themselves, and all impressions are related to all other impres‑ sions in some way or other, making all forms of bundling equally possible, and leaving it unanswered what picks out any one of them. There are three ways Hume might have resolved this problem. Two can be immediately dismissed: (i) taking each bundle to exist over a different time from all the others, and our apparent communication with one an‑ other to be an illusion arising from the fact that our ideas of one another are in pre‑established harmony; (ii) supposing that there is only one bun‑ dle, constituting myself. It is more likely that Hume would have preferred to confess to being confronted with an insoluble difficulty than accept ­either a “monadology” or solipsism. The third way is more promising. There was nothing compelling Hume to maintain that there are percep‑ tions that are not in space. He objected that a smell or sound could not be of a circular or square figure (T 1.4.5.10), but many of the things that exist in space are disposed like mist or stars or are so smoothly blended with their surroundings that it is impossible to say where one leaves off and the other starts. In all these cases, things are disposed in space without having a determinate shape. Hume himself allowed that things can be said to be

46  Impressions in space provided the whole is situated with respect to other bodies so as to answer to our notions of contiguity or distance (T 1.4.5.10). That does not mean the thing has to have any extension. Coloured points have no extension (T 1.2.1.4). Hume could have treated tastes and smells as tactile sensations specific to the mouth and nose regions of the tactile field. Experimental evidence has since established that sounds are originally located in the tactile space between the ears (Pöppel 1988, 11). Hobbes reduced passions to tactile sensations in the stomach, around the heart or lungs, or in the head or face. Volitions might be identified with sensations in the muscles and desires with sensations in the organs bound up with their satisfaction. As remarked at the outset of this section, thoughts and ideas can be considered to be disposed in an image space that permeates and surrounds the tactile and visual fields. We cannot have thoughts of any sophistication without engaging in an inner monologue, heard in the space between the ears as earworms are, or putting them in writing where they appear on the visual field. There is not much we experience that cannot be localized in visual or tactile space. Visual and tactile space are originally experienced as discrete, but we learn to associate what is seen on the one with what is felt or might be felt in the other, and this could be taken to mean that we learn that the two originally discrete spaces are overlaid in a larger space that contains them both. Hume would have none of this, and so he was confronted by a limitation problem. 1.8

Extended consciousness

In his Abstract of early March of 1740, Hume drew attention to the Treatise’s doctrine that “the soul, as far as we can conceive it, is nothing but a system or train of different perceptions, … all united together, but without any perfect simplicity or identity” (Ab 28). The Abstract then goes on in its own voice (though in harmony with Treatise 1.4.2.39, 1.4.5.19, and 1.4.6.4) to reject the Cartesian tenet that thought in general is the essence of the mind in favour of the doctrine that “our several particular perceptions … compose the mind,” adding “I say, compose the mind, not belong to it. The mind is not a substance, in which perceptions inhere.” The Abstract goes on to reiterate what had been said in Treatise 1.4.5 on this topic. There is no indication of any loss of confidence in Treatise 1.4.5 or 1.4.6. Eight months later, Hume published the third book of the Treatise along with the Appendix, in which he remarked that he had lost all hope of being able to “explain the principles, that unite our successive perceptions in our thought or consciousness” (Ax 20).34 He attributed his failure to two principles he could not abandon, that “all our distinct perceptions are distinct

Section 1.8  47 existences” and that “the mind never perceives any real connexion among distinct existences” (Ax 21). Having run into this obstacle, Hume became disinclined to revisit the matter. In a letter of February 1754, likely addressed to John Stewart, he disparaged the author of the Treatise for his impertinent attempts to resolve all the most intractable problems of philosophy and lamented, “I wish I had always confin’d myself to the more easy Parts of Erudition” (L1 #91, 187). The remainder of this section draws on what Hume said about complex impressions to craft a Humean position on the limits of consciousness. It supposes that it was not an option for Hume to take a collection of percep‑ tions to be “given” to an observer distinct from those perceptions. Grant‑ ing that, it takes his best alternative to rest with spatializing as much of our experience as possible. “Consciousness” is not used in any special Humean sense.35 The most likely Humean equivalent is “complex perception.”36 What is offered here is a development of what Hume wrote, not anything that he can be supposed to have thought. There is something odd about taking a feeling, such as a pain or passion, to be the object of an act of consciousness distinct from that feeling. I can remember a pain or a passion when I am no longer feeling it, but if I can be in pain or feel angry without being conscious that I am in those states, probably I am not in pain or feeling angry. Plausibly, they are one and the same. Impressions of pain or anger are “states of consciousness.” This observation can be extended to being both in pain and angry, but with a twist. If I am conscious of feeling the one without being conscious of feeling the other, probably I am not in the other state. Though my feelings are different and distinguishable and separable, my consciousness “extends,” so to speak, over them both. I can stop feeling pained while still feeling angry, or the reverse. But I feel both insofar as I am simultaneously conscious of both. My consciousness is reducible to parts, but not divided into the parts it can be reduced to. And so likewise for my complex impression. The same can be said about impressions of colour. In Hume’s language, “colour,” “blue,” “red,” and like terms designate what it feels like to be in a visual sensory state, inconceivable to those with no sense of vision. Seeing red is no different from feeling reddened. (As used here, “feeling reddened” is not being in the state a thing is in when it causes an onlooker to see red, but the state a thing is in when it sees red. Compare what an onlooker sees when attributing pain or anger to someone else with what that person feels.) Seeing red while smelling and tasting is being simultaneously conscious of, that is, being a complex impression consisting of all three impressions. For Hume, feelings of being in visual sensory states are minimally visible points disposed over a visual field. I do not just see red or feel reddened. Nor do I just see multiple points of red or feel reddened at multiple points.

48  Impressions I feel reddened over an area, usually while feeling coloured in other ways over surrounding areas, extending out to the boundaries of my visual field. Though it was not mentioned earlier, the same holds for pains. Many pains are felt at different points on different body parts, which is to say both that the feelings are disposed in space and that consciousness is not just consciousness of a single pain but of a field of spatially disposed tactile sensations, a field that has the shape of my body, or better, that is my body. By now, it may seem that the original identification of an impression with the consciousness of that impression has been rejected, first by recognizing that the impressions can be multiple and different while consciousness is single and unified, and then by allowing that some of the impressions are disposed outside of one another in space while consciousness is not only unified consciousness of each of them but the consciousness of how they are disposed. The identification of consciousness with an impression can be upheld with some stretching on both sides. Earlier, it was noted that we might be inclined to think that being in pain and being conscious of being in pain are one and the same. But the case where consciousness is reduced to a simple impression, like a pain, is rare. The identification of one consciousness with more than one impression can be sustained by taking the impression to be complex. There is no consciousness of simple impressions except in degenerate and pathological cases, as when the sentient being is very primitive (Ax 16), or something happens to block, damage, or otherwise restrict the senses. Just as impressions must be complex to accommodate an identifica‑ tion with consciousness as it normally occurs, so consciousness must be extended over space to accommodate an identification with impressions that extend over space. It is not obvious why different feelings, sensations, or impressions should have to overlap at a point or be nowhere in space before they could constitute a single conscious state. We are not so well acquainted with what consciousness requires as to be able to say that there is a contradiction in the notion of a single conscious state that is extended without detriment to its unity. There is nothing more surprising about be‑ ing simultaneously hot on the left hand and cold on the right hand than there is about being simultaneously in pain and angry. This extends to being yellowed at the centre (of the visual field) and blued towards the periphery (as when looking at the sun on a clear day). It is a matter of fact that current visual consciousness extends out to cer‑ tain bounds and simply stops there. Those bounds retain their shape and size over time. The same holds true for complex tactile impressions and tactile consciousness. The visual and tactile fields are discrete. No coloured point appears anywhere on the tactile field and no tactile sensation appears anywhere on the visual field. But, in most human beings, consciousness

Section 1.8  49 “extends” over one visual field and one tactile field (see, however, Sec‑ tion 8.2.1). We come to associate the visible appearance of our own body parts with tactile sensations felt on corresponding parts of the tactile field, and the visual experience of remote objects like fire, stones, or snow with tactile sensations we would feel if we were to touch those objects. These associations are all but constantly confirmed over time, and they lead us to consider the visual and tangible fields to be overlaid in a common space that contains them both. That consideration in turn explains why con‑ sciousness extends over just them. Hume maintained that smells, tastes, sounds, desires, passions, and vo‑ litions are nowhere in space (T 1.4.5.9–10). But, as noted at the end of Section 1.7, these impressions can be localized in body parts and so on the tactile field. Many of our thoughts, notably those concerning our unper‑ ceived surroundings, or the parts of partially occluded objects are disposed on an image field that surrounds and permeates the paired visual and tac‑ tile fields. All these impressions and ideas constitute a single, spatially ex‑ tended complex perception or momentary consciousness, corresponding to a Carnapian temporary total state. Accounting for other thoughts is challenging, but not impossible. On a Humean account, many abstract ideas involve forming spatial images, be they visual or tactile. We further find that we cannot engage in any sophis‑ ticated form of abstract reasoning without either speaking to ourselves in an inner monologue or working our thoughts out in writing. In the former case, the train of thought is disposed in the same head space where we hear tunes playing after they have been “caught in our head.” In the latter, our thoughts are projected outside of our bodies onto paper or electronic media. This account can be extended to include temporal limits. I cannot recall the first four notes of Beethoven’s fifth symphony all at once. When I try, the best I can manage is to form an idea of the sound an orchestra makes when the musicians are tuning their instruments prior to a performance. My distinct recollection takes time to occur, and over that time one note is recollected after another. As I discover by experience, I am nonetheless aware of the earlier notes and of their relative position in time. But at the point of recalling the last note, I am not recalling the earlier ones. Rather than being now aware of the earlier ones, I am earlier aware of the earlier ones and that earlier awareness is a component of a temporally extended awareness, that is, an awareness that is not confined to the present moment but extends into the recent past. As the notes occur successively over time, with earlier ones ceasing before the next one begins, the earlier conscious‑ ness of an earlier note combines with the later consciousness of a later note to constitute a unified, temporally extended consciousness of what has transpired over an interval. There are narrow limits to this temporal

50  Impressions field. As further notes are recollected, earlier notes fade away. The memory progresses like a raindrop rolling down a window, discovering new parts of the window as it goes, and leaving a continually diminishing and disap‑ pearing trail behind it. My temporal consciousness, like my act of recollecting what was tempo‑ rally sequential, takes time to occur and extends some way back in time. It has a field of its own, like the visual field, the tactile field and the image field. But the field is not a memory field. I am conscious of the very recent past in the very recent past, by way of a very recent consciousness that combines with current consciousness to form a temporally extended consciousness. Imagine sitting in a swivel chair and turning slowly around in a circle. At any one moment, your visual field extends over a portion of a surrounding sphere of coloured points. Your consciousness of points off to the left is it‑ self off to the left of your consciousness of points off to the right. That does not prevent you from having a combined consciousness that comprehends the whole field and the relations of its parts. As you turn in the chair, coloured points appear on the leading side of your visual field, previously viewed points disappear off the trailing side and the whole array of coloured points shifts to the side across the field. This might be conceived of “digitally,” as if at each moment a visual field were presented and over a succession of moments a corresponding succes‑ sion of visual fields were presented. Temporal consciousness is not this succession of visual fields. It is a fur‑ ther field extending back in a temporal dimension. It includes the earlier visual fields in a temporally extended consciousness. (The difference is that between a series of momentary visual consciousnesses and the extended consciousness of a temporal series.) As your consciousness of points on the left is off to the left of your consciousness of points on the right, your consciousness of the earlier states of the visual field is earlier than your consciousness of the present state. These temporally disposed states appear as components of a temporally distributed, but still unified consciousness stretching from the present moment some little way back into the very recent past.37 As time passes, this temporally extended consciousness changes in the way consciousness of the visual field changes. (These changes could be continuous.38) The leading edge, the present, is sharply defined, and new information is constantly appearing at that edge. As it appears, it pushes information that was earlier acquired back into the past. That past, trailing portion is not well defined. It may extend further back with circumstances, but it is only ever fractions of seconds away. Information that falls off the trailing edge is no longer part of temporal consciousness. As the field of visual impressions is surrounded by a larger image field containing ideas of previously experienced landmarks, believed to still now surround us, so

Section 1.8  51 a memory field extends beyond the trailing edge of the temporal field.39 It contains vivacious ideas of previously experienced impressions. The limitation problem is addressed by taking “my” perceptions to be those that occur within the bounds of my tactile field (my body) and an overlaid visual field, considering those fields to extend a little way into the past to include very recent successions, considering the visual field to extend some way beyond the bounds of the tactile field, and supposing that an image field permeates and extends some way beyond the sensory and temporal fields. Smells, tastes, sounds, and passions are special kinds of tactile perception, disposed at special locations on the tactile field. Ideas of occluded and surrounding perceptions permeate and surround these fields. They are “mine” by virtue of being centred on the sensory fields (T 2.3.7.2). Other ideas are projected onto the visual or tactile fields or are verbalized. Though this digression has been speculative, it does not compromise on the commitments stated prior to Hume’s loss of hope over explaining the unity of consciousness (Ax 20). The first, that we have no idea of self or substance (Ax 11), is consistent with the identification of consciousness with a spatially and temporally extended complex impression. Hume went on to observe that All perceptions are distinct. They are, therefore, distinguishable, and separable, and may be conceiv’d as separately existent, and may exist separately, without any contradiction or absurdity. (Ax 12) The distinctness of perceptions is accounted for by their being differently disposed on the sensory fields (Section 1.6). While what is distinct can be conceived as separately existent, and can exist separately, there is no absurdity in recognizing that it is surrounded and that we are conscious of it in its surroundings. When I view this table and that chimney, nothing is present to me but particular perceptions.… But this table … and that chimney, may and do exist separately. This … implies no contradiction. There is no contradiction, therefore, in extending the same doctrine to all the perceptions. (Ax 13) The table may be pressed hard up against the chimney. They do exist separately even in this case provided that “existing separately” means occupying distinct, even if immediately adjacent locations in space.40 “Extending the same doctrine to all the perceptions” means accepting that there is no

52  Impressions contradiction in accepting that all the different perceptions do occupy dif‑ ferent places and that one can be had in the absence of the other. This does not preclude the existence of spatially complex perceptions extending over a specious present. [N]o proposition can be intelligible or consistent with regard to ob‑ jects, which is not so with regard to perceptions. But ’tis intelligible and consistent to say, that objects exist distinct and independent, without any common simple substance or subject of inhesion. This proposition, therefore, can never be absurd with regard to perceptions. (Ax 14) On the account presented here, perceptions are not taken to inhere in a simple substance or subject. For this reason, Hume’s observations over Ax 15–19 are also consistent with the account presented here. [H]aving thus loosen’d all our particular perceptions, when I proceed to explain the principle of connexion, which binds them together, and makes us attribute to them a real simplicity and identity; I am sensible, that my account is very defective[.] (Ax 20) There is no principle of connection that binds simple perceptions together and gives them a real simplicity and identity. But there are complex visual and tactile impressions that consist of different, though not necessarily dis‑ tinguished or separated parts (what is distinguishable is not necessarily distinguished, and likewise for what is separable). These complex impres‑ sions extend out to certain limits and simply stop there. Everything that falls within those limits is the subject of a unified perceptual consciousness, including relations of contiguity and distance as well as their relata. Allow‑ ing for the relative mobility of the head and limbs, the limits of the sensory fields and sensory consciousness remain the same over time while what appears within them comes and goes. No further principle of connection needs to be invoked to account for the unity of consciousness within the spatial bounds of a specious present. [A]ll my hopes vanish, when I come to explain the principles, that unite our successive perceptions in our thought or consciousness. (Ax 20) Unfortunately, Hume did not go on to explain what led all his hopes to vanish when he turned to explain the unity of successive perceptions. What

Section 1.9  53 has been said about a temporal field was either not something he consid‑ ered, or not something he found acceptable, perhaps for reasons explored in the following section. 1.9

Temporally extended consciousness

Prior to Einstein’s work on the relativity of simultaneity, it was widely accepted that the past does not exist, and that time is an order of coming into being and passing away.41 Expressions of this doctrine can be found in Augustine, “Those two times, therefore, past and future, how are they, when even the past now is not, and the future is not as yet?” (Confessions 11.14); Descartes, “it does not follow from the fact that I existed a little while ago that I must exist now, unless there is some cause which as it were creates me afresh at this moment—that is, which preserves me. For it is quite clear to anyone who attentively considers the nature of time that the same power and action are needed to preserve anything at each indi‑ vidual moment of its duration as would be required to create that thing anew” (Meditations III [AT VII, 49]); Hobbes, “it is all one to say, motion past and motion destroyed, and that future motion is the same with mo‑ tion which is not yet begun” (Elements 2.7.3); Newton, “Absolute, True, and Mathematical Time, of itself, and from its own nature flows (fluit).… Absolute Space, in its own nature, without relation to anything external, remains always (semper manet) similar and immovable” (Principia 1, 9); Leibniz, “Whatever exists of Time and of Duration, perishes continually” (Leibniz‑Clarke 5.49);42 Bayle, “the preservation of creatures is a continual creation”;43 “Every [part of time] has to exist alone. Every one must begin to exist, when the other ceases to do so. Every one must cease to exist, before the following one begins to be” (Bayle Dictionnaire, 354); Locke, “There is another sort of Distance, or Length, the Idea whereof we get not from the permanent parts of Space, but from the fleeting and perpetually perishing parts of Succession” (Essay 2.14.1); Scarburgh, “even Life it self, our very Being is but a Point. For only τò νûν, This Point of Time, The In‑ divisible Instant, The present Moment, is. And what was, and what shall be, is not” (Euclid, 1705, 3); and Reid, “The past was, but now is not. The future will be, but now is not” (EIP 3.2, 258). Anyone who believes that only the present moment exists must account for why we think there was a past. The standard recourse for an empirist44 is to appeal to memory, understood as involving the storage and retrieval of traces or echoes of past experience or the ability to reproduce the past. But appealing to a presently existing trace or copy does nothing to explain why it is recognized as a trace or copy of something that once existed. Supposing that memory is accompanied by consciousness of an act of re‑ trieving a trace from a storage bin does not help. The idea of a container

54  Impressions in which information is retained over time presupposes that of endurance over time, which is the idea that needs to be accounted for. Moreover, re‑ trieving an item from storage is an act that takes time to perform. When the act is completed, the prior activity of searching the bin and retrieving something from it no longer exists. Without presupposing what needs to be explained, how memory is possible, supposing that a currently occur‑ ring trace was produced by retrieval from storage is of no help. The same might be said for the consciousness of performing the operation of making a copy, except that this account begs the question a second time by assum‑ ing we are already in possession of an idea of a prior existence and a lan‑ guage with a past tense, which allows us to identify the currently existing product as a copy of something that existed earlier. Hobbes (Human Nature 3.7) proposed that we recognize a conception as a copy by the fact that it has decayed somewhat from what it was like when first imprinted. But that is only possible if we already remember what the original was like. In the first edition of Essay 2.10, on “retention,” Locke attempted to account for memory by appealing to storage and retrieval but switched in the second edition to appealing to an ability to make reproductions, apparently because he did not want to countenance ideas continuing to exist unperceived while in storage. But he made the change without delet‑ ing or proposing an alternative to the original edition’s language of clear and muddied imprinting, searching memory banks, and more or less easily recovering information from memory banks. That talk permeates almost every paragraph of Essay 2.10 throughout its later editions. He also failed to notice that the second edition account says nothing to explain the origin of the idea that a reproduced idea has been had “before.” “Before” would have to either be an innate idea or presuppose some form of memory other than the one being described.45 Hume was as beset by this problem as Hobbes and Locke. He declared that “Experience is a principle, which instructs me in the several conjunc‑ tions of objects for the past” (T 1.4.7.3). But he did not think that I experi‑ ence the “conjunctions of objects for the past” as being off in the past in the way I experience what is off to the left of centre on the visual field by seeing it existing where it is off to the left of centre. While “space or exten‑ sion consists of a number of co‑existent parts dispos’d in a certain order, and capable of being at once present to the sight or feeling” (T 2.3.7.5), it is “impossible to recal the past impressions, in order to compare them with our present ideas, and see whether their arrangement be exactly similar” (T 1.3.5.3). It must be similarly impossible to see how past impressions are temporally disposed relative to the current ones. Hume instead described experience as “conspiring” with habit to “oper‑ ate upon the imagination” to “make me [now] form certain ideas in a more

Section 1.9  55 intense and lively manner” (T 1.4.7.3). Having previously experienced cer‑ tain states of affairs, my cognitive constitution is changed, so that now, at the current moment, I am led to form ideas that I consider, on account of their degree of liveliness, to be copies of those past conjunctions. (Though “habit” is also mentioned, it could not be necessary because we are able to recall things we experienced only once.) Hume went on to write that without this enlivening quality “we cou’d only admit of those perceptions, which are immediately present to our consciousness, nor cou’d those lively images, with which the memory presents us, be ever receiv’d as true pic‑ tures of past perceptions” (T 1.4.7.3). But the problem is worse than he presented it as being. Even with the enlivening quality, it is unclear what could give us the idea that an image is an image of what is past. If it is maintained that consciousness only extends over those perceptions that are immediately present, it contains nothing that can give us the idea that any of them is a copy of a past situation, because it contains nothing that can give us the idea that there has ever been any such thing as a past. Supposing an immediately present “image” to be endowed with an enlivening quality only explains why it is experienced as lively, not what gives us the idea that there has been a prior state of the universe.46 Yet, at Treatise 1.2.3.7, Hume followed Locke (whom he went on to cite) in declaring that “we form the idea of time” from “the succession of ideas and impressions,” so that “the same duration appears longer or shorter” according as our “perceptions succeed each other with greater or less rapidity.” It is unclear how either of them could have said this. If im‑ pressions and ideas cease to exist the moment they are past and can only be recalled by way of experiencing current copies, there can be no such thing as an experience of succession or an estimate of how many ideas have suc‑ ceeded one another over a duration. There is only ever experience of the “current” impression or idea in the series, accompanied by some number of further, simultaneously occurring ideas. It is impossible to describe that collection of simultaneously present perceptions as succeeding one another with greater or lesser rapidity, and it is unclear how we could have come to think of them as “forming” a greater or lesser interval of time.47 It is not even appropriate to describe one of the impressions as “current” and the others as accompanying it. All are equally currently present. Something must give. Maintaining that all ideas are derived from experi‑ ence, that the idea of time is formed from the succession of perceptions, and that we are only conscious of present perceptions makes it a problem how we could form the idea of time. When it is not noticed that something must give, what gives tends to give without being noticed either. It will often give in one way in one place and in another in another, depending on which side of the conflict is catching

56  Impressions attention. When discussing the origin of the idea of time, Hume followed Locke in appealing to the experience of the succession of impressions. In this case, what gives is the restriction of consciousness to the present mo‑ ment. Speaking in Treatise 1.2.3.10 of how five notes played on a flute give us the “impression and idea” of time, Hume wrote, But here it only takes notice of the manner, in which the different sounds make their appearance. The sounds are flute notes, which Hume elsewhere described as ceasing to exist the moment the breath ceases (T 2.3.9.12; repeated at DP 1.10). The manner of their appearance is coming to be and passing away in sequence, as he stressed at the end of the paragraph: “different ideas, or impressions, or objects dispos’d in a certain manner, that is, succeeding each other.” For the mind to “take notice” of this manner of appearance requires that its “notice” or consciousness extend over more than just the present mo‑ ment. This cannot mean that the present moment expands to extend over a region any more than it means that the present moment contains current retentions or memories invested with a new original impression of a degree of pastness. Hume identified time with the succession of perceptions, so if there is any succession only the last member can be the present moment. Consciousness, rather than the present moment, must extend into the past, as it extends over the visual field or the tactile body. The past cannot be non‑existent. It can only be prior to the present. Consciousness must for its part be a complex, temporally extended whole, comprised of prior as well as current states and including consciousness of how the states are ordered. Earlier, Locke had been forced to make the same concession. Having declared the parts of succession to be “fleeting and perpetually perish‑ ing” (Essay 2.14.1), and that “Duration … is the Idea we have of perishing distance, of which no two parts exist together” (Essay 2.15.12), he nonetheless wrote that we “observe” (not remember) a “train of Ideas” to “constantly succeed one another in [our] Understanding” and are fur‑ nished with the idea of succession by “Reflection on these appearances of several Ideas one after another in our Minds” (Essay 2.14.3). The theories of memory as data retrieval or reproduction that were offered in Essay 2.10 are not mentioned. Instead, reflection is said to extend its “notice” over a collection of ideas that Essay 2.14.3 describes as occurring in a “train.” (Reflection had been earlier defined at Essay 2.1.4 as “that notice which the Mind takes of its own Operations, and the manner of them[.]”) The summary paragraph, 2.14.31, declares that we are capable of observ‑ ing “how our Ideas … constantly some vanish, and others begin to ap‑ pear,” and are further capable of “observing a distance in the parts of this

Section 1.9  57 Succession.” For such observations and reflections to be possible without invoking innate ideas, our consciousness must extend back into the past, allowing us to directly see how the past ideas are positioned in the past, not just know them by means of current reproductions or stored remnants. Locke and Hume did not realize what they had been “forced, as it were, by the truth” to say (as Aristotle liked to put it when commenting on how his predecessors groped towards what he considered the correct account of things). If we can take notice or observe or reflect on what is past, even if only by a fraction of a second, then consciousness must extend back into the past. To combine with current consciousness to make up a complex state of consciousness, that past component of a unified consciousness can‑ not be nothing. It must exist albeit in the past rather than the present. The past must therefore exist, albeit in the past. Less paradoxically put, tempo‑ ral order ought to be divorced from existence assumptions, and conceived as an order of disposition in terms of before and after rather than as an order of past, present, and future existence.48 Speaking of God, Locke wrote, “he sees all things past and to come; and they are … no farther removed from his sight, than the present: They all lie under the same view” (Essay 2.15.12). For them all to lie under the same view, they must all exist within that field of view, albeit some at prior points (moments), and some at posterior. Granting that it is possible for God to see “past” (prior) and “future” (posterior) things all laid out on one field of view, there could be no contradiction in the tenet that human beings and other animals might also be able to see them laid out in that way, though with a narrower field of view, extending if only a little way into the “past” (over prior moments). Locke went so far in this direction as to write For whilst we are thinking, or whilst we receive successively several Ideas in our Minds, we know that we do exist; and so we call the Existence, or the Continuation of the Existence of our selves … Commensurate to the succession of any Ideas in our Minds, the Duration of our selves. (Essay 2.14.3) If we know that we endure, we know that our bodies and our eyes, nerves, and brain existed earlier, that is, that they extend back into the past. There can be no more absurdity or contradiction in supposing that temporally extended sense organs can give us a view of things disposed over time (that is, as before or after) than there is in recognizing that spatially extended sense organs give us a view of things disposed in space. But Locke refused to recognize this possibility, declaring that whatever God might be able to view, “what is once past [we human beings] can never recal” (Essay 2.15.12). The best we can do, for Locke, is make new copies of what once

58  Impressions was. We cannot, like a judge, summon the subject itself to appear in court, much less “recall” it in the way God does, by seeing it still standing in its place in the past. In the end, Locke could not get over the conviction that those moments of time that are earlier than the phenomenally current moment do not exist. Even God’s view of the series of earlier and later events is something that Locke described as eluding our comprehension. We attribute that power to him only because we consider an infinite being to have powers we cannot grasp (Essay 2.15.12). Locke’s account of the origin of the idea of time in the “observation” of succession and reflection on successive ideas and his implicit commitment to the duration of selves and so of consciousness are in tension with his presentist temporal metaphysics. But he never noticed the tension. Sometimes the one commitment, sometimes the other, directed his thought and writing, depending on what concern was uppermost in his mind. While following Locke in his account of the origin of the impression and idea of time, Hume’s pronouncements on the nature of time are somewhat more equivocal. The fullest is Treatise 2.3.7.5. Without having recourse to metaphysics, any one may easily observe, that space or extension consists of a number of co‑existent parts dispos’d in a certain order, and capable of being at once present to the sight or feeling. On the contrary, time or succession, tho’ it consists likewise of parts, never presents to us more than one at once; nor is it possible for any two of them ever to be co‑existent. Treatise 2.3.7.5 does not just say that no two parts of time coexist; it adds that whereas the parts of space are capable of being “at once present to the sight or feeling,” the parts of time are never presented “more than one at once.” The contrast with what is said about space makes it clear that this is a claim about the extent of consciousness, rather than about space or time, and that it is grounded on “observation” rather than “metaphys‑ ics.” Purportedly, observation establishes that consciousness is confined to a single moment while spreading out (at least in sight or feeling) over multiple parts of space. Treatise 1.2.3.10 (block quoted on p. 56) abandons the notion that con‑ sciousness is confined to a single moment, dimly recognizing that our “no‑ tice” can be distributed over more than one moment, just as visual and tactile consciousness can be distributed over more than one place. But, like Locke, Hume made this point without quite realizing what it means, and as a result sometimes maintained that consciousness is confined to one mo‑ ment and sometimes implied that it cannot be, depending on what project was at hand.

Section 1.9  59 Should Hume be criticized for failing to work out a cogent position on the origin of the idea of time? Should he be credited with having broken free, if only occasionally or inadvertently, from the prejudices that only the present moment exists, that we can only be conscious of what exists at the present moment, and that being distributed over a stretch of time or space destroys the unity of consciousness? The approach taken in what follows is to respect Hume’s expressed commitments and hold him to the consequences of those commitments. In the case at hand, that means respecting his presentist commitments and concluding that they made it incumbent on him to say more than he did about how we manage to obtain an idea of succession, consistently with his views that all ideas are derived from impressions and that no impression exists for more than a moment. Adopting this historical approach also means recognizing that, however unaccountably, Hume did think we have an idea of succession and built accounts of time and causality on that foundation. This commitment entails that we must have some access to the past: enough to constitute a complex impression consisting of one moment followed by another and so a complex impression of succession. Adopting this historical approach further means recognizing that Hume never appreciated or acknowledged the tension between recognizing that we experience succession and his declared commitment to presentism. He had no thoughts on this problem, except, perhaps, for Ax 20. It is empirically obvious that we have an idea of succession. That fact allowed Hume to proceed to offer accounts of time and causality without first providing an explanation of how the idea of succession comes about. This having been said, it is necessary, going forward, to keep an eye on whether everything Hume said about the idea of time is consistent with his presentism. Should a conflict be uncovered, we cannot presume to say how he would have resolved it. We should conclude that his thought was incomplete on this matter, resisting the temptation to think that he would have resolved it in what we now think is the best way. Hume’s world was not our world, and he may not have weighed his commitments as we now weigh them. We cannot underestimate the prevalence and the force of presentist commitments among all early modern European thinkers.49 It is not impossible that someone would have broken free of that commitment, but if Hume consciously meant to reject it, he would not have written only things that suggest the opposite (T 2.3.7.5; T 1.2.2.4; T 1.2.3.8).50 Confronted with the objection that his presentist commitments leave it inexplicable how we acquire an idea of succession, compromising those commitments may have been the last thing he would have been willing to do.

60  Impressions Notes 1 T 1.1.1.5. Hume might have done better to give the example of looking at a paint chip held against a differently coloured wall and imagining what the whole wall would look like if painted in that colour. This case addresses a con‑ cern raised by Allison (2008, 19–20). 2 Frasca‑Spada notices this intriguing comment (2002, 21). 3 EHU 2.8 (quoting T 1.1.1.10); 4.13; 12.9; Ab 5; T 1.1.1.12, 1.1.6.1, 1.2.1.4, 1.2.3.3–4, 1.2.3.15, 1.4.2.12, 1.4.2.31, 1.4.4.12–13, 2.1.5.6. 4 Allison instances a common interpretation when opening his discussion of the “elements” of Hume’s philosophy with the statement that “As species of per‑ ception, both impressions and ideas for Hume are mental particulars.… [I]n the case of impressions the weight falls on being mental” (2008, 13). Perhaps not quite, for reasons to follow. 5 A further reason for his reticence is discussed in the conclusion to this book, Section C.2. 6 Kail offers a good reason for rejecting the third (2007, 28). 7 The exchange can be found in Inquiry, 255–65. Hume’s letter to Blair was discovered by Paul Wood in 1986 and is not included in current editions of Hume’s letters. Somerville (1995, Chapters 6–9) offers a thorough study of the Hume–Reid exchange. Somerville considers Hume to have misunderstood Reid. Falkenstein (2000, 2002) maintains that while Hume’s grasp of Reid’s position was limited, he saw the main point and did have a substantive objec‑ tion to it. 8 While Reid was described in Section 1.2 as holding that sensation and percep‑ tion are concomitant effects of sensory stimulation, the Inquiry tends to say that perceptions are “suggested” by sensations, as if they were consequences of sensation, like Humean ideas, only consequences that do not copy or resemble their causes. The “Abstract” (Inquiry, 261) implies that the draft Inquiry is likely to have contained this language. (Reid dropped it from his discussion of perception in the later Intellectual Powers, which was only published af‑ ter Hume’s death.) Hume could be excused for taking this as a return to the doctrine that there are innate ideas. He may have noticed how another friend, Richard Price (more of a friend to Reid than to Hume, perhaps), had in 1758 appealed to innate ideas and (somewhat sharply) taken him to task for reject‑ ing them (Review, 18–89, esp. 63, 88–9n). Properly understood, the suggestion relation only implies the sort of nativism that Hume accepted at Enquiry 2.9n. Reid maintained that in smell, taste, hearing, and touch the conception is only suggested insofar as sensory stimulation occasions the right sensation. This makes our conceptions dependent on stimulation of the sense organs. If they are innate, it is only in the sense in which Hume took sense impressions to be innate. For a different take on this topic, see Wright (1987). 9 This is one indication that the draft Inquiry will not have included the sixth chapter on vision. Reid drew a distinction between the appearances of colour to the eye and the qualities of bodies that cause those appearances, and he as‑ serted that the vulgar are perfectly well aware of this distinction and do not suppose “appearances of colour” to be qualities of bodies (Inquiry 6.4, 86–7). He also maintained that the vulgar never use terms like “colour,” “coloured,” “blue,” and “scarlet” to refer to appearances of colour. They only use them to refer to the qualities in bodies that cause appearances of colour. Reid went so far as to describe these causes as “unknown” (Inquiry 6.4, 86  line 21).

Notes  61 Philosophers abuse ordinary language when they use colour terminology to refer to sensible qualities (Inquiry 6.5, 90). 10 He excluded colour from this list. See the previous note. 11 At the time of his letter to Blair, Hume’s most likely source for attributing this line of thinking to Reid would have been passages in the draft Inquiry cor‑ responding to the published Inquiry 2.8, 38–9; 2.9, 42 esp. lines 16–33; 5.8, 73 lines 14–30. 12 Hume would likely have found this doctrine in a passage of the draft Inquiry corresponding to the published Inquiry 5.4, 62 lines 17–27. 13 Jacovides (1999) argues that this doctrine can also be found in Gassendi and Locke. Though he hesitates to extend the case to touch as well as vision (467) and plausibly argues that Locke would not have wanted to take a stand on whether ideas are solid (480), the thesis that some ideas are extended and mo‑ bile opens the door to recognizing that some are in fact solid, as revealed by observation of how they preserve volume under collision. See the closing para‑ graphs of this section. 14 Hume said as much in the second passage cited above from his letter to Blair. For Reid’s agreement concerning the immediate objects of vision, see Inquiry 6.3, 83. 15 Reid’s thought experiment has been discussed and debated by Nichols (2007, 77–83, 94–107), Van Cleve (2015, 39–530), Falkenstein (2016), Van Cleve (2016), and Falkenstein (2023, 222–5). Nichols offers some important ob‑ servations on how the experiment might be challenged by a developmental account like the one offered at the time by Condillac, and some unsettling manuscript evidence concerning Reid’s early thought on this topic. Condillac’s approach is discussed in Falkenstein (2005). 16 Grandi (2011, 2018, 2023, 179–82) has drawn attention to the fact that Reid’s position on the use of colour terms was challenged in the early nineteenth cen‑ tury by John Fearn, who saw it as a central (and implausible) pillar on which Reid’s account of the workings of the mind rests. Even Reid’s student, Dugald Stewart, was perplexed by the doctrine. 17 Hume had trouble understanding what “act” and “action” mean if not motion from one place to another (T 1.4.5.26–7). 18 See also Abstract 28 and “Immortality.” The first (metaphysical) section of this essay shows that the mature Hume continued to endorse the principal doctrines of Treatise 1.4.5. 19 Enquiry 12.9–13 and 15; Treatise 1.4.2.44–9 and 1.4.4.6–14. 20 Inquiry 2.6. The Inquiry only refers to Hume by name on four occasions (­Inquiry 3; Inquiry 6.6, 92 and 94; and Inquiry 7, 208). In no case is Hume identified as the author of the Treatise. More than a decade before Ad was pub‑ lished, Reid had respected its request by guarding Hume’s anonymity. Hume’s description of Ad as an “answer to Dr. Reid” (Letter to Strahan of 26 October 1775 [L2 #509, 301]) is misdirected. 21 Beattie’s objections have been discussed by Falkenstein (in Hume 2011, Appen‑ dix C) and Rocknak (2013, 16–27). They are further debated in Garrett (2019, 61–2) and Rocknak (2019, 88–90). 22 To preserve fidelity to what Hume might have read, citations from Beattie’s ­Essay are drawn from the first edition of 1770, and checked for variants against the fifth of 1774. Variants other than punctuation and capitalization have been noted. The only reference to Beattie I have come across in Hume’s writings is the famous, “that bigotted, silly Fellow, Beattie” in his letter to Strahan of 26 October 1775 (L2 #509, 301). A letter from Strahan to Hume suggests that

62  Impressions Strahan and Cadell refused to publish Beattie’s book to avoid giving offence to Hume (L2 Appendix C #13, 360). 23 URC, 114–16. The passage has been translated in Sassen (2000, 153–4). See also her commentary at 20–21. 24 Rocknak (2013, 15–16) appeals to Beattie‑style considerations to conclude that Hume could not have meant to say that ideas differ from impressions only in vivacity. 25 In both 1770 and 1774, Beattie’s footnote reads “Treatise of Human Nature, vol. I. p. 1. 2. 362” but for a slight variation in punctuation. He could only have been citing from 1739, where pages 1 and 2 correspond to the first para‑ graph of the Introduction to the Treatise. “1. 2” may be a misprint for “12.” Treatise 1.1.1.1 runs from pp. 11–13 of 1739. There is no reference to objects at Treatise 1.1.1.1, though it does open with the assertion that “All the percep‑ tions of the human mind resolve themselves into … impressions and ideas.” The target at p. 362 is likely T 1.4.2.40: “If the name of perception renders not this separation from a mind absurd and contradictory, the name of object, standing for the very same thing, can never render their conjunction impos‑ sible.” Beattie likely meant the two sentences to be taken conjointly. 26 Hume spoke of three at T 1.2.4.23 but two of them, juxtaposition and the general appearance of things, can be assimilated. 27 Hume could have mounted the same reply to objections raised more recently by Rocknak concerning a splintery idea (2013, 23) and Van Cleve concerning a tubercular person (2016, 235). 28 Rocknak writes not just of imagining warmth, but of imagining being at the beach or sitting before a fire (2013, 23; 2019, 89). We imagine these visual sur‑ roundings to try to boost ourselves into managing to form the tactile idea of warmth but are not always successful. Sometimes we say we are just too cold to be able to even imagine what it is like to be warm, and we mean it. 29 The union of successive perceptions is more of a problem on the assumption that the earlier perceptions no longer exist. 30 One way to distinguish fantasy from reality is that the fantasy, being projected before us onto the same field occupied by current sense impressions, cannot compete with those impressions. See Section 1.5 at the end. 31 Some of these points and others over the remainder of this section have been made before by Garrett (2011, citing Stroud [1977, 136–40] and Garrett [1981]). Garrett identifies Basson (1958), Pears (1975), Beauchamp (1979), Strawson (2001), and Kail (2007) as proponents of similar views. Haugeland (1998) and Cottrell (2015) might be added to that list. None of these scholars say entirely the same thing and some may view what is said in this and the fol‑ lowing sections with the sentiment James attributed to Wundt, Helmholtz, and the sailor whose horse managed to get its hoof caught in its stirrup (Principles 1.6, 171). 32 Here, and over the remainder of this section, I grant for the sake of argument that there are “real” relations that might act like gravity to pull things together. If relations are only abstract ideas, and cannot exist apart from being formed by minds, the problem is exacerbated. 33 “I think it is that Hume’s account not only doesn’t explain personal bundling, but that, when pushed to the limit, it doesn’t even allow for it” (Haugeland 1998, 70). 34 Thiel (2011, 402) has suggested that Hume may have written the Appendix concurrently with the Abstract. He maintains that it is not surprising that the

Notes  63 Abstract does not bring up the reservations noted in the Appendix because it was written to promote the Treatise and defend it against criticism, and claims that the Abstract focuses on features of the Treatise 1.4.5 account that the Ap‑ pendix leaves intact. But the thesis that our several perceptions compose the mind (Ab 28) is not left intact by a declared commitment to two principles that are said to make it impossible to explain how our successive perceptions could be united in our thought or consciousness (Ax 20, 21). 35 Hume referred only infrequently to consciousness, and what he said about it changed over time. Appendix 20 declares that “consciousness is nothing but a reflected thought or perception,” likely motivated by an earlier comment that “thought alone finds personal identity, when reflecting on the train of past per‑ ceptions” (Ax 20). Enquiry 7.14 and 7.21 use “consciousness” in opposition to “sensation” in the way Locke used “reflection” in opposition to “perception.” “Immortality” defines consciousness as “that system of thought … formed dur‑ ing life” (Miller, 591–2). 36 Section 1.6. The account of consciousness presented here has some affinity with one proposed by Dainton (2000, 2008, chs. 2–3). A minor difference concerns the prospects for accounting for the synchronic unity of consciousness by ap‑ pealing to regions of space. Compare Dainton (2008, 34–9) with the remainder of this chapter, Section 1.7 at the end, Section 8.2.1, and Falkenstein 1986. This is not the place to address Dainton’s objections to what he calls “the space‑thesis.” Doing so would require emphasizing the extent to which we manage to unify all our experiences within a single space despite the challenges confronting doing so (Section 8.2.1), and the extent to which we experience a loss of self identity and sense of fragmentation when we fail in that enterprise (Falkenstein 1986). 37 This needs to be carefully understood. Dainton at one point writes of us expe‑ riencing “contents passing through our consciousness” (2008: 66). Likely, he does not mean it (see 2008: 74 on what it means for experience to be self‑uni‑ fying). On the account presented here, consciousness is not an abiding thing through which contents flow, like a person standing in a field looking through a pair of binoculars at a passing train. The contents (the parts of the train, including not just the cars, but the knuckles joining them) are constitutive of consciousness. They are “sticky.” They adhere to constitute a temporally ex‑ tended whole that is more than the sum of its parts (consciousness of a series rather than a series consciousnesses). They are also transitory. They emerge at the present, move into the past, and die off, within fractions of a second. As a result, consciousness is continually changing: growing into the future while dying into the past. 38 Supposing time to be discrete if and only if there are moments between which there are no further moments, dense if and only if between any two given mo‑ ments there is a further moment, and continuous if and only if any two parts that do not have a further part between them share a part in common, I take this to be what Dainton 2008: 65 means to propose. Hume will have wanted to maintain that time is discrete, but see Section 2.2.1 39 On the model of imagined spatial surroundings, it is tempting to think of short‑term memory as a field that extends some way back into the recent past beyond the past limit of the temporal field. But this cannot strictly be correct. Even short‑term memories are concurrent with current experience. They still take time to execute, however, and according to Hume always begin with pre‑ sent experience. See Treatise 2.3.7.2, 4–5, and 7–8.

64  Impressions 40 Taking things to exist separately provided that measures can be taken to set them at a distance from one another reduces “do exist separately” to “could exist separately.” 41 Thomas (2023) has discovered that four of Hume’s later contemporaries devel‑ oped concepts of what is now called a specious present, but she stresses that all of them also accepted presentism. Coherence would demand that anyone with those commitments be a retentionalist, like William James, but that would presume that they thought their commitments through. 42 Clarke did not contest this observation, though he was not one to let any point he did not like escape his censure. Thanks to Dante Dauksz for drawing this passage to my attention. 43 Dictionnaire, 204. At this point (art. Pyrrho, note B), the sceptical Abbé is speaking, but his remark is made without reference or justification, as if it were common knowledge. It is not contested by the “learned Divine” whose reflec‑ tions conclude the note. 44 Following Hatfield (1990, Appendix 1), consider “empiricism” to refer to a theory of knowledge (all non‑trivial knowledge is based on sensory experi‑ ence) and “empirism” to refer to a theory of psychogenesis (all thoughts orig‑ inate from sensory experiences). “Meaning empirism” refers to a theory of meaning (all meaningful terms denote ideas that can be traced back to sense impressions). 45 Reid, Intellectual Powers 3.7, 285, made a similar observation, extended to Hume at 3.7, 287. 46 A similar point is again to be found in Reid, Intellectual Powers 3.7, 288–9. Costa writes, “I don’t see that Hume has the resources to explain how a single idea (which may be a complex idea, but must be such that all of its parts are experienced contemporaneously) can represent a temporal sequence of events. It can, perhaps, represent simultaneously the content of those events, but it cannot represent them as a temporal sequence” (1990, n9). 47 Appealing to causal relations between the perceptions is not an option. Causal relations are only identified by first experiencing regularities in temporal suc‑ cession. Hugh Mellor has proposed a clever alternative, but it would have been unthought of at the time. See Section 3.1. 48 Reid, whose criticisms of Locke and Hume have been incidentally noted up to now, failed to recognize that Locke could have fallen into thinking of con‑ sciousness in this way. Instead, Reid dug in his heels and declared that the past can only be known by memory (Intellectual Powers 3.5, 270, 271). He could afford to do so because he had a different account of memory to offer, one invested in the thesis that it is possible to conceive what has no existence. Granting that much, memory can be said to involve a conception of what no longer exists (Falkenstein 2023, 223–31 and Section 1.2 pp.23–4). 49 Thomas (2018, 202–3) has drawn attention to an early eighteenth‑century English debate between John Jackson and Edmund Law in which Jackson ap‑ pears to infer that absolutism about time entails what we would today call eternalism, which Thomas describes as the doctrine that past, present, and fu‑ ture time all exist, though only the present exists “now.” Thomas suggests that “Morean absolutists [notably including More, Newton, and Clarke, as well as Jackson himself] are committed to eternalism.” She does not, however, sug‑ gest that they recognized this commitment, or would have accepted it if they did. On the contrary, she observes that it was “neglected.” She conjectures that the neglect occurred because “the debate over eternalism did not reach its

Bibliography  65 current formalized stage until the twentieth century.” Supposing this is not just repeating that eternalism went unconsidered before the twentieth century, the implication is that eternalism could not be coherently articulated prior to the conceptual advances that led to the notion that time might only be an order of before and after. In that case, it might be better to take More, Newton, and Clarke to have had views about the divine nature that are in unnoticed tension with their presentism. 50 Baxter maintains that Hume did mean to reject presentism (2008, 20–1). As evidence, he cites Treatise 1.2.3.8 “’Tis evident that time or duration consists of different parts: For otherwise we cou’d not conceive a longer or shorter duration.” Baxter glosses this as “Successive parts of time exist, just not all presently” (21). But Hume did not write that it is evident that time or dura‑ tion consists of different existing parts any more than he wrote that time or duration consists of different perishing parts. He only wrote that it consists of parts, leaving us to look elsewhere for evidence of whether the parts exist or perish. Parts of either sort can compose a longer or shorter duration. Hume’s immediately following sentence is “’Tis also evident, that these parts are not co‑existent.” This is naturally read as asserting that one part must cease to ex‑ ist before the other begins to exist, which is how Bayle, Dictionnaire, 354 had glossed it.

Bibliography Ainslie, Donald. 2015. Hume’s True Scepticism. Oxford: Oxford University Press. Allison, Henry. 2008. Custom and Reason in Hume: A Kantian Reading of the First Book of the Treatise. Oxford: Oxford University Press. Basson, A. H. 1958. David Hume. Baltimore: Penguin. Baxter, Donald L. M. 2008. Hume’s Difficulty: Time and Identity in the Treatise. London: Routledge. Beauchamp, Tom L. 1979. “Self Inconsistency or Mere Self Perplexity?” Hume Studies 5: 37–44. Costa, Michael J. 1990. “Hume, Strict Identity, and Time’s Vacuum.” Hume Stud‑ ies 16: 1–16. Cottrell, Jonathan. 2015. “Minds, Composition, and Hume’s Scepticism in the ­Appendix.” Philosophical Review 124: 533–69. Dainton, Barry. 2000. Stream of Consciousness: Unity and Continuity in C ­ onscious Experience. London: Routledge. Dainton, Barry. 2008. The Phenomenal Self. Oxford: Oxford University Press. Falkenstein, Lorne. 1986. “Spaces and Times: A Kantian Response.” Idealistic Studies 16: 1–11. Falkenstein, Lorne. 2000. “Reid’s Account of Localization.” Philosophy and Phe‑ nomenological Research 61: 305–28. Falkenstein, Lorne. 2002. “Hume and Reid on the Perception of Hardness.” Hume Studies 28: 27–48. Falkenstein, Lorne. 2005. “Condillac’s Paradox.” Journal of the History of Phi‑ losophy 43: 403–35. Falkenstein, Lorne. 2016. “Dualism and the Experimentum Crucis.” Philosophy and Phenomenological Research 93: 212–17.

66  Impressions Falkenstein, Lorne. 2023. “The Intellectual Powers of the Human Mind.” In Scot‑ tish Philosophy in the Eighteenth Century, volume 2, edited by Aaron Garrett and James A. Harris, 215–54. Oxford: Oxford University Press. Frasca‑Spada, Marina. 2002. “Hume on Sense Impressions and Objects.” In His‑ tory of Philosophy and Science, edited by Michael Heidelberger and Friedrich Stadler, 13–24. Dordrecht: Kluwer. This volume is also listed as Vienna Circle Institute Yearbook 9 (2001). Garrett, Don. 1981. “Hume’s Self‑Doubts about Personal Identity.” Philosophical Review 90: 337–58. Garrett, Don. 2006. “Hume’s Naturalistic Theory of Representation.” Synthese 152: 301–19. Garrett, Don. 2011. “Rethinking Hume’s Second Thoughts about Personal Iden‑ tity.” In The Possibility of Philosophical Understanding: Reflections on the Thought of Barry Stroud, edited by Jason Bridges, Niko Kolodny, and Wai‑hung Wong, 14–40. New York: Oxford University Press. Garrett, Don. 2019. “What, in the World, Was Hume Thinking? Comments on Rocknak’s Imagined Causes.” Hume Studies 45: 59–68. Grandi, Giovanni B. 2011. “The Extension of Colour Sensations: Reid, Stewart, and Fearn.” Canadian Journal of Philosophy 41(S1): 50–79. Grandi, Giovanni B. 2018. “On the Ancestry of Reid’s Inquiry: Reid, Stewart, and Fearn.” In Common Sense in the Scottish Enlightenment, edited by C. B. Bow, 77–106. Oxford: Oxford University Press. Grandi, Giovanni B. 2023. “Theories of Perception.” In Scottish Philosophy in the Eighteenth Century, volume 2, edited by Aaron Garrett and James A. Harris, 144–214. Oxford: Oxford University Press. Hatfield, Gary. 1990. The Natural and the Normative: Theories of Spatial Percep‑ tion from Kant to Helmholtz. Cambridge: MIT Press. Haugeland, John. 1998. “Hume on Personal Identity.” In Having Thought: Es‑ says in the Metaphysics of Mind, edited by John Haugeland, 63–71. Cambridge: Harvard University Press. Hume, David.  2011. An Enquiry Concerning Human Understanding. Edited by Lorne Falkenstein. Peterborough: Broadview. Jacovides, Michael. 1999. “Locke’s Resemblance Theses.” The Philosophical Re‑ view 108: 461–96. Kail, P. J. E. 2007. Projection and Realism in Hume’s Philosophy. Oxford: Oxford University Press. Lennon, T. M. and Stainton R. J. 2008. The Achilles of Rationalist Psychology. Dordrecht: Springer. Mijuskovic, Ben Lazare. 1974. The Achilles of Rationalist Arguments. The Hague: Martinus Nijhoff. Nichols, Ryan. 2007. Thomas Reid’s Theory of Perception. Oxford: Clarendon Press. Pears, David. 1975. “Hume’s Account of Personal Identity.” In Questions in the Philosophy of Mind, edited by David Pears, 208–23. London: Duckworth. Pöppel, Ernst. 1988. Mindworks: Time and Conscious Experience. Boston: Har‑ court Brace Jovanovich.

Bibliography  67 Rocknak, Stefanie. 2013. Imagined Causes: Hume’s Conception of Objects. Dordrecht: Springer. Rocknak, Stefanie. 2019. “Reply to My Critics.” Hume Studies 45: 77–93. Sassen, Brigitte. 2000. Kant’s Early Critics. Cambridge: Cambridge University Press. Somerville, James. 1995. The Enigmatic Parting Shot: What was Hume’s “Com‑ pleat Answer to Dr Reid and to That Bigotted Silly Fellow, Beattie?” Altershot: Avebury. Strawson, Galen. 2001. “Hume on Itself.” In Exploring Practical Philosophy: From Action to Values, edited by D. Egonssen and others, 69–94. Aldershot: Ashgate. Stroud, Barry. 1977. Hume. London: Routledge. Thiel, Udo. 2011. The Early Modern Subject: Self‑Consciousness and Personal Identity from Descartes to Hume. Oxford: Oxford University Press. Thomas, Emily. 2018 Absolute Time: Rifts in Early Modern British Metaphysics. Oxford: Oxford University Press. Thomas, Emily. 2023. “The Specious Present in English Philosophy 1749–1785: Theories and Experiments in Hartley, Priestley, Tucker, and Watson.” Philoso‑ pher’s Imprint 23. https://doi.org/10.3998/phimp.1281 Van Cleve, James. 2015. Problems from Reid. New York: Oxford University Press. Van Cleve, James. 2016. “Replies to Falkenstein, Copenhaver, and Winkler.” Phi‑ losophy and Phenomenological Research 93: 232–45. Wood, Paul B. 1986. “David Hume on Thomas Reid’s An Inquiry into the Human Mind on the Principles of Common Sense: A New Letter to Hugh Blair from July 1762.” Mind 95: 411–16. Wright, John P. 1987. “Hume versus Reid on Ideas: The New Hume Letter.” Mind 76: 392–8.

2 Finite divisibility; manners of disposition; points

Early modern philosophical psychologists tended to be Cartesian dualists. They accounted for how we perceive spatial qualities and relations by ap‑ pealing to innate ideas or operations performed on sensations that they took to be purely qualitative and disposed only in time. There were two exceptions. Hume and Kant took space to be a “manner” or “form” in which sensations are given or “intuited.” For Hume, the thesis that visual and tactile sensations are disposed in space is packaged with the thesis that empty space is inconceivable. The former does not entail the latter. Hume grounded it on a prior rejection of the infinite divisibility of space. This chapter argues that Hume had a good case for taking our visual and tactile perceptions to be only finitely divisible, but not for taking space to be only finitely divisible. It further argues that his principal reason for re‑ jecting the conceivability of empty space, an argument from the non‑entity of unqualified points, rests on a composition fallacy. Hume’s independent case for considering space to be a “manner of disposition” of coloured and tangible points made a plausible, though neglected contribution to the theory of spatial perception. 2.1

The development of Hume’s thought about space and time

The conclusions reached in Chapter 1 concerning what Hume had to say about the spatial and temporal extension of complex impressions, con‑ sciousness, and mind are further supported by considering what he had to say about space and time. Book 1 part 2 of the Treatise is largely devoted to this topic. It begins by arguing that our ideas and impressions of space and time are divisible into finite numbers of “perfectly simple and indivis‑ ible” parts (T 1.2.1.2). It then argues that the same must be true of space and time (T 1.2.2). Taking a turn, it concludes that since the simple and indivisible parts of space and time are not extended, they must have “other qualities” (T 1.2.3, title). Rather than being “separate or distinct ideas” (T 1.2.4.2), they are ideas of “the manner or order in which objects exist” DOI: 10.4324/9781032677880-3

Section 2.1  69 (T 1.2.4.2). Treatise 1.4.2.2 then adds, “Or, in other words, ’tis impossible to conceive either a vacuum and extension without matter, or a time, when there was no succession or change in any real existence.” Little of this remains in the Enquiry. It contains two footnoted expres‑ sions of a commitment to the theses that there are ideas and impressions, which it calls “images … present to the fancy or the senses” (EHU 12.18n) that are indivisible, and that none of our ideas of quantity is infinitely divisible. The first note also claims that the indivisible images must be smaller than anything that is further divisible. Yet they cannot be aggre‑ gated without increasing the size of the aggregate in proportion to the number of added parts. [T]here are … parts of extension, which cannot be divided or lessened, either by the eye or imagination. These images, then, which are present to the fancy or senses, are absolutely indivisible, and consequently must be allowed by mathematicians to be infinitely less than any real part of extension; and yet nothing appears more certain to reason, than that an infinite number of them composes an infinite extension. How much more an infinite number of those infinitely small parts of extension, which are still supposed infinitely divisible. (EHU 12.18n) [A]ll the ideas of quantity, upon which mathematicians reason, are nothing but particular, and such as are suggested by the senses and im‑ agination, and consequently, cannot be infinitely divisible. (EHU 12.20n) These are claims just about images and ideas. Their application to exten‑ sion has been downgraded to a rhetorical question. The Enquiry’s discussion of the missing shade of blue also involves an implicit commitment to the thesis that visual impressions are disposed in space (Section 1.3) and with that to the thesis that our idea of space is an idea of the manner or order in which component impressions are disposed in a complex impression. In the first two (1748 and 1750) editions of the Enquiry, the 12.20 footnote was supplemented with the assertion that our ideas of greater, lesser, and equal are far from being exact and determinate enough to un‑ derwrite mathematical demonstrations of infinite divisibility. This is a point that had been made at Treatise 1.2.4.17–33 and that had served as a foundation for later remarks about the relation between arithmetic and geometry in Treatise 1.3.1. Hume’s reiteration of it was dropped from the third edition of the Enquiry of 1756 and does not appear in any subse‑ quent authorized edition.1 Hume may have deleted the comment for the

70  Finite divisibility; manners of disposition; points sake of consistency with the body of Enquiry 12.18, which emphasizes the plausibility of the reasoning that supports the doctrine of infinite di‑ visibility. But he may also have developed other reservations about the deleted remark. There is evidence that something happened in the 1756–7 period that made Hume lose confidence in large portions of what he had said about space and time in the Treatise.2 In a letter to Strahan of 25 January 1772 (L2 #465, 252–4), he reported that he had been working on a recasting of Treatise 1.2 (described as a dissertation “on the metaphysical Principles of Geometry”) for inclusion in his Four Dissertations, eventually published in 1757. He went on to write that a chance meeting with an amateur math‑ ematician, Philip, 2nd Earl of Stanhope (1714–86),3 had convinced him that the dissertation was too defective to merit publication. He wrote to his publisher of the time, Miller, that he would not publish it and, going by current evidence, recovered and destroyed such manuscripts and copies as were extant. This was a drastic move. It indicates some serious rethinking.4 The fact that Hume continued to include the passages block quoted earlier in editions published after 1756 indicates that he continued to be comfortable with the thesis that our impressions and ideas of space and time are only finitely divisible, and by implication (as further indicated by his continued republication of the discussion of the missing shade of blue) that our visual and tactile perceptions are spatially extended. But by 1757, he had likely lost confidence in much of the rest of what he had said in Treatise 1.2.2–5. In what follows, it is argued that this loss of confidence was justified as regards the infinite divisibility of mathematical and physical space and time (Section 2.4), temporal passage in the absence of change (Section 3.3) and vacuum (Section 2.6 and Chapter 5). Hume had good reasons to continue to insist on the finite divisibility of our spatially and temporally extended impressions and ideas (Section 2.2) and to maintain that our ideas of space and time are ideas of the manners or orders in which our perceptions ap‑ pear (Sections 2.5 and 2.8). 2.2

The finite divisibility of our ideas and impressions

Surprisingly for someone with Hume’s empirist commitments,5 Treatise 1.2 opens by considering our spatially and temporally extended ideas with‑ out first asking what impressions those ideas are copied from. An appeal to the empirist principle that ideas ought to be understood by appeal to impressions only comes up late, in Treatise 1.2.3.1. The inversion might be explained by Hume’s view of his audience. Con‑ jecturally, he wanted to address Platonist mathematicians (Wilson 2008,

Section 2.2  71 254–5, 256–7), who reject the principles that all ideas are copied from sen‑ sations and that all general terms refer to particular ideas. One of Hume’s two arguments for the former principle reduces to challenging opponents to give an example of an idea that has not been copied from an impression (EHU 2.6; T 1.1.1.8). Ideas of pure extension, numbers, and other geo‑ metrical objects are high on the list of potential examples. Hume may have thought it better to refute infinite divisibility without relying on empirist tenets. The refutation nonetheless appeals to a brute fact, though one that the Platonists could not deny. (BF): Brute fact ’Tis universally allow’d, that the capacity of the mind is limited …: And tho’ it were not allow’d, ’twould be sufficiently evident from the plainest observation and experience. (T 1.2.1.2) (BF) is not very controversial.6 There is a limit to how much detail can be included in the idea of a familiar city (T 1.1.1.4), to how many games of chess anyone can play blindfolded, and so on. Imagining a uniformly col‑ oured tabletop, and speculating that our mental image could be infinitely divisible only establishes that our mental capacities could be infinite. Op‑ posed to that possibility is the fact that we cannot imagine an infinitely variegated tabletop. It would be giving up the argument to object that the imagined parts might be so small and confused that we cannot distinguish them, even though they are still there, or to object that even though only a finite number of parts can be distinguished, those parts are potentially further divisible. These concessions recognize that our imaginative capaci‑ ties are limited. (BF) is also all Hume needed to draw the conclusion he most wanted to draw. (FD): Finite divisibility It requires scarce any induction to conclude from hence, that … by proper distinctions and separations we may run up [“the idea, which we form, of any finite” quantity7] to inferior ones, which will be perfectly simple and indivisible. (T 1.2.1.2) If any given part could be divided into yet further parts, it would contain more information than the mind could grasp and consequently could not be grasped in all its detail. Summing up:

72  Finite divisibility; manners of disposition; points (Summary) In rejecting the infinite capacity of the mind, we suppose it may arrive at an end in the division of its ideas; nor are there any possible means of evading the evidence of this conclusion. (T 1.2.1.2; compare EHU 12.18n) Hume remarked in passing that (Remark) ’Tis also obvious, that whatever is capable of being divided in infinitum, must consist of an infinite number of parts, and that ’tis impossible to set any bounds to the number of parts, without setting bounds at the same time to the division. (T 1.2.1.2) (Remark) plays no role in the demonstration of (FD), which, according to (Summary), follows immediately from (BF).8 Hume was particularly interested in the case where the idea is that of space or time. As discussed in Sections 1.3–5, for Hume an idea “of” space is a spatially extended idea. The argument of Treatise 1.2.1.2 establishes that it must “run up to” inferior ideas that cannot be further divisible, that therefore cannot themselves be extended, and that therefore cannot them‑ selves be spaces. The same holds for “any finite [quantity]” (T 1.2.1.2) and so for time. Hume went on to observe that the ideas we form of arbitrarily small things, like a thousandth or ten‑thousandth part of a grain of sand, can‑ not be greater or lesser than one another, even though we assign different numerical values to their sizes. The ideas of a grain of sand, of a thou‑ sandth part of a grain of sand, and of a ten‑thousandth part of a grain of sand are all the same: a pinprick or brownish colour point (T 1.2.1.3). We could not have an idea of a grain of sand set beside a thousandth part of the grain of sand. Were the idea of a grain of sand that of a pinprick or brownish colour point, the thousandth part would be too small to be imaginable. Were the thousandth part imagined as a pinprick or brownish colour point, it would defy the limits of our imaginative capacities to form an image of each of a thousand such tactile or visible points disposed adja‑ cently over three dimensions to constitute a grain of sand. We distinguish between a grain of sand and a thousandth part of a grain of sand only by distinguishing between numbers assigned to ideas, not by distinguishing between ideas.9 Having addressed the Platonist mathematicians, Hume turned to con‑ sider whether we have infinitely divisible impressions (T 1.2.1.4). This

Section 2.2  73 remains a live question because there is some evidence that impressions can be much more complex than ideas (T 1.1.1.4). The ensuing discussion accepts that there is an external world consisting of extended objects set at various distances from an observation point and sending rays of light to that observation point. It further draws a distinc‑ tion between these objects and “images or impressions” experienced when looking at them from the observation point, either with the naked eye or through microscopes and telescopes. Hume was not at this point inter‑ ested in sceptical objections to this background picture. (Our knowledge of an external world only begins to come into question at Treatise 1.2.6.) He drew on the background realist ontology to aid us in attending to our visual impressions. We are asked to place a spot of ink on a piece of paper and retreat from it. We discover that the impressions constitutive of the spot continually diminish up to a point at which any further retreat causes the spot to disappear. This indicates that, just prior to the disappearance, we were experiencing a visual impression that cannot be further divided. Hume was aware that there is a degree of vagueness to both the distance and the image apprehended as the distance increases or diminishes. Put a spot of ink upon paper, and retire to such a distance, that the spot becomes altogether invisible; you will find, that upon your return and nearer approach the spot first becomes visible by short intervals; and after‑ wards becomes always visible; and afterwards acquires only a new force in its colouring without augmenting its bulk; and afterwards, when it has encreas’d to such a degree as to be really extended, ’tis still difficult for the imagination to break it into its component parts, because of the uneasiness it finds in the conception of such a minute object as a single point. (T 1.2.4.7) Hume’s argument would be circular if it depended on the assumption that the space between the observer and the spot is finitely divisible (Cottrell 2019, 89). His recognition that there are distances over which the im‑ pression only appears to change in clarity, distinctness, and brightness before appearing to augment in size and that the augmentation in size is not always accompanied by an augmentation in the number of distinctly identifiable parts entails that there are changes in distance that are not accompanied by any discernible change in the number of impressions. Ac‑ cepting that it takes some considerable change in distance to produce a noticeable change in the number of impressions, it makes no difference whether the distance to the ink spot is infinitely divisible or not. Even if the distance increases by infinitesimal amounts, the impression does not undergo infinitesimal diminution.

74  Finite divisibility; manners of disposition; points Hume also considered the objection that the impression only disappears because the object has been removed beyond the distance at which it re‑ flects enough light to affect us. He dismissed this objection on the ground that increasing the strength of the illumination does not make the object appear any larger. At the distance at which an object subtends thirty sec‑ onds of arc or less, strengthening the light coming from it does not make it appear to subtend a larger angle; it only makes it look brighter. The view through a land telescope does lead us to experience a larger image. (That through an astronomical telescope only leads us to see more stars.) We identify the larger telescopic image and the minimally visible point that appears to the naked eye with the same external object. However, the ques‑ tion concerns the minimally visible impressions that are constitutive of the telescopic image. They disappear while retreating with the telescope before an eye (T 1.2.1.4). The comparison of the telescopic and naked eye experiences (T 1.2.1.5) teaches us that objects that consist of a great number of parts can produce simple and indivisible images. This might lead us to think that the objects have parts that are smaller than our simple and indivisible impressions. Strictly, all that follows is that our sense of vision does not always do a good job of distinguishing the parts of the objects around us. It can compress what is vastly complex and compounded into a muddy point. But, Hume maintained, while our simple impressions might be inadequate representations of large and complex things, they are adequate representa‑ tions of the smallest possible things. Nothing can be smaller than a thing that has no parts into which it can be further divided. This discussion, indeed, the whole of Treatise 1.2.1, shows that Hume conceived of ideas and impressions as things that represent extension by being either extended or disposed outside of one another. Ideas are de‑ scribed as things that may be “run up” to simple and indivisible “infe‑ rior ideas” (T 1.2.1.2), hence as things that are either “minimal” or have “sub‑divisions” (T 1.2.1.3). The idea of a grain of sand is called an image of that object (T 1.2.1.3). The “impressions of the senses,” later called “images or impressions” are described as things that can be “reduc’d to a minimum” at which point they are “incapable of any farther diminu‑ tion” (T 1.2.1.4). The senses are said to give us “images of things” that are “disproportion’d” to those things because they “represent as minute and uncompounded what is really great and compos’d of a vast number of parts” (T 1.2.1.5). We are said to take these images to be “equal or nearly equal,” presumably in size, “to the objects,” when in fact we labour un‑ der the difficulty of not being able to “[enlarge] our conceptions so much as to form a just notion” of an insect a thousand times less than a mite (T 1.2.1.5).

Section 2.2.1  75 2.2.1 The finite divisibility of our temporally extended impressions

Hume neglected to argue that our temporally extended impressions are only finitely divisible. This is a serious omission. Finite divisibility is the premise for his conclusions that the parts of time “are inconceivable when not fill’d with something real and existent” and that “’tis impossible to conceive … a time when there was no succession or change in any real existence” (T 1.2.4.2). The premise cannot rest on an appeal to the finite capacities of the mind. That appeal only establishes that we cannot form arbitrarily complex ideas, not that we cannot have arbitrarily complex impressions. When making the case for space, Hume appealed to (i) a spatially ex‑ tended impression; (ii) a process that produces obviously smaller impres‑ sions; and (iii) the fact that the process has a limit, beyond which the impression does not shrink any further but simply disappears. Unfortu‑ nately, there are theoretical and technical limitations to a parallel experi‑ ment for time. Theoretically, the experience of visual after‑images and pains makes it evident that impressions have a certain “latency.” They can last for some time after the stimulus that produced them has ceased. Per‑ haps all of them do to some extent. This does not mean that they could not become any briefer. But it does mean that we cannot determine how brief they can become simply by exposing ourselves to successions of increas‑ ingly brief stimuli. We instead need to shutter the impressions somehow. Hume might have envisioned producing a series of clicks by inserting a card into a rotating, spoked wheel and determining at what speed the clicks are no longer heard distinctly. Or, in a variant on his case of the wheeling coal (T 1.2.3.7), he might have envisioned producing flashes of light in increasingly rapid succession (say by rotating a fan blade before a lantern) and determining the speed at which we see a steady light.10 How‑ ever, these experiments may only measure the latency of a flash or sound, not the duration of the minimally perceptible silence or period of darkness. The difference is illustrated by a striking variant on click/flash experiments later described by Helmholtz. This experiment also suggests that there may be technical limitations to establishing just how brief a perceptible tempo‑ ral interval can be. Here also belong the characteristic effects of intermittent illumina‑ tion, which are produced best by the regular periodic sparks of an electro‑magnetic induction coil with a rotating armature.… Every sin‑ gle spark of this apparatus lasts for a very short interval and seems instantaneous as compared with the movements of material things; and yet the light from these brief sparks is strong enough during this

76  Finite divisibility; manners of disposition; points extraordinarily brief moment to make a perceptible impression on the retina. Illuminated by a single electric spark, all moving objects appear instantaneously at rest. Of course, the eye can perceive them only as they were at the moment when they were illuminated. As to their posi‑ tion before and after this moment, it has nothing to go by.… If a series of electric sparks succeed each other at very short intervals, stationary objects illuminated in this way look just as they do by steady illumination in continuous light, whereas moving objects appear mani‑ fold. That is, each single spark reveals the moving object in the position in which it is at the given moment; and as all these impressions last a little time, they are all present simultaneously and cause the moving object to appear multiple. The quicker the movement, the farther apart the images of the body will be, because the distance it goes during each interruption of the light gets greater. (PO 2 §22, 211) In Helmholtz’s experiment, the illumination appears constant, and station‑ ary objects appear as they normally do for the expected reason: the visual impressions produced by each earlier spark last over the brief period when the illumination is interrupted, producing the appearance of an enduring object. This includes the stationary laboratory walls, which appear con‑ stantly illuminated. But moving objects do not appear to move. Instead, multiple images of the moving object appear simultaneously at separate locations along the path of motion. The latency period of the impressions only causes the images to appear simultaneously. It does not “close the gaps” in the trajectory. The eye must be sensitive to the extremely brief pe‑ riods when they are not illuminated, even though it sees no darkness. And how could it do otherwise, Helmholtz asked, given that it was never able to see the object illuminated at the intermediate positions? More worrisome is the implication that we lack an effective means of determining how brief a sensible stimulus or sensible gap between stimuli could become. Hume could know none of this, but he might have seen another draw‑ back to trying to establish the finite divisibility of our temporally extended impressions by experiments. They yield a positive clock value for the mini‑ mally perceptible sound or flash of light, thereby suggesting that any un‑ changing impression that lasts for longer than that must be a monotonous succession of briefer parts, just as any ink spot larger than the minimally visible one must be a composite of minimally visible parts. Hume had no problem accepting that spatially extended impressions might be composed of identically coloured parts (T 1.2.3.4–5), but he denied that what he called “unchangeable” objects endure (T 1.2.3.7–11, 1.2.5.29, 1.4.2.29).

Section 2.3  77 2.3

The Georgian thesis

Hume’s conclusion in Treatise 1.2.1 is variously stated as “the idea, which we form of any finite [quantity], is not infinitely divisible, but … may [be] run up … to inferior ones which will be perfectly simple and indivisible” (T 1.2.1.2); “the imagination reaches a minimum, and may raise up to itself an idea, of which it cannot conceive any sub‑division, and which cannot be diminish’d without a total annihilation” (T 1.2.1.3); “’tis plain, that the moment before it vanished the image or impression [produced by an increasingly distant ink spot] was perfectly indivisible” (T 1.2.1.3); “Nothing can be more minute, than some ideas, which we form in the fancy; and images, which appear to the senses; since these are ideas and im‑ ages perfectly simple and indivisible” (T 1.2.1.5); and “we can form ideas, which shall be no greater than the smallest atom of the animal spirits of an insect a thousand times less than a mite” (T 1.2.1.5). While he said that our ideas and impressions can be simple, indivisible, minimal, not subject to diminution, and no greater than any arbitrarily small object, he did not say that they have no size. Hume did say, though not in quite so many words, that our simple and indivisible ideas are unextended. [L]et us take one of those simple indivisible ideas, of which the com‑ pound one of extension is form’d.… ’Tis plain it is not the idea of extension. (T 1.2.3.13–14) But it does not follow that they have no size. Hume (like many others at the time) defined “extended” as “having parts set outside of parts”. (George 2006, 147–8). ’Tis plain [one of those simple indivisible ideas, of which the compound one of extension is formed] is not the idea of extension. For the idea of extension consists of parts[.] (T 1.2.3.13) A real extension … can never exist without parts, different from each other[.] (T 1.2.4.3) In light of this definition, saying that a part is unextended only implies that it has no parts of its own, not that it has no size. But in what sense would something that has no part have size?

78  Finite divisibility; manners of disposition; points Rolf George (2006) has argued that Hume considered the simple and indivisible parts of extension to have angular magnitude. “Hume never meant to say that the ‘simple, indivisible ideas’ are unextended … in the sense of subtending an angle of 0°, especially since he and his contem‑ poraries never used the term ‘extension’ to describe the angular size of a visual image” (149).11 George bases this conclusion on the visual theory of Hume’s day, and on pronouncements concerning the angular size of minimally visible points made in works Hume will have studied, such as Locke’s ­Essay (2.15.9), and works it is possible he will have known about or been educated in “like other schoolboys of a seafaring nation” (George, 2006, 146), such as those of Hooke, William and Samuel Molyneux, and William Porterfield. George further appeals to what Hume had to say. [T]he minimal size of the ink spot after one has retreated from it is easily triangulated. Hume implied that the minimum can be measured when he said that for time there is “no exact method of determining the proportions of parts, not even so exact as in extension” (T 1.2.4.24). Evidently, then, the smallest parts of extension can be measured using methods like those of Hooke, though without his specious precision. They are “lesser impressions, that are indivisible to the eye or feeling, and may be call’d impressions of atoms or corpuscles endow’d with colour and solidity” (T 1.2.3.15). They have a “degree of minuteness” that cannot be reduced (T 1.2.2.5) and cannot be “diminish’d without total annihilation” (T 1.2.1.3). (George 2006, 148–9) George might have added that at Treatise 1.2.3.17 and 1.2.4.2 Hume spoke of indivisible parts being “fill’d” with something real or existent. If the parts have no capacity to hold anything, it is odd to speak of them being filled. Hume may of course have been speaking metaphorically, but if he wanted to drive home the point that the parts do not enclose any volume it is strange that he would have chosen to employ a metaphor that is so much at variance with what he purportedly wanted to emphasize.12 It might be objected that if simple and indivisible parts had angular mag‑ nitude their left sides would have to be distinctly placed from their right sides, which would mean that they consist of parts set outside of parts. In reply, George denies the premise. [T]his argument turns on the kind of mathematical thinking Hume en‑ deavoured to denounce. The question is not if, in some abstract sense, minima have parts, but whether they have visible parts. (2006, 149)

Section 2.4  79 Angular separation and visible distinction are not the same. It is an em‑ pirical question how much separation is required before two stimuli are seen in distinct places, and the findings do not support [the conclusion that a left side removed by 30 seconds of arc from a right side is seen in a distinct place from the right side]. The so‑called minimum resolvable, seeing two stimuli as separate, is not as small as the minimum visible. For normal sight the smallest visible black dot on white paper subtends 10ʺ–35ʺ, the separation required for seeing two stimuli as separate is 30ʺ–60ʺ. (2006, 150) As George notes, this point was made by Hume’s contemporary, James Jurin, albeit in an essay on distinct vision published only in 1738 in Robert Smith’s Opticks (2006, 155, 153).13 George further invites us to draw a thin black straight line on white paper and retreat from it until just before it vanishes. “[T]his line has visible length with distinguishable parts, but no visible width, no separable edges even if the width subtends a measurable angle upon the eye” (2006, 149). George’s thesis, that Hume never meant to say that the simple parts of space and time have no size, is too well defended to be dismissed. In what follows, it is recognized as a plausible account of what Hume may have thought, alongside the “null” thesis that he considered the simple and in‑ divisible parts of space and time to lack any magnitude.14 2.4

The finite divisibility of space and time

Having argued that our ideas and our visual impressions are only finitely divisible, Hume went on to conclude that the same holds for space and time (T 1.2.2, title).15 His principal argument for this conclusion (T 1.2.2.1–2) draws on what might be called the adequacy premise. (A‑prem): Adequacy premise [O]ur ideas are adequate representations of the most minute parts of extension[.] (T 1.2.2.1) (A‑prem) rests on the claim of Treatise 1.2.1.3 and 1.2.1.5 that Nothing can be more minute, than some ideas, which we form in the fancy; and images, which appear to the senses; since these are ideas and images perfectly simple and indivisible. The only defect of our senses is, that they give us disproportion’d images of things, and represent as

80  Finite divisibility; manners of disposition; points minute and uncompounded what is really great and compos’d of a vast number of parts. (T 1.2.1.5) Accordingly, “thro’ whatever divisions and subdivisions we may suppose [‘the most minute parts of extension’] to be arrived at, they can never become inferior to some ideas, which we form” (T 1.2.2.1), making the latter “adequate representations” (T 1.2.2.1) of the former. (If they are inadequate, it could only be because the most minute parts of extension are larger than our simple ideas.) Hume appealed to the further principle that (A‑prin): Adequacy principle Wherever ideas are adequate representations of objects, the relations, contradictions and agreements of the ideas are all applicable to the objects[.] (T 1.2.2.1) Drawing on (A‑prem) and (A‑prin), he argued that, since our simple and indivisible ideas and impressions cannot be packed together without mak‑ ing a package that augments in magnitude in direct proportion to each thing that is added, no finite space or time could contain infinitely many simple and indivisible parts. (EA): Equal augmentation But that [the supposition “that a finite extension contains an infinite number of parts”] is absurd [?], I easily convince myself by the consid‑ eration of my clear ideas. I first take the least idea I can form of a part of extension, and being certain that there is nothing more minute than this idea, I conclude, that whatever I discover by its means must be a real quality of extension. I then repeat this idea once, twice, thrice, &c. and find16 the compound idea of extension, arising from its repetition, always to augment.… [A]nd were I to carry on the addition in infinitum, I clearly perceive, that the idea of extension must also become infinite. Upon the whole, I conclude … that no finite extension is capable of containing an infinite number of parts; and consequently that no finite extension is infinitely divisible. (T 1.2.2.2) Presumably, “repeat[ing] this idea once, twice, thrice, &c” means forming copies of the initial idea and imagining each further copy being disposed as closely as possible to the previously imagined ones.17

Section 2.4  81 (EA) does not come off well on either the Georgian or the null theses. According to the Georgian thesis, visible minima all have the same non‑0 angular magnitude, so an infinite number of them must compose an infi‑ nite extension. However, the Georgian thesis draws a distinction between having parts and having visible parts (George 2006, 149, 150). The ad‑ equacy premise (A‑prem) is false on George’s account. While it is analytic that nothing that has parts can be more minute than a thing that has no parts, it is not analytic that nothing can be more minute than a thing that has no visible parts.18 Other things (such as portions of the space in which minimally visible impressions are disposed or points in such a space) could have proportionably decreasing angular magnitudes or no size of any sort, and it is not absurd that a finite extension should contain infinitely many such things. This might be cited as a reason to reject the Georgian thesis, but (EA) is not supported by the null thesis, either. On the null thesis, the simple and indivisible parts of our ideas and impressions have no magnitude. But then it cannot be absurd that adding them to one another would produce a magnitude that does not constantly augment with each additional part. As originally noted by Falkenstein (2006, 62–4), (EA) can be defended by appeal to what might be called the discreteness thesis (T 1.2.4.6).19 (D): Discreteness I ask any one, if [they see] a necessity, that a colour’d or tangible point shou’d be annihilated upon the approach of another colour’d or tangi‑ ble point? On the contrary, [do we] not evidently perceive, that from the union of these points there results an object, which is compounded and divisible, and may be distinguish’d into parts, of which each preserves its existence distinct and separate, notwithstanding its contiguity to the other?… A blue and a red point may surely lie contiguous without any penetration or annihilation. (T 1.2.4.6) (D) asserts that a blue and a red point can “lie contiguous” to one another (touch in the sense of having no gap between them) without any penetra‑ tion or annihilation (without overlapping at a point of contact, which for differently coloured points would amount to the annihilation of one, or their mutual penetration to produce a point of a mixed colour). This is equivalent to maintaining that the disposition of coloured and tangible points is governed by an immediate successor relation.20 Proceeding in any given direction from any given point, there is always an immediately next point that lies in that direction. Each step to the next point marks an augmentation of a line equal to that produced by the step to the prior

82  Finite divisibility; manners of disposition; points point. There must be an equal augmentation because (i) no gaps are being stepped over (so no increasingly small gaps are being stepped over), (ii) no point has any size (so later ones are not smaller than earlier ones), and (iii) no point overlaps any of the others.21 The augmentation that results under these circumstances is not produced by any addition of magnitudes pos‑ sessed by the points, because they have no magnitude to add. It is instead produced by the discrete manner in which the points are disposed.22 In a discrete space, there are not always further locations between any two given locations. Points can only be set so close to one another, like eggs in an egg crate, except there are no cradles that hold them apart. The space in which they are disposed just does not contain any intermediate places for them to be disposed at.23 The visual field models a discrete space. Because our vision has finite resolution, coloured points cannot be discerned arbi‑ trarily closely to one another. Minimally visible points nonetheless appear as if pressed hard up against one another, without any gaps.24 If they are similarly coloured, there are not even any boundaries or edges between them. They appear as an uninterrupted surface. 2.4.1 Discreteness and distinctness; whole‑part priority; measurement

Remarking on the absence of any visible distinction between immediately adjacent, similarly coloured points, Graciella De Pierris (2015, n176) has objected that appealing to discreteness “misleadingly suggests that minima pre‑exist the wholes of extension they compose and, therefore, that a fixed finite number of such minima are already there waiting to be discovered.” She proposes that, for Hume, homogeneous, simple, and indivisible points are originally confounded. They only come to be distinctly perceived by means of operations that separate them from their surroundings. In particular, the simple (unextended) minima whose confounding re‑ sults in a homogeneous appearance of extension at a given time (the darkly colored ‘points’ out of which the ink‑spot is composed) are not separately perceived as minima at this time, for they constitute the ap‑ pearance of extension only by being confused or confounded with one another. (2015, 114) She even appears to say that minima do not exist prior to being operation‑ ally distinguished from their surroundings. “[T]he exact (finite) number of minima in a given whole of extension is completely indeterminate, not simply unknown. For the notion of a minimum is defined by a temporally extended phenomenological process of successive diminution or division” (2015, 114).

Section 2.4.1  83 Earlier, Edward Slowik argued that Hume did not accept the “compos‑ ite hypothesis” that spatial magnitudes are “formed, or built up, out of the minimal visible points, such that the magnitude of the entire extended figure is the mere sum of the separate contributions of its composite mini‑ mals” (2004, 360). Instead, Slowik maintains, for Hume the parts of an originally perceived spatial magnitude are only distinguished from one another subsequently, by constructive acts of the imagination. Where dif‑ ferences in originally perceived spatial magnitudes are great, we “intuit” the difference without first discriminating and counting numbers of com‑ ponent parts or attempting a part‑for‑part superposition (2004, 362–5). Slowik bases this assertion on Hume’s view that relations of proportion in quantity and number are intuited where small numbers are concerned (T 1.3.1.2–3), and on Hume’s claim that points are too minute and con‑ founded to be counted or accurately juxtaposed (T 1.2.4.18–21). Slowik takes our “intuitive” awareness of (larger) differences in magnitude to pose an insuperable difficulty for the composite hypothesis. On the composite hypothesis,… as the number of minimal visibles is increased, the corresponding magnitude of the compound idea (or im‑ pression) is increased.… Yet, Hume’s insistence that the perception of a large visible figure cannot be shown to possess more minimal visibles (than a simultaneously perceived smaller visible figure) would seem to imply … that it is not possible to perceive the individual contribution of each minimal to the overall magnitude of the aggregate[.] (365) The kicker precedes these explanatory remarks. But, if the mind can determine magnitude without the need of a point summation, then why introduce the composite hypothesis at all? (365, see also 372–3) Slowik and De Pierris have a valid point (correcting, among others, Falk‑ enstein, 1997). For Hume, the impressions from which our ideas of space are derived could not be built up from an aggregation or summation of pre‑existing parts. They are sense impressions, after all, which means that they must be “original” (as Hume put it at Treatise 2.1.1.1) and not the products of any act of imagination. The parts of an extended impression are not psychogenetically prior to the whole. But there are both textual and theoretical reasons for recognizing that our spatially extended sense impressions are nonetheless originally given as arrays of discretely dis‑ posed parts, even if those parts are not edged off from one another by qualitative contrast.

84  Finite divisibility; manners of disposition; points Textually, Hume did not just write that “the idea of extension consists of parts” and that the “compound idea of extension … is compos’d of [‘perfectly simple and indivisible’] ideas” (T 1.2.3.14, my stress), he asked what original impression (or impressions) that idea is copied from and an‑ swered that it is copied from “impressions of colour’d points, dispos’d in a certain manner” (T 1.2.3.3) and (more tellingly) “impressions of touch … found to be similar to those of sight in the disposition of their parts” (T 1.2.3.5, stating that the impressions have spatially disposed parts). Elsewhere, he remarked that “A real extension … can never exist with‑ out parts” (T 1.2.4.3) and that “space or extension consists of a number of co‑existent parts dispos’d in a certain order, and capable of being at once present to the sight or feeling” (T 2.3.7.5). He did not write that a large visible figure cannot be shown to possess more minimal visibles than a simultaneously perceived smaller visible figure. He only wrote that we cannot determine this fact by an “exact numeration” (T 1.2.4.19) of those visibles. It is nonetheless “just, as well as obvious” that “as the proportion of the numbers [of points] varies, the proportion of the lines and surfaces is also vary’d” (T 1.2.4.19). While we are often “able at one view to de‑ termine the proportions of bodies … without examining or comparing the number of their minute parts” by relying on “the whole united appearance and comparison” (T 1.2.4.22), this very remark acknowledges that there are numerous minute parts, just ones we do not examine very closely. We can only determine the proportions of bodies at one view because the num‑ ber of minute parts comprehended within the whole united appearance makes that union of parts look as large as it does. Conjecturally, the union of the numerous minute parts is a function of their contiguity and their resemblance, resulting in the appearance of an extended, uniform area. Taken together, these remarks entail that the magnitude of wholes is a function of the number of discretely disposed parts they contain but that our assessments of magnitude are often (though not always) only vaguely determined by those numbers. In more detail: i The points do not “sum” (Slowik 2004, 360) to compose a magnitude. They could not, supposing, with De Pierris (2015, 118), that they have no magnitude of their own to contribute to a sum. They “sum to” or “compose” a magnitude only by being discretely disposed. That pre‑ supposes a whole in which they are perceived as being so disposed. ii Insofar as the points are “confounded” (T 1.2.4.19), it is only in the sense that their contiguous disposition and similarity make them ap‑ pear as a uniformly extended surface, not in the sense that they over‑ lap, which for points would mean they coincide and so could not figure as parts of an originally given extension (T 1.2.4.6).

Section 2.4.1  85 iii Points are not always confused. At the distance just before the ink spot vanishes, it must appear as a single dark point surrounded by a white area. To insist that it is still unclear how many dark points there are is to deny that the ink spot experiment proves its stated result. When a second dark point moves so close to the first that any further approach would cause the whiteness between them to disappear, and then a third dark point is so placed next to the second, the result must be an ap‑ pearance of, first, a 3‑point, and then a 5‑point long line, composed of exactly enumerated, dark and white points.25 To deny this is to deny the “finding” reported by (EA).26 iv While retreating from an ink spot or moving small objects into close proximity are operations, these are not operations of the imagination that produce ideas. They are operations of the body on the environ‑ ment creating experimental situations in which we experience complex impressions containing distinct and enumerable points. v When points are confused, they can still be vaguely enumerated. In virtue of being discretely disposed within a whole, multiples of two or more points compose a magnitude (call it a “portion”) that is increas‑ ingly easy to distinguish from portions that contain an increasingly disproportionate number of points, even when those portions are ho‑ mogeneous and disposed immediately alongside one another. vi In “intuiting” that one portion is greater, lesser, or approximately equal to another we are really vaguely perceiving that it contains a greater, lesser, or approximately equal number of discretely disposed points. The fact that we cannot exactly enumerate the points in a homogene‑ ous expanse does not imply that they are not there. This account is further supported by two theoretical considerations. i De Pierris writes that when I “gradually remove myself to the threshold distance just before the spot vanishes, what I actually perceive is a se‑ ries of ever smaller closely resembling impressions until I finally reach one that cannot be further diminished without annihilation” (2015, 113). She further writes that when we do this “we lose the appearance of continuous homogeneous extension with which we began, and we are phenomenologically presented with a single discrete unit” (2015, 117). All true, but there is more. The ink spot is a dark area on white paper. We never see it as a “single discrete unit,” if this is taken to mean that it is all that we see. As we retreat from it, the paper does not disappear. At the threshold distance, we do not just see a single discrete unit. We see an entire visual field containing (among other things that have since appeared) a homogenous white extension amid

86  Finite divisibility; manners of disposition; points which there is a minimally visible dark point. The dark point appears by way of contrast with the surrounding whiteness that hems it in on every side. Otherwise, we would have an abstract idea of some dark colour, with no shape or limits. (Even points have shape, that is, edges beyond which they do not extend. They just are nothing more than their edges, which are collapsed to a singularity.) Among other things, the ink spot experiment teaches us that individual minima can appear as distinct components of a visual field and that when they do, they appear pressed against their surroundings, without any gaps. It is only to be expected that if adjacent points were uniformly coloured, they would appear unbroken or continuous, even though they are disposed at discrete locations. ii When we perceive at first glance that one uniformly qualified shape is obviously larger than another, the opposite extremities of the larger shape must be perceived to be in different places at the same time. The larger shape must therefore consist of different portions, even if those portions are not edged off from one another by a colour contrast, scratch, or felt ridge (T 1.2.4.3 at the end, echoing Bayle, Dictionnaire, 360). If the larger shape is not a longer line, but a larger closed figure, it must have more than just one pair of opposite portions, and if it is a shape of any size, there must be some distance between its opposed peripheral portions, which means it must be perceived to have a cen‑ tral portion set between its peripheral portions. Insofar as we perceive these sectional portions to have some considerable magnitude, we must perceive them to have their own central and radial portions. We can only iterate these considerations so many times before the portions be‑ come difficult to distinguish, but as far as we can go, we discover that spatial magnitudes must contain as many parts as there are distinct places that those magnitudes occupy. 2.4.2 The failure of Hume’s finite divisibility arguments

While (D) entails that an infinite number of non‑overlapping points must compose an infinite extension, it does so at a cost. It only appeals (rhetori‑ cally) to the absence of the necessity of the opposite (“I ask any one, if [they see] a necessity”), that is, to a possibility, though one backed up by a matter of fact or “evident perception.” That matter of fact is, moreover, a matter of fact concerning coloured and tangible points. (D) only says that we evidently perceive that a blue and a red point may lie contiguous to one another without annihilation or penetration. Treatise 1.2.2.2 is no better, despite its pretension to find an “absurdity” in a finite extension with an infinite number of parts.

Section 2.4.2  87 I first take the least idea I can form of a part of extension.… I then repeat this idea once, twice, thrice, &c. and find the compound idea of extension, arising from its repetition, always to augment[.] (my stress) This is an appeal to experience of how things look or feel, or, as Hume put it, to a “finding.”27 Moreover, this finding is not grounded on an analysis of the nature of the “least ideas,” considered as adequate representations of the most minute parts of extension. Hume’s appeal to (A‑prin) fails at this point. It is not always true that where ideas are adequate representa‑ tions of objects, the relations of the ideas must be applicable to the objects. That is only true of those relations (like resemblance) that are determined by their relata. But, as Hume noted at Treatise 1.3.1.1, some relations, no‑ tably those of disposition in space, can change even while the relata remain unchanged. Even though nothing could be smaller than an indivisible idea, the discrete manner in which ideas are disposed is not dictated by their mi‑ nuteness but instead by the finite capacities of the mind in forming images, reflecting the finite powers of resolution of the senses of vision and touch (T 1.2.1). This is a fact concerning the limits of sense and imagination. It is not a necessary rule governing how the smallest possible things might be disposed because, on the null hypothesis, there is nothing about the things that prevents their being placed arbitrarily closely to one another. Having based his conclusion on this finding, or evident perception, Hume was in no position to claim that the supposition that a finite exten‑ sion may contain infinitely many parts is absurd (T 1.2.2.2). Falling back on the claim that “this conceit of a mathematical continuum is not forced on us by appearances, and there is no way to know if it is true of real‑ ity” (Baxter 2009, 119) would not perturb mathematicians, who do not consider themselves to be constrained to working with empirically given objects or states of affairs that are true of (physical) reality. And physical scientists would claim a right to appropriate the mathematics that best allows them to account for the appearances. If we must appeal to an in‑ finitesimal calculus to adequately account for the elliptical motions of the “planets” (that is, of minimally visible points of light appearing on the in‑ ner surface of a “celestial sphere” or better a visual field) then we cannot be faulted for treating them as arbitrarily small bodies moving through an infinitely divisible space, even if that is not how they look. Treatise 1.2.2 contains two further arguments for the finite divisibility of extension. (A third, Treatise 1.2.2.5, rests on the success of the others and so can be set aside.) Each has the same defect as (EA). The first, the “Malézieu” argument of Treatise 1.2.2.3, asserts that only units exist, where a unit is understood to be something “perfectly

88  Finite divisibility; manners of disposition; points indivisible, and incapable of being resolv’d into any lesser unity” (T 1.2.2.3).28 Anything that has parts is only an arbitrarily delimited collec‑ tion of units (Treatise 1.2.2.3 calls it a “fictitious denomination”), which exists in virtue of the existence of the units of which it is composed. It follows that if extension “never resolves itself into any unite or indivisible quantity” it could not exist. The second, widely assumed to have been drawn from a similar argu‑ ment in note F of Bayle’s Zeno article (Dictionnaire, 353–4n), is, like its presumed source, challenging to interpret.29 It premises that ’Tis a property inseparable from time, and which in a manner consti‑ tutes its essence, that each of its parts succeeds another, and that none of them, however contiguous, can ever be co‑existent. (T 1.2.2.4) Presumably, this means that however closely packed the parts of time may be (however “contiguous” they may be), only one of them ever exists, that one being the present moment. Given this premise, Hume concluded that “time, as it exists, must be compos’d of indivisible moments.” Otherwise, the one existing part of time, the present moment, would itself be divisible into parts, and “there wou’d be an infinite number of co‑existent moments, or parts of time,” contrary to the premise (T 1.2.2.4).30 These arguments at best establish that extension and time must contain indivisible units. If it is supposed that these units have no extension or duration of their own, it remains possible that infinitely many of them could be distributed over a finite space or time.31 There is no contradic‑ tion in the supposition that a continuum might be composed of an infinite number of points.32 Even though this result was only rigorously established in the nineteenth century, eighteenth‑century scholars understood the real number continuum, and they understood that the real numbers between 0 and 1 can be mapped onto a finite line. It would take an entirely different argument to establish that the reals do not mark infinitely many parts of a finite space or time. Hume revisited the infinite divisibility of time at Enquiry 12.19. An infinite number of real parts of time, passing in succession, and ex‑ hausted one after another, appears so evident a contradiction, that no man, one should think, whose judgment is not corrupted, instead of be‑ ing improved, by the sciences, would ever be able to admit of it. Apparently, the argument is that if time is a succession of “real parts,” where each real part must be “exhausted” or pass out of being before the

Section 2.5  89 next can begin to be, an infinite number of real parts could not be suc‑ cessively exhausted between now and a second from now. Hume did not explain why not. He may have thought that it takes some time for a real part to be exhausted, so that were there infinitely many real parts it would take an infinite time for them to all be exhausted. But since the real parts in question are parts of time, any given real part cannot take up any more time than it takes up, and so cannot take any more time to be exhausted than it takes up. For each real part in any given time interval, however much or however little time that real part takes up, there is a real part of time (that part itself) over which it is exhausted so that the total number of real parts, whether finite or infinite, is distributed over the interval without deficit or remainder. 2.5

The other qualities of our complex ideas

The title of Treatise 1.2.3, “Of the other qualities of our ideas of space and time,” is explained by the conclusion of Treatise 1.2.1, that all our ideas consist of simple and indivisible parts. The simple and indivisible parts of our ideas of space and time could not be spaces or times (which have parts). They must have “other qualities.” Our ideas of space and time could not be “separate or distinct ideas” (T 1.2.4.2). They must be com‑ plexes with those qualities. It is only at this point that Hume brought up the empirist thesis that all ideas are copies of impressions. The best way to discover the “other quali‑ ties” of our ideas of space and time is to identify the original impressions from which those ideas are derived (T 1.2.3.1). Upon opening my eyes, and turning them to the surrounding objects, I perceive many visible bodies; and upon shutting them again, and consid‑ ering the distance betwixt these bodies, I acquire the idea of extension.33 What am I considering when I shut my eyes and consider the distance be‑ tween visible bodies? Hume might have first looked at his table to see his books and papers (T 1.4.2.18), then closed his eyes, and formed an idea of the distance between the differently coloured books and papers or be‑ tween the edges of a uniformly coloured book cover or piece of paper. But what constitutes a greater or lesser distance between these “landmarks” or endpoints? Thinking, perhaps, of clearing the table of everything, Hume wrote, The table before me is alone sufficient by its view to give me the idea of extension. This idea, then, is borrow’d from, and represents some

90  Finite divisibility; manners of disposition; points impression, which this moment appears to the senses. But my senses convey to me only the impressions of colour’d points, dispos’d in a certain manner. (T 1.2.3.4) In this exemplary case, I do not see multiple objects. I just see one table. Tables have parts, some of which, their legs for instance, are distant from one another, but Hume seems to have been thinking of looking just at the top of his table, with it filling his field of view. In this case, “objects” have disappeared. The entire visual field is filled with the view of a portion of a featureless, flat tabletop and so with identically coloured “points.” But these points are not just coloured. Each has a place at which it is disposed relative to the others. Suppose that in the …composition of colour’d points, from which we first receiv’d the idea of extension, the points were of a purple colour; it follows, that in every repetition of that idea we wou’d … place the points in the same order with respect to each other[.] (T 1.2.3.5) We first receive the idea of extension from a “composition” of coloured points. If we subsequently place the points with respect to each other, it is because we are copying a “composition” in which they were “first received.” Hume concluded, (IE): Idea of extension If the eye is sensible of any thing farther, I desire it may be pointed out to me. But if it be impossible to show any thing farther, we may conclude with certainty, that the idea of extension is nothing but a copy of these colour’d points, and of the manner of their appearance. (T 1.2.3.4) Scholars have since been concerned with how Hume’s observation that the idea of extension is a copy of a “manner of appearance” of simple impressions fits with his doctrines that all ideas are obtained from simple impressions (T 1.1.1.7 and 6) and that whatever is different is separable (T 1.1.7.3).34 But Hume was likely focused on the first conjunct of (IE): on the fact that the idea of extension is the idea of coloured points. The colour is the “other quality” he was concerned to identify, not the place or manner of appearance. Hume immediately went on to observe that the points need not be col‑ oured. They could be tangible (T 1.2.3.5). They could also all be of the same

Section 2.5  91 colour or the same tangible quality (T 1.2.3.5). But there can be no spatial manner of appearance unless there are some coloured or tangible points that appear. There can be “no separate or distinct ideas” (T 1.2.4.2) of space. There is nonetheless something surprising about Hume’s declaration that visual and tangible points are “dispos’d in a certain manner” (T 1.2.3.4) or “placed in an order with respect to each other” (paraphrasing T 1.2.3.5). How does this fit with the “copy principle” that “all our simple ideas in their first appearance are deriv’d from simple impressions, which are cor‑ respondent to them, and which they exactly represent” (T 1.1.1.7)? In the exemplary case of the view of the tabletop, the impression is not simple but complex. The complex impression consists of simple and indi‑ visible parts, in this case, purple points. But the impression is not just the disordered aggregate of those parts. It is those parts placed in an order with respect to each other. The order of placement is a further feature of the complex impression, not of its parts. While each part has its own place in addition to a colour or tactile quality, its place is place with respect to other parts and so is a quality it acquires from the fact that it is given as a part of a complex impression. Not to mention the copy principle of Trea‑ tise 1.1.1.7, from which the ideas of space and time are exempt (because that principle only applies to simple ideas whereas these are complex), there is here no violation of the more general empirist principle that even our complex ideas of “inconstant” relations between simple ideas (those not derived from comparison of the relata, as discussed at Treatise 1.3.1.1) are all copied from corresponding complex impressions. It just needs to be recognized that complex impressions really are impressions, that is, imme‑ diately given wholes, and not the product of any operation performed on simple impressions, even comparison. This still raises a concern. The simple and indivisible purple points con‑ stitutive of the tabletop are separable from one another (with a jigsaw, for instance). They can therefore be separated from their manner of disposi‑ tion, which suggests that the manner of disposition ought to be separable from them, contrary to Hume’s declared aim (T 1.2.4.2). This concern is addressed by Hume’s account of “distinctions of rea‑ son” (T 1.1.7.17–18). On that account, it is possible to draw distinctions between things that are neither different nor separable provided the dis‑ tinctions refer to different ways a particular might resemble other particu‑ lars. Rather than conceive these resemblances separately, the distinction involves thinking of the particular in relation to different classes of resem‑ bling particulars. We conceive a white globe to belong either to a class of globes or a class of white particulars, thereby drawing a “distinction of reason” between the globe’s whiteness and its shape without forming dis‑ tinct or separate ideas of its whiteness and its shape.

92  Finite divisibility; manners of disposition; points The same holds for the purple points constitutive of the table. I can cut a point out of the table with a jigsaw, thereby separating it from the man‑ ner in which it is disposed relative to the surrounding purple points, but it still appears somewhere on my visual field where it is surrounded by other points. Even if I only imagine it, I still imagine it as a point and not as an abstract idea of purple. Just as I can only imagine a white globe by imagin‑ ing whiteness extending out to a certain spherical boundary beyond which it visibly contrasts with a not‑white background, so I can only imagine a purple point as a point by imagining it hemmed in on all sides by points of contrasting colour35 or of no colour.36 There is no way of conceiving a purple point without conceiving it as disposed in some manner relative to some surroundings. Likewise, I cannot conceive a manner of disposition without conceiving some coloured or tangible points disposed in that fashion. Hume applied this doctrine at Treatise 1.2.3.5. But afterwards having experience of the other colours of violet, green, red, white, black, and of all the different compositions of these, and finding a resemblance in the disposition of colour’d points, of which they are compos’d, we omit the peculiarities of colour, as far as possible, and found an abstract idea merely on that disposition of points, or man‑ ner of appearance, in which they agree. Omitting the peculiarities of colour as far as possible means imagining differently coloured points disposed to make the same figure/background contrast. Because the same sensible points may be differently disposed, the man‑ ner of disposition cannot be based on anything found in the points consid‑ ered individually. If that were the case, changing the manner of disposition would be impossible without some change in the points. But it is possible. Accordingly, relations of place in space do not depend on the relata. They are given in a complex impression. As Hume put it, spatial relations are “perceived” along with their relata. All kinds of reasoning consist in nothing but a comparison, and a dis‑ covery of those relations, either constant or inconstant, which two or more objects bear to each other. This comparison we make, either when both the objects are present to the senses, or when neither of them is pre‑ sent, or when only one. When both the objects are present to the senses along with the relation, we call this perception rather than reasoning; nor is there in this case any exercise of the thought, or any action, prop‑ erly speaking, but a mere passive admission of the impressions thro’ the organs of sensation. According to this way of thinking, we ought not to

Section 2.5  93 receive as reasoning any of the observations we may make concerning …the relations of time and place; since in none of them the mind can go beyond what is immediately present to the senses. (T 1.3.2.2) As Slowik and De Pierris have tacitly observed (Section 2.4.1), Hume cannot have been a “psychological atomist.” He did not accept that all complex perceptions are the products of higher‑order acts of association or comparison performed on previously given simples. In recognizing this fact, Hume tacitly repudiated the “Malézieu” principle that “existence in itself belongs only to unity” (T 1.2.2.3, my stress). As Hume was careful to note twice over the course of Treatise 1.2.3.4, the idea of extension is not just copied from “the impressions of colour’d points,” but of coloured points disposed or appearing in a certain manner. Whatever we might think of the Malézieu principle considered as an ontogenetic principle, it has no psychogenetic validity. Though he could not have known it, modern computing captures what Hume thought. The byte 10100110 is different from byte 01100110. It is not made different by its component bits but by the manner in which the bits are registered, that is, the manner in which they are disposed relative to one another. The registration of the bits is original. The computer can‑ not just be fed four 0s and four 1s all at once in no particular order. They need to be given in an order and the computer needs to be designed to receive and preserve that order. That manner of disposition of the bits is a further, “real” feature of the byte, which makes it a whole that is greater than the sum of its parts. Similarly, the manner of disposition or appear‑ ance of the coloured points (as constituting, say a white globe rather than a white cube) is a feature of the complex impression that cannot be derived from the component, simple and indivisible, coloured points. The coloured points do nothing to determine their manner of disposition. That manner is only given in the complex impression.37 Contrary to the Malézieu principle, existence is in this case not just the existence of the perfectly indivisible points or units, but the existence of the compound, which is not a unit but a “number” of perfectly indivisible points disposed or appearing in a certain manner. The manner of disposi‑ tion is a further, “real” quality of the complex impression, as well as of our ideas of space and time. Hume’s account of the origin of our ideas of space and time might best be described as “intuitionist,” where “intuition” is taken to refer to im‑ mediate apprehension. Unfortunately, Hume used “intuition” more nar‑ rowly to refer to immediate apprehension of relations between ideas. As block quoted above, he preferred to describe our immediate apprehension of space and time as “perception.” But “perception” is widely extended to

94  Finite divisibility; manners of disposition; points mediate cognition as well as immediate apprehension. “Sensation” is also sub‑optimal, as it is widely taken to refer just to what Hume called “simple” impressions. In what follows, “intuition” is used more broadly than Hume used it, to refer to immediate or original apprehension of all sorts, notably including direct apprehension of contiguity and distance relations.38 2.6

The other qualities of our simple and indivisible ideas; the non‑entity of unqualified points

According to Treatise 1.2.3.14, “the idea of extension consists of parts.” If there were only one indivisible part or “point” there could be no idea of extension. According to Treatise 1.3.1.1, the manner in which the points are disposed does not depend on their intrinsic qualities and so does not depend on whether they have the same or different qualities. This suggests that at least some points need not have any qualities. A point of no quality can make a better qualitative contrast with a point of some quality than two points of different qualities. (It is possible to do all of computing with bytes composed of just one quality, “turned on” or “1.” The “0” or “off” bits are defined by the absence of current flowing through that part of the register. Light rays hitting some points on the retina but not others have the same effect on the brain.) Provided there are some qualified points mixed with “unfilled” points, the latter can appear by way of contrast, though their numbers may be vague (Section 2.4.1) and their bounds indetermi‑ nate (note 35). Hume resisted this suggestion. He delved further into the question of the qualities of the simple and indivisible parts of the idea of extension over Treatise 1.2.3.12–17. There, he proposed to take one of those simple indivisible ideas, of which the compound one of extension is form’d, and separating it from all others, and consider‑ ing it apart, … form a judgment of its nature and qualities. (T 1.2.3.13) This project is suspect. It may seem unexceptionable to take a jigsaw to Hume’s table and cut a point away from the surrounding points and con‑ sider it in isolation. But, as discussed in Sections 2.4.1 and 2.5, the point is only conceived as a point (as opposed to an abstract idea of the colour purple) insofar as it is conceived as hemmed in on all sides by points of contrasting colour (or no colour). The distinction between a point and its surroundings is a distinction of reason, drawn by considering resemblances between the point in its given surroundings and points that resemble it in colour that are found in other surroundings. A point is not something that can properly be considered “apart.”

Section 2.6  95 Setting this worry aside for now, Hume proceeded to observe that, ’Tis plain it [“one of those simple indivisible ideas” separated “from all others” and considered apart] is not the idea of extension. For the idea of extension consists of parts; and this idea, according to the supposi‑ tion, is perfectly simple and indivisible. Is it therefore nothing? That is absolutely impossible. For as the compound idea of extension, which is real, is compos’d of such ideas; were these so many non‑entities, there wou’d be a real existence composed of non‑entities; which is absurd. Here therefore I must ask, What is our idea of a simple and indivisible point? (T 1.2.3.14) The reason for Hume’s worry about whether the idea is “nothing” is ex‑ plained over the course of his subsequent discussion of a dispute over the nature of mathematical points (T 1.2.4.9–15). The dispute arose from Eu‑ clid’s Elements (1, definition 1), which defines a geometrical point as “that which has no parts.” As Hume went on to discuss, mathematicians and translators often embellished this definition by observing that a geometri‑ cal point has no magnitude or dimensions: neither width, nor breadth, nor thickness.39 These embellished definitions make it seem that a point must be a “non‑entity” and raise the question of what makes it “real.” Transla‑ tors, textbook editors, and mathematicians sometimes attempted to ad‑ dress this question, offering comments that could range from fairly brief remarks to pages of closely argued text.40 But this mathematical dispute was not his topic, Hume declared. He maintained that his question concerning the nature of simple and indivis‑ ible points “has scarce ever yet been thought of,” adding “We are wont to dispute concerning the nature of mathematical points, but seldom con‑ cerning the nature of their ideas” (T 1.2.3.14). His answer to this different question is that the idea “of” a point must be an idea of vision or touch, since these are the only senses that supply us with ideas of extended things. As such, he maintained, it must be endowed with one of the qualities ex‑ perienced by those senses, colour or solidity (Hume may have been using “solidity” as a catch‑all for any tactile quality). The idea of space is convey’d to the mind by two senses, the sight and touch; nor does any thing ever appear extended, that is not either visible or tangible. That compound impression, which represents extension, consists of several lesser impressions, that are indivisible to the eye or feeling, and may be call’d impressions of atoms or corpuscles endow’d with colour and solidity. But this is not all. ’Tis not only requisite, that these atoms shou’d be colour’d or tangible, in order to discover

96  Finite divisibility; manners of disposition; points themselves to our senses; ’tis also necessary we shou’d preserve the idea of their colour or tangibility in order to comprehend them by our im‑ agination. There is nothing but the idea of their colour or tangibility, which can render them conceivable by the mind. Upon the removal of the ideas of these sensible qualities, they are utterly annihilated to the thought or imagination. (T 1.2.3.15) This is where Hume’s proposal to “take one of those simple indivisible ideas … and considering it apart … form a judgment of its nature and qualities” (T 1.2.3.13) led him astray. Were it possible to perceive a single point in isolation from any surroundings it would of course need to display the sort of qualities that vision or touch detect. But it would still have to be an individual and not a universal. As Hume acknowledged in this very pas‑ sage, it would have to be a “lesser impression,” or “atom” or “corpuscle” that is “endow’d with colour and solidity” (T 1.2.3.15). It could not be an abstract idea of colour or solidity. No “simple indivisible idea” could be truly “considered apart” and still appear as a particular (a “lesser impression,” atom, or corpuscle) rather than a universal. The most separate consideration would have to include an immediate neighbourhood of contrasting points that hems in the “lesser impression” on all sides.41 (As Hume noted at Treatise 1.2.4.19 modified by 1.2.4.6, a point can only be conceived as a point by way of contrast with its surroundings; when it is not edged off from them in that way, it blends in with them to form an extension containing indeterminately many points. See Section 2.4.1.) A simple and indivisible point immediately sur‑ rounded by contrasting points has a position in a neighbourhood of points. (The relation between a point and its position is analogous to that between a white globe and its shape [pp. 92–3, Section 5.3.5].) Its position is a fur‑ ther real quality and is essential to making it an individual. The notion that spatial points are essentially characterized by position is ancient. Aristotle maintained that “the sciences that are derived from fewer principles (for instance, arithmetic) are more exact than those (for instance, geometry) that require further principles” (Metaphysics 1.2 982a25–30). Among his many illustrations of this claim, one, given at Posterior Analytics 1.27, states that arithmetic is more exact than ge‑ ometry because arithmetic is based on units whereas geometry is based on units that have position.42 This was appreciated in Hume’s day.43 In his long commentary to his 1705 translation of Euclid’s Elements 1, Definition 1, Edmund Scarburgh attributed the thesis that a point is “A Monade, or Unity having position” to the Pythagoreans and to Proclus and observed that

Section 2.6  97 Indeed to have Position, or Situation is the only positive conception to be made of a Point: Whose Existence is in its Locality[.] (Euclid 1705, 2) A Point is to be conceived such a kind of Unity, such a Monade, or In‑ divisible Being, as in Space to possess an impartible Place, or Position. (Euclid 1705, 3) Colour or tangibility is not necessary to make points “real,” according to Scarburgh. Position or locality does that job. Issac Barrow reviewed and attacked the Aristotelian account, but only because he considered arithmetical units to be divisible, not because he wanted to deny that points have a position.44 But Aristotle and Proclus affirm that Unity (they had more rightly said Numbers) the Principle of Arithmetic, is more simple than a Point which is the Principle of Geometry, or rather of Magnitude. Because a Point implies Position, but Unity does not. [At simplicior est, inqui‑ unt Aristotles & Proclus, unitas Arithmeticæ (rectiùs numeri dixissent) principum, quàm punctum, quod est principium Geometriæ (vel mag‑ nitudinis potiùs). Quòd punctum implicet positionem, unitas non item.] A Point, says Aristotle [fn: IV. Met. cap.  6], and Unity are not to be divided, as Quantity: Unity requires no Position, a Point does. But this Comparison of a Point in Geometry with Unity in Arithmetic is of all the most unsufferable, and derives the worst consequences upon Math‑ ematical Learning. For Unity answers really to some Part of every Mag‑ nitude, but not to a Point.… A Point is rightly termed Indivisible, not Unity.… Numbers arise from the Division of Unity. [Verùm imprimìs deterrima est, & pessimas in Mathesin derivat consequentias ista puncti Geometrici cum Arithmetica unitate collatio: Nam unitas reverà parti cujuslibet magnitudinis aliquotæ respondet, non puncto … Et punctum rectè dicitur indivisibile; unitas verò nequaquam … ex unitatis divisione procreantur numeri.] (Barrow 1734, 48 / 1685, 54–5) Barrow went on to remark that a point “taken universally” (likely he meant separated from its surroundings and considered abstractly) is “indetermi‑ nate, and void of Position,” like “Unity taken the same Way.” But he added that “Unity taken particularly implies a definite Position,45 and all other particular Circumstances, as well as a particular Point” (1734, 48–9).46 It may be that Hume’s decision to distance himself from the dispute over the nature of mathematical points (T 1.2.3.14) led him to overlook the tenet

98  Finite divisibility; manners of disposition; points that points have position. Whatever its cause, the oversight undermines the argument of Treatise 1.2.3.12–16, hereafter called “the argument from the non‑entity of unqualified points.” The argument turns on the premise that the simple indivisible parts of space must be “endow’d with colour and solidity” (T 1.2.3.15) in order to “discover themselves to our senses,” be “comprehend[ed] … by our imagination,” and be “conceivable by the mind” (T 1.2.3.15). That premise is false. Hume was right to observe that a colourless and intangible point would not by itself affect the senses, and he would have been right to maintain that the experience of a sensory field en‑ tirely taken up with such points would be indistinguishable from no experi‑ ence whatsoever. But colour and solidity are not necessary to make points sensible, imaginable, or conceivable. Contrast and position are necessary. Points must be positioned in the neighbourhood of immediately contrasting points. Some points must be qualified (coloured or tangible) for there to be a contrast, but it is not necessary that all be qualified. A colourless point surrounded by coloured points is as evident to the senses as a coloured point surrounded by points of a contrasting colour, and more distinct than a point surrounded by points of the same colour (Section 2.4.1). Its posi‑ tion, as made distinct by its contrast with its coloured surroundings, is a positive quality that endows the point with all the “reality” it needs to affect the senses as a part of a larger whole that affects the senses. This makes it comprehensible by the imagination and conceivable by the mind. Scarburgh described the positions of points as being “impartible” from them, but he was thinking of geometrical points. Hume’s remark that the lesser impressions contained in the compound impression of extension might be called atoms or corpuscles implies to the contrary that he con‑ sidered them to be mobile. This need not entail that there are two kinds of being: coloured or solid corpuscles, and a network of immobile, absolute places constituting an independently existing, substantival space contain‑ ing the corpuscles. The corpuscles and the “unfilled” points could be sup‑ posed to all be mobile. Each still occupies a position relative to each of the others. All are collectively experienced, from moment to moment as parts of a whole sensory field. Hume could have gone either way on this ques‑ tion. Recognizing the conceivability (and perception) of a vacuum would not have committed him to accepting absolute or substantival space.47 This having been said, substantivalism was the more theologically sus‑ pect alternative. It is surprising that Hume would not have found it attrac‑ tive just for that reason. 2.7

Hume’s motives

Hume claimed that space and time are not infinitely divisible and that it is impossible to conceive either a vacuum or a time when there was no

Section 2.7  99 succession or change in any real existence. It is controversial why he was invested in drawing any of these conclusions. According to Norman Kemp Smith, Hume’s “main motive in denying space and time to be infinitely divisible … was his desire to vindicate for reason the right to have jurisdiction in every field of possible human knowl‑ edge” (1941, 287), and rebut those who would attempt “deliberately to discredit a faculty upon which philosophy has itself to rely in its attempts to restrain the forces, already sufficiently strong, of ignorance, fanaticism and superstition” (285). Hume had seen Port Royal argue that reason incontestably demonstrates the incomprehensible thesis that the smallest atom is infinitely divisible. It had gone on to conclude that, “it is good to tire the mind on these subtleties, in order to master its presumption and to take away its audacity ever to oppose our feeble insight to the truths pre‑ sented by the Church, under the pretext that we cannot understand them” (Port Royal 4.1, 233). Bayle (Dictionnaire, 362) had gone further, arguing that all accounts of matter are equally absurd, but that infinite divisibility is “strongest” because it “quibbles best.” When the scholastic distinctions and the jargon with which it is defended have been used up, without having been able to make this theory com‑ prehensible, one can take refuge in the nature itself of the subject and claim that, since our mind is limited, no one ought to find it odd that we cannot resolve what concerns infinity, and that it is of the essence of such a continuum that it be surrounded by difficulties that are insur‑ mountable to human beings. (360–1) As Kemp Smith observed (1941, 286), Bayle then turned this argument on those who would question religious mysteries. If a man is convinced that nothing good is to be expected from his philosophical inquiries, he will be more disposed to pray to God to persuade him of the truths that ought to be believed than if he flatters himself that he might succeed by reasoning and disputing. A man is therefore happily disposed toward faith when he knows how defective reason is. (Dictionnaire, 206, compare 372) Finding these challenges to reason too serious to be ignored, Hume sought to show that finite extension is not infinitely divisible. According to Paul Russell (2008), Hume was at least as concerned to undermine a theology developed by Newton and Clarke as to respond to Port Royal and Bayle. Perhaps to pre‑empt the concern that absolute space

100  Finite divisibility; manners of disposition; points and time would place constraints on God’s operations and manner of ex‑ istence Newton proposed that God constitutes duration and space by ex‑ isting always and everywhere (Principia, “General Scholium” 2, 390) and that space might be identified with his “sensorium” (Opticks 3, Query 31, 379). Clarke tried to turn the tables on Newton’s opponents by offering an “a priori” argument for the existence of God that appeals to the necessary existence of spatial modes or attributes, and hence of some necessarily existent substance to which they belong. [W]hen we are endeavoring to suppose that there is no being in the uni‑ verse that exists necessarily, we always find in our minds … some ideas, as of infinity and eternity, which to remove (that is, to suppose that there is no being, no substance in the universe to which these attributes or modes of existence are necessarily inherent) is a contradiction in the very terms. For modes and attributes exist only by the existence of the substance to which they belong. (Demonstration, 13) Clarke buttressed this argument by following Opticks, 379 in maintain‑ ing that space can be void of matter, which entails that matter could not itself be a necessary being (Demonstration, 20). According to Russell, this will have made it important for Hume to attack the thesis that “unfilled” or void places are conceivable (2008, 107). Russell further maintains that even Hume’s critique of infinite divisibility was directed against Clarke, in this case, against Clarke’s attempt to argue against Collins that a material being cannot think (2008, 108). In contrast, Donald Baxter has maintained that Hume was a “Pyrrho‑ nian empiricist” whose “conclusions about space and time should be read as conclusions about space and time as they appear” (2009, 121). Com‑ mentators have been wrong to assume that he was concerned with “space and time as they really are or space and time as they are presupposed to be by our geometry and physics” (2016, 174). He was concerned with distinguishing “how space and time apparently appear from how they re‑ ally appear” (2016, 174). We can be talked out of the former whereas the latter has more stability (2009, 115). But “some incautious uses of the words real and really” notwithstanding, “Hume contrasts the appearances of objects with their unknowable real natures” (2009, 116), and “means really to be talking about the external world as it is experienced via our impressions and as it is represented by our more stable ideas,” not to “pen‑ etrate into the nature of bodies” (2009, 118). Accepting this, Hume’s position on space and time is really founded on his case for scepticism about knowledge of an external world and for the thesis that we can have no ideas that are not drawn from sense experience (Baxter 2009, 121). He never meant to argue that space and time are not infinitely divisible or that a vacuum cannot exist. While he did argue that

Section 2.7  101 space and time as they “really appear” are only finitely divisible and that we cannot experience or even “really imagine” a vacuum, those arguments are only effective against realist opponents when further backed up by a claim that we have no way of getting beyond the appearances. A final account of what motivated Hume might be drawn from his re‑ mark in the introduction to the Treatise (Introduction 4) that all the sci‑ ences, even mathematics and natural philosophy, are in some measure dependent on the science of human nature, and might be changed and improved by our being better acquainted with the extent and force of our understanding. He may have wanted to cash in that remark by applying his account of impressions and ideas to the resolution of then‑controversial issues concerning the composition of extension and the existence of ab‑ solute space. Don Garrett, for instance, has proposed that Hume wanted to challenge “the intelligibility and explanatory value of Newtonian met‑ aphysicians’ conceptions of empty absolute space and empty absolute time as independent metaphysical entities or quasi‑entities” as well as the ­Cartesian view that the universe is a plenum (1997, 56–7). It is not easy to weigh these accounts of Hume’s motives. While En‑ quiry 2 and Treatise 1.1.1 begin by arguing for empiricism, Pyrrhonian arguments only come up in Enquiry 12 and in Treatise 1.4.2 and 1.4.4. In the Treatise, the claim that all our knowledge is confined to our per‑ ceptions is not broached until 1.2.6, after the arguments concerning space, time, and a vacuum have been drawn to their conclusion, which is odd if Hume intended to premise them on a Pyrrhonian approach.48 In the Enquiry, perhaps due to his encounter with Stanhope (see Section 2.1), Hume appears to have retreated from the view that mathematical demonstrations are inadequate to get behind the appearances (T 1.2.4.8 and 17–33), replacing it with the assertion that the reasoning is impec‑ cable, however “seemingly absurd” (EHU 12.20) the conclusions. Far from being attacked, the paradoxes of infinite divisibility are described as “The chief objection against all abstract reasonings” (EHU 12.18). The Treatise’s appeal to an alternative account of space as composed of simple and indivisible points is relegated to a pair of footnotes. In the first (EHU 12.18n), physical points are not invoked to avoid the paradoxes but rather figure in an argument that deepens them. (Enquiry 12.18 does not retreat from declaring the demonstrations of infinite divisibility to be “unexceptionable.”)49 In 1748 and 1750, the second footnote (EHU 12.20n) contained a passage charging that the ideas of greater, less, and equal are not exact or determinate enough to serve in demonstrations of infinite divisibility. But this passage was deleted from the third and all subsequent authorized editions of the Enquiry.50 Without the support of the attack on demonstrations, EHU 12.20n is reduced to imploring mathematicians to avoid exposing their discipline to the ridicule of the ignorant by confining their reasoning to the finite quantities suggested by the senses and imagination.

102  Finite divisibility; manners of disposition; points Kemp Smith’s account of Hume’s motives fits better with the late in‑ troduction of external world scepticism in the Treatise and with Hume’s apparent consternation over that result. But it does not sit well with Hume’s attitude to the relations between reason, scepticism, and religion in his authorized works. Scepticism is a double‑edged sword. It can just as well be used to attack superstition, enthusiasm, and more reason‑ able forms of religion. In the Dialogues, the arguments against natural religion are offered by the sceptic, Philo, and Cleanthes’ observation on “The ill use, which Bayle and other libertines made of the philo‑ sophical scepticism of the Fathers and the first Reformers” (D 1.17) goes uncontested. The Enquiry does contain a much‑cited passage comparing the doctrine of infinite divisibility to religious mysteries. No priestly dogmas, invented on purpose to tame and subdue the rebel‑ lious reason of mankind, ever shocked common sense more than the doctrine of the infinite divisibility of extension, with its consequences; as they are pompously displayed by all geometricians and metaphysi‑ cians, with a kind of triumph and exultation. (EHU 12.18) But this is a comparison, not an identification. It is meant as a reproach. Champions of the paradigm of rational science should not be behaving like priests and awing their students into amazed belief by propounding mysteries, even though they have what Enquiry 12.18  goes on to de‑ scribe as “unexceptionable demonstrations” to offer. While Hume was aware of the use that Port Royal, Bayle, the early church Fathers, and the first reformers had made of scepticism, he did not object that math‑ ematicians are aiding the cause of superstition. He may have thought that such subtleties are more likely to try the patience of ordinary peo‑ ple than inspire their awe. Superstition has better ways of advancing its claims. These considerations suggest that the mature Hume might not have con‑ sidered an attack on the doctrines of infinite divisibility or the existence of vacua to have an important role to play in his anti‑religious project. Though the mature Hume felt freer to openly attack religion and did so in Enquiry 7.21–5, 8.36, 10, and 11; Morals 9.3; “National Characters” 162n2; and throughout the Natural History and the History, infinite di‑ visibility and vacuum do not come up in those attacks. Even in the Dia‑ logues, Philo charges that the non‑existence of God is as conceivable as the non‑existence of matter (D 9.7). That is a step away from declaring exten‑ sion where there is no matter to be inconceivable. Hume may have given

Section 2.7  103 up on drawing a theological conclusion from the debate over empty space because, as Emily Thomas (2018, 206) has observed, the furore over the religious implications of Newtonian absolute space had “ceased to occupy the central philosophical stage” after the 1730s. For the mature Hume, Newton, Locke, and Clarke were “Arians or Socinians” (NHR 12.23n.78) whose heretical religious opinions could be indulgently excused in a foot‑ note, there being no longer any need to take their theological views and arguments seriously. Even in 1739, when Hume would have had more reason to attack Clarke’s argument from the necessary existence of space and time, he would have found himself in possession of other ways of combatting Clarke’s views. His rejection of the intelligibility of the notions of substance and inher‑ ence (T 1.1.6.1 and 1.4.5.2–3) scuttles Clarke’s claim that spatial and tem‑ poral “attributes or modes” must exist in some substance. Furthermore, Clarke’s claim that we cannot imagine the absence of space had already been called into question by Berkeley’s immaterialism and his claim that a being capable of vision but not touch could not discover the first elements of geometry (NTV 153–5). Hume’s own claim that “an object may exist, and yet be no where” (T 1.4.5.10) opens the door to conceiving a universe consisting of just smells, tastes, sounds, and passions, like the universe of Condillac’s statue in the early stages of its cognitive development.51 This al‑ ready scuttles Clarke’s claim that we cannot conceive the absence of space. With these sorts of considerations at his disposal, the author of the Treatise had no need to reject the existence of “unfilled” visible or tangible points to undermine Clarke’s argument for the existence of God. His tenet that anything can be the cause of anything (T 1.4.5.29–32) and his parity argu‑ ment against the antagonists of the free‑thinkers (T 1.4.5.16) also made it unnecessary for him to reject the infinite divisibility of space and time to undermine Clarke’s side in the debate with Collins. Might Hume have taken the line he did on infinite divisibility and on vacuum to establish one of his other conclusions? Only one is apparent: that a rejection of the possible existence of vacua serves as a premise in the “Berkeleian” sceptical argument of Enquiry 12.15 and Treatise 1.4.4.6–14.52 Even here there is no need to reject vacua. Having fol‑ lowed Berkeley in arguing that extension cannot be conceived in general but only as the extension of some figure, Hume only needed to observe that figures cannot be conceived without conceiving their edges and that edges cannot be conceived without qualitative contrast. Qualitative con‑ trast only requires one quality, establishing a contrast between locations where that quality is disposed and locations where it is not disposed. The “Berkeleian” argument does not require a commitment to the inconceiv‑ ability of vacua.

104  Finite divisibility; manners of disposition; points It seems that Hume did not need to reject the infinite divisibility of space and time or the conceivability of vacua for any of his other purposes. He may have only taken up these subjects to show how the science of human nature might contribute to the sciences of mathematics and natural phi‑ losophy. In that case, his encounter with Stanhope may have convinced him to abandon the project. 2.8

Hume’s achievements

On the way to arguing for the claims that space and time are not infinitely divisible, that a vacuum is inconceivable, and that unchanging objects do not endure Hume sought to prove that our ideas of space and time are finitely divisible, which entails that our visual and tactile perceptions are extended. He further established that our ideas of space and time are copied from sensory intuition of an order in which simple components of complex, visual and tactile impressions are placed in relation to each other, and an order in which all of our perceptions pass into and out of existence. These are results that fly in the face of a widely held opinion that conscious experience is simple, that minds and their thoughts and sensations are not distributed over space, and that consciousness only occupies the present moment. While Hume did not trumpet these results in Treatise 1.2, he did draw attention to them at Treatise 1.4.5, and proposed to, but was even‑ tually discouraged (under duress) from reissuing them in “Immortality.” Despite its popularity, the opinion that the mind or consciousness is not spatially extended has been difficult to sustain when it comes to details. Among others is the detail concerning how an unextended being could ex‑ perience spatial properties and relations. Descartes, one of the champions of dualism, might be read as having held that we have an idea of “intelligible extension” understood as the object of geometry. We confuse sensations of colour, pain, heat, cold, and other sensible qualities, which are our own mental states and so nowhere in space, with locations in intelligible ex‑ tension, thereby coming to perceive a variety of spatially extended objects. However, Descartes did not explain what guides this operation. Granting that colours and tactile qualities are modifications of the state of a spiritual substance, and as such are nowhere in space, what is it about the experience of a disordered, simultaneous mass of colours and tactile qualia that leads us to consider one of those sensations to be disposed at any one location in a homogenous space? It is irresistible, when confronted with such a prob‑ lem, to talk about mind‑body union. Granting that minds are connected to bodies makes it possible to talk about external objects and light rays or other emissions that impinge on parts of physically extended sense organs, which in turn affect parts of the brain. An innate mechanism can be taken to be responsible for leading us to associate sensations with the locations

Section 2.8  105 of impingements on sense organs or brain parts. This leaves the question of how a spirit that lacks the advantage of being joined to a body can perceive spatially extended and located objects. Fortunately, that is not our problem. That is, it is not our problem unless we are what Berkeley took us to be. Berkeley initiated a revolution in the theory of vision by replacing ap‑ peals to innate mechanisms with appeals to learned association. In the case of visual depth perception, for example, we learn to associate purely qualitative features accompanying visual experience, such as degrees of brightness and distinctness of the visual image, or sensations in the muscles responsible for focusing an image on the retina or bringing the two eyes to converge on a distant object, with the distance we must walk or reach to come into contact with an object, and so learn to read these features as signs of outward distance. Berkeley’s account of visual depth takes for granted that we experience visual images that are extended over two remaining spatial dimensions and that we experience tangibly measured distances in the remaining di‑ mension. However, Berkeley considered these spatial resources to be liable to further reduction. He intended his account of visual depth perception to be only a first step in the project of establishing that all that exists are spirits and their ideas. It would not have fit his project to argue at length that there is no external material world only to look within and discover that his visual sensations are disposed over two spatial dimensions and his tactile sensations over three, and so are in that sense corporeal and material. He aimed to establish the stronger thesis that we only see light and colours (NTV 156), and only feel tactile sensations that are originally given as nowhere in space. Hume had no such commitments. Taken in combination, Treatise 1.2.1, 1.2.3.2–5, 1.2.3.10, 1.2.4.5–6 and 19–22, 1.3.1.1, 1.3.2.2, and 1.4.5.15 provide what might be called an intuitionist account of spatial experience (Section 2.5). According to this account, visual and tactile impressions are divisible into simple parts that are disposed outside of one another over space, prior to any association or combination, and prior to any judgment. Spatial experience is not the product of acts of imagination or understand‑ ing. Simple visual and tactile impressions are already given as disposed in space in our original and most primitive sensory experience. Falkenstein, 1990, has argued that Hume was not alone in taking this approach. It is also found in Kant’s account of space and time as forms of intuition. Falkenstein (1994, 2000, 2005) and Grush (2007) have also ar‑ gued that Hume’s and Kant’s intuitionist account was superior to the rival empirist, nativist, and sensationist accounts offered by Berkeley, Reid, and Condillac. It is not unusual for a thinker who develops an original line of thought to fail to be fully aware of what they have achieved. But Hume did see

106  Finite divisibility; manners of disposition; points some implications. It follows from his account that things that are para‑ digmatically mental, such as colour qualia, are located in space. Visual and tactile sensory consciousness is itself distributed over space. Over Treatise 1.4.5, he trumpeted this implication as an objection to a “remarkable” argument for the immateriality of the soul. Hume considered anything extended to be divisible (T 1.2.4.3), and he appealed to the tenet that visual and tactile impressions are extended to respond to the principal argument his predecessors had employed to dem‑ onstrate the necessary unity of a thinking being (T 1.4.5.7 and 16). In mounting that argument, he was mounting an argument against the incor‑ ruptibility and immortality of the soul that cut equally against the Car‑ tesians, against their British opponents, and against Bayle (Dictionnaire, 129–34). This is a direct and devastating anti‑religious argument that takes on all comers. It has no investment in the inconceivability of vacuum or infinite divisibility of space and time. Hume was not perfect. He could overlook things and get things wrong. In the case at hand, he still made a significant contribution to the theory of perception and the theory of mind. Notes 1 This historical fact poses a challenge to De Pierris’s contention that Hume in‑ sightfully showed that a consistent version of empiricism implies that there is no exact standard of equality or congruence for the continuous magnitudes dealt with in geometry (2015, 110, 129–30, with detailed discussion over 119–22). Larvor (2023) argues, chiefly on the basis of Hume’s appeal to the (by then ex‑ ploded) “contradiction and absurdity” (EHU 12.18) in the notion of an angle of contact between a circle and its tangent that Hume was not expert in math‑ ematics and gained his knowledge of the mathematical topics he took up from a tradition of non‑expert, popular, and outdated works that drew on a handful of mathematical puzzles or purported puzzles to make points in religion and philosophy. The existence of that literature made the young Hume feel embold‑ ened to attempt a contribution of his own. Larvor’s position is only slightly challenged by the fact that the 15‑year‑old Hume made a copy of George Campbell’s treatise on fluxions (Kawashima 2004). For more on the ambiguity of this evidence, see Stewart (1990, 8–9) and Schabas and Wennerlind (2020, 59, citing a disparaging remark about conic sections at Treatise 2.3.10.4). The biographical incident Stewart mentioned is gathered from Hume’s letter of 12 November 1739 to George Carre of Nisbet (L1, 35). 2 George (2006) argues that James Jurin’s visual theory may have led him to lose confidence in elements of his account of space and time even prior to this date. 3 For more on Stanhope, see Bellhouse (2007). The description of Stanhope as an “amateur mathematician” is Bellhouse’s and is not used with pejorative intent. Stanhope wrote extensively on technical topics in the theory of probability and corresponded with other mathematicians, but he never published any of his work and declined invitations to do so.

Notes  107 4 It also indicates that the Treatise may well have contained “obvious errors” (compare Holden 2004, 30–1) that the mature Hume was no longer prepared to endorse under his own name. 5 See Chapter 1, note 44, on the distinction between empirism and empiricism. 6 But see Cottrell (2019, 88). 7 Reading, with T2, 707, “quantity” for “quality.” 8 Flew (1976, 260) finds (Remark) objectionable. Among others, Baxter (2009, 107) and Garrett (2015, 62) disagree. But worries about (Remark) can be set aside. It plays no role in establishing (FD). It bears noting that it is one thing to worry whether space or extension has potential parts. It is another to worry whether ideas have potential parts. Hume’s question at this juncture concerns the divisibility of ideas. 9 Large numbers, for their part, are only thinkable by being gathered by 10s or other exponents, and even then only for a finite number of digits (T 1.1.7.12). 10 Thomas (2023) has discovered that one of Hume’s later contemporaries, Wat‑ son, performed such experiments. 11 As this passage makes clear, George does not just mean to discuss the angular magnitude of the objects that cause visual impressions. He means to discuss the angular magnitude of our simple visual impressions (“images”). Berkeley wrote that we do not see optic lines and angles (NTV 12) and Hume belatedly agreed (Ax 22). But we do not need to see optic lines and angles to see angular magni‑ tude. We see what subtends the angle, rather than the lines that form it, and we can employ various means to measure the angle subtended by what we see. Ask what we would see if our visual fields were wider. It is natural to reply that we would see more of what lies off to the right and the left. But that is not correct. We already see as much of what lies to the left as we can under the physical con‑ ditions before us. Just as we can see galaxies arbitrarily far in front of us, if they are bright enough, we can already see them arbitrarily far off to the left. If our visual fields were wider, we would not see any more of what lies off to the left. We would see more of what lies around to the side. In the extreme, we would see what lies directly behind us. The visual fields of rabbits, shrimp, and other creatures do not extend laterally. They extend radially, like the inner surface of a sphere rather than like a plane. Our visual fields are no different. We only see a portion of the inner surface of the visual (or celestial) sphere in which we are cen‑ tred. That portion has an angular magnitude. We may not know precisely what that magnitude is, but that does not mean we do not see it as having any magni‑ tude. We can judge whether we see roughly half of everything we see when we rotate in place, and so take the left‑right angular magnitude of our visual fields to approximate 180 degrees. We can then take the distance from the midpoint of the visible equator to the left side to be approximately 90 degrees, the midpoint of that distance to be approximately 45, and so on. The means we employ for measuring angular magnitude are means for dividing great circles of a sphere on which we are centered into equal intervals. A sextant does not in first instance measure the angular separation of distant objects. It measures the angular sepa‑ ration of their images on the visual field. When Hume wrote about measuring magnitudes, he did not just appeal to determining equality of numbers of parts by juxtaposition. He referred to determining the proportions of bodies “at one view … without examining or comparing the number of their minute parts” by reference to “the whole united appearance and comparison of particular objects” (T 1.2.4.22). This is what we see when we see their angular magnitude.

108  Finite divisibility; manners of disposition; points 12 Fogelin (1985, 33) also remarks on the oddity of Hume’s metaphor, though to a different end. 13 George argues that a late reading of Jurin’s essay may have prompted Hume to abandon the thesis that whatever is distinct is separable. That thesis does not reappear in the Enquiry. The Enquiry justifies conclusions the Treatise drew from the separability principle by appeal to the thesis of meaning empirism (EHU 2.9; 7.4). Garrett (2015, 46) has claimed that the separability principle has a strong claim to be considered the second principle (after the copy prin‑ ciple) in Hume’s science of human nature. This is true of 1739, but already in Abstract 7 Hume was shifting to invoking meaning empirism to do the work he had earlier done with the separability principle. 14 De Pierris (2015, 118) offers a recent endorsement of the null thesis. 15 If it had only been Hume’s project to talk about space as it appears to the senses, he would not have written Treatise 1.2.2, which proposes to do more than that, both in its title and in the course of its arguments (compare Baxter (2009, 118–21). The fact of the finite divisibility of our impressions and ideas is already established in Treatise 1.2.1 without the need for any further argu‑ ments. Hume did write in a late note, directed in the Appendix to be appended to Treatise 1.2.5.26 that “As long as we confine our speculations to the appear‑ ances of objects to our senses, without entering into disquisitions concerning their real nature and operations, we are safe from all difficulties, and can never be embarrass’d by any question.” But that is just a comment on the advantages of that policy. It does not mean he always did or always meant to confine him‑ self accordingly. The note is appended to what he had said about vacuum, not infinite divisibility. The same holds for his comment that “at present I content myself with knowing perfectly the manner in which objects affect my senses” (T 1.2.5.26, my stress). 16 My stress, explaining the earlier interpolation of the question mark. 17 Hume’s argument has classically been represented differently. Holden (2004, 24) gives the classic version: (H1) “[W]hatever is capable of being divided in infinitum must consist of an infinite number of parts”; “Every thing capable of being infinitely divided contains an infinite number of parts” (T 1.2.1.2, 1.2.2.2). (H2) “[T]he idea of an infinite number of parts is individually the same with that of infinite extension; … no finite extension is capable of containing an infinite number of parts” (T 1.2.2.2). Therefore: (H3) “[N]o finite extension is infinitely divisible” (T 1.2.2.2). On the opposed, (EA)‑based interpretation offered in the text, (H1) is a throw‑away remark and (H2) is not a premise. (H2) is the conclusion and (H3) is a further conclusion from (H2). The status of (H2) as primary and (H3) as corollary conclusions is clearly indicated by the words that precede each: “Upon the whole, I conclude” and “and consequently.” As recognized by Baxter (1988), the argument for (H2) and so for (H3) is given where one would expect to find it, over the immediately prior portions of (EA). It is supported by (A‑prem) and (A‑prin) and turns on appeal to a brute fact concerning what we “find” when we set one simple and indivisible part as close as possible to another. The classic version neglects this argument. Holden, who has his own qualms about the classic version and also recognizes (H2) as a conclusion, bypasses (EA). He instead takes (H2) to be justified by an “ontological regress argument to simple first parts” (2004, 28). Holden is right that Hume did offer

Notes  109 such an argument, and right to draw our attention to its importance, not only for Hume but also for a host of early modern philosophers. But, as he himself notes, the argument only comes up at Treatise 1.2.2.3, “where it is clearly in‑ tended to support (H2)” (2004, 28). It is not part of what Holden describes as the “lead argument” of Treatise 1.2.2.2 (2004, 23–4). Rather than be a regress argument to simple first parts, the lead argument is an augmentation argument from an observed fact about simple first parts. The adequacy of the regress argument is considered below. 18 Ideas are implicated in this result since they are copies of visual (or tactile) impressions. 19 See also Wilson (2008, 272–83, esp. 282–3). Baxter (1988) appeals to a related thesis, focused on rejecting the possibility that minimally sensible points can be continually aggregated without surpassing any finite bounds. Baxter argues that this related thesis is supported by phenomenological considerations like those recounted in (D), but he hesitates to infer that Hume endorsed it. Al‑ lison (2008, 40–1) comes close to recognizing that the manner of disposition is discrete but then veers away from it, declaring that Hume must “fall back on sheer aggregation” (41–2). Allison draws instead on Broad’s (1961, 169) no‑ tion of an intrinsic minimum distance. In a discrete space, there is no intrinsic minimum distance. See note 24. 20 Hume was clear that the thesis holds for “homogeneous” as well as differently qualified points. Over the ellipses in (D), he asked the reader to consider differ‑ ently coloured points merely as an aid to conceiving the truth of a proposition that holds for points in general. “Let [us] aid [our] fancy by conceiving these points to be of different colours, the better to prevent their coalition and con‑ fusion.” The phrase, “coalition and confusion” refers to circumstances, like identity of colour, that make the distinction between discretely disposed points harder to notice. It does not (absurdly) refer to their being brought any closer together (see note 24). If “coalition and confusion” entailed any overlap, the extension would evaporate, as (on the null thesis) points can only overlap by complete coincidence. 21 These considerations justify and may have motivated Hume’s rejection of the possibility of a finite space being composed of infinitely many proportionally smaller parts (T 1.2.2.2n). 22 A distinction between discrete and continuous quantities was recognized in Hume’s day and earlier (Bayle, Dictionnaire, 356; Chambers, Cyclopedia art. Quantity). Hobbes wrote that two spaces are said to be contiguous when there is no other space between them, but then went on to add that they are only said to be continual when they share a part in common (Elements of Philosophy 2.7.10). Compare Jesseph (1993, 89). 23 It might be objected that referring to discreteness as a feature of the space in which points are disposed reifies space, by treating it as a substance that bears qualities or a “thing” that imposes constraints on the manner in which points are disposed. Those who have this concern might consider the constraint to arise from the finite powers of resolution of the senses and the finite capacities of the mind. But that comes at the cost of reverting to the theses of Treatise 1.2.1, taking Hume’s argument to only establish the finite divisibility of our im‑ pressions and ideas, and abandoning the stronger thesis of the title of Treatise 1.2.2. Another alternative is to take discreteness to be a brute fact about the way in which points are found to be related. But when combined with Hume’s doctrine that relations of contiguity and distance in space are not based on

110  Finite divisibility; manners of disposition; points anything in the things in space (T 1.3.1.1), it is not clear how this differs from considering discreteness to be a property of the space in which those things are disposed. It may be that the only way to account for Hume’s position on the finite divisibility of space itself, consistently with the null thesis, is to accept that he meant to argue that discreteness is a feature of space itself, at whatever cost to other things he said. 24 It is sometimes objected (recently by De Pierris [2015, 177, 177n176]) that a discrete visual space would look like it is full of holes or scattered drops of col‑ our, like a “pointillist” painting. But this is question‑begging. If a space really is only finitely divisible, there is not always a further location between any two locations, and so no gaps or holes between immediately adjacent locations. For further details, see Falkenstein (2006, 62–4). 25 While a single point can be identified as such in the midst of contrasting sur‑ roundings, it may be impossible to say whether a single point is aligned with only one other point when the points do not contrast. 26 Plausibly, the finding only holds for small numbers. An entire visual field com‑ posed of alternating dark and white points would contain too much informa‑ tion to be exactly enumerated at a single glance even though the points are all distinctly perceived. 27 As Baxter puts it, “All [Hume’s] conclusions about space and time should be read as claims about space and time as they appear. If one is going disagree with Hume on infinite divisibility, then one is going to have to take issue with his Pyrrhonian Empiricism” (2009, 121). Baxter offers this comment in defence of Hume’s position on space and time. But on the supposition that Hume meant to draw a conclusion about space and time (not just about our impressions and ideas), it is a criticism. The Enquiry appears to take a stronger line, claiming that “nothing appears more certain to reason, than that an infinite number of [‘parts of extension, which cannot be divided or lessened, either by the eye or imagination’] composes an infinite extension” (EHU 12.18n, my stress). But Hume did not say enough to justify this assertion. If the parts in question have some angular magnitude, it would indeed be certain to reason that an infinite number of them must compose an infinite extension. But then it would not be certain to reason that all parts must have some angular magnitude. If the parts have no extension, Hume was negligent in not supplying the reasoning that makes the consequence evident. If he considered it to be introspectively evident that images with no angular magnitude can nonetheless be only discretely dis‑ posed, he would have been appealing to a “finding.” 28 According to another of Hume’s sources, Barrow, the only units that are per‑ fectly indivisible are geometrical points. Any number is divisible (1734, 48– 9/1685, 54–5). For more on Malézieu, see Larvor (2023, 124–5, 126). 29 The interpretation offered here follows Baxter (2008, 28–9). 30 Bayle did not draw this conclusion. Instead, without offering any further justifi‑ cation, he drew the (invalid) conclusion that time cannot be infinitely divisible. Hume clearly wanted to draw Bayle’s conclusion as well, but Treatise 1.2.2.4 does not explicitly draw it. It only draws the conclusion that time must consist of indivisible parts. 31 Holden (2004, esp.  16–17, 28, 35) observes that Hume and many others of the time were committed to the doctrine that any parts into which a body is logically separable are distinct existents that are already embedded in the archi‑ tecture of the whole, making the whole an aggregate of independently existing parts. He also observes that this doctrine was widely and plausibly taken to

Notes  111 entail that composites depend on their parts, which, if they are composite, in turn depend on their own parts, so that “if the whole original is to exist at all, this ontological regress cannot go on forever with no ground floor. It must ter‑ minate in noncomposite elements” (35). But, as Holden recognizes (earlier on 35), granting that the whole is composed of noncomposite elements does not entail that the “regress” to those elements cannot go on forever. 32 Franklin (1994, 87), citing Grünbaum (1952) and Trudeau (1987, 23–30); Bell (2019, 105–46). Franklin’s informative paper is marred by misquotation. He reads “ideas” where Treatise 1.2.2.1 writes “these ideas” (91). Largely on the strength of this misquotation, Franklin accuses Hume of being an idealist and takes Hume’s chief mistake to be his reliance on (A‑Prin). This does not affect his conclusions concerning the contingency of (EA). 33 T 1.2.3.2. Compare Locke, Essay 2.13.2: “Men perceive, by their Sight, a dis‑ tance between Bodies of different Colours, or between the parts of the same Body.” 34 See Costa (1990, 4); Garrett (1997, 52–4, 168–9); Falkenstein (1997), and the earlier work referenced by them. 35 There is no infinite regress. Provided the surrounding points are identically qualified, they are not edged off from one another or from more remote points but blend into an extended area, as Slowik and De Pierris have rightly observed (Section 2.4.1). The area is edged off on its “inside” from the contrasting point it surrounds but, in the absence of yet more remote differently qualified points, is otherwise indeterminately extended. Infinite extension is not an option. As discussed in note 11, the human visual field corresponds to a portion of the inner surface of a sphere, not an unbounded plane. The tactile field is also “closed”; it does not extend beyond the closed surface of the body. Conjectur‑ ally, when we choose to extend our attention so far, we imagine the visual field to be embedded in a larger image field that completes the visual sphere. When we cannot populate this surrounding image field with imagined landmarks (when it is conceived as just a “dark” area) we feel lost. 36 What matters is that the point contrasts with the surrounding points. Sur‑ rounding points that are not “filled” (T 1.2.4.2) with any quality make a more effective contrast than those filled with a closely resembling quality (T 1.2.4.6 and 19). See Section 2.6 and Chapter 5. 37 Kemp Smith (1941, 273–6) recognizes this wholistic aspect of Hume’s thought about space and time but considers it to be in tension with Hume’s commit‑ ment to what he calls a composition theory (279). Wilson (2008, 259–63) sees a tension between recognizing relations of contiguity like a is next to b and Hume’s thesis that everything that is separable is distinct (T 1.1.7.3). Separat‑ ing b from a seems to change a as well, just as separating the “kickee,” Paul, from the kicker, Peter, changes Peter, who is no longer a kicker in separation from Paul. Wilson’s solution is to propose that relations can be reduced to non‑relational properties of the relata. Intriguingly, in the case of contiguity, the non‑relational property Wilson fixes on is a place or location. This has con‑ sequences for the validity of Hume’s argument over Treatise 1.2.3.12–15. If the parts have distinct locations, there is no need to invoke yet further properties to make them “real.” Wilson would likely not have considered this to be an objection, even to Hume (283–305). 38 This observation might call for a comment on how Hume’s position relates to Kant’s. Such comparisons are difficult to draw. They require first determin‑ ing what Kant meant to say, a task that is often impossible to execute to the

112  Finite divisibility; manners of disposition; points satisfaction of many and that generally requires more work than would be jus‑ tified by the potential benefits to those not engaged in dedicated Kant exegesis. In the case at hand, any comparison depends on how Kant’s use of “intuition” is understood: as referring to what is immediately apprehended prior to any operations of the understanding (Falkenstein 1995, among others), or as re‑ ferring to singular representations (Hintikka 1969, among others). If Kantian intuitions are singular representations, then there is no comparison to be made between Kant’s forms of intuition and Hume’s manners of disposition. If they are immediate representations, then Hume and Kant had variants on the same idea. For Falkenstein (1990), the Kant of the first Critique sought to ground forms of intuition on the manner in which the subject is constituted so as to be susceptible of being affected by objects (as a computer is constituted to receive and register bits in a certain order). This gives the forms the status of condi‑ tions of the possibility of experience. Hume did not ground his intuitionism on any such speculations. For Hume, it has up to now been brute factually the case that visual and tactile perceptions are simultaneously disposed over two or three spatial dimensions and that all perceptions come into being and pass away over time. Kant’s account of how the “manifold” of intuition is brought to a “unity of apperception” under “concepts of objects” also has little affinity with Hume’s tendency to take the recognition of objects for granted. Insofar as elements of a theory of object recognition can be found in Hume, they rest on instinctive operations of the imagination, not quasi‑logical conditions of the possibility of experience. This issue is further complicated by the fact that there is no agreement among Kant scholars over whether Kant thought that spatial and temporal relations are intuited prior to any operations of the understand‑ ing (Falkenstein 1995, Warren 1998), or are themselves products of higher order thought (among others Waxman 1991, Longuenesse 1998). 39 Pardies (1734, 2); Keill (1720, 23). Having scrupulously translated Elements 1 Definition 1 as “A Point is That of which there is no part,” Scarburgh (Euclid 1705, 1) glossed the definition as meaning “void of Magnitude.” He went on to say that “These kinds of Negative Definitions are sufficiently instructive; tho’ not to the Essence of the Thing defined,” and went on to fill two folio pages with further, rather astute comments. 40 Treatise 1.2.4.9–15 recognizes four of the positions that were taken, but misses the one that was most frequently adopted by mathematicians and editors of Euclid: Marks covering infinitely divisible regions of space are thrown down on paper and designated as points or lines for practical purposes. Andrew Tacquet used a bracketed insertion in his translation of Elements 1 Definition 1: “A Point is a Mark in Magnitude, which is [supposed to be] indivisible” (Euclid 1714, 1). The unnamed translator of Claude‑François Milliet Dechales French translation of Euclid conveys Dechales’ more literal translation, “A Point is that which hath no Parts” with Deschales’ following comment “This Definition must be understood in this Sense: That Quantity, which we conceive without distinguishing its Parts, or so much as considering whether or not it has any, is a Mathematical Point; which is therefore very different from those of Zeno, which were suppos’d to be absolutely indivisible, and therefore such, that we may reasonably doubt whether they are possible; but the former we cannot doubt of, if we conceive them aright” (Euclid 1726d, 2). In the same year, Henry Hill inserted his gloss right into his translation, making no effort to draw a distinction. “A Geometrical Point, is that which has no Parts, or at least is considered as such, and which of Consequence is indivisible” (Euclid 1726h, 1).

Notes  113 Keill, in contrast, maintained that “tho’ neither Superfices, nor Lines, nor Points are real Matter; yet they exist or may exist in it, as its Modes, Termina‑ tions, or Accidents” (Keill 1720, 23–4). Perhaps the most likely edition of Eu‑ clid that the young Hume could have accessed is that found in David Gregory’s Greek/Latin edition of Euclid’s extant Omnia (Euclid 1703), which was in‑ cluded in Robert Steuart’s Physiological Library (T2, 1018). Hume subscribed to the library in December 1724 (T2, 980). Gregory translated the definition as: “Punctum est, cujus pars nulla est” without adding any further observations. For more on what Hume may have learned on this topic, see Barfoot (1990, 151–60 and 183–90) and Larvor (2023). 41 There is no vicious regress here for reasons given in note 35. 42 For further discussion, see Mendell (2019) Section 7.2. 43 This is not to say that it was appreciated by Hume. But it would not have been entirely easy for him to miss. He footnoted Isaac Barrow’s Mathematical Lec‑ tures (discussed immediately below) at Treatise 1.2.4.21, and Barrow took up this topic. 44 Barrow (1685, 53–6 and 1734, 47–9). The 1734 translation misprints Barrow’s reference to Posterior Analytics 1.27 as 1.24. Hume did not specify which edi‑ tion or translation he consulted. Presumably, he would not have used the trans‑ lation if he had access to an original language edition. 45 Presumably he meant that each unit (“more rightly said” to be a number, as he earlier put it) has its own position on the number line. 46 “[N]am punctum universaliter acceptum haud minus indeterminatum est & expers positionis, quam unitas itidem sumpta universaliter: Singulariter autem accepta unitas definitam aquè positionem, & reliquas singulares includit cir‑ cumstantias, ac ipsum singular punctum” (1685, 55). 47 Hume is often described as a relationist about space and time (Fogelin 1985, 33–5; Slowik 2004, 359; Russell 2008, 106; Baxter 2009, 131; Kervick 2016, 66). He was not, except in an attenuated sense. For Hume, nothing about the things in space and time determines where they are to be found in space or time. The “parts” that appear in space and time could appear anywhere in space or time (T 1.3.1.1). It is impossible, therefore, to discover spatial and temporal relations by inspecting and comparing ideas of the things that appear in space and time to discover some circumstance that might determine where they are placed, the way we can order paint chips over three dimensions of hue, saturation, and brightness simply by comparing them with one another. Instead, we can only discover spatial and temporal relations by “perception” (T 1.3.2.2), that is, by immediate or “intuitive” experience of how simple and indivisible impressions are given as disposed over the visual and tactile fields. The experience of how multiple, simple impressions are disposed takes the form of a complex, extended and enduring impression. The components of complex perceptions include spatial and temporal relations as well as relata. The relations are a further, originally given feature, not reducible to the relata. 48 Hume could have had sceptical consequences in mind when writing things ear‑ lier in the Treatise. But to rely on unstated and unjustified consequences when justifying earlier results would have been inappropriate. 49 It also says that the doctrine of infinite divisibility shocks “common sense” and is “too weighty for any pretended demonstration to support, because it shocks the clearest and most natural principles of human reason,” but this is done by way of setting up a true sceptical equipoise of invincible arguments for opposed conclusions.

114  Finite divisibility; manners of disposition; points 50 EHU, 269. “Appearances 56–77” in the 2000 printing of Beauchamp’s EHU is a misprint. As Beauchamp correctly goes on to observe, “~56–77,” that is, all editions after 1750 are like the lemma in omitting the entire passage. 51 The project of the Traité was to determine what we learn from the senses by considering a being who is completely impassive and immobile (a “statue”) and speculating what it would learn were it endowed with each of the senses indi‑ vidually. Condillac maintained that it would learn to develop desires and pas‑ sions and all our higher‑order cognitive abilities, thereby arguing that we have no innate cognitive capacities or abilities. He began with smell, taste, hearing, and vision and maintained that it would only be after acquiring touch and the ability to move that the being would recognize anything it experiences as other than itself. Even vision, though it would display extended and spatially ordered colours, would not make the statue aware of how colours are disposed in space until touch instructs it how to look at what it sees. 52 Baxter (2016, 182–3) has alluded to this connection, though he does not sug‑ gest that Hume’s rejection of vacua serves as a premise for the Berkeleian argument.

Bibliography Allison, Henry. 2008. Custom and Reason in Hume: A Kantian Reading of the First Book of the Treatise. Oxford: Oxford University Press. Barfoot, Michael. 1990. “Hume and the Culture of Science.” In Studies in the Philosophy of the Scottish Enlightenment, edited by M. A. Stewart, 151–90. Oxford: Clarendon Press. Baxter, Donald L. M. 1988. “Hume on Infinite Divisibility.” History of Philosophy Quarterly 5: 133–40. Baxter, Donald L. M. 2008. Hume’s Difficulty: Time and Identity in the Treatise. London: Routledge. Baxter, Donald L. M. 2009. “Hume’s Theory of Space and Time in its Skeptical Con‑ text.” In The Cambridge Companion to Hume, 2nd edition, edited by David Fate Norton and Jacqueline Taylor, 105–46. New York: Cambridge University Press. Baxter, Donald L. M. 2016. “Hume on Space and Time.” In The Oxford Handbook of Hume, edited by Paul Russell, 173–90. Oxford: Oxford University Press. Bell, John L. 2019. The Continuous, the Discrete and the Infinitesimal in Philosophy and Mathematics. Cham, Switzerland: Springer. Bellhouse, David.  2007. “Lord Stanhope’s papers on the Doctrine of Chances.” Historia Mathematica 34: 173–86. Broad, C. D. 1961. “Hume’s Doctrine of Space.” Proceedings of the British Academy 47: 161–76. Costa, Michael J. 1990. “Hume, Strict Identity, and Time’s Vacuum,” Hume Studies 16: 1–16. Cottrell, Jonathan. 2019. “Hume on Space and Time: A Limited Defense.” In The Humean Mind, edited by Angela Coventry and Alex Sager, 83–95. Oxon, Routledge. De Pierris, Graciela. 2015. Ideas, Evidence, and Method: Hume’s Skepticism and Naturalism Concerning Knowledge and Causation. Oxford: Oxford University Press.

Bibliography  115 Falkenstein, Lorne. 1990. “Was Kant a Nativist?” Journal of the History of Ideas 51: 573–97. Falkenstein, Lorne. 1994. “Intuition and Construction in Berkeley’s Account of Visual Space.” Journal of the History of Philosophy 32: 63–84. Falkenstein, Lorne. 1995. Kant’s Intuitionism: A Commentary on the Transcen‑ dental Aesthetic. Toronto: University of Toronto Press. Falkenstein, Lorne. 1997. “Hume on Manners of Disposition and the Ideas of Space and Time.” Archiv für Geschichte der Philosophie 79: 179–201. Falkenstein, Lorne. 2000. “Reid’s Account of Localization.” Philosophy and Phe‑ nomenological Research 61: 305–28. Falkenstein, Lorne. 2005. “Condillac’s Paradox.” Journal of the History of Phi‑ losophy 43: 403–35. Falkenstein, Lorne. 2006. “Space and Time.” In The Blackwell Guide to Hume’s Treatise, edited by Saul Traiger, 59–76. Malden: Blackwell. Falkenstein, Lorne. 2013. “Hume on the Idea of Vacuum.” Hume Studies 39: 131–68. Flew, Anthony. 1976. “Infinite Divisibility in Hume’s Treatise.” In Hume: A Re‑evaluation, edited by D. W. Livingston and J. T King, 257–69. New York: Fordham University Press. Fogelin, Robert J. 1985. Hume’s Skepticism in the Treatise of Human Nature. London: Routledge and Kegan Paul. Franklin, James. 1994. “Achievements and Fallacies in Hume’s Account of Infinite Divisibility.” Hume Studies 20: 85–101. Garrett, Don. 1997. Cognition and Commitment in Hume’s Philosophy. Oxford: Oxford University Press. Garrett, Don. 2015. Hume. New York: Routledge. George, Rolf. 2006. “James Jurin Awakens Hume from His Dogmatic Slumber: With a Short Tract on Visual Acuity.” Hume Studies 32: 141–66. Grünbaum, Adolf. 1952. “A Consistent Conception of the Continuum as an Ag‑ gregate of Unextended Elements.” Philosophy of Science 19: 288–306. Grush, Rick. 2007. “Berkeley and the Spatiality of Vision.” Journal of the History of Philosophy 45: 413–42. Hintikka, Jakko. 1969. “On Kant’s Notion of Intuition (Anschauung).” In Kant’s First Critique, edited by Terence. Penelhum and J. J. MacIntosh, 38–53. Bel‑ mont: Wadsworth. Holden, Thomas. 2004. The Architecture of Matter: Galileo to Kant. Oxford: Clarendon Press. Jesseph, Douglas M. 1993. Berkeley’s Philosophy of Mathematics. Chicago: The University of Chicago Press. Kawashima, Yukihiko. 2004. “Is a New Material of David Hume by himself? A Treatise of Fluxions, By Mr. George Campbell: Written by David Home, 1726.” Journal of Tokyo International University 70: 163–82. Kemp Smith, Norman. 1941. The Philosophy of David Hume. London: Macmillan. Kervick, Dan. 2016. “Hume’s Perceptual Relationism.” Hume Studies 42: 61–87. Larvor, Brendan. 2023. “The Limits of the Understanding and the Under‑ standing of Limits: David Hume’s Mathematical Sources.” Research in the History and Philosophy of Mathematics 2021: 115–31. https://doi. org/10.1007/978‑3‑031‑21494‑3_7

116  Finite divisibility; manners of disposition; points Longuenesse, Béatrice. 1998. Kant and the Capacity to Judge, translated by Charles Wolfe. Princeton: Princeton University Press. Mendell, Henry. 2019. “Aristotle and Mathematics.” The Stanford Encyclope‑ dia of Philosophy, Edward N. Zalta (ed), URL = . Russell, Paul. 2008. The Riddle of Hume’s Treatise: Skepticism, Naturalism, and Irreligion. New York: Oxford University Press. Schabas, Margaret and Carl Wennerlind. 2020. A Philosopher’s Economist. Hume and the Rise of Capitalism. Chicago: University of Chicago Press. Slowik, Edward. 2004. “Hume and the Perception of Spatial Magnitude.” Cana‑ dian Journal of Philosophy 34: 355–74. Stewart, M. A. 1990. “Introduction.” In Studies in the Philosophy of the Scottish Enlightenment, edited by M. A. Stewart, 1–9. Oxford: Clarendon Press. Thomas, Emily. 2018 Absolute Time: Rifts in Early Modern British Metaphysics. Oxford: Oxford University Press. Thomas, Emily. 2023. “The Specious Present in English Philosophy 1749–1785: Theories and Experiments in Hartley, Priestley, Tucker, and Watson.” Philoso‑ pher’s Imprint 23. https://doi.org/10.3998/phimp.1281 Trudeau, Richard J. 1987. The Non‑Euclidean Revolution. Boston: Birkhauser. Warren, Daniel. 1998. “Kant and the Apriority of Space.” Philosophical Review 107: 179–224. Waxman, Wayne. 1991. Kant’s Model of the Mind. New York: Oxford University Press. Wilson, Fred. 2008. The External World and Our Knowledge of It. Toronto: Uni‑ versity of Toronto Press.

3 Time and our experience of time

When discussing the finite divisibility of space and time (T 1.2.2), and when offering the argument from the non‑entity of unqualified points (T 1.2.3.12–17), Hume made a case for space and then claimed that the same considerations apply to time (T 1.2.2.4, 1.2.3.17). But when discuss‑ ing the origin of the idea of time he offered a dedicated treatment. His dis‑ cussion is divided into two parts, only identified as such mid‑way through (T 1.2.3.9). The first (T 1.2.3.7–8) is devoted to time “in its first appear‑ ance to the mind,” and the second (T 1.2.3.10) to time as it is “conceiv’d … in the imagination.” These descriptions notwithstanding, what Hume said in the second part reveals more about the origin and content of the idea of time than what he said in the first. Section 3.1 shows that his conclusions in the second part parallel those he had drawn about space, but also deal with special problems: All perceptions, ideas as well as impressions, and passions as well as sensations are disposed in time. And time is not just an order of disposition but an order of coming into being and passing away. Sections 3.2.1 and 3.2.2 consider what the latter fact entails for the nature of succession and temporal moments. Turning to Hume’s opening discus‑ sion of “time in its first appearance to the mind” and its consequences, Sections 3.3 and 3.4 argue that Hume had no good evidence for conclud‑ ing that there can be no experience of time where nothing changes or that unchanging objects cannot endure. The chapter concludes by engaging Donald Baxter’s subtle and compelling case for a contrary assessment. 3.1

Time as it is conceived in the imagination

Hume’s study of time as it is conceived in the imagination opens with an appeal to the inverse separability principle (ISP) of Treatise 1.1.7.3. In order to know whether any objects, which are join’d in impression, be separable in idea, we need only consider, if they be different from

DOI: 10.4324/9781032677880-4

118  Time and our experience of time each other; in which case, ’tis plain they may be conceiv’d apart. [SP:] Every thing, that is different, is distinguishable; and every thing, that is distinguishable, may be separated.… [ISP:] If on the contrary they be not different, they are not distinguishable; and if they be not distinguish‑ able, they cannot be separated. (T 1.2.3.10) This is the first appearance of (SP) and (ISP) in Treatise 1.2. The earlier discus‑ sion of space (T 1.2.3.2–5) only invokes the empirist1 principle that “every idea … first makes its appearance in a correspondent impression” (T 1.2.3.1), followed by a study of the sort of impression that supplies that idea. But perceptions of every kind can be successive (T 1.2.3.6). In principle, an idea of time could be first obtained from an impression followed by a memory of that impression. Hume accommodated this possibility by appealing to (ISP): no difference therefore no separable ideas. The difference he was concerned to deny is that between the idea of time and a succession of impressions or ideas. To show that there is no difference, Hume observed that five notes played on a flute suffice to give us the “impression and idea of time.”2 In this case there is no different, sixth impression that “presents itself to … the senses” (obviously); no impression of reflection that arises in reaction to the five notes (because they need not arouse any passion or affection); and no “new original idea” that the imagination can extract from the five sound impres‑ sions (because all ideas are copied from impressions). Instead, “here [the mind] only takes notice of the manner, in which the different sounds make their appearance.” Hume had earlier said the same thing about the idea of space and the purple points constitutive of a tabletop (T 1.2.3.4 at the end). But time is not quite like space. [T]ime or succession, tho’ it consists [like space] of parts, never presents to us more than one at once; nor is it possible for any two of them ever to be co‑existent. (T 2.3.7.5) The observation that no more than one part of time is ever presented to us at once would be trivial if read as “no more than one part of time is ever presented to us at one part of time.”3 Hume might have had a purpose for making a trivial point. But at Treatise 2.3.7.5, he wanted to explain why “the consequence[s] of a removal in space are much inferior to those of a removal in time” (T 2.3.7.4 and 2.3.8.3) and why the temporal order is future‑directed (T 2.3.7.6–9 and 2.3.8.10–12). His explanations appeal to temporal passage and the direction of temporal passage. They assume that moments come into being and pass away in succession. In that context,

Section 3.1  119 “never presents to us more than one at once” (T 2.3.7.5) means, non‑­ trivially, that no more than one ever exists. The flute note example may have been deliberately designed to recognize this feature. Elsewhere, Hume remarked that in running over the notes of a wind instrument the sound is immediately lost after the breath ceases (T 2.3.9.12, repeated at DP 1.10). This might be taken to mean that the idea of time is necessarily derived from “impressions and ideas” or only ideas, because the past parts of a succession, being no longer existent, can only be discerned by inspecting and comparing currently existing ideas, perhaps to rank them in terms of some quality they display, such as their liveliness. But then time would be an order of coexistent degrees of quality, not of successive appearance. It would also be a mystery what would lead us to consider less vivacious ideas to be memories of past impressions or to think that their degree of vivacity reflects the temporal proximity or distance of their originals. It is even a mystery what would give us the idea that there “has been” any such thing as a past. Even the ability to understand what “has been” means would be unaccountable. Hume would have none of this. [N]or can the mind, by revolving over a thousand times all its ideas of sensation, ever extract from them any new original idea[.] We cannot find anything in our ideas of the five flute notes that tells us how they are (or their originals were) disposed relative to one another in time. If we did not directly perceive the successive notes as existing in succession, our situation would be hopeless. But here it only takes notice of the manner, in which the different sounds make their appearance[.] This raises a problem. For Hume, we are somehow able to “take notice,” of the manner in which the sounds appear. But in the case of time, that manner is existing in succession. Earlier sounds no longer exist when later ones occur, yet we somehow manage to not only take notice of the earlier ones but take notice of them as existing earlier. How can that notice of them as existing earlier occur only later? Hume declared that we do not, while hearing one note, simultaneously “retain” the prior note. Every part [of time] must appear single and alone, nor can regularly have entrance into the fancy without banishing what is suppos’d to have been immediately precedent. (T 2.3.7.5)

120  Time and our experience of time For what it is worth, when I try to recall the first four notes of Beethoven’s fifth symphony, the best I can manage is to have an idea that takes time to occur, over which time I imagine each note in succession. The closest I can come to imagining them all at once is imagining the sound the orchestra makes while tuning up, which is not the same thing.4 Granting that retention has no role to play there is little choice but to recognize that taking notice of the earlier sounds as existing earlier re‑ quires that consciousness extend backward from the present moment to incorporate what exists in the past. But if only one part of time ever exists, there are no past notes and no past consciousness to be incorporated with the present. Hume seems never to have noticed this tension. At this point in his writ‑ ing, the fact that we are able to take notice of what existed earlier as exist‑ ing earlier was uppermost in his mind. From that, he drew the conclusion that we experience the notes as appearing over time. Our experience takes the form of a temporally extended, complex impression. The conception of time “can plainly be nothing but different ideas, or impressions, or objects dispos’d in a certain manner, that is, succeeding each other” (T 1.2.3.10). The conception itself consists of successive parts and so takes time to oc‑ cur. We are successively conscious of the flute notes and (consciousness being sticky) thereby conscious of them as successive. Today it is easier to say that the past is only earlier than the present, not nonexistent, and that a unified consciousness can extend into the very recent past. A tension in Hume’s commitments has been resolved at the expense of one of them. The resolution is ours, not his. He could only proceed by closing an eye on one of his commitments. He had every right to do so. It is empirically evident that in witnessing the flight of birds or the sweep of a second hand over the face of a clock we are successively conscious of successive events. While it would be nice to be able to explain that fact consistently with then‑dominant presentist commitments, Hume could reflect that no one else had managed to do so, and he had no need to do so for his purposes. He could rely on the fact that we do have that consciousness and rest assured that, given that it occurs, it must be possible somehow. Hume qualified his conclusion by noting that we may afterward con‑ sider the time taken up by the successive flute notes “without considering these particular sounds.” But then we must put some other sense impres‑ sions, ideas, or impressions of reflection in their place. The idea of time is accordingly not any “primary distinct impression” but “can plainly be nothing but different ideas, or impressions, or objects dispos’d in a certain manner, that is, succeeding each other.” Recent scholarship reads this argument positively, as establishing what our idea of time is an idea of, with the answer being that it is an idea

Section 3.2.1  121 consisting of some succession of particulars.5 Hume did say as much. As discussed in Section 2.8, in doing so, he proposed an original and plausible “intuitionist” (not his term) account of temporal experience. But in his mind the principal conclusion is negative. As indicated by Treatise 1.2.3.9, by his appeal to (ISP), and by the last sentence of Treatise 1.2.3.10, he wanted to establish that we do not have a separate or distinct idea of time. There can be no manner of appearance unless there are some elements appearing in that manner. The thesis that time is a manner of ap‑ pearance was uncovered on the way to justifying this negative conclusion. An objection discussed in Section 2.5  merits review for the temporal context. Any one flute note is separable from the others, in the sense that it would be possible to hear one of them while plugging one’s ears during the rest. For a sufficiently short‑lived being, a single flute note is even separable from any succession of perceptions whatsoever. Being separable from one another, the flute notes are separable from their manner of appearance. It might seem like it should also go the other way. However, a being that only lives long enough to have a single, simple perception is in no position to draw any conclusions about the separability of its perceptions. A being that stops its ears still experiences something, if only silences (Skrzypulec 2022), before and after the flute note, and so experiences the flute note as disposed in some manner in time. No flute note is separable from some manner of disposition. Correspondingly, no manner of disposition is separable from some perceptions. 3.2

Questions about time in its first appearance to the mind

3.2.1 What constitutes a succession?

Five notes played on a flute suffice to give us an impression and idea of time. Hume would not have maintained that they are necessary to do so. Given that the idea of time is “deriv’d from the succession of our perceptions of every kind” (T 1.2.3.6), it is not necessary that what appears be sounds. Obviously, from this case, it is not necessary that what appears be a public object as opposed to something private, such as an earworm. Is it necessary that the things that appear be different? If so, different in what way? Must they be qualitatively different, or is it only necessary that they be numeri‑ cally different? Might a single thing suffice to give us an idea of time? Not if it is instantaneous, that is, temporally simple and indivisible, on the assump‑ tion that time is a form of extension and that whatever is extended is divis‑ ible into parts. But what if it endures? Would playing five identical notes staccato suffice to make them different? Might five identical notes be played legato and still be different? Might a single note held for five beats suffice to give us an impression and idea of time? Might one blast of a foghorn?

122  Time and our experience of time H. H. Price at one point wrote,6 Consider any entity which remains absolutely unchanged throughout a finite period of time—“any unchangeable object” as Hume rather oddly calls it. We say “it is the same as it was two minutes ago”. Now strictly speaking this is not true. Indeed, it is not even sense. For since our object has not changed at all, we cannot distinguish any multiplicity of succes‑ sive stages of [or?] phases within it. Thus there are no distinguishable terms between which the relation of “being the same as” could hold (for, as we have seen, it is a relation, and requires two terms at least). In fact the idea of time does not strictly apply to this unchanging entity at all. Where there is time, there must be succession; and in this entity, ex hypothesi, there is none. (1940, 39–40) Price made two questionable assertions. One is that a relation requires two terms. Not all relations are irreflexive, as when the barber shaves himself. It needs to be explained why succession is not among them, as when an officer succeeds themselves in a second term in office. The second is that we cannot distinguish any multiplicity of successive “stages” within an unchanging object. This is not so obvious that it can be asserted without further argument. We can distinguish a multiplicity of adjacent portions within a uniformly coloured tabletop. Hume claimed that “Two objects, tho’ perfectly resembling each other, and even appearing in the same place at different times, may be numeri‑ cally different” (T 1.3.1.1). He also maintained that they may be identified as earlier and later instances or appearances of the same individual.7 There is a difference in their temporal manner of disposition that could go either way.8 That difference is more primitive than either identity or individua‑ tion and can be recognized prior to making any judgement about the iden‑ tity or individuation of the temporally different instances. Indeed, noticing the difference is a prerequisite for making any such judgement. The judge‑ ment is between two contraries (identification of successive items with one another or with the same whole, individuation of successive items). It is only called for when the fundamental difference (in time of appearance) that creates the contrary alternatives is noticed.9 Suppose A at t1 is immediately followed by B at t2, that B exactly resem‑ bles A, and that B appears in the same place as A, but that C also appears at t2, also exactly resembles A, and is contiguous to B. Take A, B, and C to be fireflies. Is C an interloping twin or did A move to the side, making B the interloping twin? Might A have flown off or been eaten, so that B and C are both interlopers? B and C must be numerically distinct because each occupies a different place at the same time. But each has an equal title

Section 3.2.1  123 to be identified with A. Since identity over time must be one–one, judge‑ ments of identity must be suspended pending further investigation. Even if that proves to be indecisive, A is recognized as different from both B and C. That is why there is a question concerning which of B and C should be identified with A. In the case of time, we might throw Price’s charges of speaking untruths and speaking nonsense back at him. An object is only properly called “un‑ changing” if it lasts for more than a moment; otherwise, it is instantane‑ ous.10 We do find many of the things around us to endure over more than a moment without undergoing any change that we can perceive, like the kettle that will not come to a boil or the stoplight that will not turn green or the fire alarm that will not stop blaring or the toothache that will not stop hurting. In these cases, we are often amazed how little time has passed while we are waiting, and too frustrated with the failure of the impression to change to turn our attention to the succession of other experiences. Our sense of the succession of moments is not based on any change in our ideas. Price’s questionable assertions continue to be endorsed. Because our experience of temporality just is the successiveness of our experiences, Hume concludes that we cannot directly experience an un‑ changing object by itself as having duration. We would instead have one unchanging perception and, without any change in it being experienced, we would not have anything from which to abstract the successiveness that is essential to the Humean idea of time. (Ainslie 2015, 76, referencing T 1.2.3.10–12) Thus the idea of time is the idea of successiveness, or better, of a succes‑ sion in general. Cases of succession for Hume are any cases of replace‑ ment, or alteration, or movement. (Baxter 2009, 135) In fairness, Ainslie and Baxter do not take on these assertions as their own. Like Price, they instead attribute them to Hume.11 But the assertions are still questionable. The assertion Ainslie attributes to Hume, that unless we experience a change in a perception, we would not experience anything from which to “abstract the successiveness that is essential to the Humean idea of time,” begs the question of why experiencing a change is necessary for experi‑ encing successiveness. There is no absurdity in the thesis that a successor might be qualitatively identical to its predecessor and so no logical impos‑ sibility in the experience of a monotonous succession (think of the blast of a foghorn).12 If we cannot have such an experience, or we can, but cannot “abstract successiveness” from it, we need to be told why not. It cannot

124  Time and our experience of time just be that it is hard to distinguish qualitatively identical successors from one another. That only makes the number of successors vague, not inesti‑ mable (Section 2.4.1).13 The assertion Baxter attributes to Hume, that cases of succession are those of replacement (presumably of one thing with a qualitatively differ‑ ent thing), alteration, or movement, restricts cases of succession to cases where there is a qualitative or local difference between the successive ele‑ ments without explaining what justifies that restriction. The continuance of an unaltering thing in its place is a possibility. So, therefore, is its being its own successor.14 It needs to be explained why this has been excluded. 3.2.2 What constitutes a moment?

As discussed in Section 2.5, for Hume the elements that are disposed in space are not, in the first instance, objects, like books and tables, nor are they detachable parts of objects like table legs and book pages. In the ex‑ emplary case of the view of the tabletop, the elements are simple and in‑ divisible parts or “points” that are said to be “endow’d” or “filled” with some real quality (T 1.2.3.15, 1.2.3.17, 1.2.4.2). This is as it should be. Hume began the Treatise by talking about im‑ pressions and ideas, not about publicly observable objects. An account of how an experience of impressions and ideas is worked up into an experi‑ ence of public objects may turn out to presuppose an experience of spatial and temporal relations between impressions and ideas (Section 6.10–12). This implies that an account of “time, in its first appearance to the mind” should not take objects or events involving objects to be the things that are originally successive in time. Though Hume described the conception of time as being that of “different ideas, or impressions, or objects dispos’d in a certain manner” (T 1.2.3.10), the original experiences from which the idea of time is derived must be those of “simple” perceptions. “Simple” is in quotes because the successive impressions and ideas con‑ stitutive of a first appearance of time to the mind need only be temporally simple, that is, not further divisible into earlier or later parts. They could be both qualitatively and spatially complex. Hume thought that two differ‑ ent colours cannot occupy the same point in space without the annihilation of one of them (T 1.2.4.6). But what holds for vision does not hold for touch. A simple and indivisible hot point might also be hard and painful. Hume further declared that “an unchangeable object … produces none but co‑existent impressions” (T 1.2.3.8). Whatever we may think of the soundness of that observation, it proves that he thought that more than one impression can occupy the same moment. Given that this is the case, allowing for the finite holding capacity of the mind (T 1.2.1.2), there is no non‑arbitrary limit to how many qualitatively

Section 3.3  125 simple or spatially simple perceptions a temporally simple perception could contain. If I have my eyes open, I am conscious of the coexistence of the points constitutive of an entire visual field. I am simultaneously conscious of the points constitutive of an entire tactile field, plus various sounds, and perhaps various smells, tastes, thoughts, passions, and volitions. Our consciousness is often not as broad or inclusive as it could be. We will sometimes attend to music or speech or to what we are eating or think‑ ing of to the exclusion of almost all else. Elements of what is on the visual or tactile fields can likewise capture our consciousness to the exclusion of almost everything else. But consciousness normally extends over many perceptions at once. For any one conscious being, time consists of a single succession of momentary total perceptions, constituting “simultaneity planes” that are stacked on top of one another like a deck of cards to constitute the life experiences of that individual. Except the deck is a virtual deck. The emer‑ gence of a card from the virtual deck entails the destruction of the exist‑ ent card. The cards are not dealt out over time. The emergence of a card defines a moment of time. On this account, time is a perpetually perishing “dimension.” But this does not entail that it must be substantival or abso‑ lute (that it would exist even if there were no perceptions to be disposed on simultaneity planes, or that it extends beyond an individual consciousness to correlate the simultaneity planes of other individuals). This having been said, Hume seems to have thought that our temporal ex‑ perience comes pre‑cut into multiple successions corresponding to different series of events or to different processes: one stream for the mountains we see outside the window; another for the shadows cast by wind‑driven clouds that we see scuttling over the mountains; a third for the ticks of a clock that we hear while looking at the mountains; and so on (T 1.2.5.29). But he failed to provide an account of what divides the streams. This is another in‑ stance of his failure to consider how we descend from the experience of com‑ plex impressions to cut out ideas of publicly identifiable objects or processes. 3.3

Phenomena and an argument bearing on the answers to these questions (or not)

Hume’s discussion of “time in its first appearance to the mind” (T 1.2.3.7–8) is devoted to establishing that the idea of time cannot be obtained from an unchanging experience. As he put it, (NM): No monotony [T]ime cannot make its appearance to the mind … attended with a steady unchangeable object[.] (T 1.2.3.7)

126  Time and our experience of time This is a special conclusion about time, not drawn for space. In the visual impression from which we first derive the idea of space, the points may all be of the same colour (T 1.2.3.5). Treatise 1.2.3.7 appeals to three phenomena to justify (NM). First, some‑ one “in a sound sleep … is insensible of time.” Second, someone “strongly occupy’d with one thought, is insensible of time.” Third, “according as [our] perceptions succeed each other with greater or less rapidity, the same duration appears longer or shorter.” All that the first phenomenon shows is that when we are unconscious, we really are unconscious and so no more aware of succession, duration, or time than we are of anything else. The second phenomenon at least involves con‑ scious experience, but it is so far from being always observed to happen that the opposite is proverbial: “A watched kettle never boils.” Someone “strongly occupy’d with one thought” can be anything but “insensible of time.”15 The proverb also offers empirical evidence against the claim that the same duration (or event) appears longer or shorter according to how rap‑ idly our perceptions succeed one another. The time it takes for the kettle to come to a boil seems much shorter while distracted by the rapid succession of other ideas than while strongly occupied with the thought of the kettle. Others were troubled or unconvinced by this sort of evidence. In his notebooks, after entering, opposite a “T,” “Time train of ideas succeed‑ ing each other,” Berkeley entered, opposite an obelus, “Why time in pain, longer than time in pleasure?” (Ayers, 307 entries 4 and 7). Reid was later to report hearing a military officer, a man of candour and observation, say, that the time he was engaged in hot action always appeared to him much shorter than it really was. Yet I think it cannot be supposed, that the succession of ideas was then slower than usual. (Intellectual Powers 3.5, 273) A passage from Treatise 1.4.2.33  might be invoked to support Hume’s contrary position. When we fix our thought on any object, and suppose it to continue the same for some time; ’tis evident we suppose the change to lie only in the time, and never exert ourselves to produce any new image or idea of the object. The faculties of the mind repose themselves in a manner, and take no more exercise, than what is necessary to continue that idea, of which we were formerly possest, and which subsists without variation or interruption. The passage from one moment to another is scarce felt, and distinguishes not itself by a different perception or idea, which may require a different direction of the spirits, in order to its conception.

Section 3.3  127 But we do not always find the presence of an unchanging object restful, as Reid observed. When a man is racked with pain, or with expectation, he can hardly think of any thing but his distress; and the more his mind is occupied by that sole object, the longer the time appears. On the other hand, when he is entertained with cheerful music, with lively conversation, and brisk sallies of wit, there seems to be the quickest succession of ideas, but the time appears shortest. (Intellectual Powers 3.5, 273) Reid was not drawing on any uncommon or profound insight. Hume would have had good reason to suspect that our attitude to the object may have more impact on how much time we consider to be passing than the number of changes it undergoes. Plausibly, this is because certain attitudes make us more impatient for our circumstances to change and so more sen‑ sitive to the endurance of the undesirable circumstance or to the monotony in the succession of events. Hume went on to appeal to a fourth phenomenon. He began by citing Locke’s authority. It has been remarked by a great philosopher, [fn.: “Mr. Locke.”] that our perceptions have certain bounds in this particular [the “greater or less rapidity” with which “perceptions succeed each other”], which are fix’d by the original nature and constitution of the mind, and beyond which no influence of external objects on the senses is ever able to hasten or retard our thought. (T 1.2.3.7) He then illustrated one‑half of Locke’s point, that involving a bound to greater rapidity, with the example of a whirling coal, If you wheel about a burning coal with rapidity, it will present to the senses an image of a circle of fire; nor will there seem to be any interval of time betwixt its revolutions; merely because ’tis impossible for our perceptions to succeed each other with the same rapidity, that motion may be communicated to external objects. (T 1.2.3.7) and concluded, Wherever we have no successive perceptions, we have no notion of time, even tho’ there be a real succession in the objects. (T 1.2.3.7)16

128  Time and our experience of time But the conclusion that needs to be established is that time cannot make its appearance to the mind when attended with a steady unchangeable object. That conclusion does not follow from the evidence. All that the case of the whirling coal proves is that when the succession of our perceptions is too rapid for us to notice, it is too rapid for us to notice, so that if we see anything at all, it appears instantaneous. It does not follow that we would get no awareness of time from an immobile and unchanging object, like a watched kettle, or from a rapidly moving one that continues on a closed course (as when the coal continues whirling). None of the phenomena Hume cited establishes (NM). The first and fourth phenomena are irrelevant and the second and third are not always true. Hume went on to offer an argument for (NM). ’Tis evident that time or duration consists of different parts: For oth‑ erwise we cou’d not conceive a longer or shorter duration. ’Tis also evident, that these parts are not co‑existent: For that quality of the co‑existence of parts belongs to extension, and is what distinguishes it from duration. Now as time is compos’d of parts, that are not co‑­ existent; an unchangeable object, since it produces none but co‑existent impressions, produces none that can give us the idea of time; and con‑ sequently that idea must be deriv’d from a succession of changeable objects, and time in its first appearance can never be sever’d from such a succession. (T 1.2.3.8) This argument opens by making a great deal of the presentist principle that the parts of time are not coexistent. From this, it draws the conclusion that an unchangeable object could only produce (or consist of?) coexist‑ ent impressions. It is unclear what justifies limiting this conclusion to un‑ changeable objects. The fact that the parts of time are not coexistent has as much bearing on changing objects as unchangeable ones. If presentism entails that unchangeable objects can only produce coexistent impressions (presumably those disposed at the one part of time that happens to exist), it ought to entail that changing objects are similarly limited to only pro‑ duce coexistent impressions (only those disposed at the one part of time that happens to exist and not those that were produced at any of the earlier or later stages of the changing object’s existence). If, on the other hand, presentism does nothing to detract from the fact that changing objects pro‑ duce a succession of impressions that are (or were or will be) disposed over other, no longer existing or not yet existing parts of time, then by parity of example it should do nothing to detract from the fact that unchangeable

Section 3.3  129 objects produce a monotonous succession of exactly resembling, momen‑ tary perceptions or equivalently an enduring perception (one succeeded by itself). A further justification for (NM) amounts to throwing down the gauntlet at any opponents. But that we really have no idea [of “time without any changeable exist‑ ence”], is certain. For whence shou’d it be deriv’d? Does it arise from an impression of sensation or of reflection? Point it out distinctly to us, that we may know its nature and qualities. But if you cannot point out any such impression, you may be certain you are mistaken, when you imagine you have any such idea. (T 1.2.5.28) There is an obvious answer to Hume’s challenge. We can obtain the idea of a time without any changeable existence from an impression that en‑ dures without changing, like the impression of the proverbial watched kettle. Rather than explain what is wrong with this answer, Hume pre‑ tended that there is no answer to be given. In fairness, he was about to go on to explain why we only “fancy we have [the] idea” (T 1.2.5.29) that there is any such thing as an invariant, enduring impression. But that is a project that presupposes that we do not acquire the idea directly from invariant, enduring impressions. Treatise 1.2.5.28 only supports that pre‑ supposition with rhetorical questions and a refusal to recognize the obvi‑ ous answer to them. (NM) is not justified by the considerations Hume offered in its support, which are either irrelevant, contentious, derived from the undue restriction of principles to just some of the cases to which they apply, or based on bluster and rhetorical questions. It has no other foundation than Hume’s say‑so. It might be objected that even if Hume failed to justify (NM), it does not follow that he could not have justified it. A sympathetic reader ought to do what they can to draw on Hume’s principles to do so. This is true. But it is not easy. Consider how (NM) is justified in what is widely and justly considered to be the state‑of‑the‑art treatment of Hume’s thought on this topic. Yet another direct consequence of Hume’s account is that only succes‑ sions have duration. It is clear that objects which exist only for a brief moment do not have duration, but the difficult case is objects that are “stedfast and unchangeable” (T 1.2.3.11). These are cases “without any change or succession” (T 1.2.5.29). In such a case there is just a single

130  Time and our experience of time object, not distinct objects successively. So the idea of a steadfast object cannot serve as an idea of duration. So a steadfast object cannot con‑ vey the idea of duration. So a steadfast object lacks duration: “For it inevitably follows from thence, that since the idea of duration cannot be deriv’d from such an object, it can never in any propriety or exactness be apply’d to it, nor can any thing unchangeable be ever said to have duration” (T 1.2.3.11). Only by a fiction can one come to think of a steadfast object as having duration. So it is Hume’s view that … only successions have duration. (Baxter 2008, 21) This defence of (NM) rests on the claims that (i) the case of a steadfast or unchangeable object is a case “without any change or succession,”17 and that (ii) “In such a case there is just a single object, not distinct objects suc‑ cessively.” Both assertions present false dichotomies. Concerning (i), being without change, notably qualitative change, is not the same thing as being without succession. It needs to be proven that what is unchanging cannot occupy a succession of moments. Concerning (ii), the alternatives are not just “a single object” and “distinct objects successively.” The question is why there is not a third alternative: a single “object” identified as continu‑ ing to exist from one moment to the next. For both of these reasons, it has not been established that an unchanging object, like a watched kettle, can‑ not serve to supply the idea of duration, much less that it lacks duration. 3.4

Temporal passage in the absence of change

Hume appealed to (NM) to draw a yet more radical conclusion. Reiterat‑ ing that “the idea of duration … can never be convey’d to the mind by any thing stedfast and unchangeable,” he inferred that nothing “steadfast” or “unchangeable” can properly be said to endure.18 (ND): Non‑duration [S]ince the idea of duration cannot be deriv’d from [“any thing stead‑ fast and unchangeable”], it can never in any propriety or exactness be apply’d to it, nor can any thing unchangeable be ever said to have dura‑ tion. Ideas always represent the objects or impressions, from which they are deriv’d, and can never without a fiction represent or be apply’d to any other. (T 1.2.3.11) (ND) draws a conclusion about what “ideas” can properly or exactly be applied to what “things,” and so about what can be “said,” known, or believed. It does not draw a conclusion about what can exist (compare

Section 3.4  131 T 1.2.5.26). For Hume, the truth of (ND) follows directly from (NM). It only remains to explain what causes so many of us, “philosophers as well as … the vulgar,” to “pretend, that the idea of duration is applicable in a proper sense to objects, which are perfectly unchangeable” (T 1.2.3.11). We might wonder whether (ND) follows as easily as Hume claimed (Boehm and Cruz 2023). But even if it does not, his charge that we only pretend that “unchanging” objects endure merits examination. If sound, it might be taken to shift the burden of proof onto those who would main‑ tain that we can perceive temporal passage when nothing changes. If un‑ sound, his later account of how we get the idea of identity, which appeals to it (T 1.4.2.29), is impacted. Hume wrote, (Pretence): What leads us to pretend that “unchanging” objects endure But tho’ it be impossible to show the impression, from which the idea of time without a changeable existence is deriv’d; yet we can easily point out those appearances, which make us fancy we have that idea. For we may observe, that there is a continual succession of perceptions in our mind; so that the idea of time being for ever present with us; when we consider a stedfast object at five‑a‑clock, and regard the same at six; we are apt to apply to it that idea in the same manner as if every moment were distinguish’d by a different position, or an alteration of the object. The first and second appearances of the object, being compar’d with the suc‑ cession of our perceptions, seem equally remov’d as if the object had re‑ ally chang’d.… From [this relation and two others he went on to mention over the ellipsis] we are apt to confound our ideas, and imagine we can form the idea of a time and duration, without any change or succession. (T 1.2.5.29) Hume wrote of considering a “steadfast” object at one time and regarding the same object an hour later. This is an ambiguous pronouncement. It does not say that we look away from the object between five and six. Nei‑ ther does it say that we continuously observe it over an hour‑long period. But it should not matter. An object that is “considered” and “regarded” at two immediately successive moments is as much a test case for Hume’s claims as one that is continuously observed for an hour or intermittently observed after an hour. And (like it or not) Hume’s description does stipu‑ late that we “consider” and “regard” the “steadfast” object at two differ‑ ent clock times. Rather than inflate the case beyond necessity, grant that the “continual succession of perceptions in our mind” consists of just two immediately successive perceptions, A and B, and that there is a “stead‑ fast” object or impression, C, “considered” when A exists and “regarded” a fraction of a second later, when B exists.

132  Time and our experience of time 5:00:00.0 pm 5:00:00.1 pm

A B

C C

Treatise 1.2.5.29 raises some other problems. One is posed by Hume’s usual tendency to use “object” in contexts where he ought to be restricting himself to terms that have a phenomenologically more basic extension, such as “sensation,” “impression,” “idea,” or “perception.” We might charitably read him as using “object” broadly, to include impressions and ideas. Suppose that they are what he principally had in mind. A more serious problem is raised by Hume’s description of the extent of consciousness.19 Hume was right that when consciousness is narrowed down on one thought we can (sometimes) be insensible of the passage of time (T 1.2.3.7). But on his account, this is not that sort of case. In this case, consciousness takes time to occur and over this time A and B occur in succession. Moreover, as Hume described it, when A occurs the moment containing A is also “filled” with a “first appearance” of some “steadfast” impression, C, and when B occurs, the successive moment containing B is also “filled,” now with a “second appearance” (Hume’s words) of C. After all, we are supposed to compare C with the succession, A–B and the comparison is supposed to lead us to think that there are first and second appearances of C that are as removed from one another as A is from B (this is according to Hume’s description). That could only happen if the first ap‑ pearance were coexistent with A and the second with B. It is hard to see how there could be any pretence concerning what this means. These are circumstances under which we experience C to endure over two successive moments (or, equivalently, to consist of two, succes‑ sive but otherwise identical “appearances”).20 Put in terms of the account of temporal consciousness of Sections 1.9 and 3.1, and the account of moments of Section 3.2.2, consciousness must extend over (at least) two successive moments. At the first of these moments, there is a consciousness of the coexistence of A and C. At the second there is a consciousness of the coexistence of B and C. Consciousness consists of two successive parts, in the first of which C occurs and in the second of which C also occurs. It follows that C endures over time. At best, we might think that, had we been intently focused on C, to the point of failing to notice the A–B succession, we might have lost track of the passage of time. But then the pretence is all on the side of forming the idea that C was only momentary. When C is experienced together with the A–B succession, its two appearances do not just “seem equally remov’d, as if the object [C] had really changed”; they are equally removed, notwith‑ standing that C is unchanged. The only “confounding” going on would be the confounding of the initial “consideration” with the subsequent “re‑ gard” of C, leading Hume to pretend that it only lasted for a moment. It might be suggested that Hume only meant to draw attention to a psy‑ chological limitation under which we all labour: that unless we attend to

Section 3.4.1  133 some coincident continual succession of perceptions, we could not form the idea that C endures, even if it really does. The watched kettle proverb de‑ scribes counterevidence, but aside from that Hume could not have been con‑ tent with such a modest conclusion. (ND) insists on the stronger conclusion that nothing “unchangeable” can ever be properly or exactly said to endure. Hume’s account of “those appearances, which make us fancy we have [the] idea” (T 1.2.5.29) that “steadfast” objects endure and that time passes when nothing changes achieves the opposite of what it proposes. It identi‑ fies circumstances under which it is demonstrable that unchanging impres‑ sions must endure or, equivalently, consist of a monotonous succession.21 3.4.1 Baxter’s defence

It might be proposed that Hume nodded when he wrote of a “first ap‑ pearance” of a “steadfast” object being “considered” at five o’clock and then “regarded” at six. Since Hume said that “the idea of duration … can never be convey’d to the mind by any thing stedfast and unchangeable” (T 1.2.3.11), he must have meant to say that C is considered and regarded at just one moment. If this entails that one moment can coexist with two successive moments, then Hume must have meant to endorse that conse‑ quence as well.22 We might even propose, contrary to the account of mo‑ ments of Section 3.2.2, that he might have made sense of the coexistence of different moments by supposing that “objects” (impressions and ideas?) occupy their own individual timelines, allowing for moments on different timelines to “coexist” (in some sense). But there comes a point at which the project of offering a faithful ren‑ dition of the history of ideas prevents us from claiming that a historical figure said things they never said, even if those things are entailed by what the figure said or might justify or explain what the figure said. Hume did not write that when we perceive the first perception in the successive se‑ ries of perceptions, we also perceive the “steadfast” impression existing at a moment in a second successive series of impressions, and when we perceive the last perception in the first successive series, we still perceive the “steadfast” impression at the same moment of the second series. He instead spoke of there being first and second appearances of “the same object” at two different moments: five and six o’clock. If that was merely a metaphor, he offered no clue what he might have meant by it. On the contrary, he wrote things that indicate that he endorsed a traditional con‑ ception of time as a single succession of moments. [T]he incompatibility of the parts of time in their real existence sep‑ arates them in the imagination.… Every part must appear single and alone, nor can regularly have entrance into the fancy without banishing what is suppos’d to have been immediately precedent. (T 2.3.7.5)

134  Time and our experience of time [T]ime or succession, tho’ it consists likewise of parts, never presents to us more than one at once; nor is it possible for any two of them ever to be co‑existent. (T 2.3.7.5) ’Tis a property inseparable from time, and which in a manner consti‑ tutes its essence, that each of its parts succeeds another, and that none of them, however contiguous, can ever be coexistent. (T 1.2.2.4) ’Tis evident, that time or duration consists of parts.… ’Tis also evident, that these parts are not co‑existent[.] (T 1.2.3.8) For the same reason, that the year 1737 cannot concur with the present year 1738, every moment must be distinct from, and posterior or ante‑ cedent to another. (T 1.2.2.4) The last of these remarks may have been influenced by one made by Bayle. I say that what suits Monday and Tuesday with regard to succession suits every part of time whatsoever. Since it is thus impossible that Mon‑ day and Tuesday exist together, and since it must be the case necessarily that Monday cease to exist before Tuesday begins to exist, there is no other part of time, whatever it may be, that can coexist with another. Every one has to exist alone. Every one must begin to exist, when the other ceases to do so. Every one must cease to exist, before the follow‑ ing one begins to be. (Bayle Dictionnaire, 353–4) Locke was of the same opinion. Duration is but as it were the length of one streight Line, extended in infinitum.… [T]his present moment is common to all things, that are now in being, and equally comprehends that part of their Existence, as much as if they were all but one single Being; and we may truly say, they all exist in the same moment of Time. (Essay 2.15.11) This is the picture of the content of moments presented in Section 3.2.2. If Hume disagreed, we might think he would have made a point of it. Instead, he wrote as if he agreed. We need to have strong evidence before

Section 3.4.1  135 supposing that a figure would have abandoned a common, traditional, un‑ questioned but not unarticulated opinion without making anything of do‑ ing so and while writing things that imply that they accepted that tradition. Hume was not afraid to declare, in opposition to “the common opinion of philosophers, as well as of the vulgar” (T 1.2.3.11), that “unchanging” objects cannot properly be said to endure. But he never said that one mo‑ ment can coexist with a succession of moments. It may be that had he been confronted with the observation that what he had said implies that a single moment must coexist with distinct, succes‑ sive moments he would have been taken aback. The discovery may have led him to wonder why he was so invested in denying that “steadfast” objects can properly be said to endure, and what good arguments he had for (NM). And it may be that those reflections would have led him to decide that what he had said could not be defended without abandoning tenets he was not prepared to deny for the sake of a consequence he had no real need to insist on.23 We cannot be confident that this is not what in fact happened. Hume did not like to change his mind about things (or be seen to have done so), but he could be persuaded that some were not worth pursuing. He considered his position on vacuum and temporal passage without change to be a consequence of his position on infinite divisibility (T 1.2.4.2), and Stanhope may have convinced him that his views on that topic would do him no credit were they published (Section 2.1). This having been said, it is legitimate to propose on Hume’s behalf that his position might be supported by supposing that there are different time‑ lines, and even that it would not have been implausible for him to invoke it.24 Berkeley seems to have recognized different timelines, and Hume likely read some of the pertinent passages. Time … being nothing, abstracted from the succession of ideas in our mind, it follows that the duration of any finite spirit must be estimated by the number of ideas or actions succeeding each other in that same spirit or mind. (Principles 98) Bid your servant meet you at such a time, in such a place, and he shall never stay to deliberate on the meaning of those words: in conceiving that particular time or place, or the motion by which he is to get thither, he finds not the least difficulty. (Principles 97) There is no larger time that contains the timeline constituted by the serv‑ ant’s ideas and the timeline constituted by yours. Suppose for the sake of argument that there were. Suppose further that in this larger time, you died

136  Time and our experience of time a thousand years before the birth of your servant. It will make no differ‑ ence to what either you or your servant experience because your sensory experiences are not the effect of material bodies existing in an external world, but ideas produced in you by God. You will have your own ideas of speaking to the servant and of subsequently encountering the servant at the appointed time and place and the servant will have his ideas of hearing you speak and subsequently encountering you at the appointed time and place. The illustrative supposition of a larger time can now be set aside. For Berkeley, the supposition of a larger time is nonsense, and it also makes no sense to consider ideas in one mind to be either earlier or later than ideas in another mind. There are just independent minds, each with its own successive ideas. The minds are not in time. Times are in the minds (Hynes, 2005). However, on the revisionist Humean account being considered here, the different timelines are not split between different minds. They figure within a single consciousness or complex perception, consisting of an invariable and uninterrupted or “steadfast” impression (object), and a further suc‑ cession of perceptions. Otherwise, the comparison and confusion required by Treatise 1.2.5.29 would be impossible. This raises the question of how that comparison and confusion is possible without recognizing either that the “steadfast” impression endures over time or that it is divided into a monotonous succession of states. Respecting Hume’s claim that “steadfast” objects cannot be properly or exactly said to endure and so cannot be properly or exactly taken to oc‑ cupy more than one moment, the different timelines might be modelled as in Figure 3.1. “Steadfast” object at Line 1

Line 1 Line 2 Line 3

Later 5pm at Line 3

Figure 3.1  Baxter’s model with pointed moments

6pm at Line 3

Section 3.4.1  137 Dots represent successive moments along three different timelines. There are no temporal gaps between successive dots, though the model must include spatial gaps for the sake of perspicuity. (The distribution of dots is intended to conform to the distribution of blocks in Baxter 2008, 38.) Taking the “steadfast” object occupying a moment on the upper timeline to be coexist‑ ent with a succession on the lower timeline means that coexistence cannot be represented by verticals to the “Later” arrow. It is instead represented by the curved lines. “Later” is line relative. Dots further to the right on one line are not necessarily later than dots to their left on other lines. A model ap‑ propriate to the consequence that “steadfast” objects are momentary should have a triangular configuration, constituted by separate moments on the “Later” arrow for one line and the coexistence arrows from those moments to the moment of the “steadfast” object on another line. The model should not look like a brick wall, which risks suggesting that the “steadfast” object has earlier and later parts, which Hume meant to deny. If one perception in the successive series of perceptions, A, occurs at one moment, 5 pm, and another, B, at another moment, 6 pm, on what might be called a “lower” timeline, while the invariant and uninterrupted impres‑ sion, C, occurs at its own moment on a separate “higher” timeline, then something needs to be said about what leads us to align C with both A and B as opposed to any other moment or moments on the lower timeline. By Hume’s account, it is because C is perceived at the same moment as A and at the same moment as B. On the interpretation offered on Hume’s behalf, some moments must coexist with successions of moments, so this cannot be dismissed. (Plausibly, the coexistence relation between B and C is only discovered “after” that between A and C by the consciousness experiencing the timelines, but set this complication aside. It is taken up in Section 4.3.) Everything hangs on what determines these perceived coexistence rela‑ tions. We are not free to say just anything on this matter because Hume applied a condition to coexistence. Considering what he said on that head, it is possible to offer a formalization that demonstrates that the implication that some moments coexist with successive moments is inconsistent with two temporal axioms Hume accepted. Consider a domain of simple and indivisible parts or moments of time {a, b, c, …} governed by an irreflexive, asymmetric, transitive temporal successor relation, > (“later than”).25 (1): (2): (3):

∀x~>xx ∀x,y(>xy → ~>yx) ∀x,y,z(>xy & >yz → >xz)

(Irreflexivity: no moment is later than itself) (Asymmetry of moments) (Transitivity of moments)

Include Hume’s claim that “’Tis a property inseparable from time, and which in a manner constitutes its essence, … that none of [‘its parts’],

138  Time and our experience of time however contiguous, can ever be coexistent” (T 1.2.2.4, see also T 1.2.3.8 and 2.3.7.5). This places a necessary condition on the relation, @, of coex‑ istence. Put in the contrapositive to reflect Hume’s phraseology, the condi‑ tion is (H)

∀x,y(~=xy → ~@xy)

Finally, add the thesis that there is at least one moment that coexists with two successive moments. (5)

∃x,y,z(@xy & @xz & >yz)

(1) and (H) contradict (5), as demonstrated by the derivation on Table 3.1.26 Table 3.1 Formal inconsistency of (1), (H), and (5) 1 2 3 4 5 6 7 8 9 10 11

(1),(H),(5) Given

⊥

@ab @ac >bc =ab =ac >ac >aa ~>aa ⊥

Instance of (5) Instance of (5) Instance of (5) 2,(H) 3,(H) 5,4 6,7 Instance of (1) 8,9 2,3,4–10

Of course, the addition of yet further principles, details, and qualifica‑ tions can do nothing to change the fact that a contradiction can be derived just from (1), (H), and (5). The formalization given here follows one given by Baxter (2008, 38–9) with one exception. (H) treats coexistence as an equivalence relation, which Baxter (2016, 185) finds unacceptable.27 Baxter instead treats co‑ existence as “primitive” (2008, 37), and consistently with that approach replaces (H) with a definition of coexistence, here denominated (B), that is purely negative. (B)

∀x,y(@xy ↔ ~>xy & ~>yx)

This is not Hume’s view. Hume was committed to both (H) and (5), which together with (1) lead to a contradiction. Acknowledging what Hume said means accepting that he had inconsistent commitments on this matter. Furthermore, he never returned to this topic in his authorized pub‑ lications, perhaps for this reason. Had he chosen to do so, he would have

Section 3.4.1  139 had two choices. He could have weakened (H), perhaps along the lines Baxter proposes. But he could just as well have abandoned (5), for which he had no use. Supposing that the project is to offer a defence of Hume’s commitment to (5) rather than an interpretation that is faithful to everything he had to say, it would border on irrelevancy to object to (B) on the ground that it does not reflect the necessary condition Hume placed on coexistence at Treatise 1.2.2.4, 1.2.3.8, and 2.3.7.5. It is more pertinent to note that Hume needed something more than (B) to explain what leads us to com‑ pare and confuse a temporal sequence of events on one timeline with a temporal sequence of events on a different timeline. As a negative defini‑ tion, (B) says nothing about how events on different timelines are coordi‑ nated with one another. It does not pick out any one pair of moments on different timelines in preference to any other and so fails to account for why we would confuse the moment on one timeline with any one succes‑ sion of moments on the other timeline. A positive principle is needed to account for the confusion. It is not enough to say that coexistence is primitive. It is necessary to explain how it comes to be perceived without giving rise to the inconsist‑ ency described above. Any further details would need to be formalized and included in the consistency demonstration (the model would need to be shown to satisfy these further principles as well as (1), (B), and (5)). Baxter does gesture at two ways in which the system might be supple‑ mented. Prior to axiomatizing Hume’s account of time and demonstrating its consistency (2008, 38–9) he offers a diagrammatic model, recopied as Figure 3.2 (without the three letters inserted in some of the blocks), and comments, “The blocks represent moments. Parts of blocks do not rep‑ resent moments, nor do they represent parts of moments.… Coexistence of moments is represented by the fact that two blocks could be cut by the same vertical line” (2008, 37). Place 1 Place 2 Place 3

C A

B Later

Figure 3.2  Baxter’s model with extended moments

In this picture, B is later than A, but C coexists with both of them. How‑ ever, for the picture to work, A, B, and C must be depicted as extended blocks rather than points. This is a further feature of Baxter’s account.28

140  Time and our experience of time He maintains that even though a “steadfast” object occupies only a single moment, it “can be long‑lasting or fairly brief without being of greater or lesser duration. There are two kinds of temporal length—duration and steadfastness” (Baxter 2016, 186). It is hard to understand how something can be perceived to be “long‑­ lasting” without being divisible into earlier and later parts and so having temporal duration. Hume himself ruled it out. ’Tis evident, that time or duration consists of different parts. For other‑ wise we cou’d not conceive a longer or shorter duration. (T 1.2.3.8) A real extension, such as a physical point is suppos’d to be [or, we might add, such as a “steadfast” object is supposed to be on Baxter’s account], can never exist without parts, different from each other; and wherever objects are different, they are distinguishable and separable by the imagination. (T 1.2.4.3) As noted in Section 2.3, George makes a similar proposal that minimally visible points could have some angular size. But, when confronted with the objection that any angle can be bisected into adjacent halves, George replies that we do not see the angular size. We are at best able to calculate it based on other things we see, making it an instance of “abstract math‑ ematical thinking.” Our visual experience is just of a point, with no appar‑ ent size, but only a location. Baxter, however, denies that we must experience “steadfast” objects as uniformly and maximally brief, instantaneous occurrences, holding that we can experience them as having some sort of temporal “length” or “depth.” This raises the question of why they are not experienced to be “bisected,” that is, divided into earlier and later portions, in virtue of their superposition in the same consciousness with a succession of events. Significantly, the embellishments concerning stacked successive series, the “temporal length” of moments, and coexistence defined by verticals to the later arrow do not feature in Baxter’s formalization or his proof that his formalization has a model. (B) is not embellished with any of these fea‑ tures, but stated as a purely negative condition for coexistence. Later, Baxter introduces the notion of “co‑duration” and claims that (H) only applies to moments on individual timelines. Moments on “co‑­ enduring” timelines may “coexist” in his primitive sense (2008, 43). With‑ out the aid of the diagram, it is not clear what determines the primitive coexistence relations. There must be some justification for aligning things on one timeline with those on another. Unless it is formally reflected in (B), the amended system has not been shown to have a model.

Notes  141 The point of this discussion is not to find fault with Baxter’s subtle attempt to defend the thesis that something that is only momentary might coexist with every temporal part of something that occupies a succession of moments. It is to illustrate that there are limits to what can be established by any formalization, including the one used earlier to demonstrate inconsistency. A formal demonstration, be it of a contradiction or of consistency, is only as good as the principles it takes off from. The real question is whether those principles are all the system requires (in the case of consistency) or more than it needs (in the case of contradiction). Critics and supporters alike need to beware of taking either the fact that there is formalization that has a model or the fact that there is a formalization that yields a contradiction to be decisive. In the case at hand, the real question does not concern an abstract set of axioms governing moments and successions. It concerns which of them Hume accepted and with what justification. Notes 1 See Chapter 1, note 44, on the distinction between empirism and empiricism. 2 Unless otherwise specified, all citations in this section are from Treatise 1.2.3.10. 3 I owe this point to Maité Cruz. 4 Mellor (1998, 122–3) tries to get around the difficulty by appealing to nested memories. On his account, I can experience an event while (i) remembering the immediately prior event, while (ii) remembering that while remembering the immediately prior event I was remembering the event immediately prior to it, while (iii) remembering that while remembering that while remembering the event immediately prior to the immediately prior event I was remembering the event immediately prior to it, and so on (see also Rocknak 2013, 21). However, this recourse relies on a way of distinguishing current impressions from memories, memories of memories, and so on. The distinction cannot be based on where thoughts are disposed in time since they are taken to be contemporaneous. It cannot be based on their qualities because Hume took temporal relations to be independent of their relata (T 1.3.2.2). And it cannot be based on an appeal to causal relations because Hume considered causes to be prior to their effects (T 1.3.2.7), which means we must have noticed the fact of succession before we can identify causes. A further problem is posed by the Humean doctrine that memories are less vivacious copies of impressions. Hearing five notes played on a flute is different from hearing five notes played on a piano with the pedal down. Hume’s account of memory entails that Mellor’s case is like the latter, where the current moment of consciousness is a cacophony of the current note and fading retentions of the earlier ones. But we hear the earlier flute notes as ceasing before the later ones begin. That is, we are earlier conscious of the earlier notes and later conscious of the later notes. Earlier consciousness combines with later consciousness to form a unified, temporally extended consciousness. See Section 1.8 and Dainton 2008: 55–6. 5 See Costa (1990, 4), Garrett (1997, 52–4, 168–9), Falkenstein (1997), and Section 2.5. Hume’s insight is sometimes downplayed by scholars who hasten to add that this idea is only “abstract.” But, for Hume, there is only an abstract

142  Time and our experience of time idea of time insofar as there are multiple, particular, temporally extended per‑ ceptions that all resemble one another in being temporally extended. We never have an abstract idea of time apart from an exemplar drawn from a set of particular, temporally extended ideas. 6 Price was laying out Hume’s position. But he was not entirely unpersuaded, declaring that “Obviously we cannot find [an unchanging sense‑impression] at all” (46). For Price, there can be no enduring particulars. At best there are processes or unvariegated series (46–7). 7 “We readily suppose an object may continue individually the same, tho’ several times absent from and present to the senses; and ascribe to it an identity, not‑ withstanding the interruption of the perception, whenever we conclude, that if we had kept our eye or hand constantly upon it, it wou’d have convey’d an invariable and uninterrupted perception” (T 1.3.2.2). 8 Here I follow Garrett (2009, 437). 9 At Treatise 1.1.5.10 Hume wrote, “I consider [difference] rather as a negation of relation, than as any thing real or positive,” going on to equate numeric difference with the negation of identity. But it is hard to lend much weight to this pronouncement given that he went on to equate difference with the “real and positive” relations of distinguishability and separability (T 1.1.7.3) and included contrariety on the list of positive relations. Hume would have done better to recognize numeric difference and identity as contraries, which both resemble one another in involving disposition at different times. Both numeric difference and identity are contrary to unity. 10 Baxter has claimed that not all instants or moments or “temporal simples” or “steadfast” objects are equally “brief” (2008, 30) or “long‑lasting” (2016, 186) or “big” (2009, 144). But on an ordinary understanding, what is “not brief” or is “long‑lasting” is so because it endures over more than a moment, and what is “big” is so because it is divisible into parts. If Baxter’s claims are to be intelligible, some other meaning has to be attached to these terms. Bax‑ ter does attach one: coexistence with a succession. But, as Price pointed out, “We have to distinguish temporal parts within [what coexists with successive things], because we want to say that one part of it is contemporary with A and another part with not‑A” (1940, 46). For Price, there is nothing “fictional” about this (1940, 47). Baxter’s position is further considered in Section 3.4.1. 11 They are not alone. See, among others, Costa (1990, 3, 7), Garrett (1997, 55), and Rocknak (2013, 137). 12 Baxter (2008, 33–6) has argued that Hume did not accept the “brevity assump‑ tion” that perceptions that “seem long” are really uninterrupted successions of exactly resembling perceptions. But it is legitimate to ask whether Hume had a good reason for not accepting that assumption. 13 Hume did try to show that we only “fancy we have [the] idea” that time passes when nothing changes (T 1.2.5.29). But proposing an alternative is not the same thing as offering a refutation. The feasibility of Hume’s alternative is considered in Section 3.4. 14 This is asserted with strong intent. The question is not just why an object can‑ not be succeeded by a later “state” of itself, but why it cannot be succeeded by itself, just as a person can love themselves or shave themselves. 15 See Dainton 2008: 52 for more on the empirical evidence for the proverb. 16 Locke brought up the case of a rapidly rotating body that appears as “a per‑ fect, entire Circle of that Matter or Colour” (Essay 2.14.8), perhaps a spoked carriage wheel, but he did so in the context of arguing that motion does not

Notes  143 always give us an idea of succession or time, not that there are bounds to the greater or less rapidity with which our perceptions succeed one another. He did offer arguments for the latter conclusion as well (Essay 2.14.10–11). But he did so to establish that the imperfectly regular succession of our ideas serves as the original clock that we use to measure the passage of time (Essay 2.14.12). For Locke, the succession of our ideas is not constitutive of time, which is infinitely divisible and flows everywhere at the same pace (Essay 2.15.9). It is our first way of measuring time. Lockean “momentary” ideas are themselves necessarily enduring. They are only phenomenally simple (Essay 2.15.9 and 9n). 17 These are Hume’s words, but at Treatise 1.2.5.29, they were written about “time and duration,” not about steadfast objects. It needs to be proven that there is no succession in the case of steadfast and unchangeable objects, be it the object succeeding upon itself or otherwise identical states succeeding one another. 18 On a natural reading, what is steadfast is what remains in place despite all ef‑ forts made to remove it from that place, which entails that it must remain in that place for more than a moment. What is unchanging is similarly what lasts for more than a moment without changing. Hume must have used “steadfast” and “unchanging” in some idiosyncratic sense in which it is not true by defi‑ nition that what is “steadfast” or “unchanging” must endure. This makes it advisable to put these terms in quotes when used in his sense. 19 Costa (1990, 9) supposes that the circumstances of this case lead us to form two distinct ideas that we then confuse with one another. “The idea of a single unchanging impression [C] is closely associated with the idea of a succession of changes [A–B].… [T]his confusion of different ideas under a single cogni‑ tive disposition … represents the abstract but fictitious idea of time without change.” But for there to be a confusion, something has to lead us to think that C continues to exist when we experience B, and that it already existed when we were experiencing A. We need to somehow get the idea of C as “extended (temporally) over those changes [A–B]” as Costa puts it (9). Costa takes this achievement for granted. But granting that there is “an idea of an unchanging perception viewed at different times” (7), is granting that the “unchanging” perception endures over those two times. 20 This is a variant on the second of two objections offered by Price (1940, 46–7). Price sharpens it by noting that at 5:00.00.1 A does not exist and charges that C cannot be contemporary with both A and not‑A unless it has temporal parts. (Compare it being impossible for C to overlap both A and B without being extended.) Baxter (2008, 33) summarizes Price’s objection over a half sentence and comments, “But whether the assumption is a good thing to believe, is a question independent of whether Hume believed it.” At the point of writing, Baxter is concerned with the latter question. But in the process of answering it, the former question seems to have fallen by the wayside. Loeb (2002, 208), asks, “Why not say that, in the presence of successive changing impressions of some objects, we do have perceived, and hence perceivable, successive impres‑ sions of an unchanged object?” 21 Over the ellipsis in the earlier citation from Treatise 1.2.5.29, Hume identi‑ fied two other causes of the pretence that “steadfast” objects endure. First, we are supposed to discover that the “steadfast” object could undergo as many changes of state or position as there are perceptions in the continual succession. This presumes that we have already coordinated the “steadfast” object with the perceptions in the continual succession and so is not a true alternative. Second,

144  Time and our experience of time we are also supposed to discover that objects undergo continuous changes at the same rate when “in the presence” so to speak, of the “steadfast” object as they would if experienced together with the continual succession of percep‑ tions. This only doubles the phenomena the “steadfast” object is supposed to be coordinated with and so is again not a true alternative. 22 “Yet there is something that lacks duration, namely a steadfast object, which coexists with something that has it, namely, some succession.… So a single indi‑ visible moment coexists with distinct successive moments.” Baxter (2008, 31). 23 Chapter 4 argues that (NM) and (ND) are impediments to Hume’s account of identity (T 1.4.2.26–30), not pillars in its support. 24 Baxter (2016, 187) resists attributing any view on which time has “dimen‑ sions or complex dimensions” to Hume. He nonetheless maintains that Hume’s position at least entails that “time is structured differently in different loca‑ tions” (186–7). His (2008, 37) models temporal successions at different places two‑dimensionally, as a brick wall with brick lines containing bricks of differ‑ ent lengths, and his (2008, 37–8), addresses concerns about what makes differ‑ ent moments part of one “succession” rather than another. “Successions” there play the role of timelines. 25 Baxter (2008, 37) appeals to the same model. 26 A contradiction might just as well have been obtained between (2), (H), and (5). 27 Baxter objects that something’s coexisting with successive things does not en‑ tail their coexisting with one another, any more than something’s overlapping two adjacent things entails that they overlap one another. But this only follows for someone who accepts that something’s overlapping adjacent things means it must be divisible into adjacent parts. Anyone who wants to maintain that the overlapping object is indivisible has some explaining to do. Hume did accept that the identity of moments is a necessary condition for the coexistence of moments (H), and identity is an equivalence relation, so he is saddled with the consequences of that commitment. 28 Up to now, mention of this feature has been avoided. Baxter’s formalization and his consistency proof make no mention of this feature, which works just as well if all “steadfast” objects are instantaneous. To my mind, the system is stronger without it. Maintaining that a “steadfast” object can take up more or less time without enduring only adds a further paradox, so the first order of business is to consider what can be said in favour of Baxter’s position without drawing on it.

Bibliography Ainslie, Donald. 2015. Hume’s True Scepticism. Oxford: Oxford University Press. Baxter, Donald.  2008. Hume’s Difficulty: Time and Identity in the Treatise. London: Routledge. Baxter, Donald.  2009. “Hume’s Theory of Space and Time in Its Skeptical Con‑ text.” In The Cambridge Companion to Hume, 2nd edition, edited by David Fate Norton and Jacqueline Taylor, 105–46. New York: Cambridge University Press. Baxter, Donald. 2016. “Hume on Space and Time.” In The Oxford Handbook of Hume, edited by Paul Russell, 173–90. Oxford: Oxford University Press. Boehm, Miren and Maité Cruz. 2023. “Time for Hume’s Unchanging Objects.” Philosopher’s Imprint 23: Article 16. https://doi.org/10.3998/phimp.1626

Bibliography  145 Costa, Michael J. 1990. “Hume, Strict Identity, and Time’s Vacuum.” Hume Stud‑ ies 16: 1–16. Dainton, Barry. 2008. The Phenomenal Self. Oxford: Oxford University Press. Falkenstein, Lorne 1997. “Hume on Manners of Disposition and the Ideas of Space and Time.” Archiv für Geschichte der Philosophie 79: 179–201. Garrett, Don. 1997. Cognition and Commitment in Hume’s Philosophy. Oxford: Oxford University Press. Garrett, Don. 2009. “Difficult Times for Humean Identity?” Philosophical Studies 146: 435–43. Hynes, Darren. 2005. “Berkeley’s Corpuscular Philosophy of Time.” History of Philosophy Quarterly 22: 339–56. Loeb, Louis E. 2002. Stability and Justification in Hume’s Treatise. Oxford: Ox‑ ford University Press. Mellor, D. H. 1998. Real Time II. London: Routledge. Price, H. H. 1940. Hume’s Theory of the External World. Oxford: Clarendon Press. Rocknak, Stefanie. 2013. Imagined Causes: Hume’s Conception of Objects. New York: Springer. Skrzypulec, Blazej. 2022. “Silence Perception and Spatial Content.” Australasian Journal of Philosophy 100: 524–38.

4 Identity

Hume claimed that our attribution of duration to unchanging objects arises from a fiction of the imagination. He further maintained that our notion of identity is also acquired by means of this fiction. This chapter charges that Hume’s account of how we acquire the notion of identity is impeded and ultimately frustrated by his appeal to the fiction, and by his background commitment to the thesis that unchanging objects cannot properly be said to endure. To justify this charge, his account of identity is presented by means of elliptical quotation. References to a fictitious duration of un‑ changing objects are dropped while retaining the rest of what he had to say. The resulting “free” Humean identity theory is shown to be preferable to Hume’s own account. This chapter concludes by arguing that Hume’s appeal to the fiction serves no larger sceptical purpose in Treatise 1.4. 4.1

Hume’s forensic conception of identity

When discussing identity, as when discussing most other ideas, Hume was not concerned with the meaning or use of words. In particular, he was not concerned with identity in the signification of terms (as when saying that “Cicero” denotes what “Tully” denotes) or with identity considered as a relation between words and objects (as when saying that “Cicero” denotes the one person who wrote de Officiis). He was, as usual, concerned with the origin of ideas. His aim was to identify the original impressions from which the idea of identity is derived.1 He began by maintaining that the idea cannot arise from a single “ob‑ ject” (or impression). If all we meant by saying that an object is the same with itself were that “the idea express’d by the word, object, were no ways distinguish’d from that meant by itself; we really shou’d mean nothing” (T 1.4.2.26). In writing this, Hume was maintaining that identity is ir‑ reflexive. Whenever we speak of something being identical with itself, we mean something else by “itself” than simply that very thing. This is not as radical as it seems. Hume’s point was that he was not concerned with DOI: 10.4324/9781032677880-5

Section 4.1  147 identity understood as sameness in all respects. He maintained that same‑ ness in all respects is properly given a different name: “unity.”2 Hume added that the idea of identity does not arise from a multiplicity of “objects” (or impressions). Two different “objects” can no more give us the idea of identity than a single “object” because “the mind always pronounces the one not to be the other” (T 1.4.2.27). It may seem that this leaves no source for the idea of identity. But this assumes that if identity is to be found anywhere it can only be among units, considered either each in relation to itself or in relation to other units. It might instead be understood as a relation that holds between units and some larger whole. Think of the faces of an opaque, solid body or the suc‑ cessive appearances of an enduring object. In these cases, all of the whole cannot be perceived at once. While “the mind” always pronounces one face or appearance to not be the other, it can identify multiple faces or appearances with the same whole. It thereby identifies the faces or appear‑ ances with one another, as parts or instances of the whole.3 Hume proceeded to make this very observation, though limited to the temporal case. (HS): Hume’s (abbreviated) solution to the difficulty concerning identity To remove this difficulty [what gives us the idea of identity given that it does not arise from a single impression and does not arise from a number of distinct impressions], let us have recourse to the idea of time or duration. (T 1.4.2.29) We cannot, in any propriety of speech, say, that an object is the same with itself, unless we mean, that the object existent at one time is the same with itself existent at another. By this means we make a difference, betwixt the idea meant by the word, object, and that meant by itself, without going the length of number, and at the same time without re‑ straining ourselves to a strict and absolute unity. (T 1.4.2.29) Going by as much of this paragraph as has been quoted,4 the kind of iden‑ tity Hume was concerned with is not term identity, or term/object identity, or sameness in all respects, but what has since been called “genidentity,” that is, identity over time (Grünbaum 1967, 57). Put in eighteenth‑century terms, this is the sense of identity that Locke labelled “Forensick” (Essay 2.27.26). It is the sense that is applied on the judgement day, when the Judge “identifies” the finite spirits then existent with those who performed the deeds of prior times. As Garrett (1997, 163) has noted, this sense of identity was a matter of some concern in the period just before Hume wrote.

148  Identity There is of course a problem with taking Hume to have solved his dif‑ ficulty by appealing to forensic identity. He maintained that unchanging objects cannot properly be said to endure (T 1.2.3.11).5 It follows that they could not properly be considered to appear over different times. There are no temporally distinct appearances to identify with them or thereby with one another. However, in a passage inserted where the ellipses occur in (HS), Hume dismissed this problem. (E): Hume’s excuse for supposing that unchanging objects endure I have already observ’d, [fn.: Part 2. Sect. 5.] that time, in a strict sense, implies succession, and that when we apply its idea to any un‑ changeable object, ’tis only by a fiction of the imagination.… This fiction … almost universally takes place; and ’tis by means of it, that a single object, plac’d before us, and survey’d for any time without our discovering in it any interruption or variation, is able to give us a notion of identity. (T 1.4.2.29) According to Hume, it makes no difference that unchanging objects cannot properly be said to endure because almost nobody ever thinks properly about them and it is the common conception of identity that needs to be accounted for. (E) makes it clear that Hume was concerned with “a single object, plac’d before us, and survey’d for any time,” and so with forensic identity attributions. And he was not just concerned about such attributions in cases where the object moves or alters or is not constantly observed, but in cases where an uninterrupted and in‑ variant object is considered, however improperly, as existing “for any time.” Indeed, the latter case is the paradigmatic one for Hume. (E) claims that (Pretence), the explanation Hume had earlier offered for why we feel impelled to pretend that unchanging objects endure (T 1.2.5.29, see Section 3.4), also explains what gives us our notion of identity.6 However, Chapter 3 argues that (ND), Hume’s claim that unchanging objects cannot properly be said to endure, has no other foundation than his say‑so (Sections 3.3 and 3.4). It further argues that (Pretence), which claims that we pretend that unchanging objects endure because we confuse them with a succession of coexistent perceptions, presupposes that we re‑ ally do experience unchanging objects as enduring over the time taken up by the coexistent perceptions; otherwise, it would be inexplicable why we purportedly confuse the one with the other (Section 3.4). Given these chal‑ lenges, it may do Hume’s philosophy more of a service to consider what he

Section 4.2  149 would have been able to say about the origin of the idea of identity without appealing to (ND) and (Pretence).7 In Sections 4.2 and 4.3, Hume’s account of identity is presented by means of elliptical quotation. His references to the merely fictitious dura‑ tion of unchanging objects are omitted, with some small adjustments to the remainder of what he wrote consequent on those omissions. The resulting identity theory is referred to as “free Humean identity theory.” Free Hu‑ mean identity theory draws only on principles Hume can be supposed to have accepted, but is free of the Humean doctrines that unchanging objects cannot properly be said to endure and that they are only supposed to do so because we imagine them to participate in the changes of some coexist‑ ent succession of perceptions. Free Humean identity theory is also Locke’s theory of identity. Section 4.4 studies Hume’s own identity theory and argues that the free theory is preferable. 4.2

Free Humean identity theory (T 1.3.2.2)

When identity is understood forensically, that is, as paradigmatically con‑ cerned with determining which currently existing soul performed which earlier deed, it follows immediately that it must be a one–one relation that ranges over earlier and later instances of “objects” or impressions. For any given moment of time, identity holds between exactly one instance present at that moment and exactly one instance at each earlier or later moment. It continues to hold for as long as that is the case. If there comes to be a time when there is no contender for the status of the later instance, or more than one, identity ceases, or identity attributions are at least suspended pending disambiguation. If there was an earlier time when there was no contender, or more than one, identity only arises after that time or identity attribu‑ tions are at least suspended pending disambiguation. This is the essential logical feature of the notion of forensic identity. The one–one condition solves “Hume’s difficulty” concerning how iden‑ tity could be a medium between unity and number. As a relation between earlier and later instances, forensic identity involves multiple, temporally distinct instances. But, as a one–one relation between earlier and later in‑ stances, forensic identity also involves a kind of oneness. No instance ex‑ tant at one time can be identified with more than one instance extant at any other time. On the assumption that the individuation of members of the same species requires existence at different places at the same time, a one–one relation between earlier and later instances rules out the possibil‑ ity of an instance being related to numerically distinct individuals. Forensic identity therefore constitutes a middle ground between number and unity. Hume himself put it best.

150  Identity We cannot, in any propriety of speech, say, that an object is the same with itself, unless we mean, that the object existent at one time is the same with itself existent at another. By this means we make a difference, betwixt the idea meant by the word, object, and that meant by itself, without going the length of number, and at the same time without re‑ straining ourselves to a strict and absolute unity. (T 1.4.2.29) But this (trivial) logic of the forensic identity relation is not enough for a theory of identity. It leaves an unanswered question: What leads us to con‑ sider an earlier or later instance to be a contender for identification with a given instance? Hume first answered this question at Treatise 1.3.2.2. In what follows that passage has first been butchered to extract a partial answer. The full passage, which follows the butchered version, deals with the identity of “objects” that have been observed only intermittently, but it also adds an important condition to the basic answer. We readily suppose an object [to be] individually the same … and as‑ cribe to it an identity … whenever … it [conveys] an invariable and uninterrupted perception. We readily suppose an object may continue individually the same, tho’ several times absent from and present to the senses; and ascribe to it an identity, notwithstanding the interruption of the perception, whenever we conclude, that if we had kept our eye or hand constantly upon it, it wou’d have convey’d an invariable and uninterrupted perception. Like Treatise 1.2.1–5, this passage is unapologetically written from an ex‑ ternal realist viewpoint, where it is supposed that there are external objects that affect the senses over time, thereby “conveying” enduring perceptions that could possibly be interrupted by the motion of the object out of range of the sense organ. The butchered version of the passage identifies two con‑ ditions under which we are led to suppose that our successive perceptions are conveyed by or constitutive of “individually the same” object. They must be invariant and uninterrupted. The full passage adds a counterfactual qualification concerning cases where the succession is interrupted. In these cases, Hume wrote, we sup‑ pose the earlier and later instances to be individually the same provided we have reason to think that, had we kept an eye or hand constantly on the object, there would have been no variation or interruption. This remark draws attention to a third condition on all identity ascrip‑ tions. For Hume, vision and touch are the two senses that inform us of

Section 4.2  151 how things are disposed in space. Insofar as keeping an eye or hand con‑ stantly on the object is important for identity ascriptions, identified objects must be disposed in space as well as time.8 Hume was not alone in pursuing this line of thought. Locke began his own discussion of identity over time with the following remark on individuation. (IP): Individuation principle When we see any thing to be in any place in any instant of time, we are sure, (be it what it will) that it is that very thing, and not another, which at that same time exists in another place, how like and undistinguish‑ able soever it may be in all other respects[.] (Locke, Essay 2.27.1) Objects may be “specifically” different from one another. That is, they may be members of different species. In that case, they are distinguished by the “specific differences” that they display. But there can be multiple members of the same species, members that to our eyes exhibit no differences in quality, like fireflies. For Locke, such individuals are distinguished from one another by their disposition in space at a time.9 Locke added that what he called the “relation” to the place and time of a thing’s origin determines its identity for as long as it lasts (Essay 2.27.2–3). Presumably, that relation is the history of motion and rest linking the thing at any subsequent time and place in its existence with its place and time of origin. Hume made the same point when taking the idea of identity to depend, if only counterfactually, on constant observation. Constant obser‑ vation reveals where the object is at each successive moment. Change of place should therefore have no effect on the identity of an object. Hume said as much: “Motion to all appearance induces no real nor essential change on the body, but only varies its relation to other ob‑ jects” (T 1.4.5.27). Motion does not affect the identity of corpuscles or the masses of matter they constitute (T 1.4.6.8). However, keeping an object under constant observation assumes we have some way of verifying, with each passing moment, that the object has not changed under our eyes or within our grasp. Locke gave an answer to what more is necessary. (Hume was later to give the same answer.) (PI‑L): Locke’s principle of identity And in this consists Identity,10 when the Ideas it [the identified object?] is attributed to [when the ideas that are attributed to it?] vary not at all from what they were that moment, wherein we consider their former existence, and to which we compare the present. (Essay 2.27.1)

152  Identity For Locke, as for Hume, identity over time requires invariance. But this appears to create a vicious circle. Invariance is consistent with numeric difference (T 1.3.1.1, (IP)). To address that, Locke and Hume appealed to constant observation. But that requires identifying things over time, which requires invariance (Compare Stroud 1977, 103–4; O’Shea 1997, 184–7). The circle can be avoided by appealing to the one–one condition.11 When an object is kept under constant observation, the one–one condition re‑ quires that, with each passing moment, there are no rivals. In line with the uninterruptedness and invariance conditions, rivals are equally proximate to the prior location of the thing being identified (minimizing the chance of an interruption in motion), and equally similar to it.12 Consider a judge who experiences an element (for Locke, an idea, for Hume a component of the complex impression constitutive of a visual or tactile field). Defeasibly, with each passing moment, the identical element is the one that is (i) most proximate to the place of the original element and (ii) otherwise indistinguishable from the original element. The more time that passes between two views the less reliable the identification becomes. At the stage where the views are “uninterrupted,” that is, immediately successive, maximal proximity and maximal resemblance provide a fairly reliable criterion. In the case where there is no later element that bears any remarkable re‑ semblance to the original element, the original element has disappeared or been destroyed. In the case where there is no such earlier element, the later element has only just emerged. In the case where two or more otherwise indistinguishable elements appear at locations equally accessible from the location of the original, it is uncertain which, if any, to identify with the original. In the case where there is exactly one later element that resembles the original, but one or more otherwise indistinguishable earlier elements occupy locations that are equally accessible relative to the location of the later element, it is uncertain which, if any, of the earlier elements to iden‑ tify with the later one. The difficulty of precisely determining the closest resemblance and distance adds a further level of uncertainty. Forensic identity attributions ought to be one–one beyond a reasonable doubt. A weaker resemblance or greater distance need not frustrate identity attributions, provided there are no rivals. However, in cases where the ob‑ servation is interrupted, the greater the time between views or the greater the distance between earlier or later candidates, the greater the chances that one of the four complications (no predecessors, rival predecessors, no successors, or rival successors) has arisen without being noticed. We realize this and proportion our confidence in our identity attributions accordingly. Learning can come to play a role in identity attributions. Once we have identified objects, and further identified species of object, and studied how objects of a species behave over time, and how objects of one species

Section 4.3  153 succeed objects of another species, and so have discovered causal relations, we can draw on that information to further refine our identity ascriptions. These enhancements enable us to be more confident of whether a replica could have been substituted for the original in cases where the view has been interrupted. Whenever we discover such a perfect resemblance, we consider, whether it be common in that species of objects; whether possibly or probably any cause cou’d operate in producing the change and resemblance; and according as we determine concerning these causes and effects, we form our judgment concerning the identity of the object. (T 1.3.2.2) However, despite Hume’s suggestion to the contrary (T 1.3.2.2), the ap‑ plication of causal inference to enhance identity attributions is a further development. We can only discover what is “common in that species of objects” or “whether possibly or probably any cause cou’d operate in pro‑ ducing the change or resemblance” after we have identified the species of object. That identification presupposes having identified at least some of the individuals characteristic of that species. That identification is in turn based on inference from maximal resemblance and proximity from mo‑ ment to moment. Causality presupposes identification; identification does not presuppose causality (Price 1940, 46; see also 6–9). When identification by maximal resemblance and proximity is success‑ ful from moment to moment, the elements that are identified later cannot be individuated from those that were identified earlier. Individuation re‑ quires that the individuated things exist at different places at the same time whereas these exist at different times. They are, moreover, not qualitatively different and so not specifically distinct. In virtue of their lack of individual and specific distinctness, a succession of identified, merely momentary el‑ ements is a single (identical), enduring object. The difference between a succession of momentary, otherwise exactly resembling elements and an enduring object is merely nominal. Retrospectively identifying a single per‑ ception to have endured is simply other words for progressively finding relations of one–one invariance and proximity to hold between a succes‑ sion of maximally brief, closely resembling perceptions. The distinction between “endurance” and “perdurance” collapses. 4.3

The two surveys (T 1.4.2.29–30)

According to Kemp Smith, at Treatise 1.4.2.29, Hume “[virtually recanted] his first, casual treatment of identity in T 1.1.5 and 1.3.1, where it is re‑ garded as a genuine, non‑fictitious special type of relation” (1941, 475n).

154  Identity Kemp Smith attached this observation as a note to a citation of the last two sentences of (HS), which suggest no such thing. He would have done better to appeal to (FP), a passage cited in Section 4.4. Aside from (FP), another insertion to Treatise 1.2.4.30 to be noted momentarily, and an odd conclu‑ sion to what is below titled the “prolonged survey,” the identity theory of Treatise 1.4.2 falls into line with that of Treatise 1.3.2.2 (and 1.1.5.2 and 1.3.1.1). At Treatise 1.4.2.30 Hume reiterated what he had said at Treatise 1.3.2.2 about the criteria for identity, being careless, as usual, over the distinction between identity and individuation.13 (PI‑H): Hume’s principle of identity Thus the principle of individuation is nothing but the invariableness and uninterruptedness of any object, thro’ a …14 variation of time, by which the mind can trace it in the different periods of its existence, without any break in the view, and without being oblig’d [by the discovery of rivals] to form the idea of multiplicity or number. The details of how we apply these criteria are further worked out over the course of a portion of Treatise 1.4.2.29 that was earlier elided from (HS). It draws inferences from two ways of “surveying” the points of time. One of the surveys is instantaneous, the other prolonged. [The instantaneous survey] [W]hen we consider any two points of … time, we may … survey them at the very same instant; in which case they give us the idea of number, both by themselves and by the object; which must be multiply’d, in order to be conceiv’d at once, as existent in these two different points of time[.] Hume’s claim that we may survey any two points of time at the very same instant is puzzling. Elsewhere, he maintained that “time or succession, tho’ it consists … of parts, never presents to us more than one at once” (T 2.3.7.5). We cannot even imagine two different times at the same time. [T]he incompatibility of the parts of time in their real existence sep‑ arates them in the imagination.… Every part must appear single and alone, nor can regularly have entrance into the fancy without banishing what is suppos’d to have been immediately precedent. (T 2.3.7.5) A survey of successive points of time that takes only an instant should be impossible.15 For Hume, the idea of extension is an extended idea.16

Section 4.3  155 This ought to hold for temporal as well as spatial extension. The idea of two successive times ought to be an idea that takes time to occur. An instantaneous survey of successive moments could still take the form of a spatial model, like the spatial models found in Section 3.4.1. Such models map ideas disposed over successive times (or tokens representative of them) onto a line in space. Because the instantaneous survey models what is disposed over parts of time as coexistent at different places at a moment, it individuates those contents. It makes it appear as if the contents exist at different places at the same time, thereby satisfying (IP), Locke’s individuation principle. In doing so, it also paints a “perdurantist” picture, on which what exists over time consists of a succession of uniformly and maximally brief temporal parts. In fact, the contents do not exist at different places at the same time. They are only modelled as being so. The appearance that they are individuated is, to use a Humean phrase, a “fiction.” Hume did not say so himself, however. He only observed that the instantaneous survey gives us the idea of number. Whereas the instantaneous survey makes the contents of moments ap‑ pear individuated, the prolonged survey identifies them (or some of them, supposing they meet the right conditions) with an enduring object. [The prolonged survey] Or on the other hand, we may trace the succession of time by a like succession of ideas, and conceiving first one moment, along with the object then existent, imagine afterwards a change in the time without any variation or interruption in the object; in which case it gives us the idea of [1739:] unity. (T 1.4.2.29) 1739 concludes that the prolonged survey gives us the idea of “unity,” but that is impossible.17 As Hume described it, the prolonged survey reveals (or can reveal under the right circumstances) an invariant and uninterrupted object. “Invariant” and “uninterrupted” are descriptions that cannot be applied to what is “conceive[d] first” as occupying “one moment.” What is merely momentary could change or be interrupted at the next moment. It is only when the survey lasts for more than one moment that we dis‑ cover whether what is being observed is invariant. We become sensible that our previous ignorance has been relieved by the observation of something that did not exist before: either a continuation of the previous, momentary object, or its replacement by something else. Either way, tracing the suc‑ cession of time by a like succession of ideas means having as many ideas in succession as there are successive moments of time. If there is only one moment of time, we make no progress in determining whether the object

156  Identity is invariant. If there is more than one, and the ideas disposed at those moments are relatively invariant, then by (PI‑H) they can be judged to be identical. Nothing turns them into a unit. It might be objected that Hume wrote that while we “conceiv[e] first one moment, along with the object then existent,” we only “imagine afterwards a change in the time without any variation or interruption in the object” (my stress). But he also wrote that we imagine it afterwards. Imagining is a form of conceiving, so we still have two successive thoughts; otherwise, we are performing the instantaneous survey. When we have the second thought, we are conscious of something that is invariant (except perhaps for a change in vivacity, if we really are only imagining, but supposing we are concerned with making a true identity attribution we will be looking for an impression), but that is present at the current moment of consciousness. There is here a succession of otherwise identical perceptions. A few lines earlier, Hume wrote that the two surveys are produced by “a single object, plac’d before us and surveyed for any time without our discovering in it any interruption or variation.” The progressive survey is described as itself taking time to occur (“we … trace the succession of time by a like succession of ideas”). The later stages of the survey are said to lead us to “discover”18 something that we did not know when the survey started: a second view of the initially experienced object that proves it to be invariant and uninterrupted. The instantaneous survey is necessarily retrospective. Events must have run their course before we can survey them in an instant. The prolonged survey also has a retrospective aspect. At any given moment after the first, the previously experienced moments figure as components of a temporally extended complex impression. But the prolonged survey is also progres‑ sive. The progress of the prolonged survey is like the progress of the visual field as it sweeps from left to right through a static environment. New information appears on one side, sweeps across the field, and disappears off the opposite side. The field must extend some little way backwards in time on pain of it being impossible for us to have any experience of suc‑ cession (see Sections 1.8 and 3.1). As the prolonged survey progresses, we discover whether what is given at each successive moment includes some content that is invariant relative to content given at the immediately prior moment and in the proximate neighbourhood. Those revealed invariance and proximity relations lead us to judge, as of the later moments and so retrospectively, that some of the successive content has remained the same over time. We do not individuate the successive contents because we do not represent them as existing at different places at the same time. We instead identify them as instances or appearances of an enduring whole. On this account, the prolonged survey does all the work in giving us the idea of identity. It progressively represents what exists at each successive moment as a successor and it retrospectively gives us the idea that some

Section 4.4  157 of what is found at the successive moments is invariant and uninterrupted and maximally proximate to what is found at the preceding moments and so revelatory of an enduring whole. Hume himself described the instantaneous and prolonged surveys as giv‑ ing us the ideas of number and unity, respectively. However, the multiplic‑ ity arising from the instantaneous survey is fictional whereas unity is only temporary, appearing just at the first moment of the prolonged survey. The two surveys are better described as giving us the fictional idea of perdur‑ ance and the true idea of an individual that endures over time. On this account, identity, uncovered by the prolonged survey, is a true “medium” between unity and number. The medium is not just flipping back and forth between a number discovered by the instantaneous survey and a unity dis‑ covered by the prolonged survey. 4.4

Hume’s identity theory

Hume opened his discussion of the idea of identity by declaring that it can‑ not be derived from either a single “object” or from multiple “objects.” He initially proposed to remove this difficulty by having recourse to the idea of time (T 1.4.2.29). But, given his prior commitments, this recourse seems impossible. According to Treatise 1.2.3.11, unchanging objects can‑ not properly be said to endure over time, whereas according to Treatise 1.4.2.27 changing objects cannot be identical. Since there can be no “me‑ dium” between changing and not changing, any more than between unity and number, recourse to time would appear to provide no hope of account‑ ing for identity. Free Humean identity theory deals with this problem by allowing that unchanging objects can properly be said to endure. Hume himself, however, proposed a different solution. (FP): Fictitious participation I have already observ’d, [fn.: Part 2. Sect. 5.] that time, in a strict sense, implies succession, and that when we apply its idea to any unchangeable object, ’tis only by a fiction of the imagination, by which the unchange‑ able object is suppos’d to participate of the changes of the co‑existent objects, and in particular of that of our perceptions. This fiction of the imagination almost universally takes place; and ’tis by means of it, that a single object, plac’d before us, and survey’d for any time without our discovering in it any interruption or variation, is able to give us a notion of identity. (T 1.4.2.29) This passage discovers a medium, but it is not a medium between unity and number. It is a medium between changing and not changing. The medium

158  Identity is being unchanging but fancifully supposed to participate in the changes of other things. We can have recourse to time after all, because even though unchanging objects cannot properly be said to endure, we “fancy we have [the] idea” (T 1.2.5.29) that they do, and this “fiction” almost universally takes place. Having discovered this alternative medium, Hume had to explain how it gives us the idea of identity. This puts pressure on his account of the two surveys, particularly on the prolonged survey. The strain appears by com‑ paring the conclusion of (FP): ’tis by means of [this fiction], that a single object, plac’d before us, and survey’d for any time without our discovering in it any interruption or variation, is able to give us a notion of identity with that of the prolonged survey: we may … imagine afterwards a change in the time without any variation or interruption in the object; in which case it gives us the idea of unity. According to (FP), surveying an uninterrupted and invariant object for any time is able to give us a notion of identity. According to the prolonged sur‑ vey, imagining a change in the time without [experiencing] any variation or interruption in the object gives us the idea of unity. Since the project is to account for what gives us the idea of identity, it is possible that Hume or someone in the printer’s office might have inadvertently substituted “unity” for “identity,” as the Nortons have suggested (T2, 783). This having been said, Hume would have had some reason to prefer “unity” to “identity”, and an imagined change in time to a survey that does take time. The prolonged survey describes us as “trac[ing] the succession of time by a like succession of ideas.” According to (FP) the succession of time is marked by “the changes of the co‑existent objects, and in particular of that of our perceptions.” We trace the succession of time by first having one perception and then having a different perception. We then “suppose” the “unchangeable” object “to participate of [these] changes.” What would lead us to make such a supposition? Hume might have anticipated a likely answer: when we first have one perception, A, we also experience the “unchangeable” object, C. When we subsequently have the second perception, B, we must continue to experi‑ ence C on pain of being led by experience to conclude that it has disap‑ peared. There could be no credible supposition of the participation of C with the change of our perceptions unless we continue to be conscious of C while being conscious of B. In being conscious of two successive complex

Section 4.4  159 impressions, A + C and B + C, we are twice conscious of C, once at an earlier moment in the company of A, and then at the subsequent moment, in the company of B. At the second moment and only then, we “discover,” according to (FP) the later instance of C to be unchanged from the earlier instance and so identify them, thereby taking C to have endured over two successive moments. It seems that our impression and idea of the “un‑ changeable” object must occur over at least as many successive moments as there are successive, changing perceptions, on pain of our failing to witness the participation of the former with the changes of the latter as we trace the succession of time by a like succession of ideas. Seeing this answer coming up, but being unwilling to allow that “un‑ changing” objects can properly be said to endure, Hume may have done what he could to block it. He wrote that when we perform the prolonged survey we conceive “first one moment, along with the object then exist‑ ent” but then only go on to “imagine” that the time changes without ex‑ periencing any change in the object. Apparently under the pressure of the conclusion that the change in time is only imagined, Hume denied that there could be any earlier and later instances of the unchanging object to be identified. We could only obtain the idea of unity, not that of identity. But then where does the idea of identity come from? Hume’s only subsequent answer to what gives us the idea of identity is given in the sentence immediately following his account of the prolonged survey. Here then is an idea, which is a medium betwixt unity and number; or more properly speaking, is either of them, according to the view, in which we take it: And this idea we call that of identity. (T 1.4.2.29) With this pronouncement, Hume admitted that, on his account of the two surveys, we do not, “properly speaking” obtain an idea that is a medium be‑ tween unity and number. We only manage to obtain an idea that alternates between these incompatible alternatives depending on the view (Baxter 2008, 63, 66–7, following Price 1940, 40–1 and Fogelin 1985, 72–3). But Hume was not done. Treatise 1.4.2.29 closes with a further change of course. We cannot, in any propriety of speech, say, that an object is the same with itself, unless we mean, that the object existent at one time is the same with itself existent at another. By this means we make a difference, betwixt the idea meant by the word, object, and that meant by itself, without going the length of number, and at the same time without re‑ straining ourselves to a strict and absolute unity.

160  Identity Having just said that the idea of identity arises from alternating between views of the object as multiplied and as a unit, Hume returned to his original proposal to “remove this difficulty” by having “recourse to the idea of time” (outset of T 1.4.2.29) and declared that when we ascribe identity to an object, we mean that it has endured over time without undergoing any change. The surveys were supposed to explain what gives us that idea. The prolonged survey could have been invoked to that end. But Hume did what he could to avoid presenting the prolonged survey as a solution and instead declared that we are left with a quite different notion of identity, one on which the word names an alteration between contrary ideas. Nothing has been said about how that notion could produce the idea of being the same as itself at different moments. We are left baffled by proclamations that cannot be integrated to form a coherent theory. This is a difficult subject and Hume did what he could with it, making considerable progress notwithstanding the impediment he laboured under. Rather than try to fit all the pieces together, it may do him more of a service to draw on a portion of what he said to show how he might have provided a better account. A Humean account of identity does not depend on (FP) and does not require considering the time over which an unchanging object is surveyed to be “supposed” or “imagined.” Hume’s commitment to (ND), the thesis that unchanging objects cannot properly be said to endure, requires those additions. (FP) is not a pillar of his account identity. (ND) is a hurdle that tripped him up on the way to giving a better one. 4.5

Identity in the context of scepticism with regard to the senses

It might be objected that (FP) is an important part of a larger project (Price 1940, 44–5). Hume’s discussion of identity is a step in an argument that aims to establish (BB),19 the thesis that the belief in an external world is based on “trivial qualities of the fancy, conducted by … false supposi‑ tions” (T 1.4.2.56). But while Hume could have appealed to (FP) when giving his reasons for (BB), he did not. The only “false supposition” that he mentioned is the “gross illusion … that our resembling [but interrupted] perceptions are numerically the same” and so “uninterrupted, and … still existent, even when they are not present to the senses.” (FP) cannot have been important to him because it justifies (BB). The (purportedly false) supposition that unchanging objects endure pales in comparison to the gross illusion that interrupted perceptions are uninterrupted. Hume’s commitment to (ND) and (FP) is best explained as the effect of a psychological mechanism he deserves credit for having recognized.

Notes  161 the imagination, when set into any train of thinking, is apt to continue, even when its object fails it, and like a galley put in motion by the oars, carries on its course without any new impulse. (T 1.4.2.22) Having been drawn to the conclusion that we cannot conceive an exten‑ sion where there is nothing visible or tangible, Hume was set into the train of thinking that an analogous conclusion must follow concerning time. He fixed on (ND). Notes 1 O’Shea (1997, 175–80) offers an instructive comparison of Hume’s view of identity with the “post‑Fregean” views of Quine, Russell, and Wittgenstein. 2 As O’Shea (1997, 176) has noted, citing Treatise 1.4.2.26, “‘imply’d’ in an af‑ firmation of identity is a difference in content.” Baxter (2006, 123–4, 2008, 63) disagrees, maintaining that identity is flatly incompatible with difference. 3 Price (1940, 47–8) accuses Hume of overlooking this point. Baxter (2006, 123–4, 2008, 63) maintains that Hume’s adherence to the doctrine that unchanging objects cannot properly be said to endure makes it impossible to address what Baxter calls “Hume’s difficulty”: the difficulty of finding a medium between unity and multiplicity. Baxter follows Price (1940, 40–1) in maintaining that, for Hume, identity is a “muddled” (Price) or “mongrel” (Baxter) concept. We are at best able to alternate between strictly and properly thinking of an object as a durationless unit and falling back into the incorrect idea that it successively exists at different moments and so is a multiplicity. Garrett (2009, 437) tries to save Hume from the “difficulty” by maintaining that, for Hume, the things to which we ascribe identity figure as parts of enduring complexes. (The point that Price accuses Hume of overlooking is one Garrett maintains Hume insisted on.) But Garrett’s position is complicated by his recognition that Hume denied that unchanging objects can be properly said to endure. Given that recogni‑ tion, Garrett can only replace Baxter’s attribution of a “mongrel” concept of identity to Hume with an appeal to ignorance of matters of fact that, if known, would have to be resolved on the side of either unity or multiplicity. (Compare Baxter’s reply to Garrett [2009, 452].) Baxter’s position has also been contested by Rocknak (2013, 124–49). Rocknak accepts that, for Hume, the idea of time cannot be derived from a series of exactly resembling perceptions (134–5). But she does not accept that this tenet made it impossible for him to offer a coher‑ ent account of the identity of objects. The objects just have to be “imagined” rather than perceived (138–48). Section 4.4 endorses Baxter’s thesis that Hume is not able to solve his difficulty. However, Sections 4.2 and 4.3 argue that Hume could have provided a better account, one on which our perceptions (not imagined objects) endure and are identified over time. 4 The break marks a lengthy omission between the opening and closing sentences of the paragraph. The omitted portions and the reasons for omitting them are discussed in due course. 5 In keeping with Treatise 1.4.6.8, which declares that motion does not destroy the identity of a “mass of matter,” this chapter speaks only of “unchanging” objects. The use of “steadfast” is avoided because of its association with immobility.

162  Identity 6 Commentators have accordingly focused on whether Hume was able to make this case. Costa (1990, 9–10) argues that (Pretence) offers Hume’s solution to the difficulty concerning identity posed by Treatise 1.2.4.26–7. Price (1940, 40–1), Fogelin (1985, 72–3), and Baxter (2006, 123, 2008, 63) maintain that it accounts for why we think as we do about objects, but is not able to resolve the difficulty, as it leaves us thinking inconsistent things. 7 Compare Price, who after first presenting (1940, 38–41) and then critiquing (46–7) Hume’s account of identity proposes to reconsider it in the light of a “revised and truly Humian notion of Identity” (48). Price’s revised account ap‑ peals to a temporally and qualitatively continuous series. 8 As Hume observed at Treatise 1.4.6.13, when we take sounds, pains, passions, or thoughts to be identical over time we usually do so by identifying a spa‑ tially located, enduring cause (foghorns, injured body parts, objects of desire or aversion, printed words, mental images, and so on). While the cause is identi‑ cal over time its effects are, Hume claimed, only “specifically identical.” We “confound” the specifically identical effects with causes that are numerically identical. As Kant was later to put it, space alone is permanent whereas time and everything in inner sense is in constant flux (B 291). This makes a reference to space essential for attributions of either identity or individuation over time, putting Kant in a position to offer a “refutation of idealism” by appealing to the tenet that mental phenomena can only be identified over time by reference to what exists in space. Conjecturally, Hume’s (and Locke’s) background commit‑ ment to presentism (Section 1.9) drove them to be one with Kant in this insight. 9 Even when discussing the identity of persons, Locke never recognized that con‑ sciousness could float freely, outside of any finite spirit or body. (Lockean finite spirits are located in space and individuated accordingly [Essay 2.27.2].) Per‑ sonality can at best be transferred from one spatial thing to another, like data from one hard drive to another. The theoretical possibility of multiple copies poses a genuine problem. Hume included spatially disposed perceptions in the bundles constitutive of selves. 10 Locke’s comma might have better been replaced with a colon. 11 It might be avoided in other ways as well, perhaps by insisting that unchang‑ ing objects do not endure. But Hume clearly did place the uninterruptedness, invariance, and eye/hand conditions on identity attribution, and he will have taken the one–one condition to go without saying. This makes it appropriate to invoke these four conditions as features of a “free” but still Humean identity theory. 12 Determining proximity does not require perceiving absolute locations in space. The elements that are judged to be identical over time are always components of a visual or tactile field. When perceivers are motionless in a largely stationary but variegated environment, that environment acts as an enduring backdrop against which the positions of mobile objects can be defeasibly determined. When that is not the case, as when everything is in random motion, identity becomes impossible to determine. Notoriously, conditions do not always allow us to make reliable identity attributions. They may even make it impossible to form a conjecture. We are at the mercy of our circumstances, and the world does not always make it easy for us to make reliable conjectures about what is going on. 13 Hume opened his discussion of identity by proposing to “explain the prin‑ cipium individuationis or principle of identity” (T 1.4.2.25). Hume’s “or”

Bibliography  163 notwithstanding, individuation and identity are not the same. Individuation has to do with what differentiates a thing from all others of its kind whereas (forensic) identity has to do with what makes it the same with itself over time. Hume had no sensitivity to the difference. 14 Consistently with the project of presenting a free identity theory, the word “supposed,” has been deleted where the ellipsis occurs. 15 This is not to say that consciousness could not extend over prior parts of a suc‑ cessive series. For then, ex hypothesi, it is not confined to an instant. The nature of our consciousness of the past is a problem on all accounts. See Sections 1.8, 1.9, and 3.1. 16 Consider Treatise 1.2.3.14, “The idea of extension consists of parts” supple‑ mented by Treatise 1.2.3.5, “in every repetition of that idea [of the ‘composi‑ tion of colour’d points, from which we first receiv’d the idea of extension’] we wou’d … place the points in the same order with respect to each other.” Other evidence: “This idea of Rome I place in a certain situation on the idea of an object, which I call the globe” (T 1.3.9.4). “To say the idea of extension agrees to any thing, is to say it is extended” (T 1.4.5.15). “We have no idea of any quality in an object, which does not agree to, and may not represent a quality in an impression” (T 1.4.5.21). 17 The Nortons argue that the 1739 printing is suspect (T2, 783). Baxter argues for the contrary (2008, 107n.27). See Section 4.4 for further discussion. 18 “[A] single object, plac’d before us, and survey’d for any time without our discovering in it any interruption or variation, is able to give us a notion of identity” (T 1.2.4.29). 19 Bad belief.

Bibliography Baxter, Donald L. M. 2006. “Identity, Continued Existence, and the External World.” In The Blackwell Guide to Hume’s Treatise, edited by Saul Traiger, 114–32. Malden: Blackwell. Baxter, Donald L. M. 2008. Hume’s Difficulty: Time and Identity in the Treatise. London: Routledge. Baxter, Donald L. M. 2009. “Replies to Perry, Falkenstein, and Garrett.” Philo‑ sophical Studies 146: 425–33. Costa, Michael J. 1990. “Hume, Strict Identity, and Time’s Vacuum.” Hume Stud‑ ies 16: 1–16. Fogelin, Robert J. 1985. Hume’s Skepticism in the Treatise of Human Nature. London: Routledge and Kegan Paul. Garrett, Don. 1997. Cognition and Commitment in Hume’s Philosophy. Oxford: Oxford University Press. Garrett, Don. 2009. “Difficult Times for Humean Identity?” Philosophical Studies 146: 435–43. Grünbaum, Adolf. 1967. Modern Science and Zeno’s Paradoxes. Middletown: Wesleyan University Press. Kemp Smith, Norman. 1941. The Philosophy of David Hume. London: Macmillan. O’Shea, James R. 1997. “Fictitious Duration and Informative Identity in Hume’s Treatise.” Manuscrito 20: 169–211.

164  Identity Price, H. H. 1940. Hume’s Theory of the External World. Oxford: Clarendon Press. Rocknak, Stefanie. 2013. Imagined Causes: Hume’s Conception of Objects. New York: Springer. Stroud, Barry. 1977. Hume. London: Routledge.

5 The conception and perception of a vacuum

Hume described “our system concerning space and time” as consisting of two theses. The second, that it is impossible to conceive a vacuum, is described as a consequence of the first, that our perceptions are only finitely divisible (T 1.2.4.1–2). The connection is the argument from the non‑entity of unqualified points (Section 2.6). The parts, into which the ideas of space and time resolve themselves, become at last indivisible; and these indivisible parts, being nothing in themselves, are inconceivable when not fill’d with something real and existent. (T 1.2.4.2) Hume returned to the topic of vacuum in Treatise 1.2.5. His project there was not to offer a further argument for the conclusion he had already drawn. It was to explain what we really experience when we “falsly im‑ agine” (T 1.2.5.14) we are perceiving or conceiving a vacuum.1 Though the argument from the non‑entity of unqualified points is objectionable, and Hume’s conclusion that vacuum is inconceivable therefore unproven (Section 2.6), this subsequent effort requires scrutiny. It appears to show that we can sometimes perceive coloured points without perceiving them to be disposed in space. Had Hume been able to prove this, he would have validated Reid’s criticisms of his account of impressions and ideas (Sections 1.2–1.4). There is much at stake here (Section 5.3.6). 5.1

Hume’s strategy

Hume’s attempt to show that we conceive something else when we imagine we are perceiving or conceiving a vacuum is carried out over four stages. First, he argued that someone who has a functioning sense of vision would not get an impression of vacuum were they in an environment

DOI: 10.4324/9781032677880-6

166  The conception and perception of a vacuum “entirely depriv’d of light,” even if they were to turn their eyes on every side (T 1.2.5.5). Second, he argued that someone placed in an environment where noth‑ ing touches them would not get an impression of vacuum even if they were to swing their limbs without encountering anything (T 1.2.5.6). Third, he argued that darkness and motion cannot give us an idea of vacuum when mixed with something visible or tangible (T 1.2.5.7–13). Fourth, he argued that the mixture of darkness or motion with some‑ thing visible or tangible nonetheless leads us to falsely imagine we can form the idea of a vacuum (T 1.2.5.14–21). It is controversial how Hume’s argument proceeds over the third and fourth stages, and whether he was successful. On one interpretation, re‑ ferred to below as the Berkeleian interpretation (because it accounts for the idea of vacuum in the way Berkeley’s New Theory of Vision accounts for the idea of visual depth), when lone bodies appear in the midst of utter darkness or are felt without feeling any resistance in their immediate sur‑ roundings, we experience eye and muscle sensations that are not disposed in any relevant way in space, but that are common to experiences of bodies separated by some visible or tangible extent. The resemblance induces us to imagine that the lone bodies are also separated by that visible or tangi‑ ble extent. But since we do not immediately see or feel anything between the lone bodies, we judge that they must be separated by an extent we do not see or feel. This judgement comes to be so quickly and easily made that we think we perceive an invisible and intangible distance between them, by analogy with the way Berkeley thought we perceive visual depth.2 On a rival interpretation, referred to below as the “Third Distance” interpretation, we immediately perceive a special kind of distance, “in‑ visible and intangible distance,” distinct from vacuum, and we confuse the two.3 In what follows it is argued that Hume most likely intended to offer a Berkeleian argument, but that on either interpretation he failed to demonstrate that we conceive something else when we claim to perceive or conceive a vacuum. As a first step to assessing these rival interpretations, consider how Hume understood and used the crucial terms that figure in his conten‑ tious references to invisible and intangible distance: motion (which bears on intangibility), darkness (which bears on invisibility), distance, and, of course, vacuum. 5.2

Hume’s terminology

Hume offered two definitions of vacuum, one phenomenological, “space, where there is nothing visible or tangible,”4 and the other physical, “exten‑ sion without matter.”5 In what follows, the phenomenological definition

Section 5.2.2  167 is preferred, consistently with his claim to only be concerned with the appearances of objects to our senses (T 1.2.5.26). 5.2.1 Motion and darkness

Motion is commonly defined as change of place. But the “motion” Hume was concerned with is the feeling that accompanies free motion of a limb. He maintained that this feeling cannot convey the idea of extension. The feeling might have successive parts and so give us the idea of time, but these parts “certainly are not dispos’d in such a manner, as is necessary to convey the idea of space or extension” (T 1.2.5.6). Hume further wrote that the idea of darkness is not a “positive” idea (T 1.2.5.5). When we say that there is darkness, we are not describing what we see. We are confessing that we are unable to see.6 ’Tis evident the idea of darkness is no positive idea, but merely the nega‑ tion of light, or more properly speaking, of colour’d and visible objects. (T 1.2.5.5) It is in order to read “no positive idea” as just meaning “no idea” or “nothing positive.” (The implied contrast with a negative idea is moot.) Hume was saying that we have no idea (or impression) of darkness. We do find ourselves in situations where there is no light. In these situations, we do not see darkness. We simply do not see. A man, who enjoys his sight, receives no other perception from turning his eyes on every side, when entirely depriv’d of light, than what is com‑ mon to him with one born blind; and ’tis certain such‑a‑one has no idea of either light or darkness. (T 1.2.5.5) By itself, this is not a significant point.7 It reduces to the trivial proposition that those who have no visual experience really do have no visual experi‑ ence, and so do not have visual experiences of space where there is nothing visible (vacuum) any more than visual experiences of anything else. But that does not make it a useless point. It is needed as a premise in a further argument offered over Treatise 1.2.5.11–12. 5.2.2 Distance

Treatise 1.1.5.1 identifies the idea of distance as the idea of a relation that, like all other such ideas, is obtained by comparing objects. Somewhat in‑ consistently, but rather more suggestively, Treatise 1.3.1.1 notes that there

168  The conception and perception of a vacuum are two kinds of relations, “such as depend entirely on the ideas, which we compare together” and “such as may be chang’d without any change in the ideas.” The relations of contiguity and distance are said to be of the latter sort. [T]he relations of contiguity and distance betwixt two objects may be chang’d merely by an alteration of their place, without any change on the objects themselves or on their ideas; and the place depends on a hun‑ dred different accidents, which cannot be foreseen by the mind. This remark implies that the relations of contiguity and distance only ap‑ ply to things that have “place” relative to other things. Importantly, the places are not determined by qualities of the placed things, like degrees of brightness or heat or other qualities but by something extrinsic to them: the order in which they are displayed in a spatially extended complex impression. At Treatise 1.2.4.6, Hume further remarked that the placement would have to be discrete (Section 2.4). This allows for contiguity to be defined as immediate adjacency. Things are contiguous when they neither overlap at a point nor have a further point between them (T 1.2.4.6). This is by definition possible in a discrete space. Distance is not just non‑contiguity (contrary to Kervick 2016, 65). It is non‑contiguity that comes in different amounts. Things, even things ap‑ pearing amid utter darkness, can be more or less distant from one another (T 1.2.5.10). This means that distance cannot be divorced from some means of perceiving the greatness of distance, even if only vaguely (contrary to Kervick 2016, 66–7 and note 14). Fundamentally, for Hume, distance is a measure of a portion of space or time (Section 5.2.3). He considered it “just, as well as obvious” that the distance between non‑contiguous ob‑ jects is measured by the number of intermediate points.8 While determin‑ ing distances by counting the number of points can be impractical, due to their smallness and confusion, where the difference in the number of com‑ ponent points is large enough, we can estimate differences in magnitude or distance within certain tolerances (T 1.2.4.22, Section 2.4.1). 5.2.3 Textual evidence

The Treatise’s 100+ references to “distance” and “distant” are almost all to a measured portion of space or time, that is, one qualified as being proximate or remote. Hume’s examples include the distance between an observer and an ink spot (T 1.2.1.4; T 1.2.4.7); the invariant distance be‑ tween the walls of a chamber (T 1.2.5.3); the distance between home and a person who receives news of events at home or is travelling back home

Section 5.2.3  169 (T 1.3.8.5); the distance at which “men of good families, but narrow ­circumstances” set themselves from home and family to reduce the con‑ tempt they feel from those around them (T 2.1.11.14–15); and the differ‑ ent effects of remoteness and proximity in space and time on the passions, on title to property, and on our attention to evil and benefit (T 1.3.8.5; T 2.3.7–8; T 3.2.3.9; T 3.2.7.5; T 3.2.8.1; T 3.3.1.15). In all these cases, the specified effects are supposed to covary with distance considered as a meas‑ ure of space or time. As Treatise 1.2.4.19–22 makes clear, that measure is determined by estimating or (under special circumstances even counting) numbers of intermediate places or positions (Section 2.4.1). Treatise 2.2.10.10 does remark that “any great difference in the de‑ grees of any quality is call’d a distance by a common metaphor.” The outstanding example is the “greater distance” discovered in the case of the missing shade of blue (T 1.1.1.10), though, over the course of Treatise 2.2, Hume was more concerned with distances between social ranks and classes and distances related to other forms of superiority and inferiority between people. There are also references to distant notions, ideas, or re‑ semblances. These are made distant by not belonging to the same species and more distant by not belonging to the same genus or higher genus. Treatise 1.3.13.3 mentions another sense of “distance,” that between the premises and the conclusion of a proof or probability, which is taken to be a function of the number of intermediate conclusions that must be drawn on the way to reaching the ultimate conclusion. All these senses of distance rest on spatial imagery. We pick up blue colour swatches and dispose them along a line in order of their degrees of quality and denominate them distant or proximate depending on the number of intermediate swatches on the line. We draw charts with the Kings and Queens on top and the aristocrats lower down and so on, and recognize distances between the ranks depending on how many intermedi‑ ate ranks there are on the chart. We draw branching charts with the high‑ est genus on top and ranks of species and species of species below it and denominate ideas “distant” depending on the number of laterally interme‑ diate branches between them on these charts. We write down demonstra‑ tions on paper, one line for each premise or intermediate inference, and denominate the ultimate conclusion more or less distant from the premises depending on the number of intermediate lines. Hume’s account of the two ways in which the moments of time might be surveyed (T 1.4.2.29, Section 4.3) further suggests that even distance in time can only be summed up by being modelled by distance in space. There is only one possible exception to the rule that distance is either a number of contiguous places, positions, or points in space or metaphori‑ cal distance based on spatial imagery: the references in Treatise 1.2.5 to invisible and intangible distance. And they are only an exception on the

170  The conception and perception of a vacuum assumption that “invisible and intangible distance” is not just other words for “space, where there is nothing visible or tangible.” “Invisible and intan‑ gible” and “not visible or tangible” are logically equivalent phrases. This leaves any difference in the meaning of “invisible and intangible distance” and “space, where there is nothing visible or tangible” to be expressed by the terms, “distance,” and “space.” But there is nowhere else in the Trea‑ tise where distance is taken to be anything other than a measure of space, a measure of time, or a metaphorical distance grounded on spatial imagery. If there is any distinction to be drawn between “invisible and intangible dis‑ tance” and “space, where there is nothing visible or tangible” it would have to be gleaned from the context in which the phrases are used. It cannot be gleaned from ordinary meaning or from Hume’s usage outside of Treatise 1.2.5. 5.3

Motion, darkness, and distance over Treatise 1.2.5.5–13

The first three of the four stages of Hume’s account of what we conceive when we falsely imagine that we are conceiving a vacuum (Section 4.1) occupy Treatise 1.2.5.5–13. Stages 1 and 2 argue that where there is noth‑ ing visible (in “darkness”) or nothing tangible (where there is “motion” without resistance) there can be no “impression of extension without mat‑ ter” (T 1.2.5.5) and nothing “dispos’d in such a manner, as is necessary to convey the idea of space or extension” (T 1.2.5.6). The third, compris‑ ing Treatise 1.2.5.7–13 argues that darkness and “motion” cannot give us “the idea of extension without matter, or of a vacuum” even when mixed with things that are visible or tangible (T 1.2.5.7). 5.3.1 Stage 1 (Treatise 1.2.5.5)

It is incumbent on those who maintain that we can see empty space to say what a visibly empty space would look like. There can only be one answer. Since space or extension consists of contiguously disposed parts, a visibly empty space would be a space composed of parts that are not “filled” with any colour, that is, what is commonly referred to as a “dark” space. As noted in Section 5.2.1, Hume’s counterpoint was to maintain that, when “entirely deprived of light” (T 1.2.5.5), we have no visual expe‑ rience whatsoever. The “darkness” we speak of in this case instead names the failure to have any visual experience. That would include a failure to have experience of space of any sort. 5.3.2 Stage 2 (Treatise 1.2.5.6)

Plausibly, Hume thought that just as a visibly empty space would have to be a space composed of parts that are not “filled” with any colour, so a tangibly empty space would have to be a space composed of parts that are

Section 5.3.3  171 not “filled” with anything solid, taking “solidity” to refer to a feeling of pressure or resistance. In keeping with this conception of tangibly empty space, Hume asked us to envision a being “supported in the air, and … softly convey’d along by some invisible power.” The reference to being “supported in the air” is plausibly construed as a reference to not feeling the earth pressing up on any body part. The reference to being “softly convey’d along,” together with a subsequent reference to “invariable” motion indicates that, though the being moves, it does not feel any surface contact, acceleration, friction, pushes, or pulls. Hume allowed that the being might have other tactile sensations. He spoke of the being moving its limbs “to and fro” and feeling “in that case a certain sensation or impression,” presumably, a sensation of stretching or fatigue. The conclusion Hume drew from this case parallels the one he drew from the visual case. The being who is “softly convey’d along” is “sensible of nothing [external to themselves], and never receives the idea of exten‑ sion, nor indeed any idea.” Even if they move their limbs, they only feel sensations or impressions “the parts of which are successive to each other, and may give … the idea of time” but do not feel anything that can “con‑ vey the idea of space or extension.” 5.3.3 Stage 3 (Treatise 1.2.5.7–13)

There are challenges to showing that darkness and motion cannot give us an idea of empty space when “mixed” with something visible or tangible. Visually, given that “darkness” refers to the absence of any visual experi‑ ence whatsoever (Section 5.3.1), a “mixture” of darkness with something visible would have to reduce to just an experience of what is visible, with no surplus. But then the Cartesian reply to the case of the evacuated cham‑ ber (Principles 2.18, discussed at Treatise 1.2.5.3) is endorsed. Were eve‑ rything visible between the walls of a chamber removed, the walls would have to appear to be contiguous, contrary to Hume’s intentions to estab‑ lish that we perceive something other than either contiguity or an interme‑ diate vacuum in this case. Tangibly, at any one time, we experience multiple muscle sensations and other tactile impressions disposed over a tactile field that corresponds to our own bodies. That experience supplies us with all the resources we need to obtain an idea of extension, with no need to appeal to limb motion. Having had that experience, only one portion of the tactile field needs to be unstimulated or grow numb or become unresponsive for us to get the sense that the remaining tactile sensations are separated by a space where there is nothing tangible. Hume had to address both objections. Costa (1990, 15n.10), Kervick (2016, 76), and Cottrell (2019, 219, 219n29) dismiss consideration of the

172  The conception and perception of a vacuum tactile case in favour of considering just what Hume had to say about vi‑ sion. But a demonstration that we see something else when we think we see a vacuum is worthless if it has to be conceded that touch supplies us with impressions and ideas of unoccupied places. Beginning with the visual case, Hume asked us to consider two points appearing amidst entire darkness. In order, therefore, to know whether the sight can convey the impres‑ sion and idea of a vacuum, we must suppose, that amidst an entire dark‑ ness, there are luminous bodies presented to us, whose light discovers only these bodies themselves, without giving us any impression of the surrounding objects. (T 1.2.5.8) He then observed, (ED): Evident distance ’tis evident, that when only two luminous bodies appear to the eye, we can perceive, whether they be conjoin’d or separate; whether they be separated by a great or small distance; and if this distance varies, we can perceive its encrease or diminution, with the motion of the bodies. But … the distance is not in this case any thing colour’d or visible[.] (T 1.2.5.10) “’Tis evident,” Hume wrote. But what makes it evident? (ED) is followed (beginning at its ellipses) by a conjectural answer. (SV): Sensible vacuum But as the distance is not in this case any thing colour’d or visible, it may be thought that there is here a vacuum or pure extension, not only intelligible to the mind, but obvious to the very senses. “[I]t may be thought that” indicates that Hume did not mean to endorse the answer. Beginning a new paragraph, he wrote, This is our natural and most familiar way of thinking; but which we shall learn to correct by a little reflection. (T 1.2.5.11) It is not immediately clear what natural and most familiar way of thinking “This” refers to. While Hume wanted to reject the consequent of (SV), that consequent is premised on what was said over the remainder of (ED). How much of what was said by (ED) falls under the scope of the “correction”

Section 5.3.3  173 Hume proposed to give? This question is best answered by identifying what earlier pronouncements are corrected by the reflections Hume went on to offer. He proceeded to write, We may observe, that when two bodies present themselves, where there was formerly an entire darkness, the only change, that is discoverable, is in the appearance of these two objects, and that all the rest continues to be as before, a perfect negation of light, and of every colour’d or vis‑ ible object. (T 1.2.5.11) This is what we would expect him to say given what he had said about dark‑ ness (Section 5.2.1). When two luminous bodies appear where there was formerly an entire darkness, they could not appear as being surrounded or separated by darkness, which is not something that is experienced. It might seem that in this case, the bodies would default to appearing contiguous, but that does not fit with Hume’s later claim that bodies can change place without undergoing any alteration (T 1.3.1.1). That implies that there could be nothing in the bare appearance of the bodies that could dictate that they are either contiguous or distant. They are just there, without be‑ ing anywhere in space in relation to one another, like smelling smoke while hearing the shriek of a fire alarm amidst entire darkness.9 Hume continued, This is not only true of what may be said to be remote from these bod‑ ies, but also of the very distance, which is interpos’d betwixt them; that being nothing but darkness, or the negation of light; without parts, without composition, invariable and indivisible. (T 1.2.5.11) The observation that the “distance, which is interpos’d” between the bod‑ ies is without parts, without composition, invariable and indivisible is a correction to (ED)’s claim that we can perceive whether the luminous bod‑ ies are separated by a greater or lesser distance that increases or diminishes with their motion. The bodies could not be observed to be in motion, and they could not be separated by a greater or lesser distance. They could not be observed to be in motion because there is nothing they could be mov‑ ing relative to. They could not be moving relative to the periphery or the centre point of the visual field, neither of which appears. And they could not be moving relative to one another because the darkness between them is invariable and indivisible, not having parts or composition, and so is not liable to increase or diminution. What we “naturally and most familiarly” consider to be darkness interposed between them is, consistently with what

174  The conception and perception of a vacuum was said at Treatise 1.2.5.5, what the blind see. The blind see nothing, so we see nothing. We do not see any distance or extension between the two bodies, and so we cannot see any distance that is greater or lesser. In saying that “the very distance, which is interpos’d betwixt them” is “nothing but darkness” Hume was saying that there is no distance interposed between them, since darkness is not anything that we experience. Now since this distance causes no perception different from what a blind man receives from his eyes, or what is convey’d to us in the darkest night, it must partake of the same properties: And as blindness and darkness afford us no ideas of extension, ’tis impossible that the dark and undistinguishable distance betwixt two bodies can ever produce that idea. (T 1.2.5.11) Hume here wrote three things about the distance that we most naturally and familiarly think is evidently perceived between the lone, luminous bodies. The first is that it has no effect on us. It causes no perception different from what the blind receive from their eyes and the blind receive no perceptions from their eyes. The second is that it must partake of the properties a blind person receives from their eyes, that is, it must not have any properties. The third is that it does not afford us any idea of extension between the two bodies. In writing that “’tis impossible that the dark and undistinguishable distance betwixt two bodies can ever produce that idea” it is not obvious that Hume meant to say that there is any such thing as a dark and undistinguishable distance between the two bodies. He could just as well have meant to say that such distance as we naturally and familiarly think is evidently perceived between the luminous bodies, being dark and undistinguishable, could not produce the idea of extension. This reading is entailed by the definition of darkness of Treatise 1.2.5.5. But then how could we have come to think it “evident” that we can perceive any sort of distance between the bodies, as (ED) admits? (SD): Sole difference The sole difference betwixt an absolute darkness and the appearance of two or more visible luminous objects consists, as I said, in the objects themselves, and in the manner they affect our senses.… [T]hese [effects] produce the only perceptions, from which we can judge of the distance. But as these perceptions are each of them simple and indivisible, they can never give us the idea of extension. (T 1.2.5.12) According to (SD) the appearance of the luminous bodies involves more than just the appearance of those two coloured points, but what more has

Section 5.3.3  175 nothing to do with contiguity or distance or disposition in space. Hume went into more detail over the ellipses in (SD). The luminous bodies are said to affect parts of the eye and to cause it to move in certain ways.10 These “manners of affection” produce quasi‑tactile perceptions that are internal to the eye. From them “we can judge of the distance” (my stress) between the bodies. The distance is not immediately seen but only inferred, as visual depth is not immediately seen but only inferred. Hume underscored this point by writing, “But as these perceptions are each of them simple and indivisible, they can never give us the idea of ex‑ tension.” Being “simple and indivisible,” the perceptions are not individu‑ ally extended. We cannot perceive extension in any one of them and Hume did not consider a perception of extension to arise from their aggregation. We instead learn to associate them with extensions comprised of coloured points (T 1.2.5.15). They come to function as distance cues: things that are not themselves in space but that lead us to judge that distance is present. Hume was invoking an analogue to the Berkeleian account of visual depth perception, according to which we learn to associate purely qualitative features of two‑dimensional visual impressions with distances measured by touch over a third dimension. The only distance that is mentioned over the course of Treatise 1.2. 5. 10–12 is visible distance, comprised of coloured points disposed in space. But that distance can be either directly perceived or only judged to be present, the way visual depth is only judged to be present without being directly perceived. When we judge that the distance is present but do not see it, and the judgement has become so habitual, quick, and easy that we do not notice we are making it, we are tempted to infer that we are seeing distance or space where there is nothing visible comprising that distance. But that is a false judgement leading to a conclusion that the argument from the non‑entity of unqualified points (Section 2.6) shows to be absurd. There can be no such thing as an invisible distance or space where there is noth‑ ing visible. There are only purely qualitative distance cues. (SD) completes Hume’s correction of (ED). According to (ED), it is “evident” that “we can perceive” a distance between lone bodies. This “evident” perception is mediate perception, like the perception of visual distance along the depth dimension. (SD) “corrects” (improves on) (ED)’s coarse observation by identifying what we really experience that leads us to first judge that there is some distance and then falsely imagine that we perceive it. Hume proceeded to “illustrate” this line of thinking by applying it to the tactile case (T 1.2.5.13). As he had earlier characterized it, this case asks whether we obtain an idea of vacuum from perceiving one thing “by the feeling; and after an interval and motion of the hand or other organ of sen‑ sation, another object of the touch” (T 1.2.5.9). The illustration picks up

176  The conception and perception of a vacuum by declaring any “distance or interval” supposed to be interposed between the objects to be only “imaginary.” It does not explain what leads us to im‑ agine that there is a distance (that is done later over Treatise 1.2.5.15–17). It instead focuses on explaining why the experience of motion does not constitute an immediate perception of distance. The explanation appeals to the earlier point that moving the limbs without meeting anything tangi‑ ble can only give us a muscle sensation. That sensation can give us a notion of time but does not consist of parts that are “dispos’d in such a manner, as is necessary to convey the idea of space or extension” (T 1.2.5.5). Modify‑ ing this case by supposing that we felt something before having the muscle sensation, and feel something else after having it, does nothing to change the fact that the muscle sensation conveys no idea of space or extension. Admittedly, Hume did not conclude that the muscle sensation gives us no idea of distance. He only concluded that it gives us no idea of extension. And as that sensation is not capable of conveying to us an idea of exten‑ sion, when unaccompany’d with some other perception, it can no more give us that idea, when mix’d with the impressions of tangible objects; since that mixture produces no alteration upon it. (T 1.2.5.13) But it would be going beyond the evidence to infer from this merely neg‑ ative assertion that Hume meant to draw the positive conclusion that bookending the muscle sensation with earlier and later tactile sensations does convey the idea of distance. Not to mention Treatise 1.2.3.2, which states that the idea of extension is acquired from the experience of dis‑ tance, Hume began Treatise 1.2.5.13 by declaring the distance that we consider to be present in this case to be only “imaginary.” The story about what leads us to imagine it has yet to be told. The only references to distance to be found in Treatise 1.2.5.5–13 are references to visible and tangible distance. This distance is either directly seen or felt, or only judged or imagined to be present. According to a false but “natural and most familiar way of thinking” (T 1.2.5.11), that is, of judging or imagining, it is evident that we can directly perceive whether two bodies, appearing where there was before an entire darkness, are separated by a “dark” distance, that is, by a vacuum. This natural and familiar way of thinking is corrected by the reflection that there is no such thing as a visual experience of darkness. The distance that appears evident to us is not a vacuum evident to the very senses but a distance that is only judged to be present. The tactile case is a good illustration of what must be also happening in the visual case. In that case, we can more easily appreciate that there is nothing between the felt objects. There are only muscle sensations that are nowhere in space but that are felt after feeling

Section 5.3.4  177 one object and before feeling the other. Because these muscle sensations are also felt when moving over intermediate objects, association leads us to imagine that there is a space between the remote objects as well, even though there is none. In both the visual and the tactile cases, we draw a false inference to an absurd conclusion that there is an invisible or intan‑ gible distance. 5.3.4 The “Third Distance” interpretation

Hume’s discussion over Treatise 1.2.5.10–13  lends itself to another in‑ terpretation. He wrote that it is “evident” that we can perceive whether the lone bodies are conjoined or separated (T 1.2.5.10) leaving it unclear whether this perception is immediate or mediate. While he went on to say that this evident distance is “nothing but darkness …; without parts, with‑ out composition, invariable and indivisible” he nonetheless described it as “interpos’d betwixt” the bodies and referred to “the dark and indistin‑ guishable distance betwixt two bodies” (T 1.2.5.11). At Treatise 1.2.5.12 and 13, he further confined himself to remarking that darkness and motion cannot give us ideas of extension. He did not explicitly extend the claim to distance. These turns of phrase make it possible to read Hume as meaning to say that we can immediately perceive lone bodies to be separated by some third kind of distance distinct from both sensible distance (intermediate coloured or tangible points) and empty space. This “Third Distance” reading gains support from Hume’s later references to two kinds or spe‑ cies of distance, one visible or tangible, the other invisible and intangible (T 1.2.5.15–18). However, the Third Distance reading does not sit well with Hume’s declarations that when two lone bodies appear we see nothing but those two bodies (T 1.2.5.11 and 12) and that when we move from one body to another without encountering any resistance our tactile sensations are only successive or distant in time (T 1.2.5.6 and 13). The Third Distance reading also does not sit well with the fact that the “corrective” that Hume offered to (ED) only denies that we see distance over space (extension) without describing any alternative distance that we do see (T 1.2.5.11); with his declaration that we only judge of the distance rather than see it (T 1.2.5.12); or with his declaration that distance is only imagined (T 1.2.5.13). The most serious challenge to the Third Distance reading is Hume’s meaning empirism (Abstract 7 citing T 1.2.3.2, reiterated at EHU 2.9 and 7.4), which demands that all meaningful words and phrases denote ideas that have been compounded from simple impressions.11 Anyone talking about a third kind of distance distinct from both “space, where there is

178  The conception and perception of a vacuum nothing visible or tangible” and filled space needs to identify the original impressions from which the idea purportedly named by the word “dis‑ tance” is copied. Merely saying what the distance is not, or using negative terms like “unextended,” “invisible,” and “intangible” does not identify the original impressions or how they are presented.12 The difficulty of reconciling a Third Distance reading with meaning empirism can be illustrated by considering some Third Distance readings, both early and late.13 C. D. Broad (1961) offered a Third Distance reading that bears some similarity to one offered recently by Dan Kervick (2016). I think that Hume’s two senses of “distance” can be expounded as fol‑ lows. Consider any two points, A and B. Then there are two quite dif‑ ferent, but closely interconnected facts or possibilities to be considered about them. (i) There is the direct relation of spatial separation between A and B. This, being a relation, is indivisible into parts, but it may be greater or less. It might be compared with the ratio between two num‑ bers, e.g. between 13 and 2. (ii) There may be a stretch or sequence of points collinear with A and B and falling between them. This might be compared with the sequence of integers between 2 and 13. I think that Hume’s two senses of “distance” are just the relation of spatial separa‑ tion, and the stretch or sequence of intermediate points which together make up the straight line joining two points. (Broad 1961, 173) Broad’s analogy neglects Hume’s distinction between relations that are determined by the relata and relations that are determined by where the relata are placed (T 1.3.1.1). For Hume, distance is a relation of the second sort. The ratio between 13 and 2 is not. There is a way (or two) to deter‑ mine a ratio between 13 and 2: either divide 13 by 2 or divide 2 by 13. There is something about 13 and 2, namely, the quantity of units they con‑ tain, that determines their ratio by means of the operation of division. This same thing also determines the “stretch or sequence of points … between them.” Broad’s two “quite different, but closely interconnected facts” are not different at all, but simply two ways of describing the same thing. In contrast, there is nothing about two coloured points, compared with one another, that determines their distance from one another, the way there is something about 2 and 13 that determines their ratio and the stretch between them. “[T]he relations of contiguity and distance betwixt two objects may be chang’d merely by an alteration of their place, without any change on the objects themselves or on their ideas” (T 1.3.1.1).

Section 5.3.4   179 What, then, are we seeing when we see the relation of distance to be greater or less? Broad’s answer does not meet the meaning empirist challenge. Let us now consider the visual experiences which correspond to these two senses of “distance.” Suppose you were to look up at the heavens on a pitch‑dark night and to see two stars. You would then be aware of two nearly punctiform visual sense‑data which are spatially separated, but are not joined by a stretch of intermediate visual sense‑data. Here we have the visual experience of distance in the first sense, without dis‑ tance in the second sense. (Broad 1961, 173) According to Broad, visual experience of distance in the first sense is awareness of lone bodies, like stars appearing in utter darkness, as be‑ ing spatially separated, without being separated by coloured points (or “visual sense‑data,” as he put it). This description of “distance in the first sense” does not distinguish it from an experience of an intermediate vac‑ uum. If being aware of stars as “spatially separated” means anything, it means being aware of stars as separated by space. If the stars so separated have no “sense‑data” or coloured points between them then they are sepa‑ rated by a space that contains no coloured points, that is, a space where there is nothing visible. A space where there is nothing visible is a vacuum (T 1.2.5.1). Hume wanted to deny that we can conceive any such space, be it a “thing” or a “non‑thing,” substantival or relational, positive or priva‑ tive. (His definition of a vacuum as “space, where there is nothing visible or tangible” is not qualified or restricted in any further way.) To salvage Hume’s position, Broad would have needed to come up with a better account of the original impressions from which we copy the ideas we purportedly have when we use the phrase “visual experience of dis‑ tance in the first sense.”14 Advancing some 60 years, Jonathan Cottrell (2019, 218) has written that, bodies can be arranged distantly in virtue of being separated by some‑ thing spatial—an extended composite of coloured or tangible points, or what Hume calls a “real extension.” But In Hume’s view, this is not the only way in which bodies can be arranged distantly.… [T]wo luminous bodies in an otherwise dark environment may be arranged distantly.… However, between two such bodies, there

180  The conception and perception of a vacuum is no “real extension”—no extended composite of colored or tangible points. Hume therefore speaks of a second “kind of distance,” which he calls “invisible and intangible distance.” Like Broad, Cottrell does not explain what makes things distant in this sec‑ ond way.15 He just says that it is Hume’s view that there is a second way, and, like Broad and Kervick, he says what does not make things distant in this way: that “Hume misleadingly speaks as if ‘invisible and intangible distance’ were a third thing located between the two luminous bodies”; that “Hume denies that there is anything located between the two lumi‑ nous bodies in a case of ‘invisible and intangible distance’” (219); and that “‘invisible and intangible distance’ is a mere negation or absence” (220). There must be more to it than this. Hume maintained that the luminous bodies can be more or less distant from one another (T 1.2.5.10). Given that the “invisible and intangible distance” can vary, we are entitled to ask what it looks like to see two bodies that are separated by a greater invisible and intangible distance as opposed to a lesser one. Cottrell does say that two bodies are arranged distantly when they are arranged in such a way that a third body could be placed between them without overlapping or displacing either of the original two (218). But he does not say what it looks like for them to be arranged in that way, except in the case where there is such a third body between them. In default of any further explanation, this manner of arrangement can only be known after the fact. We would need to first attempt to insert a line of visible objects between two lone bodies to see if it would fit without overlap or displace‑ ment. If we succeeded, we would still not know whether a yet longer one would fit. Having found the longest one that would fit, we would not know if it would continue to fit after having been removed. The lone bod‑ ies could be moving. Suppose we try to insert one line of visible bodies and it fits. Then we try to insert a longer line and it still fits. What tells us whether we picked too short a line to start or whether the bodies are re‑ treating from one another? It gets worse. Suppose we try to insert one line and it fits. What tells us that the lone bodies remained in place as the line was inserted as opposed to being displaced by the insertion? Something is displaced when it is moved out of its place. It can only be moved out of its place if it was in that place to begin with. But on Cottrell’s account, we can only tell what places lone bodies are in after we have inserted a line of visible or tangible bodies between them. If we saw something else that told us how distant the two bodies are without having to first place a line of visible objects between them, or something that we could learn to associate with the insertion of a line comprised of a certain number of visible points, these problems would not arise. But Cottrell says nothing about what that further datum might be.16

Section 5.3.4  181 Following Hume (T 1.2.5.26n), Cottrell maintains that this is a virtue of his account, due to its rigorous empiricism. [I]n [Hume’s] view, there are cases of “invisible and intangible dis‑ tance,” where two bodies are arranged distantly but have no spatial thing located between them. “[W]e find by experience” that bodies can be arranged in this way (T 1.2.5.25): for example, when we perceive two “separate” luminous bodies in an otherwise dark environment, we perceive them to be arranged distantly (that is, to be spatially related but non‑contiguous) and to have no spatial thing located between them (T 1.2.5.10). And in Hume’s view, “experience and observation” is “the only solid foundation we can give” to the science of man (T Intro 7). (226) Appeal to experience and observation is a fine thing, provided we are capa‑ ble of accurately describing what we find by experience to be the case (and disposed to accurately describe what Hume claimed we find by experience to be the case17). Cottrell appeals to an experience in which we perceive two bodies to be “arranged distantly” (226) that is, “spatially related but non‑contiguous” (226), that is, “arranged in such a way that a third body could be placed between them, without overlapping or displacing either of the original two” (218). This experience is well described. However, Cottrell goes on to stipulate that the bodies have no “spatial thing” located between them. With this additional stipulation, the experi‑ ence is no longer accurately described because it is not clear how two bod‑ ies can be experienced to be “spatially related but non‑contiguous” if they have “no spatial thing” located between them. At a minimum, more needs to be said about what the term “spatial thing” denotes. At 218, Cottrell stipulates that to be separated by “something spa‑ tial” is to be separated by an extended composite of coloured or tangi‑ ble points. Anyone on either side of the debate over the conceivability and perceivability of vacua can agree with this. Those who consider vacua to be empirically obvious will maintain that two luminous bod‑ ies that are not separated by any extended composite of coloured or tangible points can very well be spatially related but non‑contiguous, provided they are separated by a space where there is nothing visible, that is, an “invisible distance,” these being words to the same effect. For these people, the experience Cottrell means to describe is exactly as he describes it, with the proviso that the term “spatial thing” be re‑ stricted to colours and tactile qualities and not extended to include the unoccupied spatial locations that must, in this case, lie between the two luminous bodies for the experience to be coherently described. These unoccupied locations also need to be included in the description of the

182  The conception and perception of a vacuum experience and recognized to be inevitably experienced by way of posi‑ tion and contrast, on pain of accepting that the luminous bodies are perceived to be contiguous. Elsewhere (212) Cottrell includes what he calls “positive vacuum” among spatial things. “Positive vacuum” is defined as “a genuine thing that has spatial properties, and yet is immobile, indivisible, and penetra‑ ble” (206). This is not a clearly significant addition. The word “thing” (whether qualified with “genuine” or not) does not have such ponderous meaning that detaching it from the word “spatial” makes any difference to the dispute. Those who consider vacua to be perceivable and therefore conceivable have no investment in maintaining that vacua are “things” or “genuine” things. They are only invested in maintaining that vacua are perceivable, and so conceivable, by way of relative position and contrast with visible or tangible points. In the case at hand, they are only invested in maintaining that two luminous bodies that have no extended composite of coloured or tangible points between them can only be perceived to be spatially related but non‑contiguous if they are perceived to be separated by unoccupied locations or places, that is, by a space where there is noth‑ ing visible or tangible. Hume himself wrote that he had no interest in the ontological status of empty space and none in whether empty space is a thing or a “genuine thing.” His concern was only with whether we can perceive or conceive it (T 1.2.5.25–6). The only “things” mentioned in his definition of vacuum as “space, where there is nothing visible or tangible” are “things” that are not in space. The definition does not stipulate that the space itself must be a thing of any sort. Plausibly, Hume meant to deny that we can immedi‑ ately perceive or conceive any sort of unoccupied place or space between sensible points, be that place or space labelled “positive” or “privative,” substance or relation, real thing or non‑entity. Cottrell has said that we experience luminous bodies that have no ex‑ tended composite of visible or tangible points between them as being ar‑ ranged distantly and has said this means being arranged in such a way that a third body could intervene without displacement or overlap. If this does not mean that the bodies are perceived to be separated by an unoccupied space at least as large as the body that might be made to intervene between them, we need to be told what it does mean, on pain of giving up on de‑ scribing what we experience that leads us to consider the bodies to be dis‑ tantly arranged rather than contiguous, and arranged at a certain distance rather than a greater or lesser one. The “meaning empirist” challenge must be met: what are the original impressions from which the ideas denoted by the words “arranged distantly” are copied?18

Section 5.3.5  183 5.3.5 Meeting the meaning empirist challenge

Those who maintain that vacua are perceivable can meet the meaning empirist challenge.19 The definition, “space, where there is nothing vis‑ ible or tangible,” refers to space or extension. The idea of space or exten‑ sion is copied from impressions of points disposed outside of one another (T 1.2.3.4). This is as deep as it gets. Anyone wanting to know what “set outside of” means needs to be referred to an apartment building, or a checkerboard, or a necklace. These are the kinds of complex impressions from which complex ideas of parts set outside of parts are copied, and collections of these complex impressions and ideas are the basis for our abstract ideas of space and extension. The idea of a space where there is nothing visible or tangible is a minor modification. It is a copy of a complex impression consisting of parts set outside of parts, where one or more of the parts is not occupied by any‑ thing visible or tangible. Drill a hole through a white wall looking out on total darkness and retreat to the point where the hole appears as a mini‑ mally visible point of darkness. The point has a position on the visual field and appears better by way of contrast with its white surroundings than if it were coloured. It is a spatially located point that is not “filled” with anything visible, that is, a vacuum. Coloured points can be distinguished from their surroundings and sepa‑ rately conceived. So can this point of darkness. If this seems objectionable, consider what it means to “separately conceive” a point of any sort. It does not mean to consider a colour that has no shape or size. That would be an abstract idea of colour, and Hume denied that we have such abstract ideas. We cannot conceive a colour without having the idea of a particular shape and size. That means conceiving it to have edges. There can be no edges where there is no contrast. A colour patch can only be conceived as sur‑ rounded by an area of a contrasting colour or of no colour (which makes just as effective a contrast). A coloured point is no different. It does not extend out to infinity but has limits. It is just that its limits are collapsed to a singularity. It is only conceivable as hemmed in on all sides by points of a contrasting colour or of no colour.20 To “separately” conceive a coloured point is not to conceive it alone, but to conceive it in different surround‑ ings, that is, in separation from its given surroundings. All this is true of unoccupied points. They can only be conceived as surrounded by coloured points, or as disposed alongside one another to compose shapes that ap‑ pear against a surrounding, coloured background.21 Having looked at a drill hole in a white wall, we can paint the wall differ‑ ent colours and find the same unoccupied point at the same place on differ‑ ently coloured visual fields. Or we can shift our viewing position and obtain

184  The conception and perception of a vacuum complex ideas of an unoccupied point at different positions on a white vis‑ ual field. We can compare these complex impressions with other complex impressions of uniformly coloured visual fields on which there is a single point of either a contrasting colour or no colour and note resemblances in where the highlighted point is located relative to the periphery and the cen‑ tre of the visual field. The set of such resembling ideas is the foundation for the abstract idea of that position on the visual field. The idea of a position is accordingly not the idea of a property or quality of a point considered individually, but an abstract idea based on a collection of visual fields all of which highlight a point at a particular position relative to the sides and the centre of the visual field. Position on the tangible field is similar. Those defending a Berkeleian reading of Treatise 1.2.5.10–13 can also meet the meaning empirist challenge. On the Berkeleian reading, we learn to associate motion sensations and feelings in the affected parts of the organs with distances vaguely defined by numbers of coloured or tactile points. When lone‑body cases feature those cues, we imagine the asso‑ ciated coloured or tactile distance. But because we see (or feel) no such thing, we falsely imagine that there must be an “invisible and intangible distance” between the lone bodies, much as we falsely imagine that the ob‑ jects of our two‑dimensional visual experience are set at various distances away from us on a depth axis. If the Third Distance interpretation is to address concerns about mean‑ ing empirism, it can only be by finding evidence for an alternative, positive account of invisible and intangible distance in what Hume said over the ensuing Treatise 1.2.5.14–21.22 5.3.6 Hume’s dangerous concession to Reid

On the Berkeleian reading of Hume’s account, visible objects appearing amidst utter darkness and tangible objects felt after an interval of unim‑ peded motion are originally perceived as nowhere in space. The Third Distance reading of Hume’s account of the mixed cases (Section 5.3.4) leads to the same result, though for different reasons. While it allows that visual impressions appearing amidst utter darkness and tac‑ tile impressions felt after an interval of unimpeded motion are originally perceived as distant from one another, it must deny that the distant impres‑ sions have space between them, on pain of running afoul of the phenom‑ enological definition of vacuum as space where there is nothing visible or tangible. As noted in Section 5.2.3, “nothing visible or tangible” and “invisible and intangible” are logically equivalent phrases. Accordingly, the only difference between the phrases, “invisible and intangible distance” and “space, where there is nothing visible or tangible” is in the words, “space” and “distance.” Affirming that there is an invisible and intangible

Section 5.4  185 distance between impressions while denying that they are separated by a vacuum entails denying that there is any space between them. The distance must be some third sort of distance that has nothing to do with how im‑ pressions are disposed in space. In principle, impressions could display it even though they were nowhere in space. In Sections 1.2–1.4, it was shown that Hume and Reid had opposed views of the nature of mental representation. Hume maintained that it is evident to introspection that phenomenal colour qualia, pains, feelings of heat and solidity and other tactile qualia are all given as disposed at loca‑ tions on sensory fields. Reid maintained, to the contrary that our sensa‑ tions of colour and pain are extensionless and locationless states of feeling experienced by an unextended Cartesian mind. On both the Berkeleian and the Third Distance interpretations of Trea‑ tise 1.2.5.5–13, Hume allowed that visual and tactile qualia can sometimes appear without appearing to be either contiguous or remote in space. This is a dangerous concession. If our visual and tactile sensations could some‑ times be nowhere, they could always be that way. Our contrary opinions could be nothing more than juvenile preconceptions grounded on a confu‑ sion of our internal sensations with direct perceptual apprehension of the qualities of external objects, as Descartes and Reid maintained. In saying what he did about vacuum, Hume was compromising his posi‑ tion on mental representation. 5.4

Berkeleian and Third Distance interpretations of Treatise 1.2.5.14–21

Having argued that “motion and darkness, either alone, or attended with tangible and visible objects, convey no idea of a vacuum” Hume went on to claim that they are nonetheless “the causes why we falsly imagine we can form such an idea” (T 1.2.5.14). The ensuing Treatise 1.2.5.15–21 tries to identify how that happens. Hume made this case over four stages. First, Treatise 1.2.5.15, following on 1.2.5.12, argues that the man‑ ner in which the senses are affected can be the same when experiencing lone‑body pairs and when experiencing bodies that have visible or tangible objects interposed between them. Presumably, pairs of visible or tangible impressions that are separated by a line of minimally visible or tangible im‑ pressions affect the senses differently depending on how many minimally visible or tangible impressions there are on that line. When experiencing pairs of lone bodies, one or another of these effects on the senses also arises, establishing an association between that “lone‑body” experience and the resembling “filled‑interval” experience. Second, Treatise 1.2.5.16–17 argues that “lone‑body” and “­filled‑­interval” experiences that have been found to resemble one another in this primary

186  The conception and perception of a vacuum way are discovered to be related in two further ways. The two kinds of experience are “convertible” into one another, and they are found to have the same effects on the intensification and remission of degrees of quality. Third, Treatise 1.2.5.19 observes that “wherever there is a close relation betwixt two ideas, the mind is very apt to mistake them, and … use the one for the other.” Fourth, Treatise 1.2.5.21 applies that general principle to argue that the convertibility and resemblance of the lone‑body and filled‑interval cases induce us to substitute the idea of a distance, which is not consider’d as either visible or tangible [the idea of a vacuum?], in the room of extension, which is nothing but a composition of visible or tangible points dispos’d in a certain order. This “substitution” is immediately declared to be a “mistake.” It is what leads us to falsely imagine we have formed the idea of a vacuum. In pursuing this line of argument, Hume began to speak of two kinds or species of distance, one invisible and intangible, the other visible or tangible. At Treatise 1.2.5.17, visible and tangible distance is described as “mark’d out by compounded and sensible objects,” invisible and intangi‑ ble distance as “known only by the manner, in which the [supposedly?] distant objects [more precisely, the lone bodies] affect the senses.” Above, these were described as “filled‑interval” and “lone‑body” experiences. They will continue to be described using those labels, eschewing Hume’s labels. As of Treatise 1.2.5.14, Hume still needed to explain how “motion and darkness” (the lone‑body experiences) give us any idea of distance, be it comprised of sensible points, comprised of unfilled points, or be it of some other sort. It would be presumptuous to use labels that suggest that has already been done. Everything hangs on what Hume said over Treatise 1.2.5.15–17, where he gave his reasons for saying that the lone‑body and filled‑interval cases are closely related. Depending on what he was able to establish over those paragraphs, it will follow that either (i) we can immediately perceive lone bodies to be separated by a third kind of distance, distinct from either filled distance or a vacuum, (ii) “invisible and intangible distance” refers to a visible or tangible distance that we do not see or feel but that distance cues lead us to judge or imagine to intervene between lone bodies, or (iii) Hume failed to establish that there is any distance in the lone‑body cases other than that comprised by a directly perceived empty space.

Section 5.4.1  187 The following sections argue that Hume took Treatise 1.2.5.15–17 to establish (ii), but that his arguments fail to establish either (i) or (ii). To show that Hume took Treatise 1.2.5.15–17 to establish (ii), what Hume had to say about conversion (T 1.2.5.16) and intensification and remission of degrees (T 1.2.5.17) is taken up first. Though this does not reflect his presentation order it is done to justify the earlier assertion that the second stage of Hume’s case (T 1.2.5.16–17, where the appeals to conversion and intensification or remission of degrees are made) presupposes his first stage appeal (T 1.2.5.15) to manners of affection. Third Distance interpretations often take Hume’s appeal to conversion to be foundational, neglecting that it presupposes what he said about manners of affection. 5.4.1 Conversion

At Treatise 1.2.5.16, Hume concluded that “an invisible and intangible distance may be converted into a visible and tangible one, without any change on the distant objects.” What he meant by “conversion” is that the “extent of visible objects interpos’d betwixt” two given objects may be placed between two other objects that do not have anything visible be‑ tween them “without any sensible impulse or penetration, and without any change on that angle, under which they appear to the senses.”23 As Baxter has put it, “one could convert the invisible … distance between [two] stars into visible … distance merely by holding a strip of white paper so that the distant stars appear adjacent to each end of the strip” (2009, 140). In the tactile case, the muscle sensations attendant on moving the hand between two objects not previously felt to have any other solid or tangible objects interposed between them may also be felt while feeling an “interpos’d im‑ pression of solid and tangible objects.” Conversion presupposes more than that. Suppose we see one pair of stars in utter darkness, hold up a strip of paper, and mark off one distance on the strip; then see another pair, hold up the strip, and mark off another distance; and then look back at the first pair. Do we need to hold up the strip again to see whether they have moved or not? If we can only say how much of the strip will fit after holding up the paper, then when we simply see two stars, we are not going to think they are convertible into any one distance as opposed to any other.24 Hume wanted to draw the stronger conclusion that when we first see the stars (or feel one object and then another) we experience something that is common to that pair of objects as compared to only certain other pairs. The other pairs must all be separated by the same visible or tangible distance. Any separated by a greater or lesser visible or tangible distance

188  The conception and perception of a vacuum could not share the commonality. He also made it clear what the com‑ monality is. We find by experience, that two bodies, which are so plac’d as to affect the senses in the same manner with two others, that have a certain ex‑ tent of visible objects interpos’d betwixt them, are capable of receiving the same extent.… In like manner, where there is one object, which we cannot feel after another without an interval, and the perceiving of that sensation we call motion in our hand or organ of sensation; experience shows us, that ’tis possible the same objects may be felt with the same sensation of motion, along with the interpos’d impression of solid and tangible objects attending the sensation. (T 1.2.5.16, my stress) The common element is not a distance. It is, in the visual case, a certain manner in which the senses are affected and, in the tactile case, a certain feeling accompanying motion of the sense organ. At Treatise 1.2.5.12 and 13, Hume stressed that these common elements are not them‑ selves extended. They are simple and indivisible sensations that may be successive in time but that have no relevant spatial relation to one another. The phrase “invisible and intangible distance” is not mentioned until the concluding sentence of Treatise 1.2.5.16, where it is introduced with “that is.” This suggests that its meaning is to be gleaned from the earlier parts of the paragraph. Plausibly, Hume’s point over Treatise 1.2.5.16 was that we notice a similarity in the manner in which certain pairs of lone bodies affect the eyes and the manner in which pairs of bodies sepa‑ rated by a certain extent of visible objects affect them. We make a similar discovery regarding touch. The analogy between the cases leads us to infer that, since there is a visible or tangible distance between the bodies in the filled‑­interval cases, there is some likelihood that there is a distance between the lone bodies as well, just one that we cannot see or feel.25 We then discover that the lone‑body cases can be converted into cases that display precisely the visible or tangible distance we find in the associated filled‑interval cases. The discovery of convertibility presupposes a prior resemblance. Having been discovered, the convertibility further enhances the associative disposi‑ tion induced by the resemblance. But it is not the prior partner. Were it not for the original resemblance in the manner of affection and sensations ac‑ companying motion, the convertibility could not be discovered, and there would be nothing to prompt us to judge that there is a distance between the lone bodies that is not seen or felt.

Section 5.4.3  189 5.4.2 Intensification and remission

The same can be said of the third relation. We may observe, as another relation betwixt these two kinds of dis‑ tance, that they have nearly the same effect on every natural phænome‑ non. For as all qualities, such as heat, cold, light, attraction,  &c. ­diminish in proportion to the distance; there is but little difference observ’d, whether this distance be mark’d out by compounded and sensible objects, or be known only by the manner, in which the distant objects affect the senses. (T 1.2.5.17) As so described, this will not work. Suppose one distance is “mark’d out by compounded and sensible objects” (impressions), while another is known only “by the manner in which the senses are affected.” These two ways of measuring distance do not share a common metric. That leaves us with no way to determine whether a distance measured in one of these ways is greater or lesser than one measured in the other. It cannot be that heat, cold, light, attraction, and so on diminish in the same way over both distances, since that is supposed to be a resemblance we discover after comparing how they behave over roughly equal instances of distances of each sort. For any such comparison to be possible, there has to be a com‑ mon metric. Charitably, Hume was thinking of what he had already said at Treatise 1.2.5.15: that the manner in which bodies affect the senses is the common metric. Bodies separated by distances marked out by compounded and sensible impressions also affect the sense in the same manner as certain pairs of lone bodies. When we encounter lone bodies that affect the senses in that manner, we judge or imagine them to be separated by an invisible or intangible distance (that is, a space where there is nothing visible or tangible) equal to that separating the endpoints in the filled‑interval case. We then discover that qualities intensify or remit over both distances in the same way. But we only think that there is an invisible or intangible dis‑ tance between lone bodies because we are tempted to do so by the analogy between the way our senses are affected by certain pairs of lone bodies and other pairs of bodies separated by a filled interval.26 5.4.3 Manner of affection

Judging from the passages in which the phrase “invisible and intangible distance” was first introduced (T 1.2.5.16–17), it appears that Hume never

190  The conception and perception of a vacuum meant to suggest that we immediately perceive an invisible and intangible distance between lone bodies. We infer there is a distance we cannot see or feel (a vacuum) because lone‑body cases and filled‑interval cases are analogous in other ways. These relations betwixt the two kinds of distance will afford us an easy reason, why the one has so often been taken for the other, and why we imagine we have an idea of extension without the idea of any object either of the sight or feeling. For we may establish it as a general maxim in this science of human nature, that wherever there is a close relation betwixt two ideas, the mind is very apt to mistake them, and in all its discourses and reasonings to use the one for the other. (T 1.2.5.19) The two kinds of distance that we confuse with one another are those identified and discussed over the course of Treatise 1.2.5.16–18: “that dis‑ tance, which conveys the idea of extension,” namely, visible or tangible distance, “and that other, which is not fill’d with any colour’d or solid object” (T 1.2.5.18), namely, invisible and intangible distance. We do not confuse invisible and intangible distance with vacuum. We confuse it with visible or tangible distance. All we ever manage to judge to exist, and so to conceive, is visible or tangible distance. But because we cannot see or feel the invisible or intangible distance we judge to be there, we “imagine we have the idea of extension without the idea of any object either of the sight or the feeling.” We are drawing a verbal inference without being able to conceive what we are talking about. We use words for ideas, because they are commonly so closely con‑ nected, that the mind easily mistakes them. And this likewise is the rea‑ son, why we substitute the idea of a distance, which is not consider’d either as visible or tangible, in the room of extension, which is nothing but a composition of visible or tangible points dispos’d in a certain order. (T 1.2.5.21) This mistake is caused by a resemblance in the manner in which the senses are affected by lone bodies and bodies separated by a filled interval. It is further supported by discovering that cases that have been associated by this resemblance are convertible and have the same effect on the in‑ tensification and remission of degrees. But those supporting discoveries presuppose that particular lone‑body cases have already been associated with particular filled‑interval cases due to the resemblance in their manner

Section 5.4.4  191 of affecting the senses; otherwise, the further analogies would not appear. Everything hangs on whether Hume was able to demonstrate that there is a resemblance in the way instances of the two sorts of cases affect the senses. He was. But he was not able to make his case without tacitly presuppos‑ ing what he most wanted to deny: that remote lone bodies are directly per‑ ceived to be separated by a space where there is nothing visible or tangible. 5.4.4 Angular separation

We may observe, that two visible objects appearing in the midst of ut‑ ter darkness, affect the senses in the same manner, and form the same angle by the rays, which flow from them, and meet in the eye, as if the distance betwixt them were fill’d with visible objects, that give us a true idea of extension. (T 1.2.5.15) At Appendix 22, Hume noted an error at Treatise 1.2.5.12 “where I say, that the distance betwixt two bodies is known, among other things, by the angles, which the rays of light flowing from the bodies make with each other.” He went on to write, “’Tis certain, that these angles are not known to the mind, and consequently can never discover the distance.” However, his retraction is not extended to two subsequent references to optic angles at Treatise 1.2.5.15 and 16. This may have been an oversight, or it may have been deliberate. Suppose it was deliberate. In that case, Hume must have had some idea of what leads us to think that one pair of objects subtends the same angle as another. An angle formed by light rays meeting in the eye touches only a point of the eye, regardless of whether it is a large or small angle, and the light rays making the angle are not themselves seen. But, as discussed in Chapter 2 note 11, anyone who turns completely around in a static en‑ vironment discovers that the momentary visual field occupies an angular portion of the circle of the horizon. The size of that angle is proportional to the extent of the horizon visible at any given moment. A somersault further establishes that the entire visual field is a portion of a visual sphere. The angular size of an image on the visual field is proportional to the portion of the visual field that the image takes up. This entails that visual objects appearing in the midst of utter darkness could only be thought to form (approximately) the same angle as objects separated by a number of minimally visible objects if they occupy approximately equal portions of the visual field and so are separated by same distances. But subtending the same angle was supposed to be the means whereby we judge that objects are separated by the same distance.

192  The conception and perception of a vacuum Either Hume meant to include all his references to optic angles within the scope of his Appendix 22 retraction or the Treatise 1.2.5.15 and 16 appeals to optic angles are viciously circular. 5.4.5 Motion of the eye from one object to another

Scratching the reference to optic angles, the opening sentence of Treatise 1.2.5.15 is reduced to the claim that lone bodies affect the senses in the same manner as they would if the distance between them were filled with visible objects. Treatise 1.2.5.15 says no more about what this manner is. Treatise 1.2.5.12 offers more detail. The sole difference betwixt an absolute darkness and the appearance of two or more visible luminous objects consists, as I said, in the objects themselves, and in the manner they affect our senses. The … motion that is requir’d in the eye, in its passage from one to the other; and the different parts of the organs, which are affected by them; these produce the only perceptions, from which we can judge of the distance. It is not clear how eye motion alone could contribute to a judgement of distance. The case where two lone bodies appear is a case where the eye sees both at once. There is no call for it to move from one to the other, and if it does move, it is unclear what difference it would see regardless of how it moves or how far it moves, unless it moves so far as to cause one of the bodies to disappear. Otherwise, however it moves, it continues to see the same two bodies. It could not see them changing position, because there is nothing relative to which they could change position. On Hume’s account of the experience of darkness, we do not see darkness extending out to a visual field boundary, beyond which we see nothing. Seeing darkness is already not seeing anything and so not seeing anything that would be a periphery. Not seeing a periphery, we can have no sense of a centre point of a field of view relative to that periphery. We cannot see the lone bodies move relative to either a periphery or a centre point. We could not even tell whether the bodies remain in the same position relative to one another given that we do not see anything between them. Granting that different eye motions produce different sensations, it is unclear what difference the presence of these sensations makes to the appearance of the lone bodies. We would simply experience having various eye muscle sensations without being able to pick on any one of them as the one that accompanies a mo‑ tion of the eye from one body to the other. Conjecturally, Hume wrote what he did because he could not escape thinking that we are always aware of a bounded visual field, whether “filled” or not, and have a sense of where coloured points are positioned

Section 5.4.6  193 relative to the sides and the centre of that field. But that is to tacitly presup‑ pose awareness of the very thing he wanted to deny: perception of a space where there is nothing visible. We might try to help Hume out by speculating that there is something like a single preferred point on the visual field and that we always strive to move our eye so that the object of attention falls on this preferred point.27 But before we could identify the sensation accompanying a motion that brings the visual image of a lone body to the preferred position, we would need to know where the lone body is, and where the preferred position is, and that is impossible. When a lone body appears, it is all that appears. We do not see where it is in relation to a periphery or any other point on a visual field including a preferred point. Worse, if the body is not already at the preferred point, that point is “unfilled” by anything visible and so marks the location of a vacuum, so we cannot be taken to see it or where it is. We could only see the lone body while perhaps feeling we are not seeing it optimally. As we moved our eyes we would feel various eye muscle sensations, without seeing any consequent variation in the lone body. Even if the occurrence of one of those sensations were to be followed by the feeling that the lone body is now seen optimally, we would have no idea where the body was originally positioned and so have no idea of any special prior appearance of the body to associate with the most recent motion sensation. Identifying a particular sensation that goes along with “moving the eye from direct contemplation of one thing to direct contemplation of another” (Baxter 2009, 139) would be impossible. It seems possible only because we are all along tacitly sup‑ posing that lone bodies occupy positions relative to a centre point of an otherwise empty visual field, that is, that they occupy a space where there is nothing else that is visible. When we reject that background assumption, as we must if we consider vacuum to be inconceivable, it becomes apparent that eye motions alone could not lead us to localize lone bodies. 5.4.6 Perceptions produced by organ parts (local signs)

Perhaps Hume never meant to say that eye motion sensations can do the job on their own. They work together with the other kind of simple and in‑ divisible perceptions mentioned at Treatise 1.2.5.12, perceptions produced by the affected parts of the organs. But what are these perceptions? Noth‑ ing would be gained if they were just colour sensations. Colour sensations are lone bodies unless the visual field is filled with them, in which case we are not considering a lone‑body case. Eye motion sensations experienced together with lone colour sensations cannot give us an idea of the distance between those lone colour sensations (Section 5.4.5). Hume may have thought that when we see a coloured point, we also get a unique sensation specific to the affected part of the eye. Theories

194  The conception and perception of a vacuum postulating the existence of an additional set of visual sensations, called “local signs” or “specific nerve energies,” became popular in the nine‑ teenth century, when they were often invoked in tandem with eye mus‑ cle sensations in accounts of how we localize colour sensations (Hatfield 1990, Chapter 4). But they had precedents in Hume’s day. An early instance of a local sign theory was developed by Molyneux (­Dioptrica Nova, 289–90). Molyneux supposed that we innately consider impingements on our body parts to come from a direction orthogonal to the surface of the body part at the impacted point, as proven by the old trick of reaching around behind someone to tap them on the opposite shoulder. In line with this hypothesis, when light rays touch a point on the concave surface of the retina, or the eye is removed from its socket and someone reaches in with a stick to poke the flesh in the interior cavity, we feel upper impacts to come from a direction below us, lower impacts from above, rightward from the left, and so on. On this account, we must feel which body part is being touched, and so must have feelings specific to the affected body part. These feelings are not just tactile, in the sense of being produced by impingements on the skin or flesh, but visual, in the sense of being produced by impingements on the retina. But on this account, we also innately associate these feelings with directions in an ambient space. Two lone bodies, one throwing light at the upper part of the retina, the other at the lower, would be thought to occupy remote positions in a space where there is nothing visible. Rather than refute the possibility of directly experiencing a vacuum, Molyneux’s local sign theory entails it. In his Opticks of 1738, Robert Smith attempted to wed a local sign theory with Berkeleian commitments by maintaining that we need to learn where the local signs are located. The Opticks is likely too late to have in‑ fluenced what Hume published in 1739, but it shows what it was possible for someone to think at the time. Let us now consider a little by what regular steps and observations a person in this case [adult, no vision since birth, and newly made to see], might learn to know the places … of objects. Since he cannot direct his eye to look at any particular object, whose place he knows by feeling, at first we must suppose his eye at rest; and when he has learned to know his hand, or his finger end, let him move it gently upwards and downwards. During this motion he cannot help perceiving some sort of alteration in the visible appearance, occasioned by the corresponding motion of the picture of his finger over different parts of the retina. Then by carefully observing and remembering what sort of sensation was perceived when his finger was in any particular place, suppose above his eye; whenever the like sensation shall again be excited, by another

Section 5.4.6  195 picture of the same or of a different object, falling upon the same place of the retina, wherever it be, he will conclude that this object, whose place is unknown, is above his eye, or in the place where he formerly held his finger. (Opticks 1.5.135, 44–5, my stress) On this account, stimulation of a part of the retina does not just “occasion” us to experience colour patches like those constitutive of the visual image of a finger; it also supplies us with sensations (“alteration[s] in the visible appearance”) specific to the part of the retina that has been stimulated. Pro‑ vided a colour patch has been identified as an image of a movable personal body part, moving that body part, and so changing the position of its image on the retina, changes the sensations that accompany the view of that image. This makes it possible for us to learn to associate the sensations accompany‑ ing images with different places the moving limb is felt to be in, and so come to see images as occupying places signified by the sensations. Hume would have had good reason to be leery of this account. Were it workable, it would be too powerful. It does not just explain how we local‑ ize lone bodies; it explains how we localize colours on a fully illuminated visual field. It is a rival to Hume’s account of space as a manner of dispo‑ sition. On Smith’s account, the only manner of disposition is time. Spa‑ tial relations are derived from successions of sensations. Someone newly made to see would not immediately perceive coloured points as disposed in space, even on a fully illuminated visual field. Coloured points are states of immaterial minds, “occasioned” by events on the retina, as Descartes, Berkeley, and Reid maintained. The idea of space replaces that of the miss‑ ing shade of blue as the outstanding counterexample to the thesis that all ideas are copied from impressions, to the satisfaction of Platonist math‑ ematicians everywhere. This should not be surprising, because Smith was a Berkeleian, pursuing the constructivist programme of the New Theory of Vision to its most radical conclusion. Fortunately for Hume, it would not have been clear that Smith’s account is workable.28 As Smith stated it, the account presupposes that the subject has come to be aware of the places of objects by means of touch. Not to worry about how it might have achieved that much, Smith took the hard part for granted when assuming that a subject with no immediate aware‑ ness of how coloured points are disposed in space could “[learn] to know his finger end.” The visual image of a finger end is constituted by many coloured points. Provided there is enough illumination to make those points ap‑ pear, the contents of the area surrounding the finger end should be visible as well, occasioning coloured points of their own. The same holds for the

196  The conception and perception of a vacuum area surrounding that area, out to the bounds of the visual field. Except, for Smith, we do not originally see any bounds. We originally see just thousands of coloured points, each permeated with its own local sign. The points are not experienced as anywhere in space and so not experi‑ enced as distributed over a surface, much less as collected together in any particular way. Likewise for the local signs. When Smith’s subject moves its finger upwards, the motion reveals pre‑ viously occluded parts of the environment while occluding previously seen parts. But what the subject sees is just a snowstorm of thousands of differ‑ ently coloured points, some randomly disappearing and others randomly appearing, some randomly changing their local signs and others keeping them. It is far too sanguine to declare that under these circumstances the subject could learn to know its finger end, or come to see anything as dis‑ posed in space. Advances in modern computing may incline us to shrug our shoulders over this, but at the time it would have seemed an impossibly large and complex computational task. The infant who is first learning to see needs to “carefully observe and remember” (paraphrasing Opticks 1.5.135, 45) all these changes. These reflections on the rival local sign theories of Hume’s day suggest that his appeal to perceptions produced by the affected parts of the organs may have been offhand. Had he thought about it, he could not have been attracted to a theory that, developed one way, entails that we have innate ideas of empty space, and developed the other way undermines his own account of the origin of the idea of space. Hume had an intuitionist alternative to offer to both Molyneux’s na‑ tivism and Smith’s empirism. He maintained that we are not confronted by Smith’s gargantuan computational task, not because we are innately constituted to experience impingements as originating from specific loca‑ tions in an ambient space, but instead because coloured points are origi‑ nally experienced as disposed in space. We are affected by resemblance and proximity, and so see resembling, contiguous coloured points as constitu‑ tive of shapes of a particular size disposed at particular positions over a visual field. Hume would have to say that, were it supposed that we do experience local signs, we would not need to learn how they are disposed relatively to one another by moving a finger around. We would immediately see them disposed alongside one another, as we immediately see the purple points of a table disposed alongside one another (T 1.2.3.4–5). The first thing we would learn is that, unlike coloured points, these eye‑part sensations are immobile. Being permanently disposed at locations on a visual field, each eye‑part sensation would have a unique position as a distinguishing fea‑ ture. Two lone bodies would appear as contiguous or distant, and distant

Section 5.4.6  197 to a greater or lesser extent just in virtue of the eye‑part sensations that accompany them, with no eye motion required. But before cheering that this is the solution to Hume’s problem, we need to consider what he should have said about the nature of these eyepart sensations. Rejecting Smith’s rival empirist account, he would have had to say that each is originally experienced at a unique location rela‑ tive to the oblong, nose-indented periphery of the visual field, and that we learn that each has remained fixed to that location up to now. They are therefore sensations of a different sort from colour sensations, which are mobile. Elsewhere, Hume only recognized one kind of visual sensation, colour sensations. If vision (or retinal impact) acquainted us with any further sen‑ sations, we should be able to say something about them. Are they all differ‑ ent from one another or alike but varying in degrees? Supposing they vary in degrees, do they vary over one, two or more dimensions (the way col‑ ours vary in hue, brightness, and saturation), and how do they vary? Most intense at the centre of the visual field and dimmest at the periphery? Most intense at the extreme left shading to least intense at the extreme right? Scattered in no particular order? The fact that we are at a loss to describe them in any of these ways indicates that we have no experience of any qualities of purported eye‑part sensations. They would have to have some apparent features if they are to do their job unless we are unconscious of them but nonetheless cognitively affected by them. However, Hume main‑ tained that every impression appears in its true colours and that we cannot be mistaken or confused about any of them (T 1.4.2.7). Hume would have to insist that the only feature observed to be possessed by eye‑part sensations is their location on the visual field, which is essen‑ tial to them and inseparable from them. In being aware of lone bodies as coincident with eye‑part sensations we would be aware of lone bodies as disposed at remote places on an otherwise empty field. This is just what we would expect if the eye‑part sensations just are feelings of what parts of a spatially extended retina are being touched by light, and those parts are in fact spatially separate from one another. Our experience of local signs (sup‑ posing we have any) reduces to a direct perception of visual locations, there being nothing more to them. Far from offering an alternative to directly perceiving empty space, a Humean local sign theory would entail that direct perception. With local signs, Hume would be in no position to deny the perceivabil‑ ity and so conceivability of visibly empty space. Without them, he had no account to offer of what would lead us to consider lone bodies to be visibly distant from one another, were they not all along perceived to be disposed at either end of a space where there is nothing visible.

198  The conception and perception of a vacuum 5.4.7 Feeling motion from one tangible object to another

The sensation of motion is likewise the same, when there is nothing tangible interpos’d betwixt two bodies, as when we feel a compounded body, whose different parts are plac’d beyond each other. (T 1.2.5.15) When discussing what could lead us to judge that lone bodies are separated by a tangible distance, Hume only appealed the sensations accompany‑ ing motion. It is unclear how we could get any notion of distance from those sensations, even when we are moving in a filled space and “feel a compounded body, whose different parts are plac’d beyond each other” (T 1.2.5.15). Hume declared that while the sensations have parts that are successive to each other and may give us the notion of time, those parts “certainly are not dispos’d in such a manner, as is necessary to convey the idea of space or extension” (T 1.2.5.6). The sensations that arise from motion are “not capable of conveying to us an idea of extension” whether they occur alone or are bookended by touching two remote tactile objects (T 1.2.5.13). Even if the limb moves over a variegated surface, one part of the limb feels first one part of the surface, then another, and it remains ob‑ scure what leads us to consider the parts to be adjacently disposed in space. But then it should not matter that the sensation of motion is the same when moving between two lone bodies and when moving over a surface. As the latter sensation does not lead us to think of any distance, the resemblance between the two cannot lead us to think there is any distance in the former case. Hume should have been aware that our experience of tactile space, whether filled or unfilled, is based on how tactile sensations are simultane‑ ously disposed, not on motion (Section 5.3.3). 5.5

Hume’s fortunate failure

Hume maintained that when we conceive all the air and subtle matter be‑ tween the walls of a chamber to have been annihilated, we do not conceive vacuum between its walls (T 1.2.5.3).29 We instead conceive eye or limb motions from one wall to the other and feelings of parts of the eye or body affected by the walls. Because these motions and sensations are common with those experienced in filled‑interval cases, we infer by argument from analogy30 that the distance we see in the filled‑interval cases is also present between the walls of the chamber, even though we do not see or feel it. But this is a verbal inference to what Hume considered an absurd conclusion, not something we actually conceive. Similarly, when we think that lone‑body and filled‑interval cases are convertible, for example, that a white paper strip would fit between two

Section 5.6  199 lone bodies without impulse or penetration, it is not because we see an invisible distance of that magnitude between them. It is again because the motion of the eyes from one body to the other and the feelings in the parts of the eyes are like those we experience in filled‑interval cases, leading us to wrongly infer from the analogy that there is a similar, but unseen interval in the lone‑body cases. Hume’s appeal to an analogous manner in which the senses are affected in lone‑body cases is an appeal to impressions that either do not exist (sen‑ sations specific to particular locations on the sense organs, Section 5.4.6) or cannot be associated with filled‑interval cases (motion sensations, Sec‑ tions 5.4.5 and 5.4.7). He failed to demonstrate that we conceive or per‑ ceive something other than a vacuum in the evacuated room, conversion, or other lone‑body cases. It is just as well that he failed. In Section 5.3.5, it was argued that Hume made a dangerous concession to Reid when arguing that “motion and darkness, either alone, or attended with tangible and visible objects, con‑ vey no idea of a vacuum or extension without matter” (T 1.2.5.14, sum‑ marizing the outcome of T 1.2.5.5–13). He allowed that colours need not appear anywhere in space. This threatens his imagistic account of mental representation. Hume’s failure to account for why we only falsely imagine that lone coloured points appear in the midst of an empty space goes some way to mitigate the consequences of what would otherwise have been a damaging concession to Reid. It shows that the intuition that visual and tactile sensa‑ tions are originally perceived as disposed in space cannot be easily dismissed. 5.6

Unoccupied extension in the Enquiry

The Enquiry contains a single sentence that might be read as a reference to vacuum. Speaking of the unintelligibility and even absurdity of suppos‑ ing we can form ideas of primary qualities by abstraction, Hume wrote that “An extension, that is neither tangible nor visible, cannot possibly be conceived: And a tangible or visible extension, which is neither hard nor soft, black nor white, is equally beyond the reach of human apprehension” (EHU 12.15). This pronouncement occurs in the context of Hume’s appropriation of Berkeley’s argument that the primary qualities of extension and solidity cannot be conceived apart from sensible qualities like hard, soft, hot, cold, white, and black. In repeating that argument Hume remarked that “noth‑ ing can save us from this conclusion, but the asserting, that the ideas of those primary qualities are obtained by Abstraction” (EHU 12.15). The extension that is referred to here is an extension conceived apart from all sensible qualities, not an unoccupied extension located between or around

200  The conception and perception of a vacuum coloured points. Hume’s appeal to the Berkeleian argument does not re‑ quire denying the experience of unoccupied spaces between or around hard or soft, hot or cold, black or white points. That case does not conflict with his conclusion that the primary qualities cannot be conceived apart from any relation to the secondary qualities.31 Hume may have meant to insist that an unoccupied extension cannot be conceived, even by way of position relative to and contrast with oc‑ cupied locations. If so, he will have persisted in adhering to the doctrines originally propounded by the author of the Treatise. But he will not have offered an effective new argument for doing so. Notes 1 At Treatise 1.2.5.1, Hume announced that his project would be to answer three objections “which I shall examine together, because the answer I shall give to one is a consequence of that which I shall make use of for the others.” (The third objection is the one answered as a consequence of the answer worked out to deal with the first two.) Hume answered the first two objections by arguing that we are really thinking of something else when we think we are conceiving a vacuum (T 1.2.5.22–3). This answer presupposes that vacuum is inconceiv‑ able and is only called for on that supposition. It begs the question of why we should accept that vacuum is inconceivable, and properly so, in Hume’s mind, because that question is answered by the argument from the non‑entity of un‑ qualified points. 2 Hume went so far as to include one of the standard Berkeleian visual distance cues on the list of factors inducing us to imagine we have an idea of an invisible distance, writing “This annihilation leaves to the eye, that fictitious distance, which is discover’d … by the degrees of light and shade” (T 1.2.5.23). This must have been a slip since the annihilation he was referring to is that of all the air and subtle matter between the walls of a chamber. If any light remained after the annihilation, the side walls of the chamber would be visibly separated by the coloured points constitutive of the back wall, contrary to the hypothesis. The slip indicates that Hume was thinking of a Berkeleian account, according to which visual cues lead us to “imagine we have an idea” (T 1.2.5.19) of a distance that we do not immediately see. 3 Third Distance interpretations can be found in, among others, Broad (1961, 171–3), Baxter (2009, 138–41), Kervick (2016), and Cottrell (2019). The Berkeleian interpretation was first proposed by Boehm (2012), who justifies it with a careful engagement with Baxter’s third distance interpretation. That engagement is examined in Falkenstein (2013). Rather than devote space to reiterating points from Falkenstein 2013 here, interested readers are asked to retrieve it as an appendix to this chapter. 4 T 1.2.5.1, 1.2.5.10, 1.2.5.26n (mentioning an absence of anything “that by an improvement of our organs might become visible or tangible”). 5 T 1.2.4.2, 1.2.5.5, 1.2.5.7, 1.2.5.9 (mentioning “body” rather than “matter”), 1.2.5.14, 1.2.5.22. 6 Kervick 2016, 70, 71–2. Contrast Skrzypulec (2022, 524), who writes: “We do not see anything behind our head. We also do not see anything in a completely

Notes  201 dark room. Although both of these situations can be described as cases of not seeing, there is a significant difference between them. In the first case, we do not see in the sense of a lack of visual experience; but in the second case we have a visual experience that presents darkness filling the whole visual field (Sorensen 2004; Wright 2012).” Hume assimilated the second case with the first, denying that we have any such thing as a visual experience of a field of darkness. 7 But see Sorensen (2011), Wright (2012), Skrzypulec (2022), and the earlier literature referenced by them. 8 T 1.2.4.19. It was known in Hume’s day that infinite divisibility entails that a longer line must contain as many points as a shorter line. (Bayle, Dictionnaire, 366–7, appealed to lines drawn from each point on the circumference of a circle to its centre, which must each cut a concentric, smaller circle at a unique point. Variations on this geometrical demonstration were common.) This fact motivates the modern mathematical distinctions between dimension, length, cardinality, and measure of point sets (see Grünbaum 1952, 292–8). Hume’s rejection of infinite divisibility (T 1.2.1) means that he had no need for such distinctions. Attributing a reliance on them to him would be uncalled for as well as anachronistic. 9 Kervick (2016, 72) disagrees, block quoting as evidence a passage in which he claims that Hume wrote that the luminous bodies are “disposed in such a manner as to be separated by some distance or other.” Kervick’s passage does not occur in the paragraph Kervick references, T 1.2.5.12, or anywhere else in Hume’s corpus. (ED) does claim that we can perceive whether the luminous bodies are separated by a greater or lesser distance, but it does not attribute this separation to a manner of disposition, and it is controversial whether Hume meant to endorse (ED). 10 At Appendix 22, Hume scratched a further appeal to angles formed by light rays. 11 See Chapter 1, note 44, on the distinction between empirism and empiricism. 12 Kervick (2016) is an extreme example. He defines distance negatively, as the absence of contiguity (65). He maintains that elements can be more, less, or equally distant (65), but never says what makes distances greater or lesser in cases where bodies are not separated by anything visible or tangible. He only says what does not make distances greater or lesser in these cases (66, 67, 73, n.23; note particularly n.14, which fails to specify how the function d deter‑ mines the real number value of pairs). Even his apparently positive notion of a “relational void” is defined only negatively as constituted by “systems of objects standing some distance from one another without anything at all lying between them, not even empty immaterial space” (2016, 67). We are not told what does make them appear distant. 13 See Boehm (2012) and Falkenstein (2013) for further discussion of Third Dis‑ tance readings. 14 Hume had a different account, which Broad ignores (T 1.2.5.12–15, Section 5.3.5 at the end) and which might tendentiously be called “Berkeleian.” It ap‑ peals to purely qualitative feelings in the eyes and muscles that are associated with visible or tangible distances, not to awareness of spatial separation where there are no intermediate coloured or tangible points. 15 Cottrell (2019, 231, 222–3) rejects the Berkeleian view that we only judge or imagine them to be distant on the basis of motion sensations or feelings in af‑ fected parts of the sense organs.

202  The conception and perception of a vacuum 16 Cottrell argues that Hume’s concept of invisible and intangible distance had a pedigree in the notion of “privative vacuum” developed in debates over absolute space current in Britain in the period just before Hume wrote the Treatise (Cottrell 2019, 212–13). Given the dim view that writers from Bayle (Dictionnaire, 380–1) to Alexander (1956, xlii) have taken of the quality of this line of thought it might be more charitable if Hume could be excused from having participated in it. Comparison of Hume’s account with the visual theories of Molyneux, Berkeley, and Robert Smith may be more profitable. See Section 5.4.6. 17 At Treatise 1.2.5.25, Hume did not write that we find by experience that bodies can be arranged distantly but have no spatial thing located between them. He wrote that we find by experience that “the bodies may be plac’d in the same manner, with regard to the eye, and require the same motion by the hand in passing from one to the other, as if divided by something visible and tangible.” The “invisible and intangible distance” referred to in the following sentence is a distance associated with these effects on the senses. Later, at Treatise 1.2.5.26n, he declared that “The real nature of this position of bodies is unknown. We are only acquainted with its effects on the senses, and its power of receiving body.” Experience reveals a common manner of affecting the senses, not a distance. 18 Cottrell is inclined to charge his opponents with equivocation (2019, 221n.32). But it is only possible to equivocate when there is more than one meaning that can be attributed to the term in question. Defining the second sense merely by stipulating what it does not include is not an acceptable recourse. 19 Broad (1961, 171–2) described Hume’s thesis that a vacuum is inconceivable as “a platitude.” He claimed to be unable to find that he had visual images that are extended but completely without colour, or tactual images that are extended but completely lack any tactual quality. Garrett (1997, 54–5) makes a similar assertion. But to demand that the experience be “completely” without colour or tactual quality is too easy. The serious question is whether we can conceive and perceive unoccupied positions between occupied ones. 20 Of course, coloured points may be experienced as surrounded by points of the same colour. But, as Hume recognized (T 1.2.4.19), in that case, they are not so readily distinguished. See Section 2.4.1. 21 There is no infinite regress here for reasons given in Chapter 2 note 35. 22 Apart from Treatise 1.2.5.16, Kervick (2016) and Cottrell (2019) have little to say about this portion of Treatise 1.2.5, preferring to look to modern math‑ ematics or the early eighteenth‑century English debate over absolute space for support for Hume’s position. But on a Berkeleian reading, 1.2.5.14–21, sup‑ plemented by 1.2.5.12 are foundational. 23 At Appendix 22, Hume retracted a reference to angles formed by light rays made at Treatise 1.2.5.12. He did not retract the reference to optic angles made at Treatise 1.2.5.16. See Section 5.4.4. 24 This is not the only problem with Baxter’s proposal. Strips of different lengths can be aligned with the stars by adjusting the distance at which they are held from the eye. 25 This form of “argument from analogy” differs from the one Hume recognized at Treatise 1.3.12.25. In the latter case, we infer analogous causes or effects from analogous effects or causes. In the former, we infer that since two things are alike in a number of ways, they will prove to be alike in a further way as well. Hume considered this alternative form of argument from analogy to be

Notes  203 noncausal, which is why he supplied an alternative, physiological explanation for it (T 1.2.5.20). 26 Even with this patch, it is not clear that Hume’s account works. Baxter (2009, 140) has cleverly tried to help him out by proposing a case where one of two bodies is luminescent while the other reflects light from it. We think that the fainter the light reflected by the second object, the further it must be from the luminescent object. But that inference is complicated by Hume’s apparent com‑ mitment to the thesis that we do not perceive visual depth (T 1.2.5.8). Imagine that as a matter of fact, the reflective object is not laterally distant from the luminescent object but set much further away from us than the luminescent object. The reflective object is as faint as it is because light from the lumines‑ cent object must pass away from us to reach the reflective object, and then be reflected from that object all the way back to us. The distance the light travels would be the same if the reflective object were much closer to us but set at a greater lateral distance from the luminescent object. If we have nothing else to go on but the faintness of the light, we should be in no position to correlate the strength of illumination with variations in visible distance, whether the distance is marked by intermediate visible points or by the manner in which the eye is affected. Baxter speaks of seeing the two objects “at the shore, against the backdrop of a lightless sky, or in the woods, against the backdrop of mist and trees,” (2009, 140), but these are not lone‑body cases. We are conveniently seeing other things that serve as distance cues. We do not have an experience of an invisible distance if there is a backdrop shoreline or treeline between them. 27 Baxter (2009, 139) writes that “the sensations of moving the eye from direct contemplation of one thing to direct contemplation of another are the same whether the distance is invisible and intangible or is visible or tangible.” “Di‑ rect contemplation” means moving the eye (directing the optic axis) to bring the image of an object to fall on a preferred portion of the visual field. 28 James (Principles 2.20, 900–12) offers a characteristically entertaining later view of the difficulties attendant on the “mystical‑sounding” empirist attempt to account for how a perception of space is “produced” or “evolved” from a merely temporal succession of purely qualitative sensations. The mechanisms invoked by the empirists are “association” and “mental chemistry.” Associa‑ tion “[produces] nothing, but only [knits] together things already produced in separate ways.” And mental chemistry (appealing to the analogy of water emerging from oxygen and hydrogen) is covert Kantian nativism invoked by theorists who insist on their empirist credentials (901–2). 29 To be precise, he wrote that “keeping strictly to the two ideas of rest [of the walls] and annihilation, ’tis evident, that the idea, which results from them, is not that of a contact of parts, but something else; which is concluded to be the idea of a vacuum.” But the remainder of Treatise 1.2.5 is directed to “tak[ing] the matter pretty deep” to “answer [this] objection” by showing how a differ‑ ent conclusion ought to be drawn. 30 See note 25. 31 The objection that Hume’s separability thesis entails that a vacuum in or around sensible points would have to be separately conceivable from those points is addressed in Sections 2.5 and 5.3.5. Millican (2002, n.37) and George (2006, 149, 157–8, 158–9) have pointed out that the separability thesis is not endorsed in the Enquiry. The Enquiry replaces appeals to the separability prin‑ ciple with appeals to meaning empirism.

204  The conception and perception of a vacuum Bibliography Alexander, H. G. 1956. “Introduction.” In The Leibniz‑Clarke Correspondence, edited by H.G. Alexander, ix–lv. Manchester: Manchester University Press. Baxter, Donald L. M. 2009. “Hume’s Theory of Space and Time in Its Skeptical Context.” In The Cambridge Companion to Hume, 2nd edition, edited by David Fate Norton and Jacqueline Taylor, 105–46. New York: Cambridge University Press. Baxter, Donald L. M. 2016. “Hume on Space and Time.” In The Oxford Hand‑ book of Hume, edited by Paul Russell, 173–90. Oxford: Oxford University Press. Boehm, Miren. 2012. “Filling the Gaps in Hume’s Vacuums.” Hume Studies 38: 79–99. Broad, C. D. 1961. “Hume’s Doctrine of Space.” Proceedings of the British Acad‑ emy 47: 61–76. Costa, Michael J. 1990. “Hume, Strict Identity, and Time’s Vacuum.” Hume Stud‑ ies 16: 1–16. Cottrell, Jonathan. 2019. “Hume’s Answer to Bayle on the Vacuum.” Archiv für Geschichte der Philosophie 101: 205–36. Falkenstein, Lorne. 2013. “Hume on the Idea of Vacuum.” Hume Studies 39: 131–68. Garrett, Don. 1997. Cognition and Commitment in Hume’s Philosophy. New York: Oxford. George, Rolf. 2006. “James Jurin Awakens Hume from His Dogmatic Slumber: With a Short Tract on Visual Acuity.” Hume Studies 32: 141–66. Grünbaum, Adolf. 1952. “A Consistent Conception of the Continuum as an ­Aggregate of Unextended Elements.” Philosophy of Science 19: 288–306. Hatfield, Gary. 1990. The Natural and the Normative: Theories of Spatial Percep‑ tion from Kant to Helmholtz. Cambridge: MIT Press. Kervick, Dan. 2016. “Hume’s Perceptual Relationism.” Hume Studies 42: 61–87. Millican, Peter. 2002 “The Context, Aims, and Structure of Hume’s First Enquiry.” In Reading Hume on Human Understanding, edited by Peter Millican, 27–65. Oxford: Oxford University Press. Skrzypulec, Blazej. 2022. “Silence Perception and Spatial Content.” Australasian Journal of Philosophy 100: 524–38. Sorensen, Roy. 2004. “We See in the Dark.” Noûs 38: 456–80. Sorensen, Roy. 2008. Seeing Dark Things. Oxford: Oxford University Press. Wright, Briggs. 2012. “Darkness Visible?” Australasian Journal of Philosophy 90: 39–55.

6 Belief Normativity; objects

This chapter is concerned with Hume’s account of what causes belief. Briefly, belief arises from a superior degree of vivacity. It can also depend on association. This chapter begins by looking at Hume’s position on vi‑ vacity and vivacity bearers (Section 6.1). It then considers how vivacity is transferred to associated ideas (Sections 6.2–6). Hume recognized that we are often determined to have beliefs that are not “received by philosophy” (Section 6.7). It is argued that he was more concerned with motivating philosophical belief than with justifying it and that he recognized an en‑ counter with sceptical arguments as an effective motivator (Sections 6.8 and 6.9). The concluding sections of the chapter charge that he made a serious mistake in maintaining that belief can only be extended beyond the bounds of sense experience and memory by causal inference. He could have recognized that customary contiguity and customary identity can cause belief and should have recognized that they play an essential role in the recognition of publicly observable objects (Sections 6.10–12). 6.1

Impressions and ideas

One way to distinguish sensation from memory, imagination, and under‑ standing is by maintaining that sense is receptive whereas the other facul‑ ties are constructive. The materials for cognitive processes are first given in sense experience. To remember, imagine, or conceive is to reproduce or modify these originally given materials. Hume adopted this distinction at the outset of Treatise 2, where he de‑ scribed “Original impressions or impressions of sensation” as being “such as without any antecedent perception arise in the soul” (T 2.1.1.1), adding that “’Tis certain, that the mind, in its perceptions, must begin somewhere; and that … there must be some impressions, which without any introduc‑ tion make their appearance in the soul” (T 2.1.1.2). However, identifying original perceptions with sense impressions makes a leap. There are those who maintain that understanding intuits its own DOI: 10.4324/9781032677880-7

206  Belief original perceptions (Descartes Meditations 2, 3, and 5; Price Review 2 and 3). Identifying sense impressions with original perceptions requires first establishing that there are no “innate ideas” (Locke Essay 1.4) or “in‑ tellectual intuitions” (Kant Kritik). That in turn requires some other way of distinguishing sensation and reflection from understanding. An obvious alternative distinction appeals to what we experience as an immediate consequence of sensory stimulation, as determined by exposing and shutting off a sense organ or moving it to different stimulus environ‑ ments. This pragmatic distinction does not work for passions or internal body sensations, such as hunger. But in the remaining cases, it enables us to determine whether a perception is due to the senses. Hume alluded to this way of distinguishing between sense impressions and ideas at Enquiry 2.7 and Treatise 1.1.1.9. Hume distinguished passions and other internal sensations from ideas by appealing to sincere, current, self‑attribution. “A man, in a fit of anger” makes a sincere, current self‑attribution; “one who only thinks of that emotion” refrains from doing so (EHU 2.2). Someone who is told of an‑ other person who is in love has “a just conception of his situation” (EHU 2.2), but not attributing that passion to themselves means they do not feel it themselves. Hume appealed to this way of understanding the distinction between the products of internal sense and imagination when arguing for empirism (Chapter 1 note 44) regarding passions, writing that “A man of mild manners can form no notion of inveterate rage or cruelty” (EHU 2.7) The attribution of the passion to others is impossible unless one has at some point felt compelled to sincerely attribute it to oneself. All our refer‑ ences to passions gain their meaning from what we have felt in ourselves. The same holds of hunger, pain, and other internal states. These “pragmatic” criteria for distinguishing the products of external and internal sense from those of imagination or understanding are all Hume needed to justify his version of the empirist principle that all ideas are based on impressions (EHU 2.5; T 1.1.1.7). He could challenge his opponents to provide an example of an idea that cannot be traced back to an impression or operation on impressions (EHU 2.6; T 1.1.1.5). And he could premise that those who lack the use of a sense organ, who have never used the organ in the right stimulus circumstances, or who have never sin‑ cerely self‑attributed an internal sensation or passion can form no idea of what they have missed (EHU 2.7; T 1.1.1.9). However, Hume was not satisfied with a pragmatic criterion for distin‑ guishing between impressions and ideas. He wanted to identify an intrin‑ sic feature that distinguishes impressions from ideas. He had an ulterior motive for doing so. That intrinsic feature was to figure in his account of belief. As such, it would have to apply to some ideas as well as all impres‑ sions, though not to the same degree.

Section 6.1  207 Faced with a similar challenge, Berkeley had appealed to three criteria for distinguishing “ideas of sense” from “ideas of imagination.” Ideas of sense occur independently of the will (Principles 1.29). They have a “steadiness, order, and coherence” not found in ideas of the imagination (Principles 1.30). And they are more “strong, lively, and distinct” than those of imagination (Principles 1.30). These are fallible criteria for distinguishing between reality and illusion, not definitions of “idea of sense” and “idea of imagination.” By definition, an idea of sense is an idea produced in us by God, an idea of imagination one we produce (possibly involuntarily). Hume had no use for Berkeley’s definition. And he could not appeal to either of Berkeley’s first two criteria. On his account, many ideas occur in accord with principles of association, which operate independently of volition. For the same reason, he could not say that impressions are steady, ordered, and coherent whereas ideas are capricious. This leaves the third, phenomenological criterion, strength, liveliness, and distinctness. Hume had reason to rely on a superior degree of liveliness as a criterion for distinguishing impressions from ideas of all sorts. When we remember, we do not think we have been caught in a time loop and are back confronting the same series of events. Part of what sensitizes us to the difference between remembering and sensing is that we remember with our eyes open and our other senses delivering information. The remembered experience does not live up to the concurrent sensory experience. Similarly, when we imagine, we imagine while sensing and what is imagined falls even further from living up to what is concurrently sensed. We find it hard to come up with good words to describe how it falls short. It has nothing to do with brightness of colour or clarity and distinctness. It is unique, incomparable, and indescribable. But it is evident to all of us. Calling it a difference in liveliness or vivacity will do. These terms are so vague that they can be appropriated for special use. (By that token, appealing to a difference in strength or visual distinctness will not do. These terms have associations with brightness and distinctness. They are avoided in what follows, even though Hume preferred to use “force” with “vivacity.”) Hume had no need to consider vivacity to be definitive of impressions. He could take it to be a fallible criterion for identifying them. The Enquiry adopts what might be called a “faculties first” reading of the distinction between impressions and ideas. It takes a distinction between the faculties and an awareness of which perceptions are delivered by which faculties for granted and then asks us to notice the further fact that perceptions of one of the faculties, external or internal sense, are generally more vivacious than those of the others. Every one will readily allow, that there is a considerable difference between the perceptions of the mind, when a man feels the pain of

208  Belief excessive heat, or the pleasure of moderate warmth, and when he af‑ terwards recalls to his memory this sensation, or anticipates it by his imagination. These faculties may mimic or copy the perceptions of the senses; but they never can entirely reach the force and vivacity of the original sentiment. (EHU 2.1) In contrast, the first book of the Treatise puts the phenomenology first. All the perceptions of the human mind resolve themselves into two dis‑ tinct kinds, which I shall call impressions and ideas. The difference betwixt these consists in the degrees of force and liveliness, with which they strike upon the mind, and make their way into our thought or consciousness. Those perceptions, which enter with the most force and violence, we may name impressions; and under this name I comprehend all our sensations, passions and emotions, as they make their first ap‑ pearance in the soul. By ideas I mean the faint images of these in think‑ ing and reasoning[.] (T 1.1.1.1) Here, originally appearing sensations, passions, and emotions are identi‑ fied as impressions on the ground of their vivacity. It is suggested that ideas are all “faint images of these in thinking and reasoning,” but the suggestion is as of yet undemonstrated. What is clear is that thoughts lack vivacity. A distinction between the faculties is forged by the phenomenol‑ ogy (T 1.1.3.1).1 Taking the difference between impressions and ideas to be defined by vivacity gives rise to some well‑known problems.2 The “faculties first” ap‑ proach taken in the Enquiry has been comparatively neglected.3 It might be objected that the Enquiry’s approach is too realist. It saddles Hume with a commitment to the existence of sense organs stimulated by an external environment and a body with its own internal sense. Those preferring to absolve Hume of any realist commitments have a recourse. It feels and looks like something to close our eyes or move our hands, so a “faculties first” distinction can be cashed out in terms of the regular consequences of these ways determining whether a perception is a product of outer sense. Admittedly, we might only imagine that we have closed our eyes, moved our hands, or otherwise restricted the flow of stimuli to our sense organs. This objection can be addressed by admitting that the pragmatic criterion is fallible. Combining it with a vivacity criterion puts us in a position to invoke one to support the other in hard cases. If I only imagine closing my eyes, what I imagine will not compete with what I continue to see with my open eyes, making it clear that I am not really closing my eyes. I may still

Section 6.2  209 be deceived, as in dreams. It is not always easy to distinguish reality from illusion. We do the best we can with the fallible criteria we have at our disposal.4 The “faculties first” account of the distinction between impressions and ideas is assumed in what follows, both because it is Hume’s only author‑ ized account and because it avoids the problems attendant on the “phe‑ nomenology first” account. On this account, vivacity is not as important for distinguishing between impressions and ideas as it is for other purposes. The role that vivacity plays in belief and other mental operations, and the role that impressions and memories play as original sources of vivacity for those operations, are of most importance. Hume only needed to maintain that impressions and memories (as picked out by a faculties first account) have some degree of an obvious but indefinable quality called “vivacity.” Ideas have none unless something special happens to endow them with it. 6.2

Vivacity, belief, judgement, and assent

Aside from three passing references in Treatise 1.3.2 and 1.3.4, “belief” is first mentioned at Treatise 1.3.5.6–7. In this initial discussion, it is not associated with causal inference to the unobserved but with the vivacity of sense experience and memory.5 Thus it appears, that the belief or assent, which always attends the memory and senses, is nothing but the vivacity of those perceptions they present; and that this alone distinguishes them from the imagina‑ tion. To believe is in this case to feel an immediate impression of the senses, or a repetition of that impression in the memory. ’Tis merely the force and liveliness of the perception, which constitutes the first act of the judgment[.] (T 1.3.5.7) This pronouncement fits oddly with Hume’s later definition of belief as “a lively idea related to or associated with a present impression” (T 1.3.7.5). Contradiction is avoided by using “in this case” at Treatise 1.3.5.7 and “most accurately defin’d” at Treatise 1.3.7.5. There is a deliberate, initial equation of belief with the vivacity of impressions and memories.6 Belief does not end up being identified only with impressions and memories. But it includes them and extends downwards over a range of less vivacious perceptions (T 1.3.13.19). Like vivacity, belief comes in degrees. As the vivacity of perceptions decreases, so does the strength of belief. It later becomes apparent that belief diminishes to zero before vivacity does (EHU 6.4; T 1.3.9.15; 1.3.10.7). There are ideas that are enlivened, but not suf‑ ficiently to be beliefs.

210  Belief This minor exception aside, Treatise 1.3.5.7 declares that belief “just is” vivacity. By implication, vivacity just is belief. It is whatever it is about per‑ ceptions that, when present to a sufficient degree, makes beliefs of them. Treatise 1.3.5.7 also identifies assent with vivacity and so with belief. It does not take assent to be consequent on belief. It takes assent to be “noth‑ ing but” the vivacity of sensations and memories, and vivacity to “[consti‑ tute] the first act of judgment.” This is perplexing. Assent is either given or not given. But belief and vivacity come in degrees. And how can a quality of a perception, its vivacity, constitute an act of judgement? On a common understanding, assent is the act of judging a proposi‑ tion to be true. But where the proposition describes a “matter of fact” that is evident to the senses or memory, there is something that is sensed or remembered that instantiates and displays that matter of fact. Hume held that assent is “firstly” nothing but the vivacity or “belief‑compelling” nature of that sense experience or memory. Our acts of assenting to propo‑ sitions that fairly describe the sensed or remembered state of affairs are consequences of this “first act of the judgment.” Hume had more to say about propositions describing matters of fact that are not evident to the senses or memory (EHU 6; T 1.3.11–12; Sections 6.5 and 6.6). In these cases, assent to the proposition is also de‑ termined by the vivacity of the perception of the matter of fact described by the proposition. However, it is further necessary that the perception of the matter of fact has acquired more vivacity than the combined vivacity acquired by the contrary propositions. (Set aside for the moment how that happens.) What might be called a “second” act of judgement is constituted by either unreflectively feeling the superior quantity of vivacity, or reflec‑ tively surveying alternative possibilities or prior occurrences (T 1.3.12.7). On this account, judgements of matters of fact are not originally discur‑ sive. They do not involve assent to a proposition or sentence of a mental language. They involve experiencing a perception more vivaciously (Sug‑ den 2021, 838–9). Acts of agreement with spoken or written sentences are consequences of this experience. There are also judgements concerning relations between ideas (EHU 4.1; T 1.3.1). They are based on the intuited or demonstrated impossibility of conceiving ideas to be differently related (T 1.3.7.3). They involve no vivacity. But they are still not discursive. Assent is based on the incon‑ ceivability of the opposite. As Hume remarked in a notorious footnote (T 1.3.7.5n), all judgement reduces to conception. Whereas judgements of relations of ideas are incontrovertible, judge‑ ments of matters of fact have a degree of certitude, proportioned to the vivacity of the perception. Assent is still either given or withheld. There is a degree of vivacity, corresponding to a 50 percent probability, above which assent is given, and at or below which it is withheld. The greater the

Section 6.3  211 vivacity beyond that point, the more resistant the judgement is to contrary evidence (Section 6.5). Hume later supplemented this positive identification of belief and vivac‑ ity with two negative claims: that belief does not involve conjoining the idea of existence with the idea of what is believed (T 1.3.7.3) and that belief in a proposition describing a matter of fact does not depend on the ideas we conceive in accord with the proposition (T 1.3.7.4). Someone who believes such a proposition and someone who disbelieves it form the same ideas. The difference between belief and simple conception does not lie in what is conceived, but in the manner in which it is conceived (T 1.3.7.3). In the Abstract (19–21), the second of these negative observations is paired with a reductio: Belief could not be a new idea joined to those we may conceive without assent, because in that case, it would be in our power to believe anything by joining the new idea to it. Appendix 2 invokes the same reductio to support the first negative observation, that belief does not arise from joining the idea of reality or existence to the simple conception of an “object.” Both works conclude that belief must be a feeling or senti‑ ment. Incidentally, they both add that it does not depend on the will. (In both works, the reductio baldly assumes that belief does not depend on our power.) Seen in this light, the Abstract and the Appendix take involunta‑ rism as a further reason for accepting that belief is a more vivacious idea, though the point is somewhat muted. In the Enquiry, it is anything but muted. The first thing Hume had to say about belief in the Enquiry is that it does not depend on the will (EHU 5.10–11, drawing on the immediately prior strong involuntarism of EHU 5.8), and the reductio is there Hume’s only reason for concluding that belief must be a feeling or sentiment. In 1739 Hume never went quite so far as to say that belief is not in our power.7 6.3

Attention and association

Vivacity is bound up with another operation, attention. All the operations of the mind depend in a great measure on its disposi‑ tion, when it performs them; and according as the spirits are more or less elevated, and the attention more or less fix’d, the action will always have more or less vigour and vivacity. When therefore any object is pre‑ sented, which elevates and enlivens the thought, every action, to which the mind applies itself, will be more strong and vivid, as long as the disposition continues. (T 1.3.8.2) Things crop up in this passage that are foreign to everything said up to this point. The Treatise contains only two prior occurrences of the phrase,

212  Belief “operations of the mind,” both found just three paragraphs earlier at Trea‑ tise 1.3.7.7, a passage only later directed for insertion at this point (T2, 677). The prior occurrences speak of a particular operation of the mind: belief. The term “disposition” is also used in a new way. Prior occurrences of that term only refer to the “disposition” of superior beings towards us and the “disposition” of perceptions in space and time. Here the reference is to a disposition of the mind, and it is puzzling what Hume meant by that. The same question can be asked about the passage’s reference to a fixation of attention. Unfortunately, Hume never addressed these questions. Plausibly, he used “operations” as a synonym for “actions.” Actions are paradigmatically in‑ stanced by body motions (T 1.4.5.27). Conjecturally, operations “of the mind” are actions involving the “motion” of perceptions. (Treatise 1.3.8.2 associates the motions of perceptions with a speculative psychophysics ap‑ pealing to motions of animal spirits.) Actions or operations of this sort include sensation (the appearance of an impression), copying impressions (the appearance of an idea), and the regular appearance of an idea conse‑ quent to the occurrence of an impression or idea. The last of these opera‑ tions, association, is introduced over Enquiry 3.1–3 and Treatise 1.1.4 and discussed over Treatise 1.3.2. It is said to be governed by three “gentle forces” or “natural relations” (T 1.1.4.1; 1.3.6.16). The occurrence of an idea or impression is generally followed by ideas that either resemble it, copy impressions or ideas that were once contiguous to it in space or time, or are causally related to it. Iteration of these operations produces a train of thought. The complexity of the momentary contents of consciousness poses a problem. Current consciousness includes impressions from all the senses. Those of vision and touch include many spatially disposed impressions. Added to them are memories, passions, and thoughts. What makes one of these components of current consciousness the focus of association and a subsequent train of thought? Reverting to Treatise 1.3.8.2, Hume said that according as attention is fixed, actions have more vigour and vivacity. Presumably, this means that according as attention is fixed on a component of current consciousness, associations involving that component occur more promptly, quickly, and extensively. Hume also said that action is stronger and more vivid whenever any object is presented that elevates and enlivens the thought. Presumably, objects that elevate and enliven “the thought” (that induce associations) are lively thoughts, that is, vivacious perceptions. Putting the two claims together, more vivacious perceptions fix attention, which is to say that they initiate more and longer associative chains.8

Section 6.4  213 6.4

Functional roles

In the Appendix, Hume observed that “some of my expressions have not been so well chosen, as to guard against all mistakes in the readers” (Ax 1), going on to focus on what he had said about the nature of belief. It would be going too far to interpret this as an expression of dissatisfaction or an occasion for second thoughts. Hume only remarked on failing to have effec‑ tively communicated his thoughts. The mistake was on the part of his read‑ ers, though it called for him to do more to explain himself. To comply, he offered a multi‑page review of his account of belief (Ax 2–9) supplemented with ampliative passages intended for insertion in the text of the Treatise. One of the ampliative insertions, Treatise 1.3.7.7, declares that “I find a considerable difficulty in [explaining belief]; and … even when I think I understand the subject perfectly, I am at a loss for terms to express my meaning.” For Hume, there is no alternative to considering belief to differ from fiction “not in the nature, or the order of its parts, but in the manner of its being conceiv’d” (T 1.3.7.7). However, he expressed frustration over being able to find words to “explain this manner.” A believed idea “feels” different from a fictitious one. He had tried to explain this feeling by “call‑ ing it a superior force, or vivacity, or solidity, or firmness, or steadiness.” However, this effort at explanation had not been very successful (perhaps given the mistakes in the readers alluded to at Appendix 1). The terms were “unphilosophical.” He then went on, in a lengthy passage that was recopied with only slight changes as Enquiry 5.12, to describe the “man‑ ner” at greater length. The inserted passage does something other than attempt to describe the feeling of belief. It treats belief as a kind of cause and appeals to its effects. There are five of them. 1 Belief draws attention. It expresses “that act of the mind, which ren‑ ders realities more present to us than fictions.” It “gives [the ideas of the judgment] more force and influence; makes them appear of greater importance; infixes them in the mind[.]” 2 It compels assent. It “distinguishes the ideas of the judgment from the fictions of the imagination.” 3 It inspires deliberation. What is believed “weigh[s] more in the thought” and has “a superior influence on the … imagination” 4 It arouses passion. What is believed has “a superior influence on the passions.” 5 It determines action. It “renders [the ideas of the judgment] the govern‑ ing principles of all our actions.”

214  Belief Appendix 3 offers a similar analysis. Beliefs (“the conceptions, which are the objects of conviction and assurance”) draw more attention than “loose and indolent reveries” (“They strike upon us with more force; they are more present to us; the mind has a firmer hold of them,” Ax 3). We “acqui‑ esce in them” (Ax 3) or consider them to be realities rather than fictions. They influence deliberation, passion, and action (we are “more actuated and mov’d by them” Ax 3). Taken in sum, these descriptions characterize belief in terms of what has been called its functional roles (Everson 1988). There is an important literature that debates the relation of functional roles to vivacity, and whether Hume’s thought about belief changed between 1739 and 1740.9 Here, following on Sections 6.2 and 6.3, it is premised that assent just is a perception with more than medium vivacity and that attention is an effect of vivacity. Assent and attention inspire deliberation and thereby arouse passion which in turn motivates action. The functional roles can therefore be identified with vivacity or its consequences. Appendix 3, Treatise 1.3.7.7 and Enquiry 5.12 emphasize “functional roles” that had figured in Hume’s account of belief from the start: assent (1739 1.3.5.7); the arousal of passion (1739 1.3.8.3; increasing “devo‑ tion” and “fervour,” that is, attention and deliberation (1739 1.3.8.4); and determining action (1739 1.3.8.6, in this case, the action of imitation). Conversely, in 1740 and later Hume retained the notion that belief dif‑ fers from fiction in its felt quality. Appendix 3 and Enquiry 5.11–12 con‑ tinue to insist that belief is a feeling or sentiment and continue to describe it as a forceful or vivacious idea. 6.5

The transfer thesis

Impressions and memories are originally vivacious or belief‑compelling. Ideas are not (T 1.1.3.1). The relations of resemblance, contiguity in space and time, and cause and effect dispose the imagination to form ideas in response to the occurrence of prior perceptions. According to Treatise 1.3.8.2, these relations also “communicate” vivacity from impressions to ideas. Enlivened ideas “diffuse” or “convey” their vivacity “as by so many pipes or canals, to every idea that has any communication with the pri‑ mary one” (T 1.3.10.7). The Treatise proposes a conveyance mechanism. When an impression gives rise to an associated idea, “The change of the objects is so easy, that the mind is scarce sensible of it” (T 1.3.8.2). In easily slipping into the place of the impression, the idea takes on its vivacity like an imposter ap‑ pearing in the clothes of the person being impersonated (T 1.4.2.41). Hume appealed to various “experiments” (T 1.3.8.3) to justify this the‑ ory (T 1.3.8.3–6; recopied at EHU 5.15–18). They show that the appear‑ ance of an impression enlivens ideas associated with it. All three of the

Section 6.5  215 associative relations have this effect. Suppose someone has lost a friend. There are various circumstances under which they receive especially lively ideas of the friend: coming across a picture; travelling to a place they once frequented together; and coming across the friend’s handiwork. These enlivened ideas arouse passion (the person who views the picture of the friend finds that “every passion, which that idea occasions, whether of joy or sorrow, acquires a new force and vigour” [T 1.3.8.3; repeated at EHU 5.15]). They inspire deliberation and draw attention (religious icons are said to enliven the devotion and quicken the fervour of the believers [paraphrasing T 1.3.8.4; repeated at EHU 5.16]). They motivate action (viewing the handiwork of a saint, or other objects handled by them, as‑ sists in imitating the exemplary life of the saint [T 1.3.8.6; repeated at EHU 5.18]). Enlivened ideas can even nudge people towards assent to the reality of the enlivened idea. After the death of any one, ’tis a common remark of the whole family, but especially of the servants, that they can scarce believe him to be dead, but still imagine him to be in his chamber or in any other place, where they were accustom’d to find him. (T 1.3.9.18)10 However, Hume maintained that only the causal relation can fully real‑ ize the last of these effects. “’Tis only causation, which produces such a connexion, as to give us assurance from the existence or action of one object, that ’twas follow’d or preceded by any other existence or action” (T 1.3.2.2). Causation is one–one. “[E]ach impression draws along with it a precise idea” (T 1.3.9.7). Resemblance and contiguity are one–many. Any given complex impression resembles many others and will often be of a sort that has been found in many different neighbourhoods. Unfortunately, Hume’s account of the influence of these circumstances is magical to the point of incoherence.11 (OMA): One–many association There is no manner of necessity for the mind to feign any resembling and contiguous objects; and if it feigns such, there is as little necessity for it always to confine itself to the same, without any difference or variation. And indeed such a fiction is founded on so little reason, that nothing but pure caprice can determine the mind to form it; and that principle being fluctuating and uncertain, ’tis impossible it can ever op‑ erate with any considerable degree of force and constancy. The mind … even from the very first instant feels the looseness of its actions, and the weak hold it has of its objects. (T 1.3.9.6)

216  Belief This is only the second occurrence of “feign” in the Treatise (the first is a few lines earlier in this paragraph), and the first of “caprice.” These are not technical terms. The talk of a “mind” that “feels” its actions or its “hold” on objects to be “weak” or “loose” is likewise metaphorical. On an ordinary understanding of the terms, “capriciously feigning” an idea is the contrary of forming an idea by association. But these ideas are said to be due to association. Hume offered a better account over Treatise 1.3.11–12. He was there concerned with explaining what determines us to proportion the strength of our beliefs in accord with the results of a mathematical calculation. We do this when playing games of chance or determining how much to stake on the outcomes of “inconstant causes,” like favourable winds. For Hume, this is an amazing operation. A calculation of the proportion of outcomes of a certain sort over the total possible outcomes (“chances”) or the total heretofore observed outcomes (“inconstant causes”) is simply the opera‑ tion of dividing one number by another. It produces a “relation of ideas.” Relations of ideas are accepted because no contrary is conceivable. They do not have or convey vivacity. But in playing games of chance or antici‑ pating the effects of inconstant causes, we form vivacious ideas (beliefs) about merely possible outcomes. Treatise 1.3.11–12 explains how this happens by proposing that the vi‑ vacity attendant on an impression (of, say, a toss of a die or the sight of a ship sailing out of port) is divided among the ideas of all the possible or prior outcomes. Where possibilities or priors are similar, their vivacity combines. Should one cluster of similar outcomes have more than half of all the vivacity, it uses a portion equal to the combined vivacity of its rivals to cancel the vivacity of the rivals and is believed with a degree of confi‑ dence equivalent to the proportion of vivacity that remains. Conviction arises when there are no rival outcomes; assurance diminishes to nothing at the point where there is a 50–50 chance of the outcome. The implication for the case of associating a perception with one of many resembling or contiguous perceptions is obvious. Because the asso‑ ciation in those cases is one–many, the vivacity of the original impression or memory is distributed to a variety of ideas, no one of which receives enough to counterbalance its rivals. The ideas are all enlivened somewhat, but not enough to produce belief. This account relies on the transfer thesis. It is not the only account to do so. Over Treatise 1.3.9–13, Hume invoked the transfer thesis as the foun‑ dation for an extensive and plausible psychology of belief formation and human motivation. However, there is some question whether he continued to be committed to it (Dauer 1999).

Section 6.6  217 6.6

The annexation thesis

1739 1.3.7.2 argues that because belief could not be the product of annexing a further idea to what is simply conceived, it must be a more vivacious man‑ ner of conception. In 1740, first in the Abstract and then some eight months later in the Appendix, Hume made it clear that a principal motivation for drawing this conclusion was the intuition that belief is involuntary (Ab 18–20; Ax 2).12 The Appendix also recognizes that 1739 1.3.7.2 had neglected an alternative. Rather than arise from an annexed idea, belief might arise from an annexed impression (Ax 4). This alternative does not challenge the intui‑ tion that belief is not up to us. Impressions cannot be produced at will. Having raised this possibility, Hume put considerable effort into reject‑ ing it (Ax 4–7). However, he changed his mind in the Enquiry. Enquiry 5.11 declares that “the difference between fiction and belief lies in some sentiment or feeling, which is annexed to the latter, not to the former.” This could not have been an accident. Portions of Enquiry 5.10–11 resem‑ ble Appendix 2 closely enough to be a paraphrase,13 and 5.12 goes on to quote 1740 1.3.7.7 (in Hume’s lifetime only published in the Appendix). Hume must have had the text of the Appendix before him while compos‑ ing Enquiry 5.11. He must have deliberately switched from denying that belief involves an “annexed impression” to affirming that it does. This is not a trivial change.14 At Appendix 3, Hume had observed that what we say about belief has bearing on whether we can find an analogy between it and other operations of the mind. If we can find an analogy, we can hope to “explain its causes” and “trace it up to more general principles.” If we cannot, “we must despair of explaining its causes” and “consider it as an original principle” (Ax 3). “This [“analogy” or “ap‑ proach” of belief to impressions] I have prov’d at large,” Hume went on to assert (Ax 9). Most likely, he was referring to the experiments he had cited to prove that the belief that arises from causal inference is analo‑ gous to the enlivening of ideas produced by resemblance and contiguity (T 1.3.8.3–12). The general rule proven by the analogous cases, that vivac‑ ity is transferred across associative relations, is foundational for Hume’s psychology of belief formation. The Hume of the Appendix seems to have thought that this develop‑ ment would be scuttled if belief were only an impression “annex’d to [the conception], after the same manner that will and desire are annex’d to par‑ ticular conceptions of good and pleasure” (Ax 4). In that case, he charged, belief would have to be due to an “original principle.” The mature Hume was comfortable invoking original principles. He did so when accounting for custom at Enquiry 5.5, for the probability of chances at 6.3, for faith in the senses at 12.7, and for humanity (replacing

218  Belief the sympathy mechanism) at Morals 5.17n. But the opportunity to con‑ firm his account of belief by showing it to be a particular case of a law that covers analogous phenomena was too important for him to give up. Enquiry 5.13 still proposes “to find other operations of the mind analo‑ gous to [belief], and to trace up these phænomena to principles still more general.” Enquiry 5.15–18 downloads the description of analogous phe‑ nomena given over Treatise 1.3.8.3–6 with no significant alteration.15 The enlivening of ideas by resemblance and contiguity is still considered to be analogous to the belief produced by causal inference, and our ability to unite all these phenomena under a covering law is still touted as confirma‑ tion of the account of causal inference. However, the covering laws identified in the two works are not the same. I wou’d willingly establish it as a general maxim in the science of human nature, that when any impression becomes present to us, it not only transports the mind to such ideas as are related to it, but likewise com‑ municates to them a share of its force and vivacity. (T 1.3.8.2) Does it happen, in all these relations, that, when one of the objects is presented to the senses or memory, the mind is not only carried to the conception of the correlative, but reaches a steadier and stronger con‑ ception of it than what otherwise it would have been able to attain? This seems to be the case with that belief, which arises from the relation of cause and effect. And if the case be the same with the other relations or principles of association, this may be established as a general law, which takes place in all the operations of the mind. (EHU 5.14) At Enquiry 5.14, there is transport (the mind is “carried” from one percep‑ tion to the “correlative” or associated one) but no mention of a transfer or communication of vivacity.16 The “steadier and stronger conception” of the associated idea appears as a brute fact. Admittedly, Enquiry 5.14 only asks a question. But, at the close of the investigations of the ensuing Enquiry 15–19, the proposed answer is confirmed. There is nothing to ac‑ count for the “strong conception” other than the transition. [W]hat is there in this whole matter to cause such a strong conception, except only a present object and a customary transition to the idea of another object, which we have been accustomed to conjoin with the for‑ mer? This is the whole operation of the mind …; and it is a satisfaction to find some analogies, by which it may be explained. (EHU 5.20)

Section 6.6  219 The most striking indication that the mature Hume may have retreated from the transfer thesis comes from what Enquiry 6 has to say about the probability of chances and inconstant causes (Dauer 1999). In the Trea‑ tise, the account of how we come to proportion belief in accord with a mathematical calculation is a showcase for the explanatory power of the transfer thesis (Section 6.5). It is absent from the Enquiry. When discuss‑ ing the case of “chances,” all that Hume said was that, when surveying the various possibilities, the mind comes upon one type of occurrence more frequently, and that this “begets immediately, by an inexplicable contriv‑ ance of nature, the sentiment of belief” (EHU 6.3). The “sentiment” is not quantified by appeal to addition and subtraction of portions of an origi‑ nally transferred amount of vivacity. How it comes to be attached to the more frequent kind of outcome is declared to be inexplicable. Similarly, when discussing inconstant causes, Hume recounted how we conceive priors in the same proportion as they have appeared in the past and then discover, upon surveying them, that a great number of views concur in one event. But once again, he made no attempt to measure the strength of the resulting belief or explain why it reflects the result of a mathematical calculation. He simply remarked that “As a great number of views do here concur in one event, they fortify and confirm it to the imagination, beget that sentiment which we call belief, and give its object the preference above the contrary event” (EHU 6.4). He declared this op‑ eration of the mind to be unaccountable on any of the received systems of philosophy and represented himself as aiming to do no more than “present hints” that might “excite the curiosity of philosophers” while illustrating the defects of all common theories.17 This is strong evidence for considering the Enquiry to have abandoned the transfer thesis, but it is not decisive. Enquiry 10.4 shows that the trans‑ fer thesis continued to have a hold on Hume. That discussion does quan‑ tify the strength of belief and does so in accord with the formula presented in Treatise 1.3.11–12. A wise man, therefore, proportions his belief to the evidence. In such conclusions as are founded on an infallible experience, he expects the event with the last degree of assurance.… In other cases …: He weighs the opposite experiments[.] In all cases, we must balance the opposite experiments, where they are opposite, and deduct the smaller number from the greater, in order to know the exact force of the superior evidence. This passage does not mention a transfer of vivacity from impressions to ideas or a distribution of vivacity over alternative outcomes. But it still speaks of balancing the alternative outcomes against one another,

220  Belief deducting the smaller number from the greater, and forming a belief that is proportioned to that result. The claim that opposed experiments cancel one another’s influence was too central to the project of Enquiry 10 to abandon. Hume needed to be able to say that when an exception to all past experience is reported, a wise person will weigh the likelihood of the oc‑ currence of that event against the likelihood of the truth of the report and, even where reports of that sort have never been found false, the outcome is “a mutual destruction of arguments” (EHU 10.13). The mechanism re‑ sponsible for proportioning belief to the evidence is still accepted. And since belief is still described as “a more vivid, lively, forcible, firm, steady conception of an object” (EHU 5.12), the mechanism works by cancelling vivacity. Enquiry 5 also contains two passages that do allude to transfer. Enquiry 5.16 speaks of sensible objects “conveying” their “influence on the fancy” to related ideas. This is a passage that was recopied from 1739 1.3.8.4. It might be objected that Hume recopied it without noticing that it needed correction.18 But between 1748 and 1777, the original text of the Enquiry underwent 10 revisions at Hume’s hand without his noticing that any al‑ teration was needed to this passage, though he did make other corrections to the paragraph that contains it (EHU, 243). Be this as it may, Enquiry 5.20 speaks (without quoting the Treatise) of “thought” moving instantly toward the idea and “conveying” to it “all that force of conception, which is derived from the impression present to the senses.” Hume was still add‑ ing new words in support of the transfer thesis in 1748 and he left them unchanged through the subsequent editions. Hume may not have rejected the transfer thesis, but simply decided not to bring it up. Enquiry 5 is split into two parts. In the first, Hume argued that we draw inferences from one object or event to another after having found similar objects or events to be constantly conjoined (EHU 5.3–6). He then added that the inference depends on some fact being present to “the senses or memory” (EHU 5.7), and concluded that “All belief of mat‑ ter of fact or real existence is derived merely from some object, present to the memory or senses, and a customary conjunction between that and some other object” (EHU 5.8). He had earlier suggested that we might not be able to explain why custom has this influence on us (EHU 5.5). But En‑ quiry 5.9 proposes to make the attempt. Readers are warned that the ensu‑ ing, second part of Enquiry 5, where this project is executed, will contain “speculations, which however accurate, may still retain a degree of doubt and uncertainty” and advised that they may pass them over if they have no taste for such things. It appears that Hume had some hesitation about bringing up even as much as he did over Enquiry 5.10–22. This does not mean that he had no commitment to the transfer thesis or no use for it. He

Section 6.6  221 was thinking of his audience and trying to strike a balance between being instructive and being engaging. Hume may have come to think that the transfer thesis is too speculative and that he was on firmer ground to appeal to an original principle. But this is dubious. Retreating from expressing a thesis is not the same thing as abandoning it. Hume wanted to say more than just that custom produces belief. When discussing miracles, he recognized that there can be opposed customs and he wanted to ascribe varying degrees of force to them and speak of a result that emerges when the force of one is subtracted from the force of the other. The transfer thesis and the account of how belief comes to be ascribed in conformity with calculations of probability that he had presented over Treatise 1.3.11–12 continue to guide his later thoughts. This makes it rather more likely that he continued to endorse the transfer thesis. As demonstrated by the history of modifications made to the multi‑ ple lifetime editions of his collected works, Hume was not much given to changing his mind about things. As he further confessed, he was also unwilling to make any alteration that might be construed as a reply to an objection raised by any of his increasingly numerous critics (L 1 #194, 360 [to George Campbell on 7 June 1762], MOL 9). It is hard to find evidence of evolution or change in his thought. Something that might seem like an innovation, for example, a shift to a functional role account of belief in the Appendix, proves on closer examination to have only been a way of emphasizing something that had been central to the account from the start. Something that might seem like it has been abandoned, like the transfer thesis, proves on closer examination to only be something Hume decided he did not need to go into, even though it continued to govern his thought. Since the transfer thesis continues to be intimated in the Enquiry, it is taken to be authorized in what follows. This leaves Hume’s switch to taking belief to be an annexed sentiment unexplained. He may simply have thought better of the reservations he had expressed in the Appendix. All of them can be addressed. Sentiments can also have varying degrees of strength, force, or vivacity and they occur independently of the will so on either account, belief is a superior degree of vivacity and is involuntary (consistently with Ax 2). Taking an annexed sentiment to be aroused by the same causes that produce sensations, pre‑ served to some extent in memory, and communicated across associative links accommodates the transfer thesis and preserves the possibility of ac‑ counting for belief by appeal to a law that also explains the analogous enlivening due to resemblance and contiguity (answering Ax 3 and 6). The notion of a “vivacious manner of conceiving” is vague and not clearly dis‑ tinct from that of feeling an annexed sentiment (answering Ax 4). The lat‑ ter is a recognized element of Hume’s philosophy (answering Ax 5 and 7).

222  Belief 6.7

Unphilosophical probability

At Treatise 1.3.9.2, Hume remarked that the analogous cases he had cited to confirm his account of belief (T 1.3.8.3–6; EHU 5.16–18) might in‑ stead be raised as an objection to it. If inferences from cause and effect are believed because association with a lively perception enlivens an idea, and this is confirmed by the fact that all three of the associative relations enliven ideas, why does association by resemblance and contiguity not pro‑ duce belief as well? Hume could have replied that contiguity and resemblance do produce belief (Sections 4.1, 4.2, 6.10, and 6.11). He chose instead to maintain that they do not communicate enough vivacity to surpass the threshold at which enlivened ideas become beliefs. His best available reason was that resemblance and contiguity disperse vivacity over more than one alterna‑ tive, leaving no one with enough to cancel the combined weight of its rivals (Section 6.5). “But not content with removing this objection,” he wrote, “I shall endeavour to extract from it a proof of the present doctrine” (T 1.3.9.8). He went on to offer instances to prove that the strength of causal inferences is enhanced by resemblance or contiguity between cause and ef‑ fect, and diminished by difference and remoteness. This remark launched him on a wide‑ranging investigation of various kinds of “unphilosophical probability”: circumstances where the strength of beliefs falls out of line with what it would be were it determined by the proportion of confirming and contrary experiments.19 1 The influence of contiguity and resemblance. Hume initiated his pro‑ ject by citing instances where contiguity and resemblance strengthen or weaken belief: Believers who have gone on pilgrimages experience their belief to be enlivened by impressions of the landscape contiguous to where the events they believe occurred (T 1.3.9.9; HE 13, 76). The hypothesis that motion can only be the effect of collision with a moving body gains disproportionate plausibility because it posits a cause that resembles the effect (T 1.3.9.10). On the contrary, when causes and ef‑ fects are dissimilar, belief is weakened or destroyed. The lack of resem‑ blance between our current lives and our ideas of an afterlife makes it difficult to sincerely believe that wickedness in this life will cause us to suffer eternal damnation (T 1.3.9.13–14). 2 Diminished vivacity sources. If belief results from a transmission of vi‑ vacity from a vivacious perception to an associated idea, it must vary in strength depending on how much vivacity is originally available to be transferred. Memories are less vivacious than impressions, and memories of older events are less vivacious than those of recent ones, so the inference from a remembered cause or effect will be less vivacious (T 1.3.13.1).

Section 6.7  223 3 Old experiments. The same effect applies to cases where the connec‑ tion between cause and effect has not been witnessed for some time. The bare passage of time fades the memory, making us less disposed to suppose that the cause will continue to be followed by its effect, even though we have not witnessed any contrary experiments (T 1.3.13.2). 4 Eloquence and the influence of passions. Passions transmit vivacity to whatever other perceptions occur along with them. Should the percep‑ tion of an effect or a cause arouse passion, more vivacity will be trans‑ mitted to the associated idea than would otherwise be the case, and the belief will be stronger than normal (EHU 10.16 and 10.18; T 1.3.10.4; T 1.3.13.10). Hume drew special attention to eloquence (the art of speaking so as to arouse the passions of an audience) and the passions of surprise and wonder (EHU 10.18, 10.16; T 1.3.10.8, 1.3.10.4). 5 Limited carrying capacity. “Distant” inferences. The “carrying ca‑ pacity” of associative links also has a role to play. The main circum‑ stances affecting carrying capacity are “imperfect experience,” where we have not encountered enough instances of a conjunction to form a strong association (T 1.3.12.2); “contrary experiments,” where the cause or effect is not always observed to be accompanied by its partner (T 1.3.12.4–24); and “analogy,” when the present impression does not perfectly resemble the usual cause or effect (T 1.3.12.25). Inferential “distance” also plays a role. An impression or memory never conveys all its vivacity to an associated idea, and when further causal inferences are drawn from that idea, they never acquire all the vivacity of that idea. The more steps in the chain, the more vivacity diminishes, making more “remote” inferences less strongly believed (T 1.3.13.3).   But this does not always happen. When the causes and effects in the chain all resemble one another, the diminution in transferred vivacity is reduced or even eliminated. We owe our trust in the testimony of ancient historians to this purportedly “unphilosophical” effect of re‑ semblance (T 1.3.13.4–6). 6 Complexity. Prejudices. Causes are typically accompanied by various other circumstances. If these “superfluous” circumstances are more no‑ ticeable and are frequently conjoined with the circumstance that is “es‑ sential” or “absolutely requisite to the production of the effect,” they can be mistaken for it (1.3.13.9). We might trust that continued experi‑ ence will eventually reveal the error, but there is a complicating factor. We do not always find cases where the effect occurs in the absence of the superfluous circumstances to outweigh cases where the two occur together, even as our experiences become more wide‑ranging. The cus‑ tomarily present superfluous circumstances still transmit vivacity to the associated ideas and so produce belief (T 1.3.13.8–9). This makes us, as Hume put it, “prejudiced” (T 1.3.13.7). Sometimes we can correct these

224  Belief prejudices by reviewing the experiments, but “’tis still certain, that cus‑ tom takes the start, and gives a biass to the imagination” (T 1.3.13.9). We find it hard to discount the vivacity that continues to be communi‑ cated from the superfluous circumstances to the idea of the effect. 7 Other sources of vivacity. Education. Hume also recognized that there are other sources of vivacity in addition to impressions, passions, mem‑ ories, and association. Our opinions of all kinds are strongly affected by society and sym‑ pathy, and it is almost impossible for us to support any principle or sentiment, against the universal consent of every one, with whom we have any friendship or correspondence. (DP 2.33) He did not go any more detail. Conjecturally, he was following Enquiry 5.11 in taking belief to be a sentiment that has been annexed to an idea and appealing to a tendency to share in the sentiments of others.   The author of the Treatise offered a singular account of the influence of the opinions of others. When we repeatedly come across an idea by happenstance, as when having it presented to us by educators and acquaintances, the frequent, undesigned repetition endows it with vi‑ vacity (T 1.3.9.17 importantly supplemented by 1.3.12.23).20 On either the early or the later account, we end up with a belief that is gathered from the opinions of others rather than from experience of a constant conjunction. Considering these points, it is not correct to say that “All belief of matter of fact or real existence is derived merely from some object, present to the memory or senses, and a customary conjunction between that and some other object.”21 There is more at work than just a customary conjunc‑ tion. Hume’s work on this topic, the subject matter of Enquiry 10.16–19, Passions, the political, economic, literary, and moral essays, the Natural History, the History, and Treatise 1.3.9–10 and 13, is something of a tour de force. In an age before the emergence of an advertising industry or the concept of propaganda (but not before organized religion or a history of the effects of “eloquence” in ancient assemblies) Hume had put his finger on many of the factors that explain how belief can arise and be sustained in defiance of the evidence (Sugden 2021). 6.8

The normativity question

Hume remarked that some forms of causal inference are “receiv’d by phi‑ losophers, and allow’d to be reasonable foundations of belief and opinion. But there are others, that are deriv’d from the same principles, tho’ they

Section 6.8  225 have not had the good fortune to obtain the same sanction” (1.3.13.1). What justifies philosophers in approving of some causes of belief while condemning others? Answer 1. Hume observed that education is not “recogniz’d by philo­ sophers” because it is “an artificial and not a natural cause [of belief], and … its maxims are frequently contrary to reason, and even to them‑ selves in different times and places” (T 1.3.9.19).22 But he had reason to distance himself from this justification.   When education delivers conflicting results, it tends to be because it does not speak with one voice. Confronted with conflicting opinions, individuals may react as they do when confronted with “inconstant causes”: feel each opinion to have a weight proportioned to how fre‑ quently it is encountered and prefer the one that outweighs the com‑ bined weight of the opposition.23 Or they may feel the opinions of those to whom they are most closely related by neighbourhood, blood, occu‑ pation, party or other associations to have greater weight.   Seen in this light, education (and “society and sympathy” [DP 2.33]) are not artificial causes of belief. Advertisers, educators, and politi‑ cians try to artificially induce people to believe something by repeatedly exposing them to some message, taking care to do so at random and unanticipated times and places (T 1.3.12.23). But their artifice arises from understanding and exploiting how we are naturally determined to form beliefs, not from doing something contrary to nature. Advertising works to the extent that it imitates natural causes. Answer 2. Hume observed that there is a distinction to be drawn between principles of the imagination that are “permanent, irresistible, and uni‑ versal; such as the customary transition from causes to effects” and other principles that are (i) “changeable, weak, and irregular.”24 Whereas the former are “the foundation of all our thoughts and actions” and neces‑ sary if we are not to “immediately perish and go to ruin” the latter are (ii) avoidable, unnecessary, and not even “so much as useful in the con‑ duct of life; but on the contrary are [iii] observ’d only to take place in weak minds.” They (iv) “may easily be subverted by a due contrast and opposition.” For these reasons “the former are receiv’d by philosophy, and the latter rejected.” The latter are still natural, “But then it must be in the same sense, that a malady is said to be natural; as arising from natural causes, tho’ it be contrary to health, the most agreeable and most natural situation[.]”   Hume made these observations about tendencies of the imagina‑ tion leading ancient philosophers to suppose that collections of related perceptions are a single perception. It is a stretch to suppose that he

226  Belief accepted that they also apply to the “unphilosophical” causes of belief catalogued in Section 5.7 (compare Loeb 2013, 313–5, 317–8).

i Far from being “changeable, weak, and irregular” Hume described education (or “society and sympathy”) as producing beliefs that “take such deep root, that ’tis impossible for us, by all the pow‑ ers of reason and experience, to eradicate them” (T 1.3.9.17). He noted that astonishment “so vivifies and enlivens the [astonishing] idea, that it resembles the inferences we draw from experience” (T 1.3.10.4), and that the vivacity produced by eloquence “is in many cases greater than that which arises from custom and experience” (T 1.3.10.8). Prejudices “influence [our] judgment, even contrary to present observation and experience” (T 1.3.13.8). ii Far from being avoidable, unnecessary and useless the “unphilo‑ sophical” influence of resemblance on causal inferences is necessary for us to place trust in any but the most recent of historical accounts (T 1.3.13.4–6); the same operations that lead us to form prejudices are the ones that correct those prejudices (T 1.3.13.12); and the be‑ lief in an external world depends (at least in the estimation of the au‑ thor of the Treatise) on “trivial qualities of the fancy” (T 1.4.2.56). iii Far from being only observed to take place in weak minds, “’tis sel‑ dom that [long chains of reasoning] produce any conviction; and one must have a very strong and firm imagination to preserve the evidence to the end” (T 1.3.13.3). Philosophy may not “receive” the diminu‑ tion in degrees of evidence that results from the temporal remote‑ ness of vivacity sources or of experiments,25 but the philosophers who make this valid observation prove to be unable to abide by it. [Y]et notwithstanding the opposition of philosophy, ’tis certain, this circumstance has a considerable influence on the understand‑ ing, and secretly changes the authority of the same argument, according to the different times, in which it is propos’d to us. A greater force and vivacity in the impression naturally conveys a greater to the related idea; and ’tis on the degrees of force and vivacity, that the belief depends[.] (T 1.3.13.1, my stress) Even if we are philosophers, it is not within our capacity to bring ourselves to believe something by mere dint of will, even when we think we ought to do so. Philosophers, like everyone else, believe what is enlivened to the extent that it is enlivened. Loss of vivacity cannot simply be countermanded by fiat. [B]elief is the necessary result of placing the mind in such circum‑ stances. It is an operation of the soul, when we are so situated, as

Section 6.8  227 unavoidable as to feel the passion of love, when we receive benefits; or hatred, when we meet with injuries. All these operations are a species of natural instincts, which no reasoning or process of the thought and understanding is able, either to produce, or to prevent. (EHU 5.8)

iv Far from being easily “subverted by a due contrast and opposition,” beliefs induced by education or inflamed by eloquence or passion are liable to prevail in the contest with reason, custom, and experi‑ ence. For lack of the requisite vivacity, beliefs weakened by temporal distance or long chains of argument will simply not be accorded the strength that philosophers suppose to be their due.

Answer 3. Hume recognized a distinction between two kinds of believers, “the vulgar” and the “wise” (T 1.3.13.12)26 The distinction arises from how believers handle complexity. As noted earlier, causes are complexes of many circumstances, not all of which are always present on the oc‑ casions when the effect occurs. Philosophers consider the proper cause to consist only of those circumstances that are present if and only if the effect occurs and the other circumstances to be accidents that could just as well be present in other cases where the effect does not occur or absent in other cases where it does occur (T 1.3.15.6). But when the “accidental circumstances” are frequently conjoined with the “ef‑ ficacious causes” (T 1.3.13.11) that regular conjunction influences the imagination and leads us to connect the former as well as the latter with the effect. Reflectively, we condemn this operation, but we nonetheless feel the vivacity transfer it induces. Sometimes the one, sometimes the other prevails, according to the disposition and character of the person. The vulgar are commonly guided by the first, and wise men by the second. (T 1.3.13.12) Presumably, the difference is due to how opposed vivacity sources can‑ cel one another in the circumstances of the person in question. A person who has the leisure, the resources, the opportunities, the inclination, and the aptitude may repeatedly discover that in cases where a merely presumed cause fails of its anticipated effect, or the effect emerges in the absence of the presumed cause, a more exact scrutiny will reveal some previously hidden circumstance that is the true concomitant of the effect (T 1.3.12.5). This experience enlivens a counter prejudice (so to speak) to the effect that “The same cause always produces the same effect, and the same effect never arises but from the same cause” (T 1.3.15.6). Provided this general rule has been sufficiently enlivened, it can outweigh the contrary prejudice acquired from regularity in the

228  Belief union of the accidental circumstances with the effect. Seeing a presumed cause fail to be followed by its anticipated effect, leads the wise person to form the vivacious idea of some hidden cause rather than ascribe a diminished probability to the likelihood of the succession.   But, for Hume, it does not follow that the wise know better. They are simply governed by a different prejudice. Mean while the sceptics may here have the pleasure of observing a new and signal contradiction in our reason, and of seeing all philosophy ready to be subverted by a principle of human nature, and again sav’d by a new direction of the very same principle. The following of general rules is a very unphilosophical species of probability; and yet ’tis only by following them that we can correct this, and all other unphilosophical probabilities. (T 1.3.13.12) The wise have experienced some success at finding hidden causes. But their success has not been perfect. According to the mature Hume, “Though there be no such thing as Chance in the world; our ignorance of the real cause of any event has the same influence on the understanding, and begets a like species of belief or opinion” (EHU 6.1). The author of the Treatise devoted a substantial Section (1.3.12) to explaining belief in the likelihood of the outcomes of “inconstant causes.” That would have been pointless work if the philosophers of his day had obtained any great degree of success at discovering the true causes of effects. The “same cause same effect” principle is just another prejudice, arising from a limited degree of success. In the preponderance of cases, philosophers are reduced to the same situation with the vulgar. Being unable to discover hidden efficacious circumstances, they are left attempting to determine the likelihood of the outcome of inconstant causes. Their supposition that there must be a hidden cause, even in these cases, is a hasty generalization from their limited success with just a few of them. It is just another prejudice.   It only gets worse. The evidence can only be yesterday’s evidence, and Hume argued convincingly that we have no justification for supposing that the future will be like the past. Whatever justification philosophers may come up with for proportioning belief to the evidence, it will be one based on the way the world has been up to now. Claiming that it is impossible to be sceptical on this topic because nature has not left it up to us what to believe is not saying anything that “the vulgar” could not say in defence of inference from recent, lurid anecdotes or implicit faith in the doctrines of demagogues and priests. Belief is not up to us, and as “ought” implies “can,” pejoratives such as “just,” “wise,”

Section 6.9  229 and “foolish” are inappropriate. At best, they might be replaced with “fortunate” and “unfortunate.” Any presumption that fortunes will not change is itself determined by circumstances, not “just.” The occasionally distant tone of Hume’s pronouncements on the topic of epistemic normativity is striking. There are causes of belief that “have not had the good fortune to obtain the … sanction” of philosophers, he wrote, even though they are “deriv’d from the same principles” (T 1.3.13.1). The ideas imprinted on us by education are not “recogniz’d by philosophers; tho’ in reality [they are] built almost on the same foundation of custom and repetition as our experience or reasonings from causes and effects” (T 1.3.9.19) The diminution of evidence with remoteness in time is “not re‑ ceived by philosophy as solid and legitimate” (T 1.3.13.1) and the neglect of older experiments “disclaim’d” by it (T 1.3.13.2) even though “this circumstance … secretly changes the authority” of arguments proposed at different times. The following of general rules “is a very unphilosophical species of probability” (T 1.3.13.12). In making these remarks about what philosophers sanction and disclaim, it is almost as if Hume were adding, in his silent voice, “for all the good it does them.” What was uppermost in his mind was that not everyone is determined to believe as philosophers are determined to think we ought to believe. What might a philosopher be determined to do about that? 6.9

Sceptical determinism

Hume’s frequently pejorative language shows that he could not confine himself to the observer’s role. (One instance is Treatise 1.3.13.12, on preju‑ dice: “when we take a review of this act of the mind, and compare it with the more general and authentic operations of the understanding, we find it to be of an irregular nature, and destructive of all the most establish’d principles of reasoning.”) This is only to be expected. He was himself sub‑ ject to the causes he had identified. His own experiences, aptitudes, and circumstances will have led him to develop prejudices against the existence of irreducibly stochastic events, the possibility of a change in the course of nature, the reliability of common opinions or of beliefs induced by passion, and so on. He will have felt “extorted … by the present view of the object” (T 1.4.7.15) to express those opinions and he will have been concerned that the members of governing parties if not the public at large (that being too much to be hoped for) adopt his opinions. This does not mean that he will have been especially concerned to justify his opinions, particularly not if the justification appeals to a long chain of argument, which his theory declares would have little effect. Like the priests and politicians around

230  Belief him, he will have thought that persuading others of his own convictions would require employing what his own account identifies as optimal meas‑ ures for enlivening ideas. This can mean applying unphilosophical causes of belief to correct other unphilosophical causes of belief (T 1.3.13.12). Largely abandoning the pursuit of philosophy, he turned to writing his‑ tories that highlight the bad consequences of failing to proportion belief to the evidence in numerous, striking, entertaining, eloquently presented cases. Though employing art and eloquence to arouse passions is an un‑ philosophical cause of belief, it is more likely to have an instructive effect than writing a treatise on the understanding.27 Hume did not just turn to this historical remedy. He concurrently advo‑ cated another. [I]f a man has accustomed himself to sceptical considerations on the uncertainty and narrow limits of reason, he will not entirely forget them when he turns his reflection on other subjects; but in all this philosophi‑ cal principles and reasoning, I dare not say, in his common conduct, he will be found different from those, who either never formed any opinions in the case, or have entertained sentiments more favourable to human reason. (D 1.8)28 The greater part of mankind are naturally apt to be affirmative and dogmatical in their opinions; and while they see objects only on one side, and have no idea of any counterpoising argument, they throw themselves precipitately into the principles, to which they are inclined; nor have they any indulgence for those who entertain opposite senti‑ ments.… But could such dogmatical reasoners become sensible of the strange infirmities of human understanding, even in its most perfect state, and when most accurate and cautious in its determinations; such a reflection would naturally inspire them with more modesty and re‑ serve, and diminish their fond opinion of themselves, and their preju‑ dice against antagonists. (EHU 12.24) Those who have once appreciated the force of sceptical arguments are changed by that encounter. They come away from it more inclined to dis‑ trust all their cognitive abilities. While their ideas are still naturally enliv‑ ened, their belief is mitigated by the memory of the considerations that can be advanced against accepting any proposition. They may be temporarily swayed, but eventually, they recall their doubts and return to their scepti‑ cal disposition. They can be more permanently swayed only by what is repeatedly impressed on them. They tend to discount speculative opinions

Section 6.9  231 and can believe only what can be confirmed by common life and experi‑ ence. If they do advance further, it is only on principles established by induction from common experience and experiment (EHU 12.25; DNR 1.9–11). Following this line of thinking, Hume would have thought that rather than attempt to convince others of the justice of our opinions, we would do better to expose them to the reasons for considering any opinion to be uncertain. As indicated by the discussions of the nature of various forms of scepticism and the virtues of some of them over Enquiry 12 and 5.1–2, Hume embraced scepticism as an effective remedy to the problem of un‑ philosophical belief. The views of the author of the Treatise are not so clear. In that work, Hume made it appear as if he had been driven into scepticism malgré lui. The introduction opens optimistically, proposing to overcome the irreso‑ lution and conflict that has hitherto infected philosophy by pursuing an inductively based science of human nature. But problems emerge, first with justifying causal inference and then with justifying demonstrative reason‑ ing. Hope is held out for discovering the principles leading us to believe in the existence of body, but they prove to be trivial qualities of the fancy conducted by false suppositions, a discovery that arouses some expressions of frustration. Hope returns with the assertion that the intellectual world is not involved in the same contradictions as the natural world, but a few months after publication, the Appendix confesses that a stricter review reveals that to be mistaken as well. Treatise 1.4.7 declares that we find ourselves confronted with a choice between an impotent understanding and an imagination that produces errors, absurdity and obscurities. It is not easy to glean a resolution from the rambling discourse that follows. One train of thought leads to a scep‑ tical recommendation. It is not possible for some people to resist delving into matters that cannot be readily confirmed by experience. They are driven to it by curiosity, ambition, and a desire to relieve the deplorable ignorance we lie under. But giving in to these passions puts them at risk of falling into dangerous superstitions. Hopefully, cognisance of this danger will lead them to conclude that “if we are philosophers, it ought only to be on sceptical principles” (T 1.4.7.11). We cannot help believing what is enlivened to a sufficient degree, be the cause of that enlivening custom or some other factor. But if we can manage to keep sceptical arguments before our eyes, we can hope that we will be determined to accept the re‑ sults of our research with diffidence and moderation and subject them to continual testing, thereby ensuring that we will only accept what is enliv‑ ened by the more stable sources of vivacity. On this reading, the Treatise is of a piece with the Enquiry and the Dialogues, despite its pretences to the contrary.

232  Belief Were the expressions of frustration and dismay found at the close of Treatise 1.4.2, and over the course of 1.4.7 just a show put on by someone who wanted from the start to lead his readers into scepticism, but hesitated to make that project clear for fear of causing them to put their guard up? Did the author of the Treatise force his hand in an attempt to lead his readers into a sceptical epoché he had from the start wanted to induce? To that end, did he refrain from drawing positive conclusions implicated by his account of belief, and draw negative conclusions that his account of belief cannot support? The remainder of this chapter argues that he did force his hand and that the mature Hume failed to correct and capitalize on that oversight. 6.10 Customary disposition in space Up to this point, this chapter has focused on Hume’s account of belief and how it gives rise to practical and epistemological problems with unphilo‑ sophical belief. The remaining sections argue that there is a serious mistake in Hume’s account (Falkenstein and Welton 2001). In both the Enquiry and the Treatise, Hume focused on just one kind of custom.29 When objects of one sort are customarily followed by objects of another sort, we form the habit of reproducing that course of events in our thinking, and when an object of either sort is vivaciously conceived, we form a vivacious idea of one of the other sort. But in both works his dis‑ cussion of the principles of association recognizes another form of custom. [A]s the senses, in changing their objects, are necessitated to change them regularly, and take them as they lie contiguous to one another, the imagination must by long custom acquire the same method of thinking, and run along the parts of space and time in conceiving its objects. (T 1.1.4.2; see also T 3.8.5 quoted at EHU 5.17) This passage speaks of objects30 that have been found to be customar‑ ily disposed in the same way in space. Mountains, valleys, trees, streets, buildings, and heavy furnishings are examples. Seeing or feeling one of these objects, we habitually form vivacious ideas of those that have been customarily found to lie to one side or another, behind or beyond. Even mobile objects are in many cases large, complex, and hard. Seeing or feel‑ ing one of them, we form vivacious ideas of its unperceived sides. “The mention of one apartment in a building naturally introduces an enquiry or discourse concerning the others” (EHU 3.3). Expanding on that point, an occupant of one of those apartments, forming the project of go‑ ing for a walk, will follow a train of thought from doors to hallways to stairwells to streets with their adjacent buildings, to parks, paths and trees,

Section 6.10  233 with their common surroundings. None of these objects is the cause or ­effect of any of the others. The train of thought and belief is determined by what has customarily been experienced to be spatially contiguous. Beliefs about what objects will be encountered next are constantly confirmed by following a route and finding the anticipated landmarks where expected. Hume occasionally remarked that, given an impression or memory, cus‑ tomary contiguity in space can produce belief in the unperceived contigu‑ ous objects. Occlusion is a particularly important case. Suppose I see the legs and thighs of a person in motion, while some interpos’d object conceals the rest of his body. Here ’tis certain, the im‑ agination spreads out the whole figure. I give him a head and shoulders, and breast and neck. These members I conceive and believe him to be possess’d of. (Appendix 4) For instance; I hear at present a person’s voice, whom I am acquainted with; and this sound comes from the next room. This impression of my senses immediately conveys my thoughts to the person, along with all the surrounding objects. I paint them out to myself as existent at pre‑ sent, with the same qualities and relations, that I formerly knew them possess’d of. These ideas take faster hold of my mind, than the ideas of an enchanted castle. (Appendix 4, my stress) The shoulders, breast, and neck are not causes of the legs and thighs. The voice is the effect of the person, but the familiar furnishings in the next room are neither causes nor effects of either the person or the sound. Belief is in these cases the result of inference from experienced or believed objects to other objects or object parts that have been found to be customarily contiguous to them in space. Though Hume made these pronouncements, he did so inadvertently, without considering how he might have exploited what he had written to develop and support his views. Worse, he maintained that, “By means of that relation [cause and effect] alone [can we]31 go beyond the evidence of our memory and senses” (EHU 4.4). The Treatise agrees. [W]e ought not to receive as reasoning any of the observations we may make concerning … the relations of time and place; since in none of them the mind can go beyond what is immediately present to the senses, either to discover the real existence or the relations of objects. ’Tis only causation, which produces such a connexion[.] (T 1.3.2.2)

234  Belief Treatise 1.3.2.2 and 1.3.9.6 offer two reasons for this restriction. (SC): Secret cause There is nothing in any objects to perswade us, that they are either always remote or always contiguous; and when from experience and observation we discover, that their relation in this particular is invari‑ able, we always conclude there is some secret cause, which separates or unites them. (T 1.3.2.2) There is something to persuade us that objects are always remote or always contiguous: experience of them being customarily remote or customarily contiguous. Given that experience, we have no need to appeal to anything “in” the objects that makes them so or any “secret cause” that holds them in place. The customary experience establishes a habit of thought that leads us, when experiencing one object at the edge of the sensory field or beside an occlusion to vividly imagine its customarily observed neighbours at their customary locations beyond the bounds of the sensory field or behind the occlusion. Custom, habit, and vivacity transfer are all that lead us to draw causal inferences from the customary succession of species of object. By parity of example, they ought to lead us to draw what might be called geographical inferences from the customary contiguity of particular objects. (SC) precedes Hume’s analysis of causality, which gives it an appearance of plausibility that is undermined by later developments. According to those developments, just as there is nothing to “perswade” us that objects are always remote or always contiguous, so there is nothing to persuade us that the same species of objects will always be followed by the same species of objects (T 1.3.14). In the case of cause and effect, we draw inferences from a perceived object to an unperceived object because we have been ha‑ bituated to follow a train of thought. We can just as well be habituated by customary conjunction in space as customary succession in time. If there is no call for any further “necessary connection” in one case, there is no call for a necessary connection in the other. In particular, there is no call to invoke causal inference to explain why certain things customarily abide in place. No explanation is necessary. The fact of the matter, established by custom, is sufficient. Hume’s second reason for discounting inferences from customary con‑ tiguity in space (T 1.3.9.6, cited as (OMA) in Section 6.5) is that conti‑ guity is too diffuse to produce belief on its own. Any species of cause or effect is associated with a single species of effect or cause. But any given object is placed in the neighbourhood of many contiguous objects, and of increasingly many the larger the neighbourhood is taken to be. Vivacity

Section 6.11  235 is dispersed over all the rivals, nothing determines us to always make the association with just one of them, and we feel a sense of caprice in fixing on just one of them. If we ever did start to form beliefs on such a basis, we would make so many mistakes that we would soon be cured of the practice. While (OMA) is true of association by (bare) resemblance, it neglects two features of association by customary contiguity in space: direction and custom. Objects are not seen in isolation but along with immediately surrounding objects, and except when we are lost, the immediately sur‑ rounding objects define known directions in space. When inferences are drawn from an experienced or remembered object to a customarily con‑ tiguous object, the inference is always drawn concerning what lies along a particular route. Other inferences could have been drawn instead, but they are inferences drawn along different routes. The various objects that might be inferred do not contend with one another for a location in space. The vivacity of any one inference is therefore not cancelled or subtracted by the vivacity of any of the others. Each route leads to just one object (if any) in the specified direction. While many objects lie at different distances along that route, more distant objects are only inferred by starting with more proximate ones and following a path out from there, the way remote causes or effects are inferred from more proximate ones (T 2.3.7.2). These inferences are as reliable as causal inferences. We do not draw inferences from the contiguity of just any objects, but only from those ob‑ jects that have been customarily found to be disposed at the same locations relative to one another over a long period of time. Where the objects are not always seen in the same locations relative to one another, our infer‑ ences are proportioned accordingly, just as our inferences from inconstant causes (T 1.3.12) are proportioned. Everything Hume said about causal inference could just as well have been said about inferences involving customary contiguity in space. 6.11 Identity In the same passage that rejects the possibility that belief might be based on customary disposition in space, Hume further maintained that “we ought not to receive as reasoning any of the observations we may make concern‑ ing identity” (T 1.3.2.2). He allowed that we can observe identity by keep‑ ing an eye or hand on an object over a period of time. We readily suppose an object [has continued] individually the same … whenever … if we [keep] our eye or hand constantly upon it, it [con‑ veys] an invariable and uninterrupted perception. (T 1.3.2.2)

236  Belief But in this case, there is no reasoning but only (temporally extended) “per‑ ception” of the object. Reasoning would be called for should the view or grip on the object be interrupted. In that case, we would have to identify its successor counterfactually, by appealing to which object we would have ended up looking at or holding had we kept an eye or hand on the original. We readily suppose an object may continue individually the same, tho’ several times absent from and present to the senses; and ascribe to it an identity, notwithstanding the interruption of the perception, whenever we conclude, that if we had kept our eye or hand constantly upon it, it wou’d have convey’d an invariable and uninterrupted perception. (T 1.3.2.2) According to Hume, such counterfactuals can only be supported by causal reasoning. But this conclusion beyond the impressions of our senses can be founded only on the connexion of cause and effect; nor can we otherwise have any security, that the object is not chang’d upon us, however much the new object may resemble that which was formerly present to the senses. Whenever we discover such a perfect resemblance, we consider, whether it be common in that species of objects; whether possibly or probably any cause cou’d operate in producing the change and resemblance; and according as we determine concerning these causes and effects, we form our judgment concerning the identity of the object. (T 1.3.2.2) Hume was right that in these cases we need assurance that the object has not been replaced with a twin. He was also right that this assurance is grounded on experience of what is common in that “species of object.” However, the experience of what is common in that species is prior to causal inference. Causal inference presupposes having first identified spe‑ cies as a condition for going on to notice regularities in their succession. This prior experience directly supports identity attributions. It must do so if causal inference is to be possible (Price 1940, 7–9). It is no longer possible to be cavalier about the difference between pub‑ licly observable objects and private impressions constitutive of our experi‑ ences of such objects. Even aside from interruptions, the “observations we may make concerning identity” (T 1.3.2.2) presuppose decisions about whether what is now under eye or in hand is the same collection of impres‑ sions that was under eye or in hand a moment ago.32 These decisions do not produce belief in the unobserved. They are decisions about how what

Section 6.11  237 is being observed now is related to what was observed a moment ago. But they still involve a form of association: one–one association by maximal proximity and resemblance (Sections 4.1–3). Suppose that there is a collection of simple impressions, C1, that, for whatever reason, catches our attention at a moment, t1.33 Because C1 is interesting, we do not want to let it fall out of sight. So, at the immediately following moment, t2, we examine the sensory field around the place, P, where C1 occurred.34 Provided we can find exactly one collection, C2, of simple impressions that bears a striking resemblance to C1, we are satis‑ fied that C1 has not eluded us and that we are still considering the same “thing.” The resemblance need not be perfect. It need only be striking, where “striking” means that there is enough resemblance to trigger an as‑ sociation and that there are no rivals, that is, no other collections of simple impressions that bear a similarly striking resemblance to C1 and that are disposed in approximately equal proximity to P. C2 need not be located at P. It is enough that it has greater proximity to P and more resemblance to C1 than any rival. Hume specified that identity requires invariance (T 1.3.2.2). But this is (i) a stronger condition than a definition of identity over time requires and (ii) a stronger condition than his account of attention requires. i Identity over time is essentially a one–one relation, not an exact resem‑ blance/coincidence relation (Section 4.2). Resemblance and proximity are determinants of the one–one mapping. Resemblance need not be exact, and proximity need not reduce to coincidence provided they still determine a one–one relation. ii Attention is not a part of the definition of the identity relation. But it determines what we want to identify, and accounts for why we make both correct and mistaken identity attributions. Attention is a function of vivacity (Section 6.3). Impressions are vivacious, and collections of impressions that arouse passions are yet more vivacious. Hume main‑ tained that vivacity is communicated from impressions or memories to associated ideas. By parity of example, vivacity should be communi‑ cated from earlier collections of impressions to later, resembling and proximate collections of impressions.35 Greater amounts of vivacity should be communicated, and so more attention should be drawn, to those collections that are most resembling and most proximate. The one–one condition is naturally satisfied because identity attributions are made when vivacity, and so attention, is not dispersed among approxi‑ mately attractive rivals or when there is no contender for our attention. Otherwise, they are made because attention is drawn even when the resemblance is not exact, or the proximity falls short of coincidence.

238  Belief We can very well notice that the resemblance or proximity is not perfect without being inclined to withdraw our identity attributions. As Hume recognized, we consider that things can change or move without detri‑ ment to their identity (T 1.4.5.27, 1.4.6.8). Over longer periods we notice regularities in the way things of a certain sort change shape, size, quality, or location while under constant observation. Eventually, seeing one form that a collection of a certain sort takes, we permeate it with ideas of the other forms we have observed that sort of collection to take over time. We give these embellished collections names: fire, river, stone, and so on. We also discover resembling collections that are differently disposed at the same time. We consider them to be numerically distinct members of the same species. Learning how many members of a species there are dispersed over the sensory field, and how they tend to move, rest, and change over time, we come to be in a position to interrupt our observation and make identity attributions across the gap with a fair degree of confidence. Causal inference does not have much, if any role to play. We do not consider what we observed before the interruption to be a cause of what we observe later. We make the identity attribution on the same grounds we did when there was no interruption, by appealing to relations of resemblance and proxim‑ ity, establishing a one–one identification. These identity attributions are foundational for the recognition of species of object on which causal infer‑ ence depends. 6.12 Objects Beck (1979, 76–8; earlier citing Price 1940, 8) offers a classic Kantian ob‑ jection to Hume’s account of causal inference. Hume disregarded the dis‑ tinction between publicly identifiable objects, like apples and tables; and private sensory experiences, like the coloured shapes seen when viewing tables set at different distances or different angles. When it suited him, as when presenting his account of how belief arises from causal inference, all his talk was about objects and constant conjunctions between species of objects. But when it did not suit him, as when setting himself up to mount sceptical arguments, he suddenly started talking about impressions and images conveyed by the senses. When speaking in the latter vein he went so far as to suggest that our experience is so chaotic, because of eye‑shutting, head‑turning, and other forms of inattention, that no degree of constancy sufficient to establish a causal inference could be discovered (Treatise 1.4.2.21). He began his investigations into the workings of the mind by talking about impressions and ideas, as if these were the elements of all our knowledge and belief. Having started in this way, he provided no account of how we might ascend from the experience of impressions to that of objects. When he considered the matter at all, he tended to reduce

Section 6.12  239 objects to impressions, leaving it unclear how we could come by adequate evidence to draw causal inferences. It took Kant to recognize that sen‑ sory experience must be subsumed under objectivating concepts before a ­regularity‑based account of causal inference can get off the ground. Hume’s views on objects have been the subject of ongoing debate. On sceptical interpretations, Hume demonstrated that we have no experience of ­ 3–5). external or publicly observable objects (Reid Inquiry 1.5, 19–20; 2.6, 3 On naturalist interpretations, he maintained that nature compels us to be‑ lieve in external objects (Popkin 1980, 103–32), but naturalist interpret‑ ers are divided over whether he ultimately meant to endorse the “vulgar” phenomenalist view that these objects just are our perceptions (a view that some [Pears 1990, 153] charge is already too sophisticated to be attributed to the vulgar) or the “philosophical” view that they are external objects rep‑ resented by our perceptions (Wright 1983, 48–59, 74–6; 2009, 153). Grene found that Hume used “object” to refer to intentional objects, perceptions, and external objects but argued that the uses can be reconciled within a con‑ sistent theory that takes the second sense to be fundamental and the third to be a “surround, from which, somehow or other, we take our perceptions to arise” but which defies theorization (1994, 176). Rocknak (2013, xiii) distinguishes between phenomenalist, intentionalist, realist, and “cognitiv‑ ist” (not her term)36 readings and proceeds to argue for the further alterna‑ tive that objects are imagined external causes of our impressions. Others have argued that objects are natural explanatory projections with practical authority (Kail 2007, 72); that Hume anticipated early twentieth‑­century critical realism (Wilson 2008); that we can form relative ideas of objects as existent but otherwise unknown causes of our perceptions (Strawson 2014, 45, 128–33); and that objects are “contents” of mental images, which serve as their “vehicles” (Ainslie 2015, 111, 114, 145–50).37 The project of this section is not to take sides in the debate over what Hume thought about objects. It is to make a proposal concerning what he should have thought about objects, had he recognized that contiguity and resemblance can produce belief (the latter insofar as it figures in invari‑ ance and so in “free” identity attributions [Sections 4.1–3 and 6.11]). This project is motivated by the conviction that Beck raised a serious objection. Anyone who appeals to causal inference to account for Hume’s position on objects ought to confront the fact that Hume at one place recognized that sensory experience is too chaotic to permit it (T 1.4.2.21). Anyone who proposes that he began his investigations by relying on a pre‑theoretical, common belief ought to confront the facts that he (i) began his investi‑ gations by talking about impressions and ideas and (ii) could not justly assume that we are capable of recognizing regularities in the succession of species of object while maintaining that belief only arises from causal inference. He needed a non‑causal account of how we manage to recognize

240  Belief species of object in the chaos of our original experience. He had no such account. As might be expected, the revisionary account that will be of‑ fered on his behalf, drawing as it does on the contiguity and resemblance of impressions, will be phenomenalist or (if impressions are objects) naïve realist. It is referred to below as the CI‑positive (contiguity/free-identity positive) account.38 Level 1: Identification of sensible quality patches. Experience is originally the experience of complex impressions, where a complex impression is an entire visual field, associated with an entire tangible field, further out‑ fitted with smells, tastes, sounds, and passions. The initial challenge for Hume is not to account for how complex impressions are compounded but for how they are partitioned. The answer is provided by resemblance and contiguity considered, not as associative relations, but as imme‑ diately perceived relations. Hume thought that passions communicate vivacity to resembling passions (DP 2.7 and 3.9; T 2.1.4.3). He could easily have recognized that it and contiguity do the same for spatially disposed impressions. Contiguous expanses of uniformly qualified im‑ pressions enhance one another’s vivacity. Enhanced vivacity is attention. Our attention is immediately drawn to areas of uniform or smoothly graded quality. It is scattered when confronted with piles of garbage. Level 2: Identification of landmarks. Identifying a sensible quality patch over time requires relating it to positions that are considered fixed over time (Section 6.11). Positions are originally fixed relative to the extremities of the tactile or visual field. If the field is moving, or too much is moving over it we are bewildered. Making sense of our experi‑ ence requires that we find ourselves in relatively simple environments. Supposing some time passes without noticeable motion, and suppos‑ ing our surroundings are not too complex, we begin to identify colour patches over time.We further distinguish between those that are mov‑ ing relative to the sides and centre of the visual field and those that are stationary.   The identification of stationary colour patches brings about a revolu‑ tion in our thinking. They acquire the status of landmarks designating fixed positions in space, replacing the centre and sides of the visual field in this role (Section 7.3). Level 3: Eye‑head motions. Inference from constant contiguity to an am‑ bient space extending beyond the bounds of the sensory fields. Hav‑ ing identified stationary colour patches as landmarks, when we move our eyes or head to track moving colour patches of interest to us, we discover that the landmark colour patches track off the visual field.

Section 6.12  241 We learn which motions push them off, and which bring them back. When we push them off, we discover new landmarks. We begin to em‑ ploy inference from customary contiguity in space, thinking, when see‑ ing one landmark, that a motion in a particular direction will lead us to see another. In infant experience, most of these motions are rotary, and so are accompanied by a sense of degree as well as direction of rotation (Section 2.3). We come to think of the visual field as a portion of the inner surface of a sphere on which we are centred and dispose vivacious ideas of remote landmarks beyond the bounds of our visual fields at appropriate locations on this sphere relative to the landmark colour patches still on the fields. That is, we believe the remote landmarks to persist at those external locations. Should we have any doubts, a quick turn back brings that portion of the ambient space back into view, pro‑ viding repeated confirmation and further strengthening the belief. This confirmation occurs scores of times a minute as we turn our eyes or head from side to side. We come to think of our visual fields as windows on a larger space that contains them. This ambient space is largely empty, in the sense that we do not conceive anything in particular occupying the distance between remote landmarks. The landmarks themselves are conceived as enduring over time without moving or changing. We have no instinct to consider that empty space cannot exist or that unchanging colour patches do not endure over time. Cognitive development would only be confused and impeded by such suppositions.   Of course, if the landmark colour patches start moving or changing, we will be confused. We are dependent on our environments to present us with some degree of stability. Some environments can frustrate cog‑ nitive development. Level 4: Eye‑hand coordination. Personal body motion. Identification of customary change in species of sensible quality patch under rotation, occlusion, and other changes in perspective. As time passes, we correct our identity attributions and our fixed position attributions against one another. Tracking the motion of colour patches of interest relative to landmark colour patches, we discover how they tend to move, grow, shrink, and change in shading, tint, or brightness over time. Moving as we do so, we discover larger and larger environments containing colour patches that remain invariant and immobile relative to one another for longer periods. We become capable of keeping ourselves oriented in the face of larger and larger scale changes in our environments.   We discover that some colour patches, later to be identified with our hands, move relative to the landmark colour patches, and that they do so synchronously with our volitions. These colour patches are quite com‑ plex, being internally variegated. But they move as a whole, and that

242  Belief enables us to identify their outer bounds. Their synchronous ­motion with our volitions makes them special. Using them to move other colour patches leads to further discoveries.   The lateral motion of a colour patch always involves occlusion. ­Either the colour patch occludes increasing portions of what lies to one of its sides, or increasing portions of its area are occluded by what lies to one of its sides. Having identified colour patches that can be moved at will, we can perform experiments with repeated motions, either with the hands or with colour patches we learn to grasp with our hands. Mov‑ ing the colour patch constitutive of a pen back and forth over a piece of paper, or between the pages of a book, produces expectations concern‑ ing what will be revealed or occluded next. We imagine the occluded parts to still be contiguous to the visible parts, in effect considering the occluded colour patches to “permeate” (or, as we later come to think of it, lie “behind”) the occluding colour patches.   Experience of rotations has different effects. Turning over the colour patch partially constitutive of a book, we see first the colour patches constitutive of its front cover, then those constitutive of its spine, and then those constitutive of its back. We uncover increasingly large por‑ tions of previously occluded colour patches constitutive of the book and occlude increasingly large portions of previously apparent ones. Mov‑ ing the colour patches constitutive of the book face laterally and then performing the rotation in the new location, and doing this repeatedly for various locations reveals a special kind of contiguity relation, having to do with what serially appears with rotation. With time and experi‑ ence, any given impression of the colour patches constitutive of any face of a familiar solid comes to be permeated with vivacious ideas of col‑ our patches constitutive of the faces that would appear under rotation around any given axis. Having an impression permeated by these viva‑ cious ideas is what we call perceiving a “solid” or “body.” It necessar‑ ily involves both a current impression and a collection of contiguously associated vivacious ideas.   A related development occurs if we walk. Approaching or receding from a colour patch, or walking around it, produces a new set of im‑ pressions. With repetition, we come to anticipate what impressions will be encountered next upon approaching, retreating, or walking around a colour patch. These anticipations are vivacious ideas. The collection of vivacious ideas arising from considering all the different motions towards, away from, or around a familiar colour patch permeates its location, investing the colour patch with yet more of the content charac‑ teristic of what might be called a “body” such as a pen, book, or table.   These inferences from contiguity are further enhanced by infer‑ ences from identity. Since identity only requires as much resemblance,

Section 6.12  243 immobility, or uninterruptedness as needed to support the one–one con‑ dition, our ideas of colour patches come to be permeated with ideas of characteristic patterns of motion and change of those colour patches. The changes characteristic of a fire slowly burning down in a fireplace are an example. Level 5: Preferred views. Publicly observable objects. By means of these developments, the experience of a body comes to involve a collection of simple impressions and a collection of associated ideas. When a body is currently experienced, its forward face in its present state is a col‑ lection of simple impressions that we have learned to distinguish from currently surrounding collections of simple impressions. It might be a colour patch or the pressure of a pen in hand. That complex impression is permeated with vivacious ideas of what would appear under mo‑ tions revealing or causing occlusion, under rotation, under changes of viewing distances and angles, under different lighting conditions, under ageing, under a magnifying glass or microscope, and so on. The current impression may be largely disregarded in favour of the vivacious idea of a preferred facet, viewed from a preferred distance and angle.39 (At the preferred distance and angle the features of most importance to us are best displayed.) Think of going online to purchase a kettle and in‑ specting various photographs of kettles. As these photographs are taken from a preferred angle and distance, so the idea of a body picks on the preferred view when the actual view falls short. In some cases, we may mistake what we see for something else because what we see arouses a vivacious idea of a preferred view that belongs to a different body.   Bodies are publicly observable objects in the sense that the view that is an impression for one viewer is an associated idea for another viewer. Publicly observable objects are not the intentional objects of thoughts or acts of mind. They are not abstract ideas. They are not projections of perceptions into an imagined external world or imagined causes of per‑ ceptions. They are not objects represented by perceptions or “contents” of perceptions considered as “vehicles.”40 They are particular collec‑ tions of impressions and ideas that are private to different viewers but that have an affinity of content. When two or more viewers are viewing “the same object” items present as a collection of simple impressions in the complex impression of one viewer are present as associated ideas in the collection of the other. Both viewers may share the same preferred view of what they are seeing, but it is possible that each viewer may have developed their own preferred view, highlighting aspects that have been most relevant in their experience. If the viewers are language users, they may nonetheless use the same description or name for what they are seeing, but that is an addendum. The CI‑positive account applies to

244  Belief non‑verbal animals. Like Hume and Reid, who never used the term,41 the CI‑positive account has no use for the notion of a “concept” which is ambiguous between a preferred view, which is private, and a name, which is public. Hume was wrong to declare that when experience shows that certain “ob‑ jects” are either always remote or always contiguous, we always conclude that there is some secret cause that separates or unites them. He was wrong to declare that we can only conclude that what we see after an interruption is the same as what we saw before if we can think that nothing would have caused it to be replaced with a replica (T 1.3.2.2). It is only because we dis‑ cover that colour patches are customarily disposed in the same ways rela‑ tive to one another and move and change in regular ways over time that we can recognize the species of object that figure in causal inferences. The “categories” are not substance, cause, and reciprocity, but resemblance, difference, proximity, customary disposition in space over time, and one– one maximal resemblance and proximity over time. Granting that we are innately disposed to recognize resemblance and proximity, there is nothing “transcendental” about the cognitive operations that lead us to recognize objects. They are empirically guided every step of the way. It is a question where this leaves Hume’s arguments for scepticism with regard to the senses. Notes 1 On this topic, the difference between the Enquiry and the Treatise is one ­between psychophysics and phenomenological psychology, not one between epistemology and psychology. Compare Qu (2020, Chapter 1). 2 See Allison (2008, 15–17) for a good statement of commonly raised concerns. For a reply see, among others Garrett (2015, 38–40) and note 4 below. 3 Rickless (2018) does not mention any prior occurrence of the “faculties first” approach in his summary of the recent literature on Hume’s distinction be‑ tween impressions and ideas. 4 It would be wrong to bring up passages such as T 1.3.5.6 and 1.3.9.19 (on the effects of repetition) or 1.3.10.9–10 (on madness) in this connection. These passages are concerned with the distinction between belief and fiction, and so with a distinction between kinds of ideas, not between impressions and ideas. 5 Enquiry 5.20 also contains a reference to “belief, where it reaches beyond the memory or senses,” implying that belief should not simply be identified with what reaches beyond memory or the senses. 6 It is a question why the later, “most accurate” definition only applies to ideas. Hume did not go into any more detail. 7 Treatise 1.3.7.7, which brings up the reductio, only appeared in 1740 in the Ap‑ pendix. 1739 2.3.1 takes actions of the mind to be necessitated, but refrains from pointing out that this applies to belief. Perhaps the closest 1739 comes to invol‑ untarism about belief is 1.4.1.7: “Nature, by an absolute and uncontroulable

Notes  245 necessity has determin’d us to judge as well as to breathe and feel; nor can we any more forbear viewing certain objects in a stronger and fuller light, upon account of their customary connexion with a present impression, than we can hinder ourselves from thinking as long as we are awake[.]” But has nature only determined us to judge or has it also determined us what to judge? Does the stronger and fuller light command assent? The Enquiry is more explicit. 8 Baier (2008, 164–5), drawing on Ainslie (2005, 151, 162), objects that vivacity could not be attention because in that case our maximally vivacious perception of ourselves would block, rather than enable our sympathy with others. But the claim here is about the nature of attention rather than of vivacity. Attention is making more prolonged and extensive associations with something that is viva‑ cious. On this account, sympathy with others should be an effect the vivacity of the perception of self, and that seems to have been Hume’s point. 9 See Marušić (2010) and the earlier work referenced there. Marušić’s conclu‑ sions are generally endorsed here, though not always for the same reasons. 10 Hume attributed this effect just to custom understood as repeated encounters, but, as he himself noted, impressions of the places where the encounters oc‑ curred are also essential. The ideas of the deceased are not enlivened by custom alone or contiguity alone, but by customary contiguity. 11 This is the earliest of five passages that Loeb (2011, 316–17) cites as evidence for a “third level” of epistemic commitment, one at which Hume endorses an account of the grounds of normative distinctions (313). However, the princi‑ pal conclusion is not normative. It identifies circumstances under which “’tis impossible [we] can ever operate with any considerable force or constancy.” Belief is not up to us. 12 This is controversial. Marušić (2011) plausibly defends the argument of 1739 1.3.7.2. 13 Appendix 2 and Enquiry 5.10–11 are not included in Beauchamp’s register of remains (EHU, lxiv–lxv), but there is a considerable affinity. 14 It also suggests that Hume continued to be preoccupied with his account of the psychology of belief at the time of writing the Enquiry, to the extent of intro‑ ducing further refinements and clarifications. 15 Beauchamp (EHU, lxv) identifies EHU 5.15–17 as a remain of T 1.3.8.3–6, but proceeds to include EHU 5.18 as well (EHU, lxvii). 16 Marušić (2011, n.2) has noticed this difference, but overgeneralizes it, claiming that Hume does not hold that a belief is an idea with a high degree of force and vivacity in the Enquiry. See, to the contrary, the references to “enlivening,” “vivacity,” and “force” over Enquiry 5.12, 5.15–18, and 5.20. The former passages recopy 1740 1.3.7.7 (added in Ax) and 1739 1.3.8.3–6, but the last makes references to vivacity and “force of conception” that are original to the Enquiry. Qu (2020, 23–7) argues that the mature Hume retreated from the psychological tenets of the Treatise. But Hume not only recopied but also reaf‑ firmed the transfer thesis in the Enquiry. 17 Bayes’s work on how new evidence affects our inferences from prior prob‑ abilities was only made public by Price in 1761, too late to have motivated Hume to have had reservations about the Treatise’s theory of mathematical probability. On 18 March 1767, Hume wrote to Price that “the Light, in which you have put this Controversy, is new and plausible and ingenious, and per‑ haps solid.” But he asked for more time to weigh it (New Letters #126, 233). There is no indication of any changes to later editions of ETSS consequent on Hume’s promised review. The topic has bearing on Hume’s position in the

246  Belief e­ ssay on miracles, on which he will not have been prepared to make any appar‑ ent concessions. For further discussion, see Falkenstein and McArthur (2013, Appendix A). 18 In 1748 recopying meant rewriting a passage word for word in long hand. It is not so easy to inadvertently omit changes in that operation. 19 Hume’s work on this subject has been largely neglected. Among the few who have considered it are MacNabb (1951, 94–100), Falkenstein (1997), and Loeb (2002, 2013, and other work referenced there). For MacNabb and Loeb, Hume’s recognition of unphilosophical probability poses an epistemological problem, which Hume resolved by appealing to the influence of general rules. For Falkenstein, philosophers are determined to believe as they do by one set of general rules and others by a different set and there is nothing philosophers can say to others to justify their set. MacNabb suggests that the justice of philo‑ sophical belief is established by the “method of challenge” (1951, 95), as if all the philosophers needed to do to silence the furious bigots was demonstrate that their beliefs are statistically well founded. Not only is this impractical, it begs the question of why we should prefer reliance on statistics as of yesterday to other foundations for belief. 20 It would be unphilosophical to reject this explanation because it does not tell a story about how vivacity is transferred from one perception to another. There is nothing to say that an idea can only acquire vivacity from a vivacious percep‑ tion, any more than that a body can only be set into motion by collision with a moving body. 21 EHU 5.8. See also EHU 5.21 which suggests that there is “a kind of pre‑­ established harmony between the course of nature and the succession of our ideas” ensuring that “our thoughts and conceptions have … gone on in the same train with the other works of nature.” Like the author of the Treatise, the mature Hume was not strictly of this opinion. EHU 16–19, the whole of the Natural History, and numerous anecdotes from the History make it clear that it is not always true. EHU 10.4 declares that “A wise man, therefore, pro‑ portions his belief to the evidence.” This would not be an achievement worthy of an honorific appellation if we were all determined to do so. 22 This is the second of five passages that Loeb (2011, 316–17) cites as evidence that Hume endorsed an account of the grounds of normative distinctions. The first was discussed at note 11. 23 “Undogmatic” acceptance of whatever opinions are most widely accepted in one’s society is a feature of the sceptical disposition, according to Sextus, ­Adversus Mathematicos 9.49 and Outlines of Pyrrhonism 1.23–4 (M 364 and PH 306–7). 24 All citations in this paragraph are from T 1.4.4.1. This is the fifth of Loeb’s five passages. 25 This is the third of Loeb’s five passages. 26 This is the fourth of Loeb’s five passages. 27 The Natural History may not have been written to advance an anti‑religious project. It is as concerned with the causes of belief in general as with spe‑ cifically religious belief, which serves merely as an example. It illustrates the unfortunate effects of “unphilosophical” causes of belief as a way of enabling us to resist those causes. In large part, it does so by employing an “unphilo‑ sophical” method: ridicule. Such episodes as those of the little suckling gods, the convert who ate God, and the encounter of the Turkish ambassador with the ­Capuchins, make this one of the most hilarious works in philosophy.

Notes  247 28 The speaker is Philo. Over Dialogues 1.12–17, Cleanthes takes issue with what Philo has said, but Cleanthes’ objection is only that Philo has no justification for excluding the design argument from legitimate speculation. 29 Education is discussed only briefly over Treatise 1.3.9.16–19, where it is dis‑ missed as “unphilosophical.” 30 Following Hume’s own usage at Treatise 1.1.4.2, this section speaks of the things that are discovered to be customarily disposed in space as “objects,” exemplified by publicly identifiable occupants of an external world like apples and tables. Hume supposed that we perceive such objects throughout his dis‑ cussion of custom‑based causal inference in both the Enquiry and the Treatise, so nothing is being presupposed to make a case about contiguity that he does not himself presuppose to make a case about cause and effect. 31 This was the wording over the five editions from 1748 to 1760. In 1764, “can we” was changed to “we can.” In the two subsequent editions of 1767 and 1768, it was changed back to “can we.” In 1770 and all later editions, “we can” replaced “can we” a second time. Beauchamp (EHU: xl–xlii) has argued that the 1770 edition was based on markups made to that of 1764, resulting in the inad‑ vertent loss of over 150 changes made to EHU in 1767 and 1768. (The evidence is further reviewed by Beauchamp in EPM: xxxi–xxxii, DP/NHR: xxix–xxxi, and E: 433–4, with reference to other works in ETSS.) Changing “can we” to “we can” removes the implication that only causal inference can lead us be‑ yond the evidence of experience or memory. Hume appears to have had second thoughts about making this change in 1767 and then inadvertently retained the 1764 change in 1770 and thereafter. Plausibly, the 1764 change, like most of Hume’s later changes, was made for stylistic rather than substantive reasons. 32 It helps if our work is not complicated by the doctrine that unchanging “ob‑ jects” do not endure over time. See Chapter 3, addressing the claims of Treatise 1.2.3.7–11. 33 An account of what might lead us to partition a momentary complex impres‑ sion, understood as an entire sensory field, into collections of simple impres‑ sions is given in Section 6.12. 34 This presupposes that we can identify P from one moment to the next. To start, the observer is stationary in a largely motionless environment and P is taken to be a location relative to the centre point and the periphery of the visual field or a location relative to the extremities of the tangible field. The less the starting position is like this, the more difficult identity attributions become. We find it difficult to orient ourselves and feel lost and confused. Experienced perceivers have discovered how to manage more complexity. But finding ourselves sta‑ tionary in largely motionless environments is a condition on beginning to make sense of our experience. 35 While he tended to write of vivacity being transmitted from impressions and memories to ideas, and from those ideas to other ideas, Hume recognized that resembling passions and passions that have a similar “tendency” also commu‑ nicate vivacity (DP 2.7 and 3.9; T 2.1.4.3). The extension to earlier and later sense impressions is natural. 36 She describes the alternative as one on which objects are “to varying degrees … imagined” and finds it propounded by Price and Kemp Smith, among others. 37 Strawson (2014, 19–21 and 32–9) offers a good account of the content/vehicle distinction and a defence of its applicability to the early modern context. The outstanding reason for thinking Hume would not have accepted the distinction is Treatise 1.4.5.19.

248  Belief 38 The account that follows is modelled on that found in Condillac’s Traité of 1754, and so on what would have been thinkable at the time. It considers just the recognition of visual objects. An account of the recognition of tangible objects ought to be provided as well but is not attempted here. 39 Price writes in this context of a “standard shape” (1940, 206). 40 That is, they are not any of these things on the revisionary, CI‑positive Humean account presented here. No position is being taken on what they are for Hume, supposing he had one to offer. 41 Based on a search for “concept” in ECCO editions of Reid’s works and the InteLex electronic edition of Hume’s complete works and correspondence.

Bibliography Ainslie, Donald C.  2005. “Sympathy and the Unity of Hume’s Idea of Self.” In Persons and Passions: Essays in Honour of Annette Baier, edited by Annette Baier, Joyce Jenkins, Jennifer Whiting, and Christopher Williams, 143–73. Notre Dame: University of Notre Dame Press. Ainslie, Donald C.  2015. Hume’s True Scepticism. Oxford: Oxford University Press. Allison, Henry. 2008. Custom and Reason in Hume: A Kantian Reading of the First Book of the Treatise. Oxford: Oxford University Press. Baier, Annette. 2008. Death and Character: Further Reflections on Hume. Cam‑ bridge: Harvard University Press. Beck, Lewis White. 1979. “A Prussian Hume and a Scottish Kant?” In McGill Hume Studies, edited by David Fate Norton, Nicholas Capaldi, and Wade L. Robison, 63–78. San Diego: Austin Hill Press. Dauer, Francis W. 1999. “Force and Vivacity in the Treatise and the Enquiry.” Hume Studies 25: 83–99. Everson, Stephen. 1988. “The Difference between Feeling and Thinking.” Mind 97: 401–13. Falkenstein, Lorne. 1997. “Naturalism, Normativity, and Scepticism in Hume’s Account of Belief.” Hume Studies 23: 29–72. Falkenstein, Lorne and David Welton. 2001. “Humean Contiguity.” History of Philosophy Quarterly 18: 279–96. Falkenstein, Lorne and Neil McArthur. 2013. “Appendix A: From Richard Price, Four Dissertations.” In David Hume, Essays and Treatises on Philosophical Sub‑ jects, edited by Lorne Falkenstein and Neil McArthur, 531–43. Peterborough: Broadview. Garrett, Don. 2015. Hume. London: Routledge. Grene, Marjorie. 1994. “The Objects of Hume’s Treatise.” Hume Studies 20: 163–77. Kail, P. J. E. 2007. Projection and Realism in Hume’s Philosophy. Oxford: Oxford University Press. Loeb, Louis E. 2002. Stability and Justification in Hume’s Treatise. Oxford: Ox‑ ford University Press.

Bibliography  249 Loeb, Louis E. 2013. “Epistemological Commitment in Hume’s Treatise.” In ­Oxford Studies in Early Modern Philosophy VI, edited by Daniel Garber and Donald Rutherford, 309–48. Oxford: Oxford University Press. MacNabb, D. G. C. 1951. David Hume: His Theory of Knowledge and Morality. Reprinted in facsimile Oxon: Routledge, 2019. Marušić, Jennifer Smalligan. 2010. “Does Hume hold a Dispositional Account of Belief?” Canadian Journal of Philosophy 40: 155–84. Marušić, Jennifer Smalligan. 2011. “Belief and Introspective Knowledge in Treatise 1.3.7.” Hume Studies 37: 99–122. Pears, David. 1990. Hume’s System. Oxford: Oxford University Press. Popkin, Richard H. 1980. “David Hume: His Pyrrhonism and His Critique of Pyrrhonism.” In The High Road to Pyrrhonism, edited by Richard A. Watson and James E. Force, 103–32. San Diego: Austin Hill Press. Price, H. H. 1940. Hume’s Theory of the External World. Oxford: Clarendon Press. Qu, Hsueh M. 2020. Hume’s Epistemological Evolution. New York: Oxford University Press. Rickless, Samuel. 2018. “Hume’s Distinction between Impressions and Ideas.” ­European Journal of Philosophy 26: 1222–37. Rocknak, Stefanie. 2013. Imagined Causes: Hume’s Conception of Objects. Dordrecht: Springer. Strawson, Galen. 2014. The Secret Connexion: Causation, Realism, and David Hume. Revised edition. Oxford: Oxford University Press. Sugden, Robert. 2021. “Hume’s Experimental Psychology and the Idea of Erroneous Preferences.” Journal of Economic Behaviour and Organization 183: 836–48. Wilson, Fred. 2008. The External World and Our Knowledge of It: Hume’s Critical Realism, an Exposition and a Defence. Toronto: University of Toronto Press. Wright, John P. 1983. The Sceptical Realism of David Hume. Minneapolis: University of Minnesota Press. Wright, John P. 2009. Hume’s ‘A Treatise of Human Nature’ An Introduction. Cambridge: Cambridge University Press.

7 Causes of the belief in bodies

This chapter shows that, on his own principles, Hume ought to have recognized that the belief in externally, independently, and continuously existing bodies is caused by sensation and by inference from customary contiguity, customary identity over time, and cause and effect. The belief takes the form of a belief in the external existence of sensible qualities. For sensible qualities to be external is for them to be disposed in space beyond the bounds of the visual or tactile sensory fields. For them to be bodies is (i) for them to be disposed to form shapes that are sharply edged off from one another by qualitative contrast and (ii) for these shapes to be visually or tangibly resistant to compression or deformation, and so solid. This positive, phenomenalist account of the causes of the belief in body is im‑ plicit in Hume’s critique of the modern philosophy. The dismissive account of Treatise 1.4.2.2–39 is objectionable at every step of the way. 7.1

Belief in “body” in the Treatise and the Enquiry

Treatise 1.4.2.2–39 seeks to determine what causes lead us to believe in externally, independently, and continuously existing objects or “body,” as Hume put it. It rejects some obvious and plausible answers in favour of a complex alternative that makes the belief seem unjustified. Enquiry 12.7 replaces the Treatise’s account with a single sentence asserting that the belief arises from “a natural instinct or prepossession.” The following paragraph describes this instinct as “blind and powerful.” There are posi‑ tive normative overtones to these pronouncements implied by the “provi‑ dential naturalism” of the earlier Enquiry 5.22. Though he abandoned it, what the author of the Treatise had to say about the causes of the belief in body has captured the imagination of generations of scholars, who have read it as an anticipation of phenom‑ enalism, neutral monism, sense datum theory, critical realism, sceptical realism, projectivism, pre‑theoretical vulgar immersion, or other theories.1 It merits consideration. DOI: 10.4324/9781032677880-8

Section 7.2  251 7.2

Sensation of bodies

The Treatise’s investigation of the causes that lead us to believe in body falls into four parts. In the first (T 1.4.2.3–13), Hume argued that bodies are not discovered by the senses. In the second (T 1.4.2.14), he argued that the belief in body is not based on reasoning. In the third (T 1.4.2.15–23), he argued that the belief can be produced by the influence of the “coher‑ ence” of our impressions on the imagination. In the fourth (T 1.4.2.24–43), he argued that it often depends upon the influence of the “constancy” of our impressions on the imagination. Hume began his discussion of why the senses cannot lead us to believe in body by giving multiple reasons for why they only lead us to experience our own internal states. Most of these reasons are not based on anything learned from the senses alone, and so are irrelevant to the question of what the senses cause us to believe.2 It is only midway through his treat‑ ment that Hume turned to consider what the senses alone teach us. (AE): Apparent exteriority [O]ur own body evidently belongs to us; and as several impressions ap‑ pear exterior to the body, we suppose them also exterior to ourselves. The paper, on which I write at present, is beyond my hand. The table is beyond the paper. The walls of the chamber beyond the table. And in casting my eye towards the window, I perceive a great extent of fields and buildings beyond my chamber. From all this it may be inferr’d that no other faculty is requir’d, beside the senses, to convince us of the ex‑ ternal existence of body. (T 1.4.2.9) Hume offered three reasons for rejecting (AE). All are found at Treatise 1.4.2.9. Three embellishments on the first reason might be drawn from things said elsewhere. The third reason is that, according to “the most rational philosophers,” we do not immediately perceive “distance or outness (so to speak).” This reason only applies to one of three dimensions. Even if the impressions constitutive of the paper, the table, and the walls are not disposed “be‑ yond” those constitutive of our hands in a depth dimension, they are dis‑ posed around our hands in the remaining two dimensions, and so are still outside them.3 Hume’s second reason for rejecting (AE) is that even though sounds, tastes and smells are thought to be “continu’d independent qualities,” they do not appear to be anywhere in “extension,” and so cannot appear to be outside our bodies. But Hume elsewhere maintained that visual and tactile impressions are exceptions to this questionable (Section 1.7) rule.

252  Causes of the belief in bodies The supposition that visual impressions are disposed in space features in the discussion of the missing shade of blue (T 1.1.1.10 reiterated at EHU 2.8). According to Treatise 1.2.3.4, the senses convey an impression of coloured points disposed in a manner that is alone sufficient to give us the idea of space. Elsewhere, Hume declared that “The first notion of space and extension is deriv’d solely from the senses of sight and feeling” (T 1.4.5.9). These tenets are crucial for Treatise 1.4.4, which charges that we only come to have the ideas of primary qualities because coloured and tactile impressions are disposed in space (Sections 1.2–5, 2.2, 2.5 and 2.6; Chapter 4 note 16). The first reason for rejecting (AE) is that: (I): Idealism properly speaking, ’tis not our body we perceive, when we regard our limbs and members, but certain impressions, which enter by the senses; so that the ascribing a real and corporeal existence to these impressions, or to their objects, is an act of the mind as difficult to explain, as that which we examine at present. (T 1.4.2.9) As flagged by “properly speaking,” Hume here changed topic to consider what something other than the senses tells us. The senses do not sense themselves and so cannot cause us to believe that impressions either enter them or are created by them. Nor is there any difficulty in explaining why they would cause us to ascribe a real and corporeal existence to some impressions. Vision and touch present impressions as disposed outside one another over two or three spatial dimensions. They further present their impressions as forming uniformly qualified shapes, some of which move through space and are observed to resist compression upon collision with one another (T 1.4.4.9 and 13). Taking “corporeal” to mean “extended and solid,” these sense impressions are corporeal. They are also real according to Treatise 1.3.9.3, which describes them as forming a “first system” of reality. Any impropriety in speaking of visual or tactile impressions as real or corporeal would have to be drawn from something other than what the senses tell us. The vivacity conveyed by that other source of information would have to be weighed against that of the senses before declaring that the senses could not cause a belief in body. The prospects for that are not good. Hume considered belief and assent to be originally ascribed to sense experience and memory (Section 6.2). As the original sources of vivacity, they are more vivacious than any idea, and what they teach us is constantly renewed. Any belief based on reasoning that is abstract, intricate,

Section 7.2  253 temporally remote, or conflicted will not survive the contest. The best it could manage would be “proof against proof” (EHU 10.11), and it is un‑ likely it could manage that. Hume offered one embellishment on (I) in an earlier passage, writing that “all sensations are felt by the mind, such as they really are” so that if they “present themselves as distinct objects” it could not be because of their “nature,” but would have to be because of their “relations and situa‑ tion” (T 1.4.2.5). He continued, Now if the senses presented our impressions as external to, and inde‑ pendent of ourselves, both the objects and ourselves must be obvious to our senses, otherwise they cou’d not be compar’d by these faculties. The difficulty, then, is how far we are ourselves the objects of our senses. ’Tis certain there is no question in philosophy more abstruse than that concerning identity, and the nature of the uniting principle, which constitutes a person. So far from being able by our senses merely to determine this question, we must have recourse to the most profound metaphysics to give a satisfactory answer to it; and in common life ’tis evident these ideas of self and person are never very fix’d nor determi‑ nate. ’Tis absurd, therefore, to imagine the senses can ever distinguish betwixt ourselves and external objects. (T 1.4.2.5–6) While our ideas of “self and person” may never be “very fix’d nor deter‑ minate,” the idea at issue is that of body. Sensory experience does draw a fixed and determinate distinction between our own body and what im‑ pinges on or surrounds it. We find that we can move some colour patches merely by volition, whereas we cannot change the state of motion of other colour patches unless we move those in our voluntary control into con‑ tact with them.4 We call those within our voluntary control “my hands,” and “my” body parts. The distinction between what we can and cannot move by an act of will draws a preliminary boundary between self and externality.5 This boundary is refined by experience of whether impres‑ sions move with the sensory fields or independently of them. When we touch something, and feel it warm, and move away, the feeling of heat disappears. But when we touch something, and burn ourselves, the pain moves away with us, and remains in a fixed spot on the tactile field, and we find we cannot shake it off. Visual after‑images and floaters are also affixed to the field. We think of what is attached to the field as personal and dependent because it moves along with our volitions to move our eyes, head, or other body parts. What moves independently is external. We think of the sensory fields as dirty windshields on an external world,

254  Causes of the belief in bodies some of what we sense being stuck to the windshield and the rest sensed through and beyond it. A second embellishment on (I) is elliptically cited by John Wright (2016, 66). [E]very impression, external and internal, passions, affections, sensa‑ tions, pains and pleasures, are originally on the same footing;… they ap‑ pear, all of them, in their true colours, as impressions or perceptions.… Every thing that enters the mind, being in reality a perception, ’tis im‑ possible any thing shou’d to feeling appear different. This were to sup‑ pose, that even where we are most intimately conscious, we might be mistaken. (T 1.4.2.7; Wright’s omissions; my brackets) According to Wright, (WC): Wright’s conditional [Hume] argues that since our sense impressions really are subjective and mind‑dependent, they cannot appear otherwise. (2016, 66; see also 2009, 238; 1983, 41–2) But Treatise 1.4.2.7 does not say that our sense impressions are subjective and mind dependent. It only says that they are all on the same footing, that they “enter the mind,” and that they are known by consciousness. The claim that everything that enters the mind is “in reality a perception” adds nothing given that “perception” is just the generic term for impressions and ideas. In the following (AE), Hume acknowledged that several impres‑ sions do appear exterior to our bodies. Accepting that the senses have no other conception of what is subjective or “mind”‑dependent than what is interior to our bodies, it follows from (WC) by modus tollens that since some impressions do not appear as internal to our bodies, they really are not subjective or mind dependent. It gets worse. At Treatise 1.4.2.6, Hume declared that the senses are in no position to draw a distinction between “self and person” and external objects. So the information supplied by the senses also puts them in no position to assert the antecedent of (WC), making a modus ponens appeal to (WC) is impossible.6 A portion of Treatise 1.4.2.7 elided from Wright’s citation makes it clear that Hume only meant to assert a biconditional, not the antecedent of (WC): [N]or is it conceivable that our senses shou’d be more capable of deceiv‑ ing us in the situation and relations, than in the nature of our impres‑ sions. For since all actions and sensations of the mind are known to us

Section 7.2  255 by consciousness, they must necessarily appear in every particular what they are, and be what they appear. (T 1.4.2.7, my stress) Contrary to Wright, Hume did not argue “from the real nature of sense im‑ pressions to a conclusion about the way they appear” (2016, 67). He got no further than arguing that their real nature and their appearance cannot be different. It is not said that sensations either are or appear as subjective or mental. That would be to put weight on the phrase “of the mind” that it is not suited to bear without more argument. Hume, like Wright, may have intended to draw on such an argument. But it would have to rest on more than appeal to what is made evident by the senses. It therefore has no place in a discussion of what the senses determine us to believe.7 Hume thirdly embellished (I) with parity considerations. Impressions fall into three classes: those of the figure, bulk, motion, and solidity of bodies; those of colours, tastes, smells, sounds, heat and cold; and “the pains and pleasures, that arise from the application of objects to our bodies, as by the cutting of our flesh with steel, and such like” (T 1.4.2.12). Hume noted that “the vulgar” consider Class 2 impressions to be “on the same footing” with Class 1. Both are supposed to have a “distinct continu’d existence.” Some philosophers consider only the Class 1 impressions to have this status, but “the vulgar” disagree, basing their opinion on what appears to the senses. (A1–2): Argument for class 1–2 identity [W]hatever may be our philosophical opinion, colours, sounds, heat and cold, as far as appears to the senses, exist after the same manner with motion and solidity, and … the difference we make betwixt them in this respect, arises not from the mere perception.… [W]hen the con‑ trary opinion is advanc’d by modern philosophers, people imagine they can almost refute it from their feeling and experience, and that their very senses contradict this philosophy. (T 1.4.2.13) Hume observed that everyone agrees that Class 3 impressions are “merely perceptions; and consequently interrupted and dependent beings” (T 1.4.2.12). But Class 3 are on the same footing with Class 2, he maintained, and since it has been established that Class 1 are on the same footing with Class 2, we ought to conclude that they are all on the same footing in being interrupted and dependent beings. But something has gone awry here. To make his case, Hume needed to appeal to the same authority at each step of the way. But (A1–2) appeals to the authority of the senses whereas his argument for Class 2–3 identity appeals to a rival authority.

256  Causes of the belief in bodies (A2–3): Argument for class 2–3 identity [T]hey [Class 2 and 3] are confest to be, both of them, nothing but perceptions arising from the particular configurations and motions of the parts of the body. (T 1.4.2.13) (A2–3) is a philosophical opinion, not something that is directly perceived by the senses. If Hume was going to make a case for the identity of perceptions in all three classes, he could not appeal to the senses to establish (A1–2) and philosophy to establish (A2–3), particularly given that the senses reject (A2–3). They present some impressions as attached to locations on sensory fields and moving with them, and others as independently mobile. Rather than identify the two classes, they distinguish them. Nor could he appeal to philosophy to establish both identifications, both because Cartesians and Lockeians reject (A1–2) and because the question concerns what we learn from the senses. Hume had no good reason for denying that the senses cause us to believe that bodies exist outside of our own bodies. He might have had some reason to charge that the belief is false, but his project was not to determine whether the belief is true, but to identify its causes (T 1.4.2.1). 7.3

Reasoning to the existence of bodies

The senses cannot cause the belief in body in its full form. They present some impressions as disposed externally to those constitutive of personal body parts, and they present some impressions as moving or at rest independently of the motion or rest of the sensory fields, but they cannot show us that impressions continue to exist beyond the bounds of the sensory fields or behind occlusions. We do believe this. The senses can only provide evidence for reasoning to that conclusion. Hume considered empirically guided (“probable”) reasoning to what exists beyond the bounds of the senses or the reach of memory to be determined by “custom,”8 that is, experience of what has been frequently found to be the case supplemented by the proclivity, when given a lively perception, to (i) form lively ideas (beliefs) that copy the complex impressions (objects) that have been customarily experienced along with that perception, and (ii) dispose those ideas at the locations in space and time where those impressions have been customarily encountered relative to their triggers. These locations need not be forward on the sensory fields. They might be locations behind an occlusion, locations in an ambient space, or locations in a past or future time (for the latter, see Treatise 2.3.7.7–9, 1.3.13.6, and 1.1.7.12).9

Section 7.3  257 I hear at present a person’s voice, whom I am acquainted with; and this sound comes from the next room. This impression of my senses immedi‑ ately conveys my thoughts to the person, along with all the surrounding objects. I paint them out to myself as existent at present, with the same qualities and relations, that I formerly knew them possess’d of. These ideas take faster hold of my mind… ’Tis the same case when I recollect … the events of any history. Every particular fact is there the object of belief. (Ax 4) Suppose I see the legs and thighs of a person in motion, while some interpos’d object conceals the rest of his body. Here ’tis certain, the im‑ agination spreads out the whole figure. I give him a head and shoulders, and breast and neck. These members I conceive and believe him to be possess’d of[.] (Ax 4) One wou’d appear ridiculous, who wou’d say, that ’tis only probable the sun will rise to‑morrow, or that all men must dye; tho’ ’tis plain we have no farther assurance of these facts, than what experience affords us. (T 1.3.11.2 alluded to at EHU 6n) I form an idea of Rome, which I neither see nor remember; but which is connected with such impressions as I remember to have receiv’d from the conversation and books of travellers and historians. This idea of Rome I place in a certain situation on the idea of an object, which I call the globe.… I look backward and consider its first foundation; its several revolutions, successes, and misfortunes. All this, and every thing else, which I believe, are nothing but ideas; tho’ by their force and set‑ tled order, arising from custom and the relation of cause and effect, they distinguish themselves from the other ideas, which are merely the offspring of the imagination. (T 1.3.9.4) In all these cases, we are caused to believe in the unperceived external exist‑ ence of things (that is, of their existence beyond the bounds of the sensory fields or behind occlusions). We believe these things to be extended and resistant to compression, that is, to be bodies. And while we do not always believe them to continue to exist in the sense of persisting over time, we do conceive them to “continue” to exist in the sense of now being where we take them to be even though where we take them to be is not in view. In the last of these cases, our belief is attributed to “custom and the rela‑ tion of cause and effect,” but not in all.10 The contents believed to fill the

258  Causes of the belief in bodies adjoining apartment are neither causes nor effects of the sound and modu‑ lation of the voice; the rest of the pedestrian’s body is neither a cause nor an effect of the legs and thighs; the events immediately preceding the rising of the sun vary from one day to the next; the events leading up to the death of different people are various. The “custom” in these cases is either cus‑ tomary disposition in space or customary behaviour of species over time. But though Hume elsewhere recognized all these forms of custom‑guided reasoning to the “continued” existence of bodies (to their still existing de‑ spite not being on the sensory fields), at Treatise 1.4.2.14 he anomalously denied that the belief in body could be based on reasoning of any sort. For philosophy informs us, that every thing, which appears to the mind, is nothing but a perception, and is interrupted, and dependent on the mind; whereas the vulgar confound perceptions and objects, and attrib‑ ute a distinct continu’d existence to the very things they feel or see. This sentiment, then, as it is entirely unreasonable, must proceed from some other faculty than the understanding. This argument confuses reasoning considered as an inferential operation determined by custom, habituation, and vivacity transfer, and so consid‑ ered as a cause of belief, with reasoning considered as giving reasons or a justification for belief. The question at hand is not whether there are justifi‑ cations for the belief in body, but whether the operation of custom‑guided inference can cause the belief in body.11 Philosophers may consider that they have better justification for their belief that perceptions are inter‑ rupted and dependent on the mind than the vulgar do for their belief that perceptions can “continue” to exist beyond the bounds of the sensory fields. But that has no bearing on the question of whether reasoning, con‑ sidered as custom‑guided inference, is an effective cause of the vulgar belief (Loeb 2002, 194–7).12 In addressing this matter, Hume further neglected to consider a point he elsewhere recognized: that beliefs enlivened by different vivacity sources can be opposed to the point where there is “proof against proof” (EHU 10.11): completely compelling reasons for opposed conclusions.13 The child or dog who hears a familiar voice coming from the next room be‑ lieves impressions constitutive of a familiar person to exist beyond the bounds of its sensory fields. The philosopher who presses the side of an eye or walks away from a table experiences something that leads them to reason that impressions only ever appear on sensory fields (EHU 12.9; T 1.4.2.45). Opposing the one set of experiences to the other produces a “mutual destruction of arguments” (EHU 10.13). It does not demonstrate that the vulgar belief is “entirely unreasonable” or proceeds from “some other faculty than the understanding” defined as empirically guided belief.

Section 7.4  259 Philosophers themselves are so strongly influenced by the causes of the vulgar belief that they are only determined to (temporarily) suspend belief in both the unperceived existence of the speaker in the next room and the force of the sceptical arguments. Inevitably, the memory of the argument from the doubled or diminishing images fades. Then, the reiterated sound of the voice produces its customary effect on our belief, even if we are phi‑ losophers (EHU 12.23; T 1.4.2.57). Hume had good reason to accept that custom‑guided reasoning causes the belief in body. 7.4

Imagining bodies

Claiming that the belief in body cannot arise from the senses or reasoning, Hume concluded that it must be owing to the imagination (T 1.4.2.14). He proposed to determine how by first identifying the features that dis‑ tinguish those of our impressions that we consider to be distinct and con‑ tinued from those that we regard as internal and perishing (T 1.4.2.15).14 He identified two: constancy and coherence. He did not explain what led him to fix on them. Hume thought that impressions and ideas of vision and touch are dis‑ posed in space (Chapters 1 and 2), so among the other features he might have mentioned is independent mobility (Section 7.2).15 He held that pas‑ sions do not exist in space and so cannot move. Pains and visual after‑­ images are fixed to a position on the sensory fields and so do not move independently. Despite our best efforts, we find we cannot put anything in front of them to make them disappear. In all three cases, the impressions have nowhere to hide. Their disappearance can only be due to their an‑ nihilation and their appearance to their creation. They exist if and only if they are forward on the sensory field. Other visual and tactile sensations, being mobile, do have somewhere to hide. They can move behind one another or track off the edge of the sen‑ sory fields, either when our eyes, head, or hands, move or when they move relative to landmarks. We discover that they do hide from us by moving to the places where we think they might have gone and finding them there. This makes us think that they continue to exist when not forward on the sensory fields. Neglecting these considerations, Hume first maintained that some im‑ pressions “have a peculiar constancy, which distinguishes them from the impressions, whose existence depends upon our perception” (T 1.4.2.18). Explaining with examples, These mountains, and houses, and trees, which lie at present under my eye, have always appear’d to me in the same order; and when I lose

260  Causes of the belief in bodies sight of them by shutting my eyes or turning my head, I soon after find them return upon me without the least alteration. My bed and table, my books and papers, present themselves in the same uniform manner, and change not upon account of any interruption in my seeing or perceiv‑ ing them. This is the case with all the impressions, whose objects are suppos’d to have an external existence[.] (T 1.4.2.18) The references to appearing in the same order and in the same uniform manner could be read as references to customary contiguity in space (Section 6.10). At this point, Hume was a step away from concluding that we draw infer‑ ences from the customary contiguity of impressions to their persistence at the same places relative to one another, and so to their continued existence be‑ yond the bounds of the sensory fields (Section 6.10). Seeing the tree, I think of the house just outside of my field of view that I have customarily observed to lie in that direction beyond the tree, and then of the mountains beyond the house. And likewise for looking at the bed and inferring the existence of the adjacent table, books, and papers. Hume did occasionally draw such infer‑ ences, as in the cases of the partially occluded pedestrian and the contents of the room occupied by the familiar speaker (Ax 4, cited in Section 7.3). But, having come within an inch of making this discovery, he got dis‑ tracted by the fact that not everything abides constantly in place. Some things move or alter, [b]ut here ’tis observable, that even in these changes they preserve a coherence[.] (T 1.4.2.19) The subject having been changed, he cited a fading fire as a paradigm example. When I return to my chamber after an hour’s absence, I find not my fire in the same situation, in which I left it: But then I am accustom’d in other instances to see a like alteration produc’d in a like time, whether I am present or absent, near or remote. (T 1.4.2.19) Here, Hume recognized that we experience a degree of regularity in the way certain species of object change position or quality over time. This natural history or “coherence,” as he called it, is discovered over the course of repeated experiences. We are “accustom’d in other instances to see a like alteration produc’d in a like time” (T 1.4.2.19).

Section 7.4.1  261 At this point, Hume was a step away from concluding that we draw in‑ ferences from customary identity over time (Section 6.11). Once again, he did occasionally draw such inferences, as in the cases of the rising sun and human mortality (T 1.3.11.12, cited in Section 7.3). But while Hume was, as Aristotle liked to put it, “forced as it were by the truth” to occasionally draw these inferences, he did not take them up into his theory of belief. Rather than recognize what he called “constancy” and “coherence” as distinct forms of custom‑guided inference, he swiv‑ elled to treat them as empirically unguided operations of the imagination. The swivel led him to subtly redefine (almost invert) his opening ac‑ counts of coherence and constancy. 7.4.1 Coherence

Having initially exemplified coherence with the (empirically discovered) customary behaviour of an object over time, Hume began to treat coher‑ ence as being “necessary to suppose” in order to provide for “dependance and connexion” between distinct species of external object.16 To begin with the coherence, we may observe, that tho’ those inter‑ nal impressions, which we regard as fleeting and perishing, have also a certain coherence or regularity in their appearances, yet ’tis of some‑ what different a nature, from that which we discover in bodies.… [O]n no occasion is it necessary to suppose, that [our passions] have existed and operated, when they were not perceiv’d, in order to preserve the same dependance and connexion, of which we have had experience. The case is not the same with relation to external objects. Those require a continu’d existence, or otherwise lose, in a great measure, the regular‑ ity of their operation. (T 1.4.2.20) Illustrating, Hume described himself seated in his chamber, hearing a squeak of hinges, and then seeing a porter bearing a letter from a distant friend. He reported being led by these experiences to think of an opening door on the other side of the chamber, stairs that the porter must have mounted to ascend to the chamber, and ferries that must have brought the letter over continents and seas. These are objects that he had customarily observed to precede the events he had just experienced. But in this case, none of them is observed. In the case of passions, any failure to observe a regularity is proof that the passions are not in fact dependent or connected. But (to make a point Hume failed to make, but that proves to be quite relevant in this case), passions have nowhere to hide whereas external objects, being mobile in

262  Causes of the belief in bodies space, have many places where they can hide. In the case of external ob‑ jects, we draw a distinction between failing to observe a regularity and the observation of its failure to occur.17 If we did not do this, but instead took the failure to observe to be the observation of a failure, our current expe‑ riences (hearing the squeak without seeing an opening door, and so on) would disconfirm any rule we might have derived from prior experience, and we would begin to think objects are less regular in their operations than we had previously found them to be. However (not having made this point about hiding places), Hume ap‑ pears to have been stuck for an account of how we would draw the distinc‑ tion between failure to observe and observation of a failure and uncertain how to proceed without one. Imagining himself continuing to look at his books and papers while hearing the hinges squeak behind his back and then being presented with the letter by the porter, he described himself inferring that the door he re‑ membered on the other side of the room must “be still in being,”18 that the stairs he remembered must not have been “annihilated by my absence,” and that “posts and ferries” and “the whole sea and continent” between himself and his friend must have a “continu’d existence.” These are infer‑ ences from effects to causes, but he declared that the phenomena of the hinge squeak, porter, and letter might be viewed “in a certain light” as “contradictions to common experience” and “regarded as objections to those maxims, which we form concerning the connexions of causes and ef‑ fects.” Of course, we do not view them in that light or regard them in that way. Rather than consider that we have experienced failures of common experience to be reproduced, we suppose that the missing objects must still exist. But why? For Hume, the answer initially comes from the most surprising of sources.19 I am accustom’d to hear such a sound, and see such an object in motion at the same time. I have not receiv’d in this particular instance both these perceptions. These observations are contrary, unless I suppose that the door still remains, and that it was open’d without my perceiving it. And this supposition, which was at first entirely arbitrary and hypothetical, acquires a force and evidence by its being the only one, upon which I can reconcile these contradictions. (T 1.4.2.20) Hume’s surprisingly rationalist talk of forming an “arbitrary and hypo‑ thetical” supposition that “acquires its force and evidence” from inference to the best explanation seems ad hoc. He nowhere else recognized infer‑ ence to the best explanation as a source of force and vivacity. Why not

Section 7.4.1  263 maintain that since “I am accustom’d to hear such a sound, and see such an object in motion at the same time,” I will be determined by habit and the mechanisms of vivacity transfer to form a lively idea of a moving door (Gomberg 1976; Loeb 2002, ch6)? Hume seems to have felt the need to explain himself more fully, because he did not close his discussion at this point. Relabelling the “arbitrary and speculative hypothesis” as a “conclusion from the coherence of appear‑ ances”20 he wrote, But tho’ this conclusion from the coherence of appearances may seem to be of the same nature with our reasonings concerning causes and effects; as being deriv’d from custom, and regulated by past experience; we shall find upon examination, that they are at the bottom considerably different from each other, and that this inference arises from the under‑ standing, and from custom in an indirect and oblique manner. (T 1.4.2.21) By now, coherence has lost its direct experiential foundation. The coher‑ ence that is spoken of here is not regularity in the way objects of a certain species, like fading fires, are observed to behave over time, but regularity in the succession of different species of object, like hinge squeaks and moving doors. Hume went on to say that this regularity is often inflated beyond what custom warrants. (ER): Extrapolated regularity Any degree … of regularity in our perceptions, can never be a foundation for us to infer a greater degree of regularity in some objects, which are not perceiv’d; since this supposes a contradiction, viz. a habit acquir’d by what was never present to the mind.… We remark a connexion be‑ twixt two kinds of objects in their past appearance to the senses, but are not able to observe this connexion to be perfectly constant, since the turning about of our head, or the shutting of our eyes is able to break it. What then do we suppose in this case, but that these objects still continue their usual connexion, notwithstanding their apparent inter‑ ruption, and that the irregular appearances are join’d by something, of which we are insensible? But as all reasoning concerning matters of fact arises only from custom, and custom can only be the effect of repeated perceptions, the extending of custom and reasoning beyond the percep‑ tions can never be the direct and natural effect of the constant repetition and connexion, but must arise from the co‑­operation of some other principles. (T 1.4.2.21)

264  Causes of the belief in bodies When we reason from causes and effects, we proportion the strength of our belief according to the proportion of confirming and contrary experiments (EHU 6; T 1.3.12). But every case where causal reasoning is called for is a case where one of the cause or the effect is not observed or remembered. In such cases, prior experience leads us to form a vivacious idea of the missing partner. But such cases are also contrary experiments, where the regularity has failed to be observed. While habituation and the vivacity of the current impression may communicate enough vivacity to the associated idea for it to become a belief, the case gets added to the register of contrary experiments. According to (ER), head‑turning, eye‑shutting, and other in‑ terferences ought to eventually undermine the influence of any fortuitously discovered series of confirming experiments. That does not happen be‑ cause we further suppose that the interruptions are only “apparent.” Earlier, Hume described this as an arbitrary and speculative hypothesis enlivened by inference to the best explanation. But now, a further factor is introduced. (II): Inertia of imagination [T]he imagination, when set into any train of thinking, is apt to con‑ tinue, even when its object fails it, and like a galley put in motion by the oars, carries on its course without any new impulse. (T 1.4.2.22) Objects have a certain coherence even as they appear to our senses; but this coherence is much greater and more uniform, if we suppose the ob‑ jects to have a continu’d existence; and as the mind is once in the train of observing an uniformity among objects, it naturally continues, till it renders the uniformity as compleat as possible. (T 1.4.2.22) Once having found a degree of regularity in the succession of events, an inertial tendency of the imagination takes over to make us resistant to diminishing our confidence in the continuation of the regularity, even in the face of disconfirming evidence. This is not as ad hoc a mechanism as inference to the best explanation. (It is borrowed from an account Hume had developed for another purpose [T 1.2.4.23–4].) But it is empirically unguided and “unphilosophical,” like the attitude that we should keep on doing things a certain way because we have always done them that way. It operates in defiance of the evidence.21 Hume could have provided a more positive account. Having identified two resources we can draw on to manage circumstances where an antici‑ pated regularity fails to occur, constancy and coherence, he fixed on the one least suited to the job (pace Loeb 2002, Chapter 6).

Section 7.4.1  265 We are always orienting ourselves in an ambient space. We discover this space to be populated with “landmarks”: visual or tactile quality patches that customarily preserve their positions relative to one another. We take our vis‑ ual and tactile fields to be mobile windows on them. We orient ourselves by reference to those of them that we see (Sections 6.10–12). Having experienced relations of customary contiguity between impressions constitutive of immo‑ bile objects like windows, fireplaces, and doors, when only some of these impressions appear on the fields, we imagine their customary neighbours oc‑ cupying their usual positions at distances behind occlusions or beyond the bounds of the fields. These inferences are invulnerable to both (ER) and (II). Space is not like time. The regularity we observe is not one object being followed by another, but one quality patch lying a certain distance away from another in a certain direction. Contrary to (ER), if eye or head turn‑ ing causes one of the patches to track off the edge of the visual field, our failure to continue to observe the patch is only what we would expect given our association of the patch with that remote location. In this case, eye and head motions confirm the belief scores of times a minute. We take a fraction of a second to look to the side and verify that the patch remains in its expected place. Should we fail to find it there, we consider the con‑ nection to have been broken and, contrary to (II), no inertial tendency of the imagination can combat that. What we learn from rapid eye motions over small distances is discovered to hold for greater and greater distances and becomes the subject for a general rule: landmarks discovered to persist in the same relations to one another in one small area will persist in those relations regardless of how far away we go from that area. Whatever may be the case for causal inference, eye/head motions cannot make customary contiguity relations appear any less constant than they are.22 Part of what leads us to designate external objects as “external” is that we ascribe them to locations in space.23 Regularities involving external objects are not just regularities in the succession of one species after an‑ other, but regularities in where the species are found. When we fail to observe a regularity in the succession of spatial objects, we are determined by customary contiguity to dispose the missing object somewhere in am‑ bient space. That supposition is empirically guided by experience of how objects of that sort tend to move or rest, and it can be empirically verified by going to look for the object in the anticipated place or places. We do not always bother with verification. But because we not only imagine the missing object but imagine it to occupy a location that does not currently fall within the sensory field, we only think we have observed a failure in the normal succession of events if we look to where we expect the object to be and fail to find it there. Hearing the squeak of hinges, we form the vivacious idea of an open‑ ing door. We see no such thing. But we do see a window, which gives us a

266  Causes of the belief in bodies vivacious idea of the fireplace we customarily observe when we turn to the right of the window, which in turn gives us a vivacious idea of the door we customarily observe when we turn to the right of the fireplace and which we now take to be behind our backs. Having also had experience of the permanence and relative immobility of windows, fireplaces, and doors we feel no need to turn and verify that the door still exists and is opening. That idea is so strongly believed as to require disproof. If we were to turn and find that it has been removed or not been opened, we would think we had experienced a failure. But when we do not look to the place where we are vivaciously imagining the external object to be located, we cannot fancy that we have experienced a failure. We can only think that we have failed to make the observation, and our belief remains undiminished. Hume’s account replaces inference from customary contiguity in space with a proclivity to imagine an imperfectly observed temporal regularity to be more perfect than experience warrants. This makes it seem like the belief in the unperceived existence of the missing object is inadequately supported by the evidence. While that advances a sceptical project, it ne‑ glects an alternative that Hume could readily have invoked consistently with his own principles. 7.4.2 Constancy

Hume did go on to consider customary contiguity in space, or what he called “constancy,” writing that whatever force we may ascribe to coher‑ ence, “’tis too weak to support alone so vast an edifice, as is that of the continu’d existence of all external bodies” (T 1.4.2.23). Loeb (2002, 177) is baffled by this dismissal,24 but Hume may only have been thinking that coherence does not account for all the cases where we believe things to continue to exist when not perceived. Coherence may account for our in‑ ference from the hinge squeak to the unperceived existence of the door, but not to the trim around the doorframe, which does not figure as a cause or effect of anything we are experiencing or remembering. Hume’s discussion of constancy again takes a turn. Having initially de‑ scribed it as constancy in the (spatial) order in which objects like moun‑ tains, houses, and trees have “always appear’d” to us (T 1.4.2.18) or been disposed relative to one another, he swivelled to take it to be invariance in the appearance of a single object over time. His investigation led him to conclude that our experiences of constancy lead us into a contradiction that we paper over with a false supposition. Rather than further ruminate on that proposal or the literature discuss‑ ing it, the following sections argue that he could have offered a more posi‑ tive account.

Section 7.4.3  267 7.4.3 Identity and interruption

Hume’s examples of constancy all involve an “interruption” in the obser‑ vation of relatively immobile objects. [A]ll those objects, to which we attribute a continu’d existence, have a peculiar constancy, which distinguishes them from the impressions, whose existence depends upon our perception. These mountains, and houses, and trees, which lie at present under my eye, have always appear’d to me in the same order; and when I lose sight of them by shutting my eyes or turning my head, I soon after find them return upon me without the least alteration. (T 1.4.2.18) [W]e have been accustom’d to observe a constancy in certain impres‑ sions, and have found, that the perception of the sun or ocean, for in‑ stance, returns upon us after an absence or annihilation with like parts and in a like order, as at its first appearance. (T 1.4.2.24) We find by experience, that there is such a constancy in almost all the im‑ pressions of the senses, that their interruption produces no alteration on them, and hinders them not from returning the same in appearance and in situation as at their first existence. I survey the furniture of my cham‑ ber; I shut my eyes, and afterwards open them; and find the new percep‑ tions to resemble perfectly those, which formerly struck my senses. (T 1.4.2.35) In the first of these passages, the interrupted things are external objects, which are distinguished from “the impressions, whose existence depends upon our perception.” In the latter two passages, the interrupted things are described as impressions or perceptions “of” external objects or as perceptions had by surveying external objects. They are nonetheless impressions or perceptions of things that are disposed in space, and so (as discussed in Sections 1.1–4) impressions or perceptions that are themselves disposed in space. In all cases, the interruption is attributed to (i) motion of the sensory field away from and back to the location occupied by the set of spatially ordered objects, object parts, or impressions, (ii) motion of the objects or impressions behind an oc‑ clusion (the sun setting), or (iii) occlusion by the eyelids. However, Hume repeatedly declared that any “interruption” in what is suddenly described as a “perception” or impression destroys the identity of that “perception” or impression.

268  Causes of the belief in bodies [W]e are not apt to regard these interrupted perceptions as different, (which they really are).… But … this interruption of their existence is contrary to their perfect identity[.] (T 1.4.2.24) [Let us] enter, therefore, upon the question concerning the source of the error and deception with regard to identity, when we attribute it to our resembling perceptions, notwithstanding their interruption[.] (T 1.4.2.32) [T]he interruption of the appearance seems contrary to the identity, and naturally leads us to regard these resembling perceptions as different from each other[.] (T 1.4.3.36) [’T]is a false opinion that any of our objects, or perceptions, are identi‑ cally the same after an interruption[.] (T 1.4.3.43) ’Tis a gross illusion to suppose, that our resembling impressions are numerically the same[.] (T 1.4.2.56) Hume was not oblivious to the terminological shift (it is discussed at Trea‑ tise 1.4.2.31 and 38). But when he made it, he seems to have forgotten that he was still talking about things that he considered to be disposed in space. The outstanding consequence of this oversight is his claim that “interrup‑ tion” is contrary to identity. What is an interruption when that notion is applied to things that are disposed in space? One answer is that it is the annihilation and subsequent recreation of a “perception,” as when a pain or passion is intermittently felt, or a smell or sound fades into and out of consciousness. But this is not the only an‑ swer, though it is often the only answer Hume was willing to recognize. A striking instance is his reference to “the perception of the sun or ocean” as being “annihilated” (1.4.2.24). When things (be they “objects” or “perceptions”) are disposed in space, there are two ways they can be interrupted. They can be interrupted in the way they occupy space over time. (Today, we say that there is a break in their worldlines.) This happens when, at the next moment, a thing is not to be found at the place indicated by its history of motion and rest up to that point. In such cases, the thing is said to have disappeared. If the dis‑ appearance resists being explained by occlusion or the discovery of some

Section 7.4.3  269 cause responsible for a change in the body’s motion (and its subsequent discovery at the newly indicated location) the thing is said to have been an‑ nihilated. Annihilation is only attributed when the disappearance cannot be explained, and sometimes we are resistant to doing so even then, wor‑ rying that we have not done a thorough enough investigation. Things that have been annihilated cease to be perceived while we continue to perceive the location or range of locations they are expected to be found at, and while having reason to believe that they have not been occluded. In Hume’s lifetime, the sun and the oceans were never interrupted in this sense. Neither were Hume’s perceptions of the sun and the oceans, on any account on which they, too, are things that can track off the sensory fields or be occluded by other perceptions. It might be objected that we have reason to believe that no spatially disposed perception (no visual or tactile perception) can exist apart from being “perceived,” so to speak, that is, apart from being forward on the sensory field. But the question at hand does not concern reasons but causes. Of course, reasons can be causes, but then the question is how they weigh in the balance against opposed causes (Section 6.6). Whatever reasons we may have for thinking that visual and tactile impressions cannot exist behind what is on the fields or beyond their bounds need to be made explicit and weighed against the causes of the contrary belief, and that has not yet happened (Chapter 8). Until the dependency of perceptions on being forward on the sensory fields is established, it is necessary to allow for the possibility that things that have not been annihilated can cease to be forward on the sensory fields, and so can cease to be “perceived.” This is a second sense of “inter‑ ruption.” Call the first sense “interruption arising from the observation of a failure” (to continue on course). Call the second sense “interruption arising from a failure of observation.” Hume defined identity as the invariableness and uninterruptedness of any object, thro’ a suppos’d variation of time, by which the mind can trace it in the different peri‑ ods of its existence, without any break in the view, and without being oblig’d to form the idea of multiplicity or number. (T 1.4.2.30) But he also maintained that things can move without losing their identity (T 1.4.6.8; 1.4.5.27). Consequently, when preceded by the observation of dropping eyelids, volitions to move or blink, or trajectories of motion in the direction of occluding objects or the bounds of the sensory field, a bare failure to observe or “perceive” an object or an impression or a “percep‑ tion,” does not prevent us from continuing to “trace it in the different periods of its existence, without any break in the view.” We just do so by

270  Causes of the belief in bodies forming ideas of it continuing to persist in its state of motion or rest on the other side of eyelids or other occlusions, or beyond the bounds of the sen‑ sory field. There is only a “break” when we open our eyes or turn to look at the place where we imagine the idea and fail to obtain the impression the idea copies. Treatise 1.4.2.40 recognizes that “An interrupted appearance to the senses implies not necessarily an interruption in the existence” and so recognizes that a “break in the view” need not imply a break in the way a “perception” occupies space over time.25 That would require a “view” of its elanguesence or its sudden disappearance from its indicated place. Since none of Hume’s examples of cases where interrupted objects or “perceptions” are considered to continue to exist are cases that involve observations of a failure, none of them are cases where the “interruption,” such as it is, contradicts our attribution of identity to the objects or “per‑ ceptions.” None of them involves a “false supposition” or “gross illusion.” They are not cases where there is an interruption in the “periods of its existence” understood as the manner in which the object or “perception” occupies space over time. It remains to offer a “free” Humean account (Sections 4.2 and 4.3) of what causes lead us to believe that our visual and tactile perceptions con‑ tinue to exist behind occlusions or beyond the bounds of the sensory fields. 7.4.4 Identity and invariance

Rightly or wrongly, we do suppose that impressions continue to exist be‑ yond the bounds of the sensory fields and behind occlusions, and Hume was concerned to determine what causes us to do so. His answer to this question appeals to two kinds of resemblance: resemblance between earlier and later views, and resemblance between a “disposition” the mind is in when viewing an unchanging impression over a “supposed” variation in time, and when viewing a series of interrupted, but resembling impressions (T 1.4.2.33–4 and 35n). According to Hume, these resemblances lead us to confound the earlier and later perceptions. We are seduced into making what he repeatedly called a “mistake.” He could easily have drawn a more positive conclusion from just a part of his evidence. When describing how earlier and later perceptions resem‑ ble one another he emphasized that they do not just resemble qualitatively. All his cases remark on a resemblance that is specifically one of “order” and “situation” of “parts.” These mountains, and houses, and trees, which lie at present under my eye, have always appear’d to me in the same order[.] (T 1.4.2.18)

Section 7.4.4  271 [T]he perception of the sun or ocean, for instance, returns upon us … with like parts and in a like order[.] (T 1.4.2.24) [A]lmost all the impressions of the senses … [return] the same in ap‑ pearance and in situation as at their first existence. (T 1.4.2.35) There is a further suggestion, explicit in the first of these passages, that this resemblance in order or situation is repeatedly observed. This suggestion is implicit in the very term “constancy.” Two experiences are the most that is required to notice resemblance (T 1.3.1.2). But any attribution of “constancy” in situation and order of parts implies a greater number. We must have repeatedly or customarily found the perceptions to be spatially disposed in the same ways relative to one another. Hume’s examples suggest that identity does not just require resemblance. It also involves appeal to how things are disposed in space (Section 4.2). At the same time, as noted earlier (Section 7.4.3), he allowed that objects (impressions and compounds of impressions) can move without loss of identity. When there is no interruption in the view, and the earlier and later impressions are qualitatively invariant, the later impressions can still appear at a different place. But then the impression commonly appears at intermediate places at intermediate times, appearing to move continu‑ ously. Our identifications over time therefore involve proximity as well as resemblance. We do not just fixate on the place occupied by the impression but scan the immediately surrounding neighbourhood with each passing moment. Confronted with an uninterrupted view of a qualitatively invari‑ ant object, the mind must still be able to “trace it in the different periods [places] of its existence, without any break in the view, and without being oblig’d to form the idea of multiplicity or number” (T 1.4.2.30). Just as this does not mean that the object cannot move, it does not mean that it cannot change in other ways. It only requires that there be no ambiguity; that is, it only requires that, with each passing moment there is no ques‑ tion which is the most closely resembling object (compound impression) at the closest distance. If the resemblance or the proximity degrades to the point where it becomes less than obvious what to pick, we come closer to being “oblig’d to form the idea of multiplicity or number.” Identity must be one–one (Sections 4.1–3). This is how it goes when there is no “break in the view.” When there is a break, our belief is the product of previously discovered general rules governing the behaviour of that species of object. We deliberately recall our prior experiences of objects (compound impressions) of that sort and

272  Causes of the belief in bodies of how they have tended to move and change while under observation for that period of time. That instructs us what places to inspect for a later ap‑ pearance and what to expect to find there. We inspect each of those places and if exactly one of them reveals what we expect to find, we attribute identity; otherwise, we declare it uncertain what has become of the object. 7.5

A missed opportunity

The author of the Treatise strained himself in his efforts to prove that the belief in body has no foundation in the senses, empirically guided reason‑ ing, or empirically guided imagination. On each of these heads, his argu‑ ments are weak, chiefly because he neglected resources provided by his own accounts of space as a manner of disposition and of association by contiguity in space. The mature Hume abandoned the Treatise’s critique of the causes of the belief in body in favour of declaring it to arise from a blind and power‑ ful natural instinct or prepossession (EHU 12.7–8). That change did not prevent him from reiterating the Treatise’s direct arguments for external world scepticism (EHU 12.9 and 12.15). Plausibly, he had decided that his remedial sceptical project could rest just on those arguments. Conjectur‑ ally, he may also have come to appreciate that the arguments of Treatise 1.4.2.2–43 were too weak to bear repeating. But before we pat him on the back on either score, we should consider that in appealing to a blind and powerful natural instinct he missed the opportunity to develop the plausi‑ ble and powerful psychology of spatial and temporal consciousness he had begun to articulate in the Treatise. It remains to consider whether the dedicated sceptical arguments of the Treatise and the Enquiry can do their remedial job even in the face of the more robust account of the causes of the belief in body offered on Hume’s behalf in this chapter. Notes 1 Close analyses of Treatise 1.4.2 can be found in Price (1940), Kemp Smith (1941), Cook (1968), Bennett (1971), Gomberg (1976), Stroud (1977), Bricke (1980), Wright (1983), Fogelin (1985), Pears (1990), Bennett (2001), Loeb (2002), Baxter (2006), Kail (2007), Allison (2008), Wilson (2008), Fogelin (2009), Wright (2009), Butler (2010), Rocknak (2013), Meeker (2013), Ainslie (2015), Winkler (2015), among others. Not everyone has found it worth study. Strawson (2014, 46n22) excuses himself, referring his readers to Wright (1983, Chapter 2). Qu (2020, 101) offers only a perfunctory survey (107–10), declin‑ ing to “spend much time fretting over various interpretative nuances.” Garrett is likewise more interested in how Hume’s sceptical commitments play out in Treatise 1.4.7 than in the details of 1.4.2.

Notes  273 2 Wright (2009, 139; 1983, 41–8) begins his account of Hume’s position over Treatise 1.4.2.3–13 by appealing to arguments that Hume only offered much later (T 1.4.2.45). The later arguments appeal to reasoning from experiments, not to what the senses alone teach us. 3 Kail (2007, 54–5) sets some stock in this argument, writing that “We do not come to believe that objects are distinct from us through the senses because strictly speaking perceptual objects—impressions—are two‑dimensional” (55). This is only true of visual impressions, not tactile impressions. Tactile impres‑ sions make us aware of a third dimension over which sensible quality points can be disposed, and association of visual with tactile experience leads us to imagine visual points to be disposed at a distance from our bodies. But, not to harp on this, imagining what is beyond our bodies in a third dimension is superfluous given that we see coloured points disposed beyond them over the other two. 4 Hume observed that we do not attribute “reality, and continu’d existence” to impressions on account of whether they are voluntary or involuntary (T 1.4.2.16). But the reference there is to voluntary or involuntary occurrence; here it is to voluntary or involuntary motion. 5 This is a distinction that comes in degrees. As Hume recognized, the concept of belonging acquires much of its meaning from experience of what moves under our control. Belonging serves as the foundation for the closely related concept of property (literally, what is proper to us). Just as I direct the motion of my hand, so I direct the motion of the ring on my finger, the shirt sleeve at my wrist, the pen in my hand, the papers I am holding and so on. The term “my” spills over onto the impressions constitutive of these bodies to the extent that the impressions constitutive of my proper body parts are in contact with them and moving them in accord with my volition, and more of it spills over the longer they remain in contact with me and the easier they are for me to move. (Immobilized limbs and hair are “mine” for this reason.) They are experienced as proper to me or, as is said, as “my property” for as long as they are within my grasp and as easily movable as my hand. This does not make them entirely part of our bodies, because we understand that they can be easily separated from us. But it makes us think that grasping, occupying, and contacting confer title to property (T 3.2.3.4), which is the original foundation that concept is based on. It is then extended by legal fictions to such things as long possession, first possession, accession, and inheritance. 6 In fairness, Wright does not suggest that the antecedent is known by the senses. But the question at hand is what the senses cause us to believe, whether rightly or wrongly. 7 Winkler’s related attempt to defend Hume’s argument against the evidence of the senses (2015, 144–9) is also challenged by these considerations, as well as others presented over the course of this sub‑section. 8 Winkler claims that Hume’s discussion of reason at Treatise 1.4.2.14 is con‑ cerned only with “reason uninflected by imagination” (2015, 149). There is no evidence for that restriction. 9 This challenges Wright’s claim that we can never form an idea of what is exter‑ nal and independent because all ideas copy impressions and we can never form an impression that is external or independent (2009, 140; see Williams 2004, 282–3 for a similar worry). Externality and independence are spatial relations (assuming “independent” means “capable of existing beyond the bounds of

274  Causes of the belief in bodies the sensory fields,” as it is modestly taken to mean throughout this chapter to avoid unnecessary suppositions about minds and mental activities), and Hume maintained that the imagination can conceive simple ideas standing in other relations than those in which simple impressions are disposed (T 1.1.1.4; 1.2.6.9). Rocknak (2013, 108) makes a related mistake. Hume allowed that we can form ideas of impressions (T 1.3.8.17, 1.1.1.11), and so believe impres‑ sions to be disposed beyond the bounds of the sensory fields. See further Trea‑ tise 1.2.6.9, where Hume maintained that we can conceive external objects that are specifically like impressions but for having different relations, connections, and durations, that is, but for being differently disposed in space and time. 10 Kail (2007, 16) maintains that since probabilistic reasoning depends on re‑ peated prior observation of conjunctions of causes with effects (T 1.3.2.2), we have to be in a position to observe the continued existence of objects before we can reason to their continued existence. But Hume could easily have rec‑ ognized other empirically guided forms of probable reasoning. As the passages cited indicate, he occasionally did so. As discussed in Sections 6.10–12, had he recognized inference from customary contiguity and from customary one–one resemblance + proximity, identifying continued objects would not have been a problem. See further Section 7.4. 11 At the outset of Treatise 1.4.2 Hume declared that “our present enquiry is concerning the causes which induce us to believe in the existence of body” (T 1.4.2.2). He further declared all inquiry into the reasons for that belief to be “vain” as “That is a point, which we must take for granted in all our reason‑ ings” (T 1.4.2.1). See Allison (2008, 230). 12 Treatise 1.4.2.14 adds other arguments that may be quickly dismissed: (i) “[W]e can attribute a distinct continu’d existence to objects without ever consulting reason[.]” The question is whether we can also make the attribution by con‑ sulting reason. (ii) “’[T]is not by [arguments], that children, peasants, and the greatest part of mankind are induced to attribute objects to some impressions, and deny them to others.” Hume elsewhere recognized that children, peasants, and even animals infer the external and independent existence of impressions by causal reasoning (EHU 5.22; see also EHU 9 and T 1.3.16). (iii) “[A]s long as we take our perceptions and objects to be the same, we can never infer the existence of the one from that of the other.… Even after we distinguish our perceptions from our objects … we are still incapable of reasoning from the ex‑ istence of one to that of the other[.]” This is irrelevant. The question is whether we can infer externally disposed perceptions from perceptions currently on the sensory fields (see Loeb 2002, 194 and 194n24, citing Bennett 1971, 321, and Pears 1990, 164–7). 13 For Hume, the most striking instance is the case where perfectly reliable testi‑ mony produces one belief, while experience of what has always been the case produces the opposite belief (EHU 10.11). 14 See note 9 on Rocknak (2013, 108). 15 Independent motion is motion of what is on the sensory field relative to the boundaries of the sensory field. This is not a causal concept, because sensible quality patches can be found to change place relative to the sides of the sensory field without having identified any regular antecedent to their motion. Causal considerations can play a role in determining dependent mobility. The causes at play there are causes of sensory field motions (muscle, balance, and pressure sensations consequent on volition or sensations characteristic of being thrown,

Notes  275 spun, or carried). Signs of sensory field motions are used as heuristics to iden‑ tify what remains fixed to a position on the field as it moves, and so what is not independently mobile. 16 Ainslie (2015, 71–4) ignores this shift, treating everything Hume went on to say about the noise/door, porter/stairs, and letter/ferries as if it had been said about the fading fire. 17 Allison (2008, 243–4) charges that any attempt to draw a distinction between the appearance and the existence of perceptions (as with observing and fail‑ ing to observe) involves a commitment to an act/object distinction, where the acts are, to put it oddly, acts of perceiving perceptions. But when purportedly “mental” content is recognized as being mobile in space this does not follow. Being observed is just being in a certain place: forward on the sensory field. Existing without being observed is being somewhere else: behind an occlusion or beyond the bounds of the sensory field. 18 All quotations in this paragraph are from T 1.4.2.20. 19 My thoughts about the import of this passage are owing to an exchange with Graham Clay. 20 Treatise 1.4.2.20 contains over 600 words between the occurrence of “coher‑ ence” cited at the outset of this section and the one cited below. After that interval the meaning of the term has shifted from “customary behaviour of an object over time” to “supposition that an object that has customarily been observed to be conjoined with one now observed is still in being even though not now observed.” 21 Loeb (2002, 197–8) denies this, but for no good reason. It is not Hume who “cannot quite bring himself to rehearse [his] official position [that the belief in body is due to the imagination rather than understanding]” but Loeb who can‑ not bring himself to accept that the inertial tendency of the imagination is only initially created by custom, not subsequently guided by it. Loeb persistently writes of what he calls “custom‑and‑galley” as “cooperating” to produce a re‑ sult. They do not cooperate. The “galley” (II) is the prodigal child who spurns its parent’s advice and goes off on its own way. 22 This is where the account given here parts company with that offered by Loeb (2002, Section 6.2, 182–7) and his sources (Price  1940, 20–2, 35–6, 62–3; Pears 1990, 169–72, 176, 182–3). They do not recognize inference from cus‑ tomary contiguity, and so think Hume had a real difficulty providing for a distinction between observation of a failure and failure to observe. Price thinks Hume could provide for the distinction just by appeal to relations of succession in time. Pears gives reasons why he could not. Loeb expresses confidence that Hume could reconstruct the difference (186n11) but does not suggest how. 23 Contrary to Hume (T 1.4.2.10), this is the most important part. 24 Loeb eventually proposes a biographical explanation (193–207). Ainslie is similarly perplexed (2015, 71, 73–4). But “weak” often means “unable to do everything,” rather than “unable to do anything.” 25 This is not contradicted by Treatise 1.2.6.4–8 which only goes so far as to say that we cannot conceive of anything specifically distinct from impressions and ideas, without commenting on whether impressions and ideas are depend‑ ent on minds. It is contested by Treatise 1.4.2.37, which observes that “the appearance of a perception in the mind and its existence seem at first sight entirely the same,” but this “seeming” is contested over the course of Treatise 1.4.2.38–40.

276  Causes of the belief in bodies Bibliography Ainslie, Donald C.  2015. Hume’s True Scepticism. Oxford: Oxford University Press. Allison, Henry. 2008. Custom and Reason in Hume: A Kantian Reading of the First Book of the Treatise. Oxford: Oxford University Press. Baxter, Donald L. M. 2006. “Identity, Continued Existence, and the External World.” In The Blackwell Guide to Hume’s Treatise, edited by Saul Traiger, 114–32. Malden: Blackwell. Bennett, Jonathan. 1971. Locke, Berkeley, Hume. Oxford: Clarendon Press. Bennett, Jonathan. 2001. Learning from Six Philosophers. Oxford: Clarendon Press. Bricke, John. 1980. Hume’s Philosophy of Mind. Princeton: Princeton University Press. Butler, Annemarie. 2010. “Hume on Believing the Vulgar Fiction of Continued Existence.” History of Philosophy Quarterly 27: 237–54. Cook, John W. 1968. “Hume’s Scepticism with Regard to the Senses.” American Philosophical Quarterly 5: 1–17. Fogelin, Robert J. 1985. Hume’s Skepticism in the Treatise of Human Nature. London: Routledge & Kegan Paul. Fogelin, Robert J. 2009. Hume’s Skeptical Crisis. Oxford: Oxford University Press. Gomberg, Paul. 1976. “Coherence and Causal Inference in Hume’s Treatise.” ­Canadian Journal of Philosophy 6: 693–704. Kail, P. J. E. 2007. Projection and Realism in Hume’s Philosophy. Oxford: Oxford University Press. Kemp Smith, Norman. 1941. The Philosophy of David Hume. London: Macmillan. Loeb, Louis E. 2002. Stability and Justification in Hume’s Treatise. Oxford: ­Oxford University Press. Meeker, Kevin. 2013. Hume’s Radical Scepticism and the Fate of Naturalized Epis‑ temology. Basingstoke: Palgrave Macmillan. Pears, David. 1990. Hume’s System. Oxford: Oxford University Press. Price, H. H. 1940. Hume’s Theory of the External World. Oxford: Clarendon Press. Qu, Hsueh M. 2020. Hume’s Epistemological Evolution. New York: Oxford ­University Press. Rocknak, Stephanie. 2013. Imagined Causes: Hume’s Conception of Objects. Dordrecht: Springer. Strawson, Galen. 2014. The Secret Connexion: Causation, Realism, and David Hume. Revised edition. Oxford: Oxford University Press. Stroud, Barry. 1977. Hume. London: Routledge. Williams, Michael. 2004. “The Unity of Hume’s Philosophical Project.” Hume Studies 30: 265–96. Wilson, Fred. 2008. The External World and Our Knowledge of It. Toronto: Uni‑ versity of Toronto Press. Winkler, Kenneth. 2015. “Hume on Scepticism and the Senses.” In The Cambridge Companion to Hume’s Treatise, edited by Donald C. Ainslie and Annemarie Butler, 135–64. New York: Cambridge University Press.

Bibliography  277 Wright, John P. 1983. The Sceptical Realism of David Hume. Minneapolis: Univer‑ sity of Minnesota Press. Wright, John P. 2009. Hume’s ‘A Treatise of Human Nature’ An Introduction. Cambridge: Cambridge University Press. Wright, John P. 2016. “Hume’s Skeptical Realism.” In The Oxford Handbook of Hume, edited by Paul Russell, 60–81. New York: Oxford University Press.

8 Reasons for scepticism about the external existence of bodies

This chapter is concerned with Hume’s “proper” arguments for external world scepticism. The proper arguments do not attack the grounds for the belief in body. They present counterevidence. They can be understood in one of three ways: as refutations, as “proof against proof,” or as failed attempts. This chapter argues that they are best understood in the sec‑ ond way. Hume’s proper sceptical arguments conclude that what he called “sensible qualities” only exist on the sensory fields of sentient creatures and that shape, motion, and solidity are inconceivable except as properties of sensible qualities. External world scepticism follows from the fact that external objects are only ever conceived as extended, solid, sensible quali‑ ties. The argument for these conclusions constitutes proof against proof because it is based on evidence from eye crossing and distancing that is counterbalanced by evidence from rotary motion and the removal of oc‑ clusions. Due to the rareness and difficulty of attending to the results of the sceptical experiments, and the reasoning required to draw inferences from them, the doubt they induce is eventually overcome by renewed experi‑ ence, but they can have a lasting effect on our cognitive disposition. 8.1

External world scepticism in the Enquiry and the Treatise

In the Enquiry, Hume argued that the belief in continuously and inde‑ pendently existing objects or “body” arises from a “blind and powerful” (EHU 12.8) “natural instinct or prepossession” (EHU 12.7). In the Trea‑ tise, he argued that it arises from the influence of the coherence and con‑ stancy of impressions on the imagination (T 1.4.2–43). In both works, he maintained that the belief takes “the very images presented by the senses to be the external objects.”1 In both works, he went on argue that this be‑ lief is factually false, drawing on similar considerations to do so: evidence that the images presented by the senses do not exist beyond the bounds of the sensory fields (EHU 12.9; T 1.4.2.45) and reasons for concluding that extension, motion (gravitation), and solidity (impenetrability) can only be DOI: 10.4324/9781032677880-9

Section 8.2  279 conceived as properties of sensible qualities (E 12.15; T 1.4.4.6–14).2 Both works remark that we might attempt to evade the former observations by proposing that the images presented by the senses are representations of external objects, and both go on to show that this proposal is unsustainable (EHU 12.10–13; T 1.4.2.46–9). In what follows, Hume’s discussions in the Enquiry and the Treatise are considered in tandem. In keeping with his observation that “Nothing but an absolute necessity can oblige an historian to break the order of time, and in his narration give the precedence to an event, which was in reality posterior to another” (T 2.3.7.7), the account of the Treatise is given first followed by comments on any differences found in the Enquiry. An exception is made in Section 8.4 for reasons noted there. 8.2

Scepticism about independent existence

In the Treatise, Hume maintained that while a natural propensity of the imagination leads us to suppose that our perceptions continue to exist when not perceived, “when we compare experiments, and reason a little upon them, we quickly perceive, that the doctrine of the independent existence of our sensible perceptions is contrary to the plainest experience” (T 1.4.2.44). Unfortunately, he had no settled or acceptable position on what perceptions are dependent on.3 At Treatise 1.4.2.45, he concluded from one of his experiments that “all our perceptions are dependent on our organs, and the disposition of our nerves and animal spirits.” Elsewhere, he spoke of dependence on what “we are conscious of in ourselves” (T 1.4.2.10), on “us” (T 1.4.2.8), on “ourselves” (T 1.4.2.5, 1.4.2.10), on “perception” (T 1.4.2.2, 1.4.2.18), and on “the mind” (1.4.2.14). Given that he elsewhere reduced minds, perceivers, and selves to bundles of perceptions (T 1.4.6.4), dismissed the distinction between acts of perceiving and perceptions (T 1.4.5.26–7), and took organs, nerves, and animal spirits to be composed of visual and tactile perceptions (T 1.4.5.19), these are viciously circular subjects of dependence. He might have done better to propose that our perceptions are dependent on the visual and tactile sensory fields. For each of us, there is clearly a bounded region of space within which visual impressions exist, and another within which tactile impressions exist. The existence of impressions beyond the bounds of these regions is only inferred or conjectured. These regions are all that “self” or “mind” refers to when considered as a perceiving subject.4 There are two opposed ways of understanding the relation between the sensory fields and what appears on them: (i) as windows on visual and tactile impressions, where some impressions are like dirt on the window glass,

280  Reasons for scepticism and so move along with the sensory field, whereas others are disposed on the far side of the window, and so existent independently of whether the window is turned in their direction, or (ii) as monitors or cinema screens that display content that only exists on the fields.5 Hume’s experiments are reasons for the second alternative. They give evidence for concluding that impressions only exist on the sensory fields, not behind or beyond them. On this reading, Hume’s case presumes that we have arrived at the ­common‑sense view of the world described in Chapters 6 and 7. We have come to identify those impressions called volitions, as “causes” (that is, regular antecedents) of the motion of those collections of impressions con‑ stitutive of personal body parts (eyes, hands, legs, and so on) and identify the tactile impressions constitutive of these body parts with the visual im‑ pressions constitutive of them. We have learned to perceive visual depth. It also helps if we allow that, even though he failed to theorize the fact, Hume was “forced, as it were, by the truth” to recognize that we have drawn enough one–one inferences from resemblance and proximity over time to have recognized multiple unchanging “objects” (collections of im‑ pressions) surrounding our personal body parts, and enough inferences from their constant contiguity to conceive them as relatively immobile landmarks populating a space extending beyond the bounds of our sen‑ sory fields. Given that background knowledge, we have further identified certain volitions as the causes of “field motions” resulting from turning the eyes or head, or walking towards, away from, or around a landmark. And we have drawn a distinction between impressions that are affixed to a location on the sensory fields, and so move with them wherever they are turned, like pains and visual after‑images, and impressions that move or rest independently of the motion of the fields. None of these things are objects distinct from impressions and ideas. However, many of them are believed to be spatially disposed behind or beyond those impressions we identify as our personal body parts. Hume’s experiments challenge this distinction. They give us reason to think that our independently mobile impressions are only independent of our volitions, not of our sensory fields. Like pains and visual after‑images, our independently mobile impressions are “field states,” not independently existing objects viewed through a window. 8.2.1 Double vision

Hume first noted that pushing an eye out of alignment with the other eye causes the visual field to double. Viewed one way, there is nothing chal‑ lenging about this result. We have two eyes each of which views ambient objects from its own position, as illustrated by the experience of looking through a (monocular) microscope with one while looking at an adjacent

Section 8.2.1  281 notebook with the other. Novices can be unsettled by this experience and will deal with it by trying to look through just one eye at a time, but eventually, the two visual fields come to appear as disposed one to the left of the other, just as we take the microscope slide and the notebook to be disposed one to the left of the other. In this case, the two visual fields are experienced as two windows, each looking out at a separate portion of what lies before us. In normal vision, the two windows are turned so that their fields overlap. (This is what makes the view through the microscope with two eyes open so disconcerting.) Even then, we find by opening and closing each eye in turn that the left eye sees more of what lies to the left, the right more of what lies to the right, that the left sees more of the left side of centrally located solids and the right more of the right side, and that the left sees more of what lies immediately to the right of an occlusion while the right sees more of what lies to the left. Considering these facts, it is not surprising that moving one eye out of alignment with the other should cause two visual fields to appear. But there is still something troubling about what we see when we induce double vision. When we press one eye with a finger, we immediately perceive all the objects to become double, and half of them to be remov’d from their common and natural position. But as we do not attribute a continu’d existence to both these perceptions, and as they are both of the same nature, we clearly perceive, that all our perceptions are dependent on our organs, and the disposition of our nerves and animal spirits. (T 1.4.2.45) When two windows look out on overlapping portions of the same area, as when looking through binoculars, what we see in the region of overlap appears single. As the windows are angled further and further away from one another there is less overlap, but what appears in the area of overlap is still seen as single. When there is no overlap, we do not see double images of the same thing, but two separate portions of what lies before us. But when we induce double vision, it seems that each eye is experiencing its own copy of what the other eye sees, with the two copies being partially overlaid. When we start to think that each eye appears to be seeing its own copy, Hume has won his point. We have as much reason to declare what the one eye sees to be a copy as we have to consider what the other sees to be a copy.6 Moreover, we have no good reason to think we are not continuing to see two copies when the eyes are turned so the two are overlaid and the appear‑ ance of doubling disappears. The experiment suggests that we only ever see projections on our visual fields, not bodies lying on the far side of a window.

282  Reasons for scepticism Hume’s argument is commonly read differently. On standard readings (Van Cleve 2008, 11; Wright 1983, 43), the argument premises that there is only one object that appears doubled. Because we do not take both ap‑ pearances to be the object, at least one must be dependent on our activity, but then because all our perceptions (including those had prior to the dou‑ bling) are of the same nature, all must be due to our activity. However, Hume’s statement does not make any claim about the external existence of a single object. Rather than be a variant on so‑called “per‑ ceptual relativity” arguments (arguments that premise that the same ob‑ ject appears differently in different circumstances) Hume’s argument only premises a change in what we see consequent on volition. On an important alternative reading (Butler 2010), Hume meant to point to a conflict between the senses and the imagination. When double vision occurs, it frustrates our tendency to imagine perceptions had before and after an interruption to be views of something that continues to ex‑ ist unperceived. We are forced to accept that they are fleeting perceptions rather than a continued object. But when Hume wrote that “we do not attribute a continu’d existence to both these perceptions” (T 1.4.2.45), his point was only backwards‑­ looking. At least one of the “perceptions” must be newly emergent. What that implies for any subsequent experience of a single “perception” is ­unclear because, at the point of writing, only the current emergence of the doubled perceptions is under consideration. The double vision experiment does not appeal to anything our experi‑ ence leads us to imagine. It draws on what we see, though on what we see in an uncommon circumstance. In the uncommon circumstance, we see double in the same way that we see single. We are not attending to what we see in any special way. We are just seeing, and seeing something rather different from what we ordinarily see. The experiment also draws on some subsequent reasoning about what we see. We reason that each eye must be seeing a copy, and some further reasoning leads us to the conclusion that everything we see is a projection on the visual field; none of it is seen as lying on the far side. Just as the experiment does not involve a disap‑ pointment of imagined continuity, so it does not involve any special act of reflection, understood as forming ideas of ideas or considering ideas or impressions as being ideas or impressions rather than objects (contrast Ainslie 2015, 121–34). Because the experiment draws on what we see in an uncommon circum‑ stance, and because it requires that we not just see it, but draw inferences from what we are seeing, it can be (and is) countermanded by what we see in more common circumstances, and by ideas enlivened to the point of belief by what we see in more common circumstances. The more common experiences lead us to draw a distinction between what moves with the sensory field and what moves independently of it, and to draw inferences

Section 8.2.2  283 from customary contiguity to the existence of sensible quality patches be‑ yond the bounds of the sensory fields and behind occlusions. Those ex‑ periences lead us to think we are seeing through the fields to what lies on the far side. It is not hard to say how the contest between the two sorts of experiments will turn out. The one that is more often repeated and that re‑ quires less reasoning will eventually win our consent. The other will startle us, and induce temporary confusion. For Hume’s remedial purposes, all that is required is that the loser “admit of no answer,” not that it “produce conviction” (EHU 12.15n). It needs to constitute “proof against proof.” It does not need to constitute disproof. 8.2.2 Distancing

The Treatise barely mentions a second experiment, writing that the opinion that our perceptions have no independent existence is “confirm’d by the seeming encrease and diminution of objects, according to their distance” (T 1.4.2.45). Amplifying greatly, when we retreat from a location, all the colour patches on the visual field decrease in size, changing shape, bright‑ ness, saturation and hue, and losing distinctness and number of minimally visible parts as they do so. The changes are gradual, so they do not pre‑ vent us from making one–one correlations over time based on resemblance and proximity. With experience, we come to mediately perceive objective size constancy, and it requires a special act of attention to notice that the colour patches are shrinking. But with continued distancing, something happens we cannot fail to notice: the colour patches move towards what is called the vanishing point, into which they disappear. There is no evidence that they have moved off the field or been occluded. They seem to shrink to nothing at the vanishing point.7 But this is how pains, smells, tastes, sounds, and other perceptions that are fixed to locations on the sensory fields behave: by elanguescence and vanishing. Once again, the distinction between what lies on the far side of the field and what is embedded in it (Section 7.2) is undermined. It has been objected that the fact that impressions change when we re‑ treat does not prove that they do not continue to exist. It just proves that they are only visible at certain distances.8 When we remove from that dis‑ tance, other impressions, visible at the new distance, appear in their place, occluding them. Something Hume said in another context offers a reply. Our passions are found by experience to have a mutual connexion with and dependance on each other; but on no occasion is it necessary to suppose, that they have existed and operated, when they were not perceiv’d, in order to preserve the same dependance and connexion, of which we have had experience. (T 1.4.2.20)

284  Reasons for scepticism Just as passions and pains appear, intensify, fade, and vanish (doing so in place in the case of pains), so visual impressions emerge from the vanish‑ ing point, grow, shrink, and disappear back into the vanishing point. In neither case is it necessary to account for these regularities in their patterns of emergence and decay by appealing to a further dimension where they remain disposed. Performing a volition to approach or retreat we form a vivid idea and belief of what will be seen next, and that belief is confirmed without needing to think that what we see next was hidden rather than emergent. The case is quite different when we see one colour patch track off to the side of the visual field or move gradually behind another. In that case, we see the colour patch move off in the direction of a remote or oc‑ cluded location. In the former case, what appears at a location simply van‑ ishes from that location consequent on a volition to approach or retreat. It does not appear to go anywhere. What moves in the direction of a place leaves us with a lively idea of it being at that place should we turn towards it. What simply vanishes leaves no lively idea or belief concerning where it might still be. Reid was later to object that the distancing argument ignores the fact that the external object really does vary as we approach to or recede from it (EIP 2.14, 180–2). He pointed out that objects have two kinds of m ­ agnitude. One is visible magnitude, often misleadingly called “apparent” magnitude, and the other is tangible magnitude, often called “real” magnitude (misleadingly suggesting that visible magnitude is not equally real). Whereas tangible mag‑ nitude is measured by juxtaposition with standard units of measure, visible magnitude is measured by the angle the object subtends from a vantage point. Necessarily, as the vantage point is more remote, this angle decreases, mean‑ ing that the object really and necessarily varies in visible magnitude (in the angle it subtends) as we recede from it, even though it is invariant in tangible magnitude. Hume might have replied that while visual magnitude is measured by an angle that varies with distance, the size of that angle is determined by another sort of magnitude. Turning our eyes as far to the right as pos‑ sible to trace the line of the horizon, followed by turning our head as far to the right as possible, followed by turning our body, eventually brings us back to our starting point, leading us to think of the visual field as a portion of the inner surface of a sphere on which we are centred. Visual objects are originally experienced as two‑dimensional projections onto this concave surface, and their angular size is a function of how much of this surface they take up (half makes a 180‑degree angle, a quarter a 90‑degree, and so on). When our distance from a visible object changes, its angular magnitude changes because it gains or loses minimally visible parts, which bloom out of or collapse into the vanishing point. We may still identify the larger and smaller appearances as being of the same object, provided

Section 8.2.2  285 the resemblance and proximity are great enough to satisfy the one–one condition on identity correlations. But if we can overcome our tendency to mediately perceive objective size constancy, and see this enduring object inflating or collapsing like a balloon (as we often can), we must think of what we are seeing as gaining or losing visible parts and so as consist‑ ing of emerging and disappearing points. The lost parts do not move off somewhere; they simply disappear; the gained parts do not move in from somewhere; they simply emerge. But when the object is, say, a ball or a cube held in the palm of a hand and moved towards or away from us, it does not gain or lose tactile parts. We find we are seeing one thing, a changing collection of minimally visible coloured points, constitutive of a growing or shrinking visual image, while feeling another, an unchanging collection of minimally tangible pressure points. Reid could ignore this implication because, on his account, visual and tactile sensations suggest conceptions of external objects that subtend different angles at different distances. But Hume maintained that vision presents us with coloured im‑ pressions disposed in space to take up “angular” portions of a visual field while touch presents tactile impressions occupying an abiding portion of the tactile field. We are led in this case to think that the images exist only on the sensory fields. Once again, it is worth pausing to remark on what sort of evidence this experiment draws on, and what it asks us to do with that evidence. Unlike the double vision experiment, the distancing experiment does not draw on simply seeing what lies before us, even if in rather special circumstances. We have learned to perceive objective size constancy and so do not think we see the table shrink as we retreat from it, at least not before it becomes too small to distinguish. But this does not mean that we need to adopt a special reflective attitude to appreciate the evidence (Ainslie 2015, 121–34). The apprentice visual artist who learns how to draw in perspective does not do so by reading a treatise on perspective, but by looking through an Alberti window and drawing exactly what they see through each grid square. The exercise enables the apprentice to separate what is immedi‑ ately seen from an overlay of associations and judgements. Aside from asking that we unlearn how to perceive objective size con‑ stancy, the experiment asks us to notice that shrinking into the vanishing point is more like the elanguasence of passions and pains, which are not thought to continue to exist when unobserved, than it is like the motion of colour patches beyond the bounds of the visual field or behind occlu‑ sions. But though it does not draw on any further reasoning, the conclu‑ sion it asks us to draw, that all colour patches exist only on the visual field, because all disappear into the vanishing point with retreat, is opposed by the conclusions we draw from motions that do not involve approach and retreat, like turning the eyes or head from side to side or reaching out

286  Reasons for scepticism to laterally remove an occlusion. Like double vision, distancing provides “proof against proof,” not disproof. This is not a case where an original quantity of vivacity, communicated by a single experience, is divided over a number of possibilities and we end up believing the one that is endowed with a quantity sufficient to outweigh the combined weight of the alternatives (T 1.3.11–12; EHU 10.4). It is a case where different experiences enliven contrary ideas, each communicat‑ ing so much vivacity that it constitutes a “proof” (EHU 10.11–13). Cases where there is proof against proof induce hesitation and suspension of be‑ lief. However, the proofs by the double vision and distancing experiments are not as evident. The one arises from special circumstances, the other is concealed by our learned perception of objective size constancy, and both experiments demand that we draw inferences from their evidence. The contrary evidence is constantly being advertised to us, and it automati‑ cally produces vivacious ideas of and so belief in the continued existence of impressions beyond the bounds of the sensory fields and behind occlu‑ sions. Even when we are impressed by the double vision and distancing experiments, we cannot resist eventually reverting to our prior conviction that some sensible quality patches continue to exist unperceived. But, in Hume’s estimation, our memory of the contrary experiment remains, leav‑ ing us more reluctant to accept all our beliefs and disposed to refrain from accepting those that are not constantly reinforced by continued experience. The remainder of what Hume had to say on this topic is not as effective, but it does not need to be. 8.2.3 Other variations; touch

Hume’s brief allusion to a “seeming encrease and diminution of objects, according to their distance” is followed by a list of other apparent altera‑ tions and changes: by the apparent alterations in their figure; by the changes in their colour and other qualities from our sickness and distempers; and by an infinite number of other experiments of the same kind; from all which we learn, that our sensible perceptions are not possest of any distinct or independ‑ ent existence. (T 1.4.2.45) The “apparent alterations in their figure” are presumably alterations aris‑ ing from walking around a location, and so are changes in the entire con‑ tent of the visual field resulting from our voluntary motions. But in this case, we have learned that colour patches are disposed before and behind one another over a third dimension, and it seems to us that the rotary

Section 8.2.3  287 motion is only revealing colour patches that were previously occluded, while occluding those that were previously in view. There is no evidence that the colour patches that we consider to persist in space on the far side of the visual field are being destroyed and replaced as a consequence of our voluntary motions. We might think of the occlusion and revelation as a change in the “appearance” of the object constituted by those colour patches, but then it is only a change in which face it is presenting to us. The variations in colour or other qualities that we experience when sick or otherwise disturbed are originally indistinguishable from those experi‑ enced when mist or coloured glass is interposed between the eye and the object, when clouds pass over the sun or when lighting conditions are oth‑ erwise changed. A similar point holds for fever, chills, weakness, or raw‑ ness. Failing association with other indications of illness, we need to move around to determine whether these changes are carried with us or whether they are environmental. If they prove to be carried with us, they are as‑ similated with visual after‑images and pains for that reason. Otherwise, we consider ourselves to be perceiving the effect of an unusual mixture of factors on the environment as it is on the far side of the sensory field. Hume’s sceptical arguments have a more serious shortcoming. Touch is not mentioned. He may have considered himself entitled to ignore it on the supposition that the tactile sensory field is coextensive with our own bod‑ ies. Granting that, he may have thought that touch is not even apparently a window on an external world. It never acquaints us with anything other than states of our own body parts. The omission of touch is nonetheless nontrivial. Many tactile sensations are experienced only by moving to particular places. That leads us to as‑ sociate them with those places and expect to find them there when we approach those places. To have such expectations is to dispose vivacious ideas that copy those impressions beyond the bounds of the tactile sensory field. It is also to believe impressions like those ideas to continue to exist at those external locations. In Hume’s day, the principal arguments against accepting that touch acquaints us with external impressions appealed to phantom pains expe‑ rienced by amputees, the experience of doubling felt when an object is placed between crossed over fingers, and water that feels hot to one hand and cold to another.9 The first is at best irrelevant and at worst counter indicative. Phantom pains, like pains of any other sort, are carried around with the amputee wherever they go and so are not considered independent just for that reason. They are not counterexamples to the principle that what moves along with the sensory field is subjective whereas what moves independently of it is objective, and therefore do nothing to call that prin‑ ciple into question. All that the phenomenon proves is that amputation of a limb is not equivalent to amputation of a portion of the tactile sensory

288  Reasons for scepticism field. That the field should extend into a space not occupied by any per‑ sonal body part should be no more surprising than that the impressions on the visual field extend beyond those constitutive of our hands and other visible body parts. Unlike crossing eyes, crossing fingers does not double tactile impres‑ sions. There are originally two differently disposed tactile impressions, but they are experienced as if they were disposed on opposite sides of two parallel fingers rather than immediately adjacent sides of crossed fingers.10 Either way, the same sensations are disposed at the same locations on the same fingers (portions of the sensory field). The tactile field has many rela‑ tively mobile limbs, and the phenomenon shows that we can be confused about how they are positioned relative to one another. (Some of the more twisted yoga poses lend themselves to the experiment as well.) It does not show that tactile impressions only exist on the sensory fields. Of course, such confusion affects inferences drawn from the apparent disposition of tactile sensations to that of objects, but this does not challenge the dis‑ tinction between tactile impressions that are embedded in body parts and those that are independently mobile. We quickly discover that heat and cold are like bright light. They can leave “after‑images.” These after‑images are attached to different portions of the tactile field (such as different “hands”) and move with them. They then mix with other impressions, producing differently compounded im‑ pressions at those locations. There is nothing in the case of submerging hot and cold hands in lukewarm water that challenges the distinction between tactile sensations embedded in the hands and tactile sensations experienced when approaching or retreating from hot or cold objects. The case only proves that they can sometimes mix, making it difficult to determine how much of the mixture is due to sensations of each sort. Hume might not have had such an easy time of it had he attempted to offer sceptical arguments against the evidence of touch. On this topic, the weight of his case for scepticism falls on what is considered in Section 8.4. 8.2.4 Independent existence in the Enquiry

The Enquiry’s treatment of arguments against the experience of an exter‑ nal world falls into three episodes (EHU 12.6, 12.7–14, and 12.15). The first dismisses arguments from illusion (or, as Hume put it, from the “im‑ perfection and fallaciousness of our organs, on numberless occasions”). The second and third restate the arguments of Treatise 1.4.2.45–9 and 1.4.4.6–14. In all these episodes, Hume spoke as a reporter, rather than as someone with his own conclusions to draw.11 The arguments of the first episode are called “topics,” in allusion to what are today called the “modes” of

Section 8.2.4  289 ancient Pyrrhonian scepticism. They are attributed to “the sceptics in all ages.” The argument of the third episode is attributed to Berkeley. The sec‑ ond episode is best considered as a dialogue between natural instinct (EHU 12.7–8), philosophy (EHU 12.9–10), and the sceptics (EHU 12.10–13). Hume initially identified “the various aspects of objects, according to their different distances” and “the double images which arise from the pressing one eye” as being among the “trite topics, employed by the scep‑ tics in all ages” (EHU 12.6). He paused to offer his own assessment of the value of these topics, writing that they are only sufficient to prove, that the senses alone are not implicitly to be depended on; but that we must correct their evidence by reason, and by considerations, derived from the nature of the medium, the distance of the object, and the disposition of the organ, in order to render them, within their sphere, the proper criteria of truth and falsehood. (EHU 12.6) However, the diminution of a table as we move away from it is subsequently (EHU 12.9) mentioned as being among what Enquiry 12.6 goes on to de‑ scribe as “other more profound arguments against the senses.” Hume’s point may have been that the sceptics of all ages have drawn the wrong conclusion from the phenomena of distancing and double vision (Butler 2008, 128). They took them to show that the senses are untrustworthy when in fact these phenomena show that the senses are not able to “produce any immediate intercourse between the mind and the object” (EHU 12.9). But when the distancing argument is restated, it is not attributed to the sceptics but to philosophy.12 It is also stated in an importantly different form from that found in the Treatise. As discussed in Section 8.2.1, the ar‑ gument of the Treatise does not appeal to a difference between perceptions and external objects. The Enquiry does. The dialogue of Enquiry 12.7–13 opens with the voice of natural instinct declaring that, This very table, which we see white, and which we feel hard [note the naïve realist description of the table as consisting of visual and tactile sensible qualities], is believed to exist, independent of our perception, and to be something external to our mind, which perceives it. Our pres‑ ence bestows not being on it: Our absence does not annihilate it. It preserves its existence uniform and entire, independent of the situation of intelligent beings, who perceive or contemplate it. (EHU 12.8) After this, the voice of “the slightest philosophy” speaks. It accepts that there is a real table that exists independently of us and that suffers no

290  Reasons for scepticism alteration from our movement away from it. But, it goes on to observe, the table that we see diminishes as we retreat from it. From this evidence, phi‑ losophy concludes that “It was, therefore, nothing but its image that was present to the mind,” that is, a fleeting copy or representation of a table that remains uniform and independent. The sceptics speak up only after philosophy has attacked natural instinct and offered a representational realist alternative.13 They challenge philoso‑ phy to justify its residual commitment to a uniformly and independently existing table. Philosophy proves to be unable to rise to the challenge (see Section 8.3). This is a significant change. What in the Treatise was affirmed in Hume’s own voice as a reason for concluding that perceptions have no independent existence is in the Enquiry presented as a reason why philosophers con‑ clude that our immediate experience only presents us with representations of independently existing objects.14 Conjecturally, Hume changed his approach because of his emergent dis‑ satisfaction with his account of personal identity. Lacking a good account of self or mind may have made him reluctant to argue that perceptions are dependent on “us” or “the mind.” This might have made it seem easier to argue that they are different from what independently existing objects are commonly assumed to be like. Alternatively, he may have been following through on the change from taking the belief in body to being the product of trivial operations of the imagination to taking it to be a natural instinct or prepossession that even philosophers are compelled to accept (compare Butler 2008, 134, 136–7). Be this as it may, because we only ever immediately see an “image” that is continually changing as we retreat, what we immediately see is not uniform and independent. In default of an appeal to immediate sensory ex‑ perience, those who are convinced we do perceive something uniform and independent are left having to find some other justification for their belief. Concluding in his own voice after surveying sceptical responses to that attempt (discussed in Section 8.3), Hume remarked that “This is a topic, therefore, in which the profounder and more philosophical sceptics will always triumph” (EHU 12.14). He left himself room to comment further on what we ought to make of this result. 8.3

Representationalism and phenomenalism

The arguments of Treatise 1.4.2.45 and Enquiry 12.9 are directed against the independent existence of sensations. Hume recognized that some might object that our belief in body is a belief in something distinct from sen‑ sations. Some philosophers draw a distinction between sensations and

Section 8.3  291 external objects. They further take external objects to cause sensations and take sensations to be signs of the presence of these causes. Over Treatise 1.4.2.47–9, Hume went on to offer further sceptical argu‑ ments against this representationalist or, as he described it, “double exist‑ ence” theory. He first objected that the inference to external causes of our perceptions cannot be justified. All reasoning from effects to causes is based on prior experience of regularity in the succession of species of object. But we have only ever experienced regularities in the conjunction of perceptions with other perceptions. We have never experienced external objects that regularly precede our sense impressions, and so are in no posi‑ tion to infer their existence by reasoning from effect to cause (T 1.4.2.47). This argument neglects to consider that representationalism could be jus‑ tified by argument from analogy. Price brings this up (1940, 134, 123–4), but misses the most relevant analogy: two‑dimensional, inverted pictures of the impressions constitutive of the world outside a camera obscura projected on the back wall of the camera. The front half of a dissected eye works the same way (Malebranche Search 1.12.3, 56–7). While Re‑ naissance and Early Modern philosophers had no personal experience of objects outside themselves being regularly followed by their visual impres‑ sions, they had analogous experience of visual impressions constitutive of objects outside a camera obscura or dissected eye being followed by im‑ pressions constitutive of resembling projections on the back wall of the camera or a screen set behind the eye. This experience led them to infer that the fore parts of our eyes also create images of what lies before them. Supposing that the sense of vision gathers all its information from what is projected onto the back of the eye (otherwise, what would be the point of having eyes designed to produce such projections?) leads to the theory that the sense of vision is originally sensitive to images that are something like the images projected on the retina in their dimensionality, orientation, con‑ figuration, and (why not also?) colouration. That theory is then confirmed by experience of how the colour patches we see grow, shrink, and change shape with changes in our distance from the objects that cast the images and the angle from which we view those objects. It was only in the Enquiry that Hume said anything that grapples with this argument. By what argument can it be proved, that the perceptions of the mind must be caused by external objects, entirely different from them, though resembling them (if that be possible) and could not arise either from the energy of the mind itself, or from the suggestion of some invisible and unknown spirit, or from some other cause still more unknown to us? It is acknowledged, that, in fact, many of these perceptions arise

292  Reasons for scepticism not from any thing external, as in dreams, madness, and other diseases. And nothing can be more inexplicable than the manner, in which body should so operate on mind as ever to convey an image of itself to a sub‑ stance, supposed of so different, and even contrary a nature. (EHU 12.11) Interestingly, this is the first objection Hume offered to representationalism in the Enquiry. A reiteration of the Treatise 1.4.2.47 evidential objection only comes up in the following paragraph. Enquiry 12.11 offers a textbook response to an argument from analogy: there are other ways the conclu‑ sion can be accounted for, and the analogy is weak because the projection of light onto surfaces is not plausibly anything like the processes responsi‑ ble for transmitting information from the retina to the brain or mind. In the Treatise, Hume went on to consider the possibility that an in‑ ference to the existence of external objects distinct from our perceptions might arise from some other operation of the imagination than causal in‑ ference. He confessed to feeling more at a loss to dismiss this possibility, writing that it “implies a negative, which in many cases will not admit of any positive proof” (T 1.4.2.48). In an attempt to address this difficulty, he first threw down a challenge to his readers. Let it be taken for granted, that our perceptions are broken, and inter‑ rupted, and however like, are still different from each other; and let any one upon this supposition show why the fancy, directly and im‑ mediately, proceeds to the belief of another existence, resembling these perceptions in their nature, but yet continu’d and uninterrupted, and identical[.] (T 1.4.2.48) He then offered a reason for thinking that this challenge could not be met. Mean while I cannot forbear concluding, from the very abstractedness and difficulty of the first supposition [“that our perceptions are broken, and interrupted, and … different”], that ’tis an improper subject for the fancy to work upon. Whoever wou’d explain the origin of the common opinion concerning the continu’d and distinct existence of body, must take the mind in its common situation, and must proceed upon the sup‑ position, that our perceptions are our only objects, and continue to exist even when they are not perceiv’d. (T 1.4.2.48) The opening sentence is hard to understand. The rest of the passage in‑ dicates that Hume meant to conclude that the utmost the imagination

Section 8.3  293 could produce by working on broken, interrupted, and different percep‑ tions would be the belief that our perceptions continue to exist when not perceived, not the belief that they are images or representations of other objects that exist independently of being perceived. Whatever Hume might have been getting at in the opening sentence, his best justification for this conclusion would have been an appeal to his own account of the effects of the associative principles that govern the operation of the imagination. Causal inference to independently existing objects having been ruled out (T 1.4.2.47; EHU 12.12), the imagination is left having to work with re‑ semblance and contiguity. Neither of those relations can lead us to do more than form ideas that copy our impressions. But is it impossible that resemblance and contiguity could associate bro‑ ken, interrupted, and different impressions to give us continued, uninter‑ rupted, and identical ideas? Was Hume too quick to dismiss the prospects for what might be called a phenomenalist account of the belief in body? On the account of Chapter 7, common people do not absurdly believe that perceptions that are broken, interrupted, and different in their exist‑ ence are uninterrupted. They believe that perceptions that are interrupted in their appearance to the senses are uninterrupted. There is no contradic‑ tion in this belief, and it is enlivened by inferences from customary conti‑ guity in space. Ordinary people see their impressions move off their visual fields. Having learned to draw inferences from the customary contiguity of landmarks, they vividly imagine an ambient space populated with custom‑ arily contiguous landmarks, extending beyond the bounds of their visual fields and they form vivacious ideas of where the impressions that moved away will now be in this ambient space, based on their observed trajecto‑ ries and speed of motion. Further having learned to visually perceive the third dimension, they have also come to mediately perceive impressions as passing behind occlusions, and they vividly imagine them continuing to ex‑ ist at locations behind these occlusions. Their beliefs of both sorts are con‑ stantly confirmed. When they turn to look in the direction the impressions moved, or reach out to remove the occlusion, they find impressions that bear one–one resemblance and proximity relations to the impressions that earlier moved out of view. They accordingly identify the earlier and later impressions and consider their belief in their unperceived continued exist‑ ence to have been confirmed. Because their ideas are enlivened by these alternative forms of custom, they believe in the continued existence of their very impressions. In doing so, they are not as unsophisticated as Hume made them out to be. They do not consider their impressions to simply be external objects. They consider them to be partially constitutive of such objects. Objects are collections of impressions, only some of which appear at any one distance, viewing angle, or moment in their natural history of motion and change.

294  Reasons for scepticism The subset of impressions that do appear typically arouses a paradigm idea that copies impressions received in preferred circumstances and dis‑ posed at a preferred distance. That idea is designated the object, while the remaining actual and possible impressions are considered ways the object can appear to be. (Representationalism emerges from this development.) Objects are public in the sense that differently positioned observers can get the same paradigm idea from different subsets of the collection of im‑ pressions, or get a paradigm idea of their own that they associate with the public name for that collection of impressions (Section 6.12).15 Though Hume did not countenance the phenomenalist theory of Section 6.12, he could have mounted a reply (Section 8.2.2). The story that has just been told works well provided the subject is immobile but for rotation; provided nothing that is seen moves out of the horopter (as the horopter is extended by rotation of the eyes or head), even if it moves beyond the field of view; and provided what lies behind occlusions is dis‑ covered by removing the occlusion rather than by moving to look behind it. But when what is visible moves fore or aft of the horopter, or the viewer adopts a different perspective, what is seen is annihilated and replaced with something else. This applies to anything that is seen. We can ap‑ proach to or retreat from or adopt a different viewing angle on anything that appears on the visual field. Phenomenalists reply that what disappears with approach or retreat is not annihilated. It only retreats into the third dimension where it is hidden behind our heads (if we are moving forward) or is occluded by impressions that were previously hidden behind our heads (if we are retreating). The hidden or occluded impressions continue to exist over all different dis‑ tances and directions. We encounter different collections of these continu‑ ing impressions depending on how we move (Price 1940, 107–13). But this is speculation. It is not enlivened by any associative principle of the imagination. When we approach, retreat, or change viewing angle what we see appears to bloom, shrivel, or deform as we move. With each passing moment, the earlier states are thought to no longer exist and the anticipated later states to not exist yet. The operative act of the imagina‑ tion in response to these phenomena is identification over time rather than inference from constant contiguity in space. That operation does not sup‑ port the phenomenalist speculation that the altered states continue to exist at other distances or in other directions. As we do not see the thing start to move anywhere, we originally and instinctively believe we are seeing the thing change or disappear, not move into hiding. Hume had a case for scepticism about the independent existence of our visual impressions. But he had no case for claiming that the contrary opinion is based on nothing more than trivial qualities of the imagination conducted by false suppositions. The experiments of Treatise

Section 8.4  295 1.4.2.45 offer “proof against proof”: evidence from double vision and var‑ iations in distance that counterbalances equally compelling evidence from rotary motion and the removal of occlusions. Hume had a further sceptical consideration to offer. 8.4

Scepticism about primary qualities in the Enquiry and the Treatise

Hume’s second proper argument proceeds on the “universally allowed” (EHU 12.15) premise that what he called “sensible” or “secondary” quali‑ ties, “exist not in the objects themselves but are perceptions of the mind, without any external archetype or model, which they represent.”16 It then charges that “If this be allowed, with regard to secondary qualities, it must also follow, with regard to the supposed primary qualities of extension and solidity” (EHU 12.15, rephrasing T 1.4.4.6).17 The Enquiry justifies this charge in a sentence. The idea of extension is entirely acquired from the senses of sight and feeling; and if all the qualities, perceived by the senses, be in the mind, not in the object, the same conclusion must reach the idea of extension[.] (EHU 12.15) Hume went on to worry that this conclusion might be evaded by maintain‑ ing that we can form abstract ideas of extension (EHU 12.15), but his ensu‑ ing discussion is almost equally abrupt. It rests on taking it to be intuitively obvious that we cannot conceive colourless or intangible extension or trian‑ gles with angles of no determinate size and sides of no determinate length. He would have done better to observe that since abstract extension is un‑ bounded and homogeneous, an abstract idea of extension would not be an idea of anything in particular. To have the idea of a particular extension is to have the idea of a particular shape of a particular size. The idea of a sized shape presupposes the idea of a boundary marking the extent of the shape. But all such ideas are ideas of contrast between locations occupied by some quality and adjacent locations not occupied by that quality. The challenge is identifying even one visible or tactile quality that is not universally al‑ lowed to exist only in the mind. (The quality would have to be visible or tactile because all ideas are copied from impressions and vision and touch are the only senses that give us impressions that are disposed in space.) Gravitational mass, inertial mass, magnetism, and modes of cohesion like hardness are defined by how shapes move and so presuppose a prior iden‑ tification of shapes. Understanding solidity as resistance to compression, it likewise presupposes shape. Apparently, only the sensible or “secondary” qualities of colour, heat, cold, pressure (the degree to which we feel pressed

296  Reasons for scepticism upon), resistance (understood as the degree to which we feel it difficult to move a shape), and pain are left. The only shapes defined by these qualities are shapes that exist on the sensory fields or a surrounding image field. It is not a stretch to attribute this line of thinking to the mature Hume because the author of the Treatise presented a variant on it (1.4.4.7–14). It takes a different start by appealing to his rejection of infinite divisibility (T 1.2.1–2) to conclude that the division of an extended thing would have to terminate at parts that have no extension. In line with the argument from the non‑entity of unqualified points (Section 2.6), these parts would have to be either coloured or solid.18 Colour being merely “secondary” and so having no “real existence” any “real” extension would have to consist of solid parts (T 1.4.4.8). Solidity is inconceivable apart from consider‑ ing two bodies being driven towards one another and resisting penetra‑ tion upon collision. One body penetrates another when it comes to occupy some portion of the extension taken up by the other. But extension needs to be defined by appeal to solidity, whereas here solidity is being defined by appeal to extension (T 1.4.4.9). The conclusion appears to have been established, but Hume went on to worry that “we naturally imagine, that we feel the solidity of bodies, and need but touch any object in order to perceive this quality” (T 1.4.4.12). His reasons for dismissing this worry are not as direct as they might have been. He began by arguing that any feeling we get from touch would have to be different from solidity, instancing someone who has no sensation in one hand, but perceives the solidity of a table just as well by seeing that hand rest on the table as by feeling the table with the other hand. (This observation follows T 1.4.4.9 in taking solidity to be evident from im‑ pact that does not result in penetration, as when one stone is held up by another.) From this, he concluded that no tactile feeling could resemble solidity (T 1.4.4.13). But this does not follow. The fact that we can see solidity without touching things implies that solidity can exist apart from any one tactile sensation, but it does not imply that multiple tactile sensa‑ tions do not instance solidity (in say, the way they are disposed relative to one another over time). Hume went on to appeal to a variant on this case to establish that no tactile feeling could represent solidity (T 1.4.4.14). When two stones rest on one another, each must resist allowing the other to enter the space it occupies, and so there must be solidity. When a hand rests on a stone there must likewise be solidity, but in this case, there is also something else, a tactile feeling. Some part of that feeling must have nothing to do with so‑ lidity, since solidity can be present without the feeling being present. But since the feeling is a simple impression, it is not divisible into parts. It can therefore contain nothing that has anything to do with solidity. This argu‑ ment rests on the same non sequitur. The fact that the feeling is extraneous

Section 8.4  297 to the case of solidity does not imply that multiple instances of the feeling could not be disposed relative to one another in such a way as to instance solidity. Hume came closer to making a pertinent observation when adding that because solidity is complex (it is made evident by the impact of contiguous bodies along with impenetrability) it cannot be represented by a simple impression (T 1.4.4.14). But might it be represented by a collection of simple impressions? The visual impressions constitutive of a hand resting on a table or two stones resting on one another are multiple and disposed in the configuration of higher and lower coloured shapes, each remaining where it is despite our vividly imagining that the one is pressing on the other. Hume admitted that these compound visual (and associated tactile) impressions constitute a complex impression of solidity. The same is pos‑ sible for aggregates of purely tactile sensations. Hume came closest to making his case in a passing remark: “except when [simple impressions of touch are] consider’d with regard to their extension [that is, the extension they constitute in aggregate]; which makes nothing to the present purpose” (T 1.4.4.14). Granting that solidity is evi‑ denced by contiguous bodies that are impelled towards one another with‑ out penetration, it can only be evidenced by multiple tactile (or visual) perceptions disposed in the configuration of shapes that do not overflow their common boundary upon contact. That makes shape foundational for solidity, ruling out the possibility that solidity could be foundational for shape. We are left confronting the fact that shape presupposes edges and edges presuppose qualitative contrast. Where touch is concerned, the only known qualities (pain, pressure, pleasure, heat, cold, muscle strain) are supposed to exist only in the mind, that is, on the visual or tactile sensory fields or a surrounding image field. This is a striking conclusion, and it strikes in more directions than one. Though it is presented with sceptical intent, to show that we have no ex‑ perience of things outside the mind, it just as well establishes that many of the things universally allowed to only exist in the mind are extended and solid. Multiple visual and tactile points are disposed in space to consti‑ tute extended quality patches and these patches occupy sensory fields over which they are observed to move towards one another and collide without penetrating. The mind must therefore contain extended, solid parts. Hume was aware of this conclusion and drew it at Treatise 1.4.5.16 and 33. There he charged that we have as much reason for being sceptical of the existence of unextended spiritual substances or minds as of the experience of bodies external to the sensory, image, and temporal fields. The one thing that cannot be doubted is the existence of coloured and tangible points that are spatially disposed over visual, tactile, and image fields in such a way as to satisfy the definitions of extension and solidity.

298  Reasons for scepticism The conclusion is drawn at Enquiry 12.16: if we accept that sensible qualities like white and hard are merely secondary (or, as is said today, that they are qualia) then we have reason to deny that anything at all like our perceptions exists outside the mind, leaving “only a certain unknown, inex‑ plicable something, as the cause of our perceptions; a notion so imperfect, that no sceptic will think it worthwhile to contend against it” (EHU 12.16). While the Enquiry does not present the argument to this conclusion in Hume’s own voice, in a footnote closure, attributing the argument to Berkeley, it is said to “admit of no answer.” 8.5

The ambiguous consequences of Hume’s sceptical arguments

At one point in his classic textbook on presocratic philosophy (1966, 183), Philip Wheelwright crafted a dialogue from two of Democritus’s fragments. PHILOSOPHY: By convention, sweet; by convention, bitter; by con‑ vention, hot; by convention, cold; by convention, colour; but in reality, atoms and void. THE SENSES: Wretched mind! Do you take your evidence from us and then seek to overthrow us? Our overthrow is your downfall.19 It is unfortunate that Hume never thought to turn Enquiry 12.7–11 into a more explicit dialogue between natural instinct, philosophy, and scepticism. Had he done so, he might have thought of a reply nature could have made to philosophy and the sceptics. It might have gone something like this: NATURE: The very images, presented by the senses, [are] the external objects. (EHU 12.8) PHILOSOPHY: [T]he existences, which we consider, when we say, this house and that tree, are nothing but perceptions in the mind, and fleet‑ ing copies or representations of other existences, which remain uniform and independent. (EHU 12.9) THE SCEPTICS: By what argument can it be proved, that the percep‑ tions of the mind must be caused by external objects? (EHU 12.11) NATURE: The two of you may argue with one another as you please. After all your talk is over, you will both be left unable to deny that

Notes  299 the very images presented by vision and touch consist of coloured and tangible points disposed in space. Some of them persist in their rela‑ tive dispositions and resist compression despite being impelled towards one another. Your visual and tactile impressions therefore constitute extended, shaped, and solid compounds. According to you, these com‑ pounds exist only in what you are pleased to call “the mind.” Very well then, that mind is an extended, shaped, and solid body. Granting, as you must, that such a body exists, why should there be no others? One way to understand body is as a collection of shaped, extended parts that prove to be solid or resistant to compression. On this understand‑ ing, Hume never raised any doubts about the existence of body. He only pointed to considerations that cause us to believe that the sensible quali‑ ties constitutive of bodies are projections on finitely divisible, bounded, sensory and image fields. They no more exist beyond those fields than the objects projected on a cinema screen exist beyond the bounds of the screen. If only for a little while, we are led to wonder whether we (or our sensory fields) are the only bodies in existence. Notes 1 EHU 12.8, compare T 1.4.2.14, 31, 36, 38, 43, 44, 50, which replace refer‑ ences to “images, presented by the senses” with “perceptions.” 2 Many commentators have avoided engaging these arguments, preferring to consider how to reconcile Hume’s sceptical conclusions with his other com‑ mitments, or how to understand his scepticism in light of his endorsement of the arguments. Those who have done important work on the arguments in‑ clude, Price (1940), Bennett (1971), Schachter (1978), Pavković (1982), Wright (1983), Pavković (1985), Martin (2000), Bennett (2001), Loeb (2002), Somer‑ ville (2006), Butler (2008, 2009, and 2010), and Ainslie (2015). Many of them treat the arguments as perceptual relativity arguments rather than (as here) dependency arguments. 3 I am grateful to Dario Perinetti for sensitizing me to this point. 4 This move evades objections to Hume’s proper sceptical arguments classically raised by Price (1940, 114–15). 5 Less anachronistically, the point would be made by appeal to a painted canvas, dreams, or a Berkeleian God. In the case of touch, the question is whether tac‑ tile sensations come in pairs, one embedded in the skin at the point of contact, the other on its surface and liable to move away from it. 6 Hume wrote of half of the objects being “remov’d from their common and natural position,” suggesting that the other half remains in the same place as previously. But there are variants on the case that do not discriminate between the copies. Holding up two fingers, one behind the other, we find that when we focus on the closer one, the more remote one doubles, along with all the other objects in its focal plane, whereas when we focus on the remote finger, the proximate one doubles, along with all the other objects in its focal plane. This phenomenon was discussed by Molyneux (Dioptrica Nova, 287–8 with

300  Reasons for scepticism illustrations on the earlier Table 1 inserted between pages 286 and 287) and other early modern visual theorists. 7 It might be objected that this would have been a phenomenon known only to early modern visual artists and visual theorists. Common people have learned to see in three dimensions, and experience objective size constancy as they ap‑ proach and retreat from things. But when the entire visual field is filled with resembling, increasingly distant objects, the expansion out of the vanishing point can be readily experienced. A walk down a straight, level, forest path or the centre aisle of any eighteenth‑century European country church is not that different from the contemporary cinematic representation of the view from the bridge of a starship racing at “warp speed” towards the centre point of a nar‑ rowing cone of star‑streaks. 8 The objection is classically raised by Price (1940, 107). Variants can be found in, among others, Bennett (1971, 347; 2001, 305) and Loeb (2002, 210–11). Winkler offers a brief and effective riposte in a footnote: “Fair enough, but [the critics] also [say] that Hume’s empirical data ‘have no force’ against the belief that some perceptions exist when nobody perceives them. The facts certainly count against the belief, and if, as Hume thinks, nothing counts in favor of it, the facts against would seem to be decisive” (2015, 163n24). 9 Descartes AT VII, 77; Principles 4.196; Malebranche, Search 1.10.3, 50; Descartes, AT 6, 142/Olscamp, 108; Locke, Essay 2.8.21. Descartes and Male‑ branche took the phantom pain phenomenon to prove that tactile sensations of all sorts exist only in the mind, which follows from the evidence, provided that the mind is taken to extend over the space occupied by the tactile sensory field. Thanks to Dante Dauksz for prompting me to address this issue. 10 The phenomenon is best experienced by crossing the 3rd finger as far as possi‑ ble over the 4th and moving a braided wire or straightened, notched paper clip back and forth at the top crossing point. It is essential that the eyes be closed while performing the experiment and it helps if someone else moves the wire. 11 At Enquiry 12.5, Hume characterized himself as reporting on a “species of scepticism, consequent to science and enquiry, where men are supposed to have discovered, either the absolute fallaciousness of their mental faculties, or their unfitness to reach any fixed determination in all those curious subjects of speculation, about which they are commonly employed.” The use of the phrase “where men are supposed to have discovered” sets him at a distance from the considerations he went on to recount. The distancing is furthered by the remainder of the paragraph. “[O]ur very senses are brought into dispute, by a certain species of philosophers.… As these paradoxical tenets (if they may be called tenets) are to be met with in some philosophers, and the refutation of them in several, they naturally excite our curiosity, and make us enquire into the arguments, on which they may be founded.” 12 This is also true of the Treatise. Though the arguments of Treatise 1.4.2.45 are commonly regarded as sceptical, a practice followed in Sections 8.1–8.2.3, they are designated as arguments offered by “reflection and philosophy” (T 1.4.2.44) against what we naively imagine to be the case. These arguments only turn out to be sceptical when the fall‑back position that we perceive rep‑ resentations of external objects is challenged. 13 Martin (2000, 218–21) astutely characterizes the philosophical voice of the Enquiry as both attacking naïve realism and drawing on the unacknowledged intuition that there must be some entity that is what we perceive, which he calls “actualism.” Representationalism is the consequence, though sceptics charge that it entails that we are divorced from reality.

Bibliography  301 14 Butler (2008) offers a detailed and insightful study of this change but (i) denies that there is a significant difference between taking the belief in body to be a natural instinct (EHU) and taking it to be produced by trivial qualities of the imagination (T), (ii) takes Hume to have offered a perceptual relativity argu‑ ment in the Treatise as well as the Enquiry, and (iii) resists recognizing Enquiry 12.7–13 as a dialogue or taking Hume to have distanced himself from the voices of the speakers. She does, however, defend a kind of dialectical reading on which natural instinct informs what philosophy says and a philosophical position develops in reaction to sceptical challenges. See especially 141–4. 15 It is sometimes objected that ordinary people are not so sophisticated (Pears, 1990, 153). But what has just been said is a description of how common people think, not a description of what they think they are thinking. It is quite possible for people to perform very sophisticated operations (such as writing idiomati‑ cally and stylistically in their natural language) without being able to offer an on‑the‑spot account of what they are doing (without being able to codify the grammatical and stylistic rules they are in fact following). We are all ordinary people for most of our lives, so we know something about what at least some ordinary people believe and why, and that qualifies us, as ordinary people to present a theory, like the one given above, that uncovers the unconscious rules governing how we do what we do. 16 EHU 12.15. Treatise 1.4.4 does not use category terms but only offers lists. The initial list at Treatise 1.4.4.3 is “colours, sounds, tastes, smells, heat and cold.” The Enquiry lists opposed positive and privative pairs: “hard, soft, hot, cold, white, black” (EHU 12.15), and it mentions only visible and tactile qualities. Setting aside these differences, the following treatment assimilates what Hume had to say over Enquiry 12.15 and Treatise 1.4.4.6–14. The discussion in each work benefits by supplementation with what is written in the other. 17 The Enquiry wisely declines to justify the antecedent, presumably on the as‑ sumption that the ensuing argument is intended only for those who would accept it (who is anyone who accepts that sensible qualities are not features of extra‑mental reality as described by the physical sciences). In contrast, the Treatise suggests that there is a satisfactory argument for the antecedent (though only one). Loeb (2002, 220–1) charges that this was done for an ul‑ terior motive. This is eminently plausible given that the dialectical context left Hume with no need to offer any argument. The antecedent was (and is) widely endorsed and what Hume went on to say is only for the many who think the antecedent is true. Butler (2009) offers an opposed account of Hume’s motives. 18 Hume’s claim that “These simple and indivisible parts, not being ideas of ex‑ tension, must be non‑entities, unless conceiv’d as colour’d or solid” (T 1.4.4.8) reiterates the argument from the non‑entity of unqualified points. But that claim plays no role in the ensuing argument. It suffices that unqualified points could not appear otherwise than by way of contrast with differently positioned coloured or solid points. 19 Diels‑Kranz fragments 68B9 and 125 as translated in McKirahan (1994, 335) and Barnes (1982, 563). The speaker labels are my own.

Bibliography Ainslie, Donald C. 2015. Hume’s True Scepticism. Oxford: Oxford University Press. Barnes, Jonathan. 1982. The Presocratic Philosophers. London: Routledge & Kegan Paul.

302  Reasons for scepticism Bennett, Jonathan. 1971. Locke, Berkeley, Hume. Oxford: Clarendon Press. Bennett, Jonathan. 2001. Learning from Six Philosophers. Oxford: Clarendon Press. Butler, Annemarie. 2008. “Natural Instinct, Perceptual Relativity, and Belief in the External World in Hume’s Enquiry.” Hume Studies 34: 115–58. Butler, Annemarie. 2009. “Hume’s Causal Reconstruction of the Perceptual Rela‑ tivity Argument in Treatise 1.4.4.” Dialogue 48: 77–101. Butler, Annemarie. 2010. “Vulgar Habits and Hume’s Double Vision Argument.” Journal of Scottish Philosophy 8: 169–87. Loeb, Louis E. 2002. Stability and Justification in Hume’s Treatise. Oxford: ­Oxford University Press. Martin, M. G. F. 2000. “Beyond Dispute: Sense‑Data, Intentionality, and the Mind‑Body Problem.” In History of the Mind‑Body Problem, edited by Tim Crane and Sarah Patterson, 195–231. London: Routledge. McKirahan, Richard D. Jr.  1994. Philosophy before Socrates. Indianapolis: Hackett. Pavković, Aleksandar. 1982. “Hume’s Argument for the Dependent Existence of Perceptions: An Alternative Reading.” Mind 91: 585–92. Pavković, Aleksandar. 1985. “Hume’s Arguments from the Relativity of Sense‑­ Perception.” International Philosophical Quarterly 25: 261–70. Pears, David. 1990. Hume’s System. Oxford: Oxford University Press. Price, H. H. 1940. Hume’s Theory of the External World. Oxford: Clarendon Press. Schachter, Jean‑Pierre. 1978. “Hume’s Argument against the Continuing Existence of Unperceived Perceptions.” Mind 87: 436–42. Somerville, James. 2006. “‘The Table, Which We see’: An Irresolvable Ambiguity.” Philosophy 81: 33–63. Van Cleve, James. 2008. “Reid on Single and Double Vision: Mechanics and ­Morals.” Journal of Scottish Philosophy 6: 1–20. Wheelwright, Philip. 1966. The Presocratics. New York: Macmillan. Winkler, Kenneth. 2015. “Hume on Scepticism and the Senses.” In The Cambridge Companion to Hume’s Treatise, edited by Donald C. Ainslie and Annemarie Butler, 135–64. New York: Cambridge University Press. Wright, John P. 1983. The Sceptical Realism of David Hume. Minneapolis: Univer‑ sity of Minnesota Press.

Conclusion Hume’s remedy

For Hume, scepticism does not pose a problem. It enables us to proportion belief to the evidence. Both philosophers and ordinary people, especially the latter, are better off for being impressed by the force of sceptical argu‑ ments. Hume came to realize that scepticism does not need to be supported by showing that some of our most important beliefs are based on trivial qualities of the imagination conducted by false presuppositions. It suffices that the causes and reasons for these beliefs be opposed by arguments of equal force and authority. Had Hume offered a more robust account of our experience of publicly observable objects, which he had the resources to do, his proper sceptical arguments would still stand as “proof against proof.” We get over some of the doubt they lead us into, but we remain less credulous for having had that experience. C.1

Reactions to external world scepticism in the Treatise

According to the Treatise, as ordinary people we take some of our visual and tactile impressions to be external objects. But experiments involving double vision and distancing teach us that all our perceptions are only states of our sensory fields. Immediately after uncovering this result, the author of the Treatise wrote that “The natural consequence of this rea‑ soning shou’d be, that our perceptions have no more a continu’d than an independent existence” (T 1.4.2.46). But this consequence has generally not been accepted. [A]s a little reflection … [shows] that [our perceptions] have a depend‑ ent [existence], ’twould naturally be expected, that we must altogether reject the opinion, that there is such a thing in nature as a continu’d ex‑ istence, which is preserv’d even when it no longer appears to the senses. The case, however, is otherwise. Philosophers are so far from rejecting the opinion of a continu’d existence upon rejecting that of the inde‑ pendence and continuance of our sensible perceptions, that tho’ all sects DOI: 10.4324/9781032677880-10

304  Conclusion agree in the latter sentiment, the former, which is, in a manner, its neces‑ sary consequence, has been peculiar to a few extravagant sceptics; who after all maintain’d that opinion in words only, and were never able to bring themselves sincerely to believe it. (T 1.4.2.50) This reaction is not surprising because belief is not up to us. It is not, in particular, a product of our reasoning. It is the product of mechanisms gov‑ erning the transfer of vivacity, which are activated by our circumstances and our history. As long as our attention is bent upon the subject, the philosophical and study’d principle may prevail; but the moment we relax our thoughts, nature will display herself, and draw us back to our former opinion. Nay she has sometimes such an influence, that she can stop our progress, even in the midst of our most profound reflections, and keep us from running on with all the consequences of any philosophical opinion.… [The “notion of an independent and continu’d existence”] has taken such deep root in the imagination, that ’tis impossible ever to eradicate it, nor will any strain’d metaphysical conviction of the dependence of our perceptions be sufficient for that purpose. (T 1.4.2.51) The author of the Treatise went on to allow that “tho’ our natural and obvious principles here prevail …, ’tis certain there must be some struggle and opposition in the case.” But that is only because our philosophical reflections have somehow been able to acquire and “retain [some] force or vivacity” (T 1.4.2.52). Everything hangs on what is supposed to happen in response to this conflict. For the author of the Treatise, philosophers seek to accommodate their opposed natural and rational impulses by proposing that our percep‑ tions are effects of enduring objects, and that they resemble their causes. After some further ruminations on the psychological factors leading them to propose this resolution, on the advantages of their proposal, and on the impossibility of giving any sound reason for it (T 1.4.2.53–5), the author of the Treatise gave vent to a temporary expression of frustration and dis‑ may (T 1.4.2.56–7). But then he went on to declare that he would ignore the problem and proceed to examine some ancient and modern systems of the external and internal world, taking for granted that both worlds exist. Carelessness and in‑attention alone can afford us any remedy. For this reason I rely entirely upon them; and take it for granted, whatever may be the reader’s opinion at this present moment, that an hour hence he

Section C.1  305 will be perswaded there is both an external and internal world; and going upon that supposition, I intend to examine some general systems both antient and modern, which have been propos’d of both[.] (T 1.4.2.57) This “in‑attention” is not entirely “careless.” Turning to consider what ancient and modern philosophers have had to say about body may dis‑ cover a solution to the conflict between natural principles and philosophi‑ cal reflection. But this proves not to be the case. Ancient systems of the external world are found to be absurd (T 1.4.3) and by denying that sensible qualities could resemble any qualities of external objects the modern system de‑ nies the possibility of any knowledge of external and independent objects (T 1.4.4, Section 8.4). The examination of ancient and modern opinions fails to uncover a resolution. The hopeful reader has had their hopes dashed a second time. The long investigation of the causes of the belief in body has not been pointless. It has taught us that we are necessitated to believe in the external and independent existence of something that does not in fact exist indepen‑ dently and externally (at least by the evidence of some experiments). Being so necessitated, we will go on doing so, regardless of the discovery we have just made. We may not like the conclusion that our perceptions do not in fact exist independently or externally, but it follows from the evidence (at least, from some of it). In the opinion of the author of the Treatise, being confronted with that conclusion and forced to accept it, however temporarily, is good for us. In all the incidents of life we ought still to preserve our scepticism. If we believe, that fire warms, or water refreshes, ’tis only because it costs us too much pains to think otherwise. Nay if we are philosophers, it ought only to be upon sceptical principles, and from an inclination, which we feel to the employing ourselves after that manner. (T 1.4.7.11) The author of the Treatise did not go into any more detail about why we ought to preserve our scepticism and philosophize only on sceptical principles. Conjecturally, it is because it does not cost those who have been impressed by sceptical arguments too much pain to think otherwise, except when it comes to things like “fire warms,” and “water refreshes.” Those who have not been impressed by sceptical arguments, in contrast, will be convinced of many things even though it costs them very little pain to think otherwise. Many of those convictions will be disproportioned to the evidence.

306  Conclusion As the studies of Treatise 1.3 have shown, belief is not up to us. It is ­ etermined by mechanisms responsible for the transmission of vivacity. Our d causal inferences are not just influenced by what we have found to custom‑ arily be the case. They are influenced by the vivacity transmitted by contigu‑ ity and resemblance and by aroused passions. They are further influenced by the degradation of vivacity over time and distance and long chains of argument. Our opinions of all kinds are strongly influenced by society and sympathy. The force of education equals and often surpasses that of causal inference. We are naturally credulous of the words of others. We cannot resist these influences unless something happens to cause us to resist them. Having thus given an account of all the systems both popular and philo‑ sophical, with regard to external existences, I cannot forbear giving vent to a certain sentiment, which arises upon reviewing those systems. I be‑ gun this subject with premising, that we ought to have an implicit faith in our senses, and that this wou’d be the conclusion, I shou’d draw from the whole of my reasoning. But to be ingenuous, I feel myself at present of a quite contrary sentiment, and am more inclin’d to repose no faith at all in my senses, or rather imagination[.] (T 1.4.2.56) Appealing to his own case as an example, Hume noted that those who have been impressed by the force of sceptical arguments find themselves determined to doubt the most basic things. That level of doubt becomes too painful to sustain when confronted by everyday conjunctions between contiguous bodies, maximally proximate and resembling earlier and later perceptions, and causes and effects. But we come away only believing what it costs us too much pain to disbelieve. Ideas enlivened by less constant sources of vivacity prove to be possible to doubt. We end up proportion‑ ing our belief to the evidence. This makes being impressed by the force of sceptical arguments an important preparative for philosophical investiga‑ tions. It is even more important for ordinary people, who are more liable to superstition and enthusiasm. Scepticism is not something to be avoided or answered, but something to be embraced as a remedy for belief that is disproportioned to the evi‑ dence. Philosophers should not seek to overcome it and the vulgar should be exposed to it.1 C.2

Reactions to external world scepticism in the Enquiry

In the Enquiry, Hume maintained that “a natural instinct or preposses‑ sion” causes us to suppose that the images presented by our senses are external objects. The distancing argument (Section 8.2.2) teaches us that

Section C.2  307 these images are only fleeting copies or representations of supposedly uniformly and independently existing things. But “the profounder and more philosophical sceptics will always triumph” (EHU 12.14) when they challenge philosophers to prove that the perceptions of the mind must be caused by, much less resemble, any such things. The Enquiry’s assessments of the “more profound arguments” (EHU 12.6) are not uniformly positive. Despite their strength, they are said to “little serve to any serious purpose” (EHU 12.15) and to “produce no conviction” but only “momentary amazement and irresolution and confu‑ sion” (EHU 12.15n). Enquiry 12.23 offers a long harangue against “exces‑ sive” or “Pyrrhonian” scepticism, repeating the charges of uselessness and unbelievability, adding the further charge of unsustainability, and conclud‑ ing that the only point of these arguments is to show the whimsical condition of mankind, who must act and reason and believe; though they are not able, by their most diligent enquiry, to satisfy themselves concerning the foundation of these operations, or to remove the objections, which may be raised against them. We are determined to accept the conclusions of operations we cannot justify. This is the hanging point flagged in Section C.1. “Excessive” sceptical arguments stand on one side, natural instinct on the other, and everything hangs on how that conflict is resolved. In the Treatise, this is when phi‑ losophy appears to propose a representationalist solution, initiating round after round of debate and irresolution. But in the Enquiry, philosophy has already been silenced by the sceptics. In the Enquiry, Hume inserted a part break, and went on in a new vein to speak of the lasting value of a “miti‑ gated” scepticism that arises when “excessive” doubts are in some measure corrected by “common sense and reflection” (EHU 12.24). Enquiry 12.24–5 allows that sceptical arguments of all sorts do more than induce momentary amazement and irresolution and confusion. They make “dogmatical reasoners … sensible of the strange infirmities of human understanding, even in its most perfect state, and when most accurate and cautious in its determinations,” thereby inspiring modesty, reserve, polite‑ ness, and toleration. They also mitigate the influence of what the Treatise called “unphilosophical” causes of belief, inspiring a “salutary determina‑ tion” to proportion belief to the evidence, that is to “such subjects as fall under daily practice and experience.” The Dialogues accounts for how this happens in more detail. [I]f a man has accustomed himself to sceptical considerations on the uncertainty and narrow limits of reason, he will not entirely forget them when he turns his reflection on other subjects; but in all his philosophical

308  Conclusion principles and reasoning, I dare not say, in his common conduct, he will be found different from those, who either never formed any opinions in the case, or have entertained sentiments more favourable to human reason. (D 1.8) [W]e could never retain any conviction or assurance, on any subject, were not the sceptical reasonings so refined and subtle, that they are not able to counterpoise the more solid and more natural arguments, derived from the senses and experience. But it is evident, whenever our arguments lose this advantage, and run wide of common life, that the most refined scepticism comes to be on a footing with them, and is able to oppose and counterbalance them. (D 1.11) These statements were made by Philo and Hume designed them to leave Cleanthes an opening to go on a harangue about the inconsistency of pur‑ suing advanced sciences while maintaining that only reasoning confined to matters of common life can escape scepticism. But Philo’s underlying point recalls the belief theory of the Treatise. Beliefs that are empirically guided are constantly reinforced by renewed encounters with vivacious impressions and reinforcement of customary associations. Reflection on the implications of double vision and distancing requires performing rare and special experiments or overcoming the acquired perception of objec‑ tive size constancy. It also requires some reasoning concerning what the experiments reveal (Sections 8.2.1 and 8.2.2). But while these weakening factors mean that excessive scepticism cannot be sustained in the face of the ongoing evidence of sensory experience, having been impressed with the excessive arguments induces a lasting distrust of our cognitive powers that enables us to resist the influence of more occasional sources of vivac‑ ity and return more quickly to our reticence should the more occasional sources be powerful enough to induce some belief. This “mitigated” scepticism saves us from superstition and enthusiasm. It keeps us moderate, polite, and tolerant. It enables us to resist the many factors leading our ideas to be enlivened out of proportion to the evidence, making it possible for us to be wise, logical, and rational. It offers a rather different remedy than the carelessness and inattention at one point pro‑ posed by the author of the Treatise. This remedy is not a remedy for a philosophical conflict but a remedy for the problem of unphilosophical belief and for superstition and enthusiasm. Considered together the Enquiry and Philo’s voice in Dialogues 1 make it clear why Hume would have thought that we ought to preserve our scep‑ ticism in all the incidents of life, and be philosophers only upon sceptical

Section C.2  309 principles (T 1.4.7.11). But might reading that position back into the views of the author of the Treatise be going too far? Might Hume’s positive view of scepticism only have been a later development? Alternatively, might Hume have abandoned the psychology of belief he had worked out in the Treatise and come to see his remedy as a remedy just for vulgar supersti‑ tion and enthusiasm? The latter possibility can be quickly dismissed. The mature Hume never abandoned the philosophical psychology of the Treatise. Morals is from start to finish a work on human moral psychology, not on the foundations of ethics (Falkenstein 2021). The Natural History is a study of religious psychology. The Essays are taken up with political, economic, religious, and moral psychology. Hume continued to be fascinated with the psycho‑ logical causes of belief and with efforts on the part of Queens, Kings, and sectaries to control social belief on a large scale. Notable instances from the History include the imagined dialogue between Pole and Gardiner over reasons for and against religious toleration (HE 37, 431–44); passages from the first edition of the History on the enthusiasm of the Protestants (1754, 7–9), the superstition of the Catholics (1754, 25–7), and the char‑ acter of the reformation in Scotland (1754, 59–61) that Hume had felt forced to suppress;2 a further discussion (more extensive in the first edi‑ tion than later ones) of James I’s attempts to moderate the enthusiasm of the Presbyterians (1754, 61–4; HE 47, 66–9); speculations about how the Puritans thought they could win adherents to their cause (HE 54, 284–5); and numerous passing reflections on such topics as why Edward I chose to expel the Jews from England (HE 13, 76), why the English army lost heart at the siege of Orleans (HE 20, 401), and why the Duke of Burgundy de‑ cided to betray Joan of Arc to the English (HE 20, 411). All these passages (and they are but a few of dozens) draw on the psychology of belief and of the passions Hume had originally worked out in the Treatise, as does the Enquiry’s essay on miracles. On the other hand, it is impossible to know for sure whether Hume’s pos‑ itive view of scepticism was only a later development. What can be said for sure is that the Treatise goes into more detail about its causes, and the En‑ quiry into more detail about its remedial effects. The Treatise comes closer than the Enquiry to recognizing that there is a problem accounting for how it is possible for a wise person to “[proportion] … belief to the evidence” (EHU 10.4) given that all belief is “the necessary result of placing the mind in such circumstances.… [and] a species of natural instincts, which no reason‑ ing or process of the thought and understanding is able, either to produce, or to prevent” (EHU 5.8). At the same time, the Treatise is more desperate to discover and defend sceptical results, even at the cost of implausibility (T 1.4.1) and neglect of readily available resources (T 1.4.2.2–43). But it does not make the connection between the problem of unphilosophical

310  Conclusion belief and the sceptical remedy as clear as the Enquiry does. The author of the Treatise took an autobiographical route, recounting how his own encounter with sceptical arguments led him first to despair and then to mitigated belief. He refrained from drawing a moral from his own case. This having been said, the Enquiry continues to be cagy about what it aims to do. Enquiry 1 does not propose to promote a sceptical crisis as a remedy for superstition. Enquiry 12 introduces a half dozen different kinds of scepticism, allowing Hume to denounce many of them and present him‑ self as being sceptical only in a qualified and “moderate” sense. While the Enquiry is ultimately forthright about recommending scepticism as a remedy for impulsive and obstinate belief, it does so only after the reader has been confronted with sceptical arguments in their full force, and even then, only by way of excuse. In this, it follows the Treatise, which only rec‑ ommends scepticism in 1.4.7.11, after the sceptical arguments have been presented. In both the Treatise and the Enquiry, Hume appears to have been appre‑ hensive about declaring his position on the remedial value of scepticism, as if he felt he needed to conceal the surgical instruments from the eyes of the patient. As a young man, concerned to win an approving audience for his work, he will have been even more reticent about his sceptical intentions. This makes it hard to be confident that his remedial project was only a later development. C.3

The fate of Hume’s remedial project

Granting that Hume wanted to propose scepticism as a remedy for “un‑ philosophical” belief, it is worth asking whether his thought could support that proposal. Earlier chapters have argued that he could have provided a more robust account of space, of the endurance of unchanging objects, and of identity than he did (Chapters 2–4). Following on that, he could have provided a more robust account of empirically guided belief than he did, one that recognizes inference from customary contiguity in space and customary one–one resemblance and proximity over time (Chapter 6). Drawing on these resources, he could have accounted for the perception of publicly observable objects (Chapter 6). He could also have presented the belief in body as grounded on operations of the imagination that are as “permanent, irresistible, and universal” (T 1.4.4.1) as causal inference (Chapter 7). These unrecognized and undeveloped features of his thought scuttle any attempt to “destabilize” the belief in body by arguing that it is caused by “trivial qualities of the fancy, conducted by … false presuppositions” (T 1.4.2.56).3 But the mature Hume had already abandoned that project, preferring to rest his case on the diminution argument (EHU 12.9) and

Notes  311 the argument from the dependence of the primary qualities on sensible ­qualities (EHU 12.15), supplemented by arguments against demonstrative and moral reasoning (Enquiry 17–22). Unfortunately, he never took the opportunity to revisit his views on space, endurance, identity, empirically guided belief, or objective experience. In failing to do so, he exposed him‑ self to the chance that his remedial project would be undone by a positive account of the causes of the belief in body grounded on his own principles. As it turns out, he could still have executed his project. The diminution argument lends itself to being presented as a pure dependency argument, establishing that we have no evidence that the images presented by the senses exist beyond the surface of the sensory fields (Section 8.2.2). As so understood, the diminution argument is contested by inference from customary contiguity, which leads us to vividly imagine images presented by the senses to continue to exist beyond the bounds of the sensory fields and behind occlusions, and this counterevidence is constantly renewed. But the distancing argument does not need to be uncontested. All that Hume required is that it suffice to induce a degree of amazement and ir‑ resolution and confusion sufficient to induce a lasting distrust of the effi‑ cacy of our cognitive powers. The best possible outcome is that the distrust should be inefficacious against empirically guided inferences and otherwise efficacious. The conclusion of the distancing argument is further supported by the fact that the things we experience to be disposed in space, colours and tac‑ tile sensations of heat, cold, and pressure, are things that we no longer be‑ lieve to be qualities of insentient bodies (indeed, not even of sentient ones). As Hume might have put it, there are two principles that we cannot either renounce or render consistent with a scientific worldview. One is that what we today call qualia exist. The other is that visual and tactile qualia are disposed in space, shaped, and solid and so not confined to some mental or intentional realm. Anyone who seeks to deny either principle is denying what is more obvious than anything else. Anyone who accepts them while accepting the scientific worldview must accept that we really do find our‑ selves in an enchanted castle, imposed upon by spectres and apparitions. Notes 1 Ainslie (2015, esp. ch4) disagrees, arguing that philosophical reflection makes it impossible to follow through on sceptical arguments whereas vulgar attitudes never bring them up. For Ainslie’s Hume, scepticism is something to be over‑ come or avoided, not embraced. 2 HE, xiv. Todd quotes only the first two of the three suppressed passages (HE, xiv–xviii). 3 Williams (2004, 281) notes that it is unlikely that the author of the Treatise thought the belief in body can be seriously destabilized.

312  Conclusion Bibliography Ainslie, Donald. 2015. Hume’s True Scepticism. Oxford: Oxford University Press. Falkenstein, Lorne. 2021. “Moral Disagreement.” In Hume’s An Enquiry Concern‑ ing the Principles of Morals: A Critical Guide, edited by Esther Engels Kroeker and Willem Lemmens, 238–56. Cambridge: Cambridge University Press. Williams, Michael. 2004. “The Unity of Hume’s Philosophical Project.” Hume Studies 30: 265–96.

Index

Note: Page numbers followed by “n” denote endnotes Aristotle 96 Ainslie, Donald 13n10, 44–5, 123–4, 239, 245n8, 275n16, 275n24, 282, 285 Allison, Henry 60n1, 60n4, 109n19, 244n2, 275n17 analogy 202n25, 217, 223 angular magnitude see measures angular separation 191–2; see also measures argument from the non‑entity of unqualified points 95–8, 165 assent 210–11 association 39, 212 256; by contiguity and resemblance 212, 215–16, 217–18, 222, 234–5, 240, 244, 258; as a feature of identity attributions 236–7; and object recognition 243 attention 211–12, 237, 240 atomism: psychological (see psychological atomism) Barrow, Isaac 97, 113n43 Baxter, Donald 13n9, 16n25, 16n27, 16n30, 65n50, 87, 100–1, 110n27, 114n52, 123–4, 129–30, 133–41, 142n10, 142n12, 161n3, 163n17, 203n26, 203n27 Bayle, Pierre 39–42, 53, 88, 99, 134 Beattie, James 34–8 Beck, Lewis White 15n20 238 Berkeley, George 7–8, 103, 105, 126, 135–6, 207, 289, 298 Berkeleian interpretation see vacuum

belief 209–11, 213–14, 217–21; causes of 214–16, 222–4, 232–8; in “chances” and the outcomes of “inconstant causes” 216, 219; covering law account of 217–21; involuntary (see determinism); functions (effects) of 213–14; normativity of 224–9; unphilosophical 1, 222–9 body 250, 278; continued, external, and independent existence of 256, 265, 273n9, 279–80, 303–4 inferred from empirically guided reasoning 256–9; perceived by the senses 251–6; our own distinguished from other bodies 253–4, 273n5 Broad, C. D. 16n25, 109n19, 178–9, 200n3, 202n19 Butler, Annemarie 282, 289, 290, 301n14, 301n17 causal inference: dependent on coherence 7, 264; dependent on prior identification of species 7, 152–3, 236, 238, 238–9; as the foundation for inference from contiguity and inference concerning identity 233–5, 215–16, 235–6; on a par with inference from constant contiguity 234–5; as a principle of association 212, 256 Clarke, Samuel 39–40, 100, 103 coherence: initially defined as customary persistence, motion, or change over time 260–1;

314  Index subsequently considered as a product of inference to the best explanation 261–3; subsequently considered as a product of an inertial tendency of the imagination 264 colour 28–29, 30, 33–4, 47, 90, 95–6, 98, 106, 165 comprehension problem: discussed 38–41; resolved 41–2 Condillac, Étienne 8, 61n15, 103 consciousness 10–11, 47–53; extent of 42–6, 48–51, 104, 125, 132; temporal (see temporal experience); unity of 39–42 constancy: initially defined as customary contiguity of numerous objects 259–60, subsequently considered as invariance of a single object over time 266 contiguity 168; customary 232–5, 240–41, 244, 258, 260, 265–6, 266, 274n10, 275n22, 282–3, 280, 293; see also association conversion 186, 187–8 copy principle 91 Costa, Michael J. 10, 16n25, 64n46, 111n34, 141n5, 142n11, 143n19, 162n6, 171 Cottrell, Jonathan 16n25, 73, 107n6, 171, 179–82, 200n3, 202n22 custom 256

education 1, 12n2, 13n7, 224, 225, 226 empirism vrs empiricism 64n44 endurance vrs perdurance 12, 121, 128–9, 136, 142n12, 142n14, 143n17, 153, 155, 156–7 endurance fiction 130–3 extension 77, 84; see also consciousness; mind; spatiality; temporal experience eye/hand condition on identity 150–1 eye motion: and the coherence of sense experience 263–5; and vacuum 192–3

Dainton, Barry 63n36, 63n37, 63n38 darkness 167, 173–4 De Pierris, Graciela 82–6, 106n1, 108n14, 110n24 Descartes, René 8, 53, 104 determinism 1–4, 14n12, 211, 217, 221, 226–7, 229–31, 304 difference 122, 142n9, 151 discreteness, density, and continuity 63n38, 81–2, 109n22, 110n24, 168, 201n28 distance 167–70, 176–7, 186; invisible and intangible 166, 169–70, 184, 186, 188, 189–90 distinctions of reason 91–2, 121 dualism 7–8, 27 duration 76, 125–41

Garrett, Don 12n3, 14n12, 15n24, 21, 61n21, 62n31, 101, 104, 142n8, 147, 161n3, 202n19 general rules 223–4, 226(ii), 227–9 George, Rolf 78–9, 106n2 Georgian thesis 77–9; see also null thesis

failure to observe vs. observation of a failure 261–2, 265–6 fiction of endurance see endurance fiction field see image field; memory field; sensory field; tactile field; temporal field; visual field “filled interval” case see “lone body” vrs “filled interval” case finite divisibility: in the Enquiry 69, 88, 110n27; of ideas 70–2; of impressions 72–4; of space and time 68, 70, 79–82, 86–89, 108n15, 110n27, 296; of temporal experience 75–6 forensic identity 147 free identity see identity, free

hardness see solidity Helmholtz, Hermann von 75–6 Holden, Thomas 108n17, 110n31 holism see psychological atomism Hume, David: attitude to 1739 account of belief 213, 309; authorized vs. anonymous works 4; commitment to the transfer thesis 218–20; conflicted account of temporal experience 52–3, 54–6, 58–9,

Index  315 118–20, 134–5, 138–9; change in attitude to sceptical arguments 307–8, 309–10; change of mind over whether belief is an annexed sentiment 217–18; continued commitment to psychological theorizing in EHU 218, 219–21, 244n1, 245n14, 245n16; doubts concerning personal identity 46–7; expertise in mathematics 106n1; indifference to the distinction between identity and individuation 154; intellectual development 4–5; loss of confidence in his account of space and time 70; motives for his views on finite divisibility, vacuum, and time when nothing changes 98–104; ongoing commitment to T’s psychology of belief 309; reluctance to change his mind 5, 25, 135, 213, 221, 245n19; reluctance to reveal his sceptical intentions 13n11, 232, 310; revised account of belief in body 250, 278 Hume’s difficulty see identity, difficulty concerning Hume’s remedy 1–4, 11, 229–32, 286, 305–6, 307–8 ideas: cold, coloured, hot, or painful 36–7; as distinct from ideas 205–9, 244n1; extended 36–8, 40, 45–6, 74, 163n16; positive vs negative 142n9, 167; solid (see solidity) idealism 7–8, 21 identity 12, 122–3, 146–63; and causality 152–3, 235–6; conditions for 122–3, 152, 162n11, 237; customary 242–4, 261, 274n10, 280; difficulty concerning 146–7, 149, 157, 159; for Hume 159; free 149–57, 270–2; and individuation 162n 13; and motion 151; and scepticism 160–1; and space 150–1; see also difference; individuation; invariance; one–one condition; uninterruptedness image field 43, 46 immortality 106 impressions 19–20; complex 9, 11, 42–6, 47–53, 84, 85–6, 91, 93,

120, 240; corporeal (see solidity); as distinct from ideas 205–9, 244n1; extended (see spatiality); external existence of 20–1; field states 280; independently mobile (see mobility, independent); original 8, 9, 20, 22, 29, 42, 83; of primary, sensible and subjective qualities 255–6, 295–8, 301n16; real and corporeal 252–3; related to minds 21–2; simple 9, 48, 51–2, 85(iv), 85–6, 183; solid (see solidity); spatially disposed 28–9, 41–2, 163n16, 251–2 independent mobility see mobility: independent individuation 122, 151, 153, 162n13; see also difference; identity indivisible parts, points, objects see points infinite divisibility see finite divisibility instantaneous survey 154–5 interruption: arising from a failure of observation 269–70; arising from unaccountable disappearance 268–9 intuition and intuitionism 93–4, 105, 120–1 invariance condition on identity 150, 151–2 invisible and intangible distance see distance, invisible and intangible involuntarism see determinism James, William 39–42, 203n28 judgement 210–11 Kail, P. J. E. 60n6, 273n2, 274n10 Kant, Immanuel 7, 8–9, 111n38, 162n8, 238–9 Kemp Smith, Norman 99, 102, 111n37, 153–4 Kervick, Dan 16n25, 113n47, 168, 171, 178, 180, 200n3, 200n6, 201n9, 201n12, 202n22 Larvor, Brendan 106n1, 110n28, 112n40 limitation problem: discussed 42–6; resolved 51 location (of sensations or lone bodies on a sensory field) 8, 104, 193; see also local signs

316  Index local signs 193–7 Locke, John 53–8, 127, 134, 147, 151, 162n9 Loeb, Louis 13n8, 13n10, 14n14, 16n28, 16n31, 143n20, 226, 245n11, 246n19, 246n22, 246n24, 246n25, 246n26, 258, 263, 264,266, 274n11, 275n20, 275n22, 275n24, 300n8, 301n17 “lone body” vrs “filled interval” case 185–6 MacNabb, D. G. C. 246n19 Malézieu argument 15n22, 87–8, 93 manner of affection 174–5, 185, 186, 188, 189–98 manner of conception 211 manners of disposition 8, 9, 40, 69, 89–94, 117–21 Marušić, Jennifer Smalligan 245n9, 245n12, 245n16 meaning empirism 64n44, 177, 179, 182–4 measures 35–6, 78, 82–6, 107n11, 284 Mellor, D. H. 141n4 memory 49–50, 207 memory field 49–50 mental representation 22–34, 185, 199 mind 43–7, 279; dispositions of 212; extended vs unextended 7–8, 9, 31–8, 41–2, 104; operations of 212; see also consciousness mobility: and identity 151, 269; independent 253–4, 259, 280, 282–3, 287 Molyneux, William 194 moments 124–5, 139–40 monotonous succession 123–4, 125–30 motion see mobility Newton, Isaac 100 non‑entity of unqualified points see argument from the non‑entity of unqualified points Norton, David Fate and Mary J. 158, 163n17 null thesis 79; see also Georgian thesis numerical difference see individuation objects 6–7, 21, 42–3, 238–44; publicly observable 243–44 one–one condition on identity 123, 149, 152

ontogenesis vrs psychogenesis see psychological atomism paradigm idea see preferred view perdurance see endurance and perdurance perception 25, 92, 93–4; see also impressions; ideas; Reid phantom pains 287–8 phenomenalism 239–40, 293–4 place 91, 92–3, 96–8, 168, 183–4, 240, 245n10, 247n34 points 77–9, 84–5, 91–2, 94, 95, 96, 183 position see place preferred view 243, 294 presentism 10, 53, 118–20, 128 Price, H. H. 122–3, 142n10, 143n20, 153, 160, 161n3, 162n7, 238, 248n39, 275n22, 291, 294, 299n4, 300n8 prolonged survey 155–6, 158–9 psychogenesis see psychological atomism psychological atomism 9, 11, 15n22, 16n27, 42–3, 48, 51–2, 82–6, 93, 91–3, 96 Qu, Hsueh 14n15, 244n1, 245n16 realism: naïve 12, 21, 73, 253–4; representational 290, 290–3, 294, 304 Reid, Thomas: ambiguity (of secondary/sensible quality terms) and univocity (of primary quality terms) 26–7; colour terminology 30, 33–4; dualism 27, 32–3; experimentum crucis 24–5, 31; on Hume’s scepticism 1, 239, 284–5; memory 23–4; modus tollens against Hume’s account of mind 33; respect for Hume’s anonymity 32; sensation and perception (and Hume’s failure to recognize the latter) 9–10, 16n25, 22–5, 33, 165, 185, 199; tactile thought experiment 29–30; temporal experience 53, 64n45, 64n46, 64n48, 126, 127 relations 16n17, 44, 83, 87, 91, 92–3, 167–8, 178

Index  317 representationalism see realism, representational resemblance see association retentionalism 119–20 Russell, Paul 99–100, 102–3 Rocknak, Stefanie 15n20, 38, 61n21, 62n24, 62n27, 62n28, 141n4, 142n11, 161n3, 239, 274n9

49; of smells, tastes, and sounds 32, 45–6, 49; of visual and tactile sensations 28–9 31–8, 41–43, 74, 163n16 Stanhope, Philip 68 steadfast objects 143n18, 161n5; see also unchanging objects succession 121–4

Scarburgh, Edmund 53, 96–8, 112n39 secondary qualities see impressions, of primary, sensible and subjective qualities scepticism: against the evidence of touch 287–8; and causal inference 2, 6; and independent existence 279–80; and double vision 280–3, 289; and external existence 1, 6, 11–12, 21, 33–4, 73, 103, 272, 310–11; and identity 160–1; and illusion 288, 291–2; and perceptual relativity 282; a remedy for unphilosophical belief (see Hume’s remedy); and variation in magnitude with distance 283–6, 289 sensation 205–9; of bodies 251–6; see also impressions, original; location sense vs. understanding 205–6 sensible qualities see impressions, of primary, sensible, and subjective qualities sensory field: as mind 279; as monitor 280; as window 241, 253, 265, 279–80, 281, 282–3, 287 separability 47, 51–2, 91–2, 94, 96, 108n13, 111n37, 117–18, 121, 183, 203n31 simple parts, points, objects see points Slowik, Edward 84–5 Smith, Robert 194–6 solidity (also hardness) 95–6, 98, 171, 252, 295–6; for Hume 31, 296–7; for Reid 24–5, 31 space 98, 109n23; ideas of 89–94; see also finite divisibility, of ideas spatial representation 7–9 spatiality: of ideas 34–6, 37, 40, 45–6, 74, 163n16; of phenomenally experienced colour qualia 33–4, 37; of thoughts and passions 45–6,

tactile field 31, 32, 42, 43, 48, 48–9, 111n35, 287–8 tactile space 46 temporal experience 10, 49–51, 53–9, 104, 118–20, 121–5, 129, 130–3 temporal field 50–1 Thomas, Emily 64n41, 64n49, 103, 107n10 time 88–9, 125, 133–5 unchanging (“unchangeable”) objects 123, 130–3, 143n18; see also monotonous succession uninterruptedness condition on identity 150, 152 unity 147, 149, 155–6, 157, 158–9 vacuum 10, 16n25, 100, 102–3, 165– 204, 241; Berkeleian interpretation 166, 174–7, 184, 186–7, 203n2; defined 166; privative 202n16; Third Distance interpretation 166, 177–82, 178, 186–7, 189–90; see also angular separation; eye motion; local signs; manner of affection Van Cleve, James 61n15, 62n27 visual angles 107n11 visual depth 32, 242, 251 visual field 30–1, 31, 32, 37, 42, 43, 47–8, 48–9, 111n35, 241, 284 visual space 32, 35–6, 46, 241 vivacity 206–9; transfer of 214–26; see also manner of conception whole‑part priority see psychological atomism Wilson, Fred 13n6, 15n20, 70, 109n19, 111n37, 239 Winkler, Kenneth 273n8, 300n8 Wright, John 239, 254–5, 273n2, 273n9