Personal identity: complex or simple?

www.cambridge.org/9781107014442

PERSONAL IDENTITY

We take it for granted that a person persists over time: when we makeplans, we assume that we will carry them out; when we punish some-one for a crime, we assume that she is the same person as the one whocommitted it. Metaphysical questions underlying these assumptionspoint toward an area of deep existential and philosophical interest. Inthis volume, leading metaphysicians discuss key questions aboutpersonal identity, including “What are we?,” “How do we persist?”and “Which conditions guarantee our identity over time?” Theydiscuss whether personal identity is “complex,” whereby it is analyz-able in terms of simpler relations such as physical or psychologicalfeatures, or whether it is “simple,” i.e. something that cannot beanalyzed in terms of more fundamental relations. Their essays offeran innovative discussion of this topic and will be of interest to a widereadership in metaphysics.

georg gasser is a scientific researcher at the Department ofChristian Philosophy, University of Innsbruck. He is the editor ofPersonal Identity and Resurrection (2010).

matthias stefan is a scientific researcher at the Department ofChristian Philosophy, University of Innsbruck. He has publishedpapers on the ontological commitments of physicalism, personalidentity and substance dualism.

PERSONAL IDENTITY:COMPLEX OR SIMPLE?

edited by

GEORG GASSER

and

MATTHIAS STEFAN

cambr idge univer s i ty pre s sCambridge, New York, Melbourne, Madrid, Cape Town,

Singapore, São Paulo, Delhi, Mexico City

Cambridge University PressThe Edinburgh Building, Cambridge cb2 8ru, UK

Published in the United States of America by Cambridge University Press, New York

www.cambridge.orgInformation on this title: www.cambridge.org/9781107014442

© Cambridge University Press 2012

This publication is in copyright. Subject to statutory exceptionand to the provisions of relevant collective licensing agreements,no reproduction of any part may take place without the written

permission of Cambridge University Press.

First published 2012

Printed in the United Kingdom at the University Press, Cambridge

A catalogue record for this publication is available from the British Library

Library of Congress Cataloging in Publication dataPersonal identity : complex or simple? / edited by Georg Gasser and Matthias Stefan.

p. cm.Includes bibliographical references and index.

isbn 978-1-107-01444-21. Identity (Psychology) I. Gasser, Georg. II. Stefan, Matthias.

BF697.P468 2012126–dc232012012191

isbn 978-1-107-01444-2 Hardback

Cambridge University Press has no responsibility for the persistence oraccuracy of URLs for external or third-party internet websites referred toin this publication, and does not guarantee that any content on such

websites is, or will remain, accurate or appropriate.

www.cambridge.org

www.cambridge.org/9781107014442

To our parents for their love and support

Contents

List of illustrations page ixList of contributors xAcknowledgments xi

Introductiongeorg gasser and matthias stefan 1

part i framing the question 19

1 Chitchat on personal identitydavid barnett 21

2 In search of the simple vieweric t. olson 44

3 Personal identity, indeterminacy and obligationryan wasserman 63

4 Personal identity and its perplexitiesharold w. noonan 82

part ii arguments for and against simplicity 103

5 How to determine which is the true theory of personal identityrichard swinburne 105

6 Against simplicitysydney shoemaker 123

7 The probable simplicity of personal identitye. j . lowe 137

vii

8 Reply to E. J. Lowesydney shoemaker 156

9 The non-descriptive individual nature of conscious beingsmartine nida-rumelin 157

part iii reconsidering simplicity 177

10 Personal identity: a not-so-simple simple viewlynne rudder baker 179

11 Is “person” a sortal term?christian kanzian 192

12 Materialism, dualism, and “simple” theories of personal identitydean zimmerman 206

13 The morphing block and diachronic personal identityhud hudson 236

References 249Index 257

viii Contents

Illustrations

Figure 1 A taxonomy of views about personal identityover time. page 45

Figure 2 Different sums of person-stages. 75

ix

Contributors

lynne rudder baker is Distinguished Professor in the Departmentof Philosophy, University of Massachusetts (USA).

david barnett is Associate Professor in the Department of Philosophy,University of Colorado (USA).

hud hudson is Professor in the Department of Philosophy, WesternWashington University (USA).

christian kanzian is Associate Professor in the Department ofChristian Philosophy, University of Innsbruck (AUT).

e. j . lowe is Professor in the Department of Philosophy, DurhamUniversity (UK).

martine nida-rumelin is Professor of the Philosophy of Language,Philosophy of Mind and the Human Sciences, Department of Philosophy,Université de Fribourg (CH).

harold w. noonan is Professor of Mind and Cognition in theDepartment of Philosophy, University of Nottingham (UK).

eric t. olson is Professor in the Department of Philosophy, Universityof Sheffield (UK).

sydney shoemaker is Emeritus Susan Linn Sage Professor ofPhilosophy, Cornell University (USA).

richard swinburne is Emeritus Nolloth Professor of the Philosophyof the Christian Religion, University of Oxford (UK).

ryan wasserman is Associate Professor and the Chair of theDepartment of Philosophy, Western Washington University (USA).

dean zimmerman is Professor in the Department of Philosophy,Rutgers University (USA).

x

Acknowledgments

The origin of the present volume dates from a conference on personalidentity at the conference center of Innsbruck University in Obergurgl,Austria, July 19–22, 2010, where preliminary drafts of the papers werediscussed. Sydney Shoemaker’s, Hud Hudson’s and David Barnett’s paperswere not presented at the conference, but were written especially for thisvolume.We wish to thank the following individuals for their role in the

production of this volume. First, we owe a debt of thanks to JosefQuitterer, Daniel Wehinger and Monika Datterl for their help in organ-izing the conference. Dean Zimmerman and Mike Rea were kind enoughto offer valuable advice on proceeding with this project after the confer-ence. Our warmest thanks to Katherine Munn for her valuable sugges-tions for improving the readability of some contributions. We also thankHilary Gaskin and Thomas O’Reilly of Cambridge University Press for allher help and thoughtful advice in preparing this book for publication.Finally, we would also like to thank all the contributors for their greatpatience during the completion of the volume.It does not happen as a matter of course that international philosophical

conferences receive a great deal of funding. Therefore, we wish to expressour deepest gratitude to the Austrian Science Fund (FWF), grant P20186–G14, whose generous support made it possible to organize the conferenceand to edit this volume. Furthermore, we are also greatly indebted to theAustrian Federal Ministry of Science and Research and to the Vice-Rectorfor Research of Innsbruck University for co-funding the accommodationexpenses of the invited speakers.

xi

IntroductionGeorg Gasser and Matthias Stefan

the p ro j e c t

One worthwhile task for philosophy is to give an overview of a wholedomain of thought and to present the conceptual relationships that char-acterize it. The domain we have striven to portray in this introduction, on aquite general level with a broad brush, is the contemporary debate aboutpersonal identity over time. We proceed as follows: First, we specify themetaphysical question of personal identity tackled in this volume: namely,what makes a person P1 at t1 identical to a person P2 at t2? Second, we discussviews which analyze personal identity in terms of bodily and psychologicalrelations. Problems associated with these theories have recently made a four-dimensional interpretation of such views quite popular. The followingsection presents this canny metaphysical alternative to traditional three-dimensional views. Finally we discuss a rather neglected approach to per-sonal identity over time, the so-called “simple view,” according to whichpersonal identity does not consist in anything other than itself; it is simpleand unanalyzable. Eric Olson once suggested that the simple view is poorlyunderstood, and therefore deserves more attention than it has received so far(Olson 2010, section 3).A specific aim of this volume was to take up this suggestion. In the first

section, “Framing the question,” the authors draw attention to the widerframework in which the question of personal identity is posed. They revealsome of the hitherto implicit background assumptions of the theories athand, as well as the explanatory demands one should expect of thetheories. The contributions of the second section, “Arguments for andagainst simplicity,” provide original in-depth analyses of arguments putforward in favor of and against the claim that personal identity is analyz-able. The last section, “Reconsidering simplicity,” contains innovative andso far rather unnoticed arguments that might strengthen the case for thesimple view.

1

the que s t i on of p e r sona l i d ent i t y

In our ordinary non-philosophical moments we take it for granted in almosteverything we do that persons persist over time: when we make plans, weassume that we will carry them out in the future. When we punish someonefor a crime, we assume that she is the same person as the one who committedit. When we regret a misdeed, we assume that we are identical with theagent who performed it. These examples indicate that the assumption thatpersonal identity is continuous is of particular importance for practicescentral to our lives, pertaining to both the treatment of ourselves and ofothers. Furthermore, to know an entity’s identity conditions is to knowwhat kind of entity it is; a fortiori, this is true of human beings: if you wantto figure out who you are essentially, you must ask which conditionsguarantee your identity over time.

Philosophers often refer to this problem as “the” question of personalidentity. But this is misleading; the question of personal identity is not asingle problem (see e.g. Rorty 1976). It is important to distinguish betweenvarious meanings of “personal identity.” For our purposes we identify four:(i) Biographical (or narrative) identity: Who am I? This question asks

how an individual understands and defines herself in light of hervalues, convictions, and aims. “Identity” in this context is a normativeor evaluative concept which incorporates an individual’s self-understanding and her broader life-plan.

(ii) Personhood: What are the conditions for personhood? This questionseeks conditions that make something a person as opposed to a non-person. It calls for necessary and sufficient conditions for personhood,such as being an intelligent, conscious and feeling agent.

(iii) Metaphysical nature: To which metaphysical category do human per-sons belong? Possible answers include the claims that human persons aretemporary stages of human organisms, thinking substances (souls),collections of temporal parts, or bundles of mental and physical states.

(iv) Diachronic (personal) identity: What makes it the case that a person Xat t1 is identical with a person Y at t2? This question seeks thepersistence conditions of persons: that is, what it takes for the sameperson to exist at different times.

Note that the question about epistemic criteria for determining the persis-tence conditions for persons must be distinguished from (iv). The questionof what it takes for a person to persist over time is different from thequestion of how to find out whether a person at one time is identical to aperson at another time. Epistemic criteria for recognizing personal identity

2 georg gasser and matthias stefan

over time must not be confused with criteria for identity itself. Generally, ofcourse, epistemic criteria – such as the continuity of psychological andphysical features – are reliable signs for tracking personal identity over time.It is easy to imagine, however, that epistemic criteria for metaphysicalidentity come apart from identity itself. Consider the case of physicaldisfigurement after a serious accident: in such a case a person can no longerbe recognized by her physical appearance, but this does not imply that thereare two different persons, one before and one after the accident.This volume is about question (iv). An adequate answer to the question

of the diachronic identity of human persons can be presented schemati-cally in the following way: “If x and y are things of kind K, then x isidentical with y if and only if x and y stand in the relation Rk to oneanother” (Lowe 2000, p. 272).If you ask yourself whether the woman at the mall is the girl you knew in

high school, then you ask yourself whether you refer to the same humanperson twice or refer once to each of two different persons. Although thequestion of diachronic identity is related to questions (i) to (iii), keepingthem separate will help avoid confusion. Still, especially question (iii) willplay a role in answering question (iv) because the metaphysical nature of anentity determines its persistence conditions.

the de b a t e a bout p e r sona l i d ent i t y

One way to introduce the contemporary debate about personal identity is todistinguish two basic kinds of account of personal identity in the sense ofquestion (iv): the complex and the simple view.The complex view analyzes personal identity in terms of simpler rela-

tions. The fact that a person persists over time is nothing more than someother facts which are generally spelled out in either biological or psycho-logical terms, or both. That is, the complex view takes talk about “whatpersonal identity consists in” literally. It aims to provide necessary andsufficient conditions for personal identity, thereby reducing it to the hold-ing of basic biological or psychological relations. Whenever these relationsobtain, personal identity obtains.The simple view of personal identity, by contrast, denies that a person’s

identity through time consists in anything but itself. Biological and psycho-logical continuitymay be regarded as epistemic criteria for diachronic identity,but they are neither necessary nor sufficient conditions for personal identity.There are no non-circular, informative necessary and sufficient conditions forpersonal identity: personal identity consists in nothing other than itself.

Introduction 3

the b i o log i c a l a p p roach

A natural idea is that a human person’s identity over time consists inbiological persistence conditions either of the entire organism or of partic-ular parts of it − for instance the brain. The former approach has becomeknown as animalism (Olson 1997b); the latter might be called the brain-based approach (see e.g. Nagel 1986, p. 40, sketches this view as the empiricalhypothesis of the self’s true nature).

Animalism assumes that the biological functioning of the human organism −that is, the persistence of the unity and interaction of metabolic processes − isessential for human beings to persist. Accordingly, a person’s identity is nodifferent from the identity of other living things like horses or mosquitoes.Her persistence does not consist in the preservation of the same matter butrather in the preservation of the same organizational biological form, since thematter constituting the organism is continually replaced.

For animalism, the identity of a functioning organism and the identity ofthe person constituted by this organism do not necessarily go hand in hand.Take, for instance, an irreversibly comatose patient. If we assume that theactual performing of higher brain functions is a necessary condition for anindividual’s mental state, and if these brain functions are absent in acomatose patient, then the patient is a living and functioning organism,but she is not a person, because this presupposes that she enjoys a mental lifeof some complexity. According to the biological approach, we are notessentially human persons, but rather human organisms or animals. Wecan lose the status of personhood while remaining us, because our humanorganism can continue to be alive.

A well-known argument for animalism is the “thinking animal problem.”It starts with the insight that human animals exist. Wherever a humanperson is, there seems to be a human animal too: wherever you sit, a humananimal sits too; whenever you work, so does an animal; whenever you arethirsty, an animal is too. The animal is most intimately related to you, sothat it is difficult to tell the difference between you and it. Olson writes:

In fact the animal seems to be mentally exactly like you: every thought orexperience of yours appears to be a thought or experience on the part of the animal.How could you and the animal have different thoughts? But if the animal thinksyour thoughts, then surely it is you. You could hardly be something other than thething that thinks your thoughts. (Olson 2007, p. 29)

How might one respond to this argument? One response is to deny thatthere are animals. A second is to say that there is an animal where you are


but that it is only you and not the animal that enjoys a mental life. Animalsare living but non-thinking beings. Finally, one could respond that there isan animal where you are, which has the same thoughts as you, but thatnevertheless you are not identical with it. You share a mental life with ananimal but not a metaphysical nature, because you are a human personwhereas the animal is not.The first option is implausible. There are no good reasons to reject the

existence of animals. The second option dissolves into a kind of dualism:one might argue that animals qua biological organisms are unable to thinkbut that their soul is the entity endowed with a mental life. A materialistalternative to this option could instead identify human brains as thinkingbeings in contrast to animals. Accordingly, a mental life can be attributedonly to a specific part of the human animal – the brain – rather than to theanimal as such.The third alternative results in what is called the “too many thinkers

problem.” If we assume that the human animal enjoys the same mental lifeas the human person but that they are nonetheless two distinct entities, thenapparently the human animal and the human person coincide spatially.Even though ordinary parlance refers to just one entity, there are in fact two.This proposal thus fails to solve the problem of coincidence. If there is ahuman animal thinking your thoughts as well as you as a human personthinking your thoughts, then why should you assume that you are thehuman person and not the thinking human animal? There is a seriousepistemological problem because you are not in a position to tell which ofthe two entities you are. In addition, someone might wonder why athinking human animal should count as an animal and not as a person.What keeps the human animal from being a human person?If this argument for animalism is sound, there is strong reason to think

that we are, at least during certain episodes of our existence, essentiallyanimals capable of thought, unless human persons are identical either tomental substances or to brains. This conclusion pertains to the metaphysicalnature of human beings, but is closely related to the question of conditionsfor personal identity: if we are biological organisms, then our identityconditions are those of biological organisms − that is, our identity consistsin the continuity of a living body over time.An alternative to animalism is the brain-based approach. It claims that

there are certain biological conditions for personal identity which, how-ever, fall short of the entire organism. It starts from the assumption thatone part of the body − the brain (or certain parts thereof ) − is of particularimportance because it produces the mental life characteristic of being a

Introduction 5

person. Damage to the brain can result in personality changes and am-nesia, and in severe cases can utterly obliterate the capacity for highercognitive functions, whereas damage to the limbs, stomach or otherorgans has no such dramatic effects. Admittedly, we might undergo apersonality change as a consequence, but these events would not affect thebrain’s capacity for maintaining a mental life.

You can even imagine that your brain could be removed from your bodyand preserved in a functional state by a complex machine. Assuming thatyour brain works as well after as before its disembodiment, you might stillbe considered the same person under these artificial conditions. A thoughtexperiment by Sydney Shoemaker (1963) underpins this line of thought.Imagine that the brain of one person, Brown, is removed and transplantedinto the body of another person, Robinson. The resulting person,Brownson, has the body of Robinson but the brain of Brown and thusBrown’s whole psychological makeup. Most of us would be inclined to saythat the newly created person Brownson is identical with the former personBrown. Derek Parfit formulates this intuition as follows: “Receiving a newskull and a new body is just the limiting case of receiving a new heart, newlungs, new arms, and so on” (Parfit 1984, p. 253).

These considerations favor the view that the functioning brain (or acertain part thereof) needs to persist for the human person to persist, butthat the entire organism is not required to do so because only a functioning(and appropriately stimulated) brain, and no other organs, is needed forproducing an individual’s mental life. Note that the brain-based approachplaces so much emphasis on the brain because the latter sustains one’smental life. Herein lies a major difference between the brain-based viewand animalism: for animalism, the fact that a human organism can enjoy amental life is of no importance in deciding identity over time. For the brain-based approach, by contrast, the brain is essential for guaranteeing one’sidentity over time because it guarantees the continuity of one’s mental life.

the p s y cholog i c a l a p p roach

As the thought experiment of the brain transplantation between Robinsonand Brown shows, some intuitions motivate a link between personalidentity and psychological continuity rather than between personal identityand the identity of the brain. Such intuitions are encouraged by anotherthought experiment involving what Shoemaker (e.g. 1984, p. 108) called a“brain state transfer device”: this device reads the states of a person P1’sbrain, writes them into the brain of person P2, and then destroys P1’s brain,


turning P1’s organism irreversibly into a human vegetable. What becomes ofP1? Is she identical to the human vegetable or did she just acquire a neworganism thanks to the brain state transfer device?Pace the brain-based approach, P1 ceases to exist once her brain is

destroyed. The line of thought pursued in the thought experiment, how-ever, provokes the question whether we should say that personal identityconsists in the continuing functioning of the human brain at all. We couldimagine that no brain is needed anymore and that anything would do thejob as long as it sustains one’s psychological life. If psychological states andtheir continuity are the mark of personal identity rather than the continuityof the brain (or some other biological fact), then P1 continues to exist byacquiring P2’s body. This approach to analyzing personal identity amountsto the claim that a human person’s persistence consists in a particularconstellation of psychological relations over time. It is known as thepsychological approach.An advocate of some version of the psychological approach must specify

what kind of psychological relations are necessary and sufficient for humanpersons to persist. Philosophers generally try to spell out psychologicalcontinuity in terms of causal connections between earlier and later psycho-logical states, such as remembering earlier experiences, forming and carryingout intentions, and holding beliefs over time. Sydney Shoemaker character-izes this approach this way:

Reverting to the “person-stage” terminology, two person-stages will be directlyconnected, psychologically, if the later of them contains a psychological state (amemory impression, personality trait, etc.) which stands in the appropriate relationof causal dependence to a state contained in the earlier one; and two stages belongto the same person if and only if . . . they are connected by a series of stages suchthat each member of the series is directly connected, psychologically, to theimmediately preceding member. (Shoemaker 1984, p. 90)

On this view, whether a person at t1 is the same as a person at t2 dependson what constitutes an “appropriate” relation of causal dependence betweenpsychological states – however “appropriateness” is specified in detail.We have been led by a series of considerations from animalism, via the

brain-based approach, to the psychological approach to personal identity.Whichever version of the complex view one prefers, either way the bio-logical and psychological approaches agree that a person’s identity over timeis definable in terms of something other than itself. They differ merely inspelling out what that something is.

Introduction 7

two prob l em s for comp l e x a p p roache s

The general claim of any complex approach is that a person’s identity overtime can be analyzed into necessary and sufficient components other thanidentity over time itself. This claim comes with problems of its own. Thissection briefly examines two of these problems, the problem of graduality andthe problem of fission.

First, the problem of graduality (Noonan 1989, pp. 128–48). If personalidentity consists in simpler items such as bodily and/or psychologicalrelations, then we can imagine situations in which these relations admit ofdegree (Lewis 1976). This leads to the question of what is the threshold thatdemarcates personal identity from non-identity. Imagine that in a series ofoperations the parts of P1’s body are discarded and replaced with parts of P2’sbody, until all parts of P1’s original body have been replaced with parts ofP2’s body.With each operation some of P1’s psychological states are also lostand are replaced by psychological states of P2. By the end, P1’s entire bodyand consequently all of her psychological states are exchanged with P2’sbody and her psychological states so that P1 is identical to P2 in terms of herbody and psychology. It is very likely that, after only a few operations, P1still exists with just minor changes in her bodily and psychological states. Atthe end of this series of operations, however, one might have the intuitionthat P1 does not exist anymore but was replaced with P2. The crucialquestion is: is there a way to indicate the sort of relations needed betweentwo person-stages of P1 so that changes in these relations do not result in herdropping out of existence by being replaced with P2? Is there a precisethreshold demarcating P1’s persistence in time?

As can easily be imagined, it is unclear how to specify the exact thresholddemarcating P1’s existence. It may well always be possible to presentexamples of deviant continuity relations which leave it undeterminedwhether or not two person-stages are continuous. It is very likely thatcomplex views go hand in hand with bodily or psychological continuityrelations that admit of degree, and it is hard, if not impossible, to specifywhat degree must obtain to guarantee personal identity. Thus the sort ofcontinuity relations that personal identity consists in is elusive.

One could argue that this might only be an epistemic problem for finiteminds such as ours, because human bodies and human psychology areextremely complex phenomena. An omniscient being, by contrast, mightbe able to specify the appropriate sort of bodily and psychological continu-ity needed for a human person to persist. But who can tell? There are no


points of reference that a clear demarcation exists. This suggestion is merelyspeculative and we will not pursue it here any further.Apart from the insight that complex views might provide us with a less

specific account of personal identity than we would wish for, there is a moreserious objection to these views. Opponents of the complex view point to anintuition to the effect that it does not make sense to think that personalidentity can be gradual. They say: whatever tomorrow brings, there is thestrong intuition that either P1 will exist or will not; of all the people existing,either she will be one of them or none will be identical with P1. Either ofthese two states of affairs will obtain and, so the claim goes, the viewmust berejected which admits of degrees between these two states. If personalidentity is determinate, it cannot be the case that there is just P1 at t1 andjust P2 at t2 while being indeterminate whether P1 is identical to P2. Maybeno one can tell whether P1 is identical to P2, but nevertheless the statementabout P1’s identity is either true or false.One way to challenge this line of argument is by appeal to well-known

puzzle cases such as Parfit’s club (1984, p. 213). Imagine that, for some years,a club exists, then that the regular meetings cease, and a few years later thatthe members of the same club start to meet again with the same rules andthe same name. Someone could ask: have the members set up a new clubsimilar to the old one? Or do they continue the same club which existsintermittently? It seems reasonable to argue that these questions can beanswered conventionally. Points can be made in favor of the thesis that theold and the new club are identical, and other points can be made in favor ofthe alternative thesis that they are not identical. Depending on the con-ditions of identity one accepts, one can legitimately hold either thesis, forthere are no right or wrong answers. It seems obvious that once it is settledwhich conditions of identity are accepted, it does not make sense tocontinue to argue about the correct answers for this case.But can we proceed in this way when it comes to human persons?

Defenders of the thesis that personal identity is determinate claim it is notabsurd to ask whether a correct answer can be given. On the contrary, itwould be absurd if a human person, facing this question for herself, thoughtit sensible to consult a general meeting or a law court. It does not makesense, the argument goes, to assume that (under certain circumstances) ageneral meeting or a law court could simply decide whether or not a personat one time is identical to a person at another. In contrast to the case of theclub, there is a pertinent intuition that whatever decision is made, thepossibility remains open that it would be the wrong one. Any decision ismade under risk, that is, because things could differ from the way the court

Introduction 9

decides. This intuition is particularly strong when one considers the matterfrom a first-person perspective. David H. Lund, for instance, writes:

I am unable to imagine being involved in circumstances under which I am neitherfully admitted nor fully excluded. The experiences occurring under these circum-stances would have to be something for me if I am to be involved in them at all, butthe suggestion that they would be something for me even though it is indetermi-nate as to whether I am having them seems simply unintelligible when one takesthe first-person perspective and reflects upon what it is to have experience. (Lund2005, p. 229)

Second, the problem of fission (Noonan 1989, pp. 149–68). The complexview does not rule out the possibility that the necessary and sufficientrelations for personal identity over time are not one-to-one but one-to-many (Shoemaker 1963; Williams 1970). Imagine the scenario of a person P1splitting up at t1 into two subsequent distinct persons at t2, P2 and P3. Theystand to the original person symmetrically in the identity-defining relation.Both P2 and P3 are fully continuous with P1 in terms of the latter’spsychological and biological characteristics. Neither from the outside northe inside can it be determined which of the two successors is identical to P1.Although the same identity-defining relations obtain between P1 and P2 aswell as between P1 and P3, P1 cannot be identical with both successors at thesame time because P2 and P3 are numerically different persons.

The proponent of the complex view must have a story to tell about suchscenarios. One option is for her to reject the presupposition which is oftencalled the “only x and y principle.” This principle says that two persons, P1at t1 and P2 at t2, are identical iff (a) they stand in the appropriate internalrelationship and (b) there is no other competing person, P3 at t2, who standsin the same relationship with P1 as P2 at t2 (see e.g. Nozick 1981, pp. 29–47).In other words, whether P1 and P2 are identical depends on whether aprocess of fission occurs. If fission occurs, identity does not obtain; if fissiondoes not occur, identity obtains.

However, this line of argument seems to imply that questions of personalidentity do not depend only on facts internal to the relation between aperson existing at different times. Instead, whether or not P1 exists after t1depends on the seemingly secondary issue of whether or not another personsimilar to P1 will exist at t2. The adoption of the “only x and y principle” hasbeen criticized as counterintuitive: what else could a person’s and hersuccessor’s being identical depend on other than facts about themselves?In addition, the “only x and y principle” appears ad hoc, its only ostensible


purpose being to exclude fission cases from the discussion about personalidentity.A related solution to this puzzle could propose that what matters to P1’s

survival is not personal identity but simply that someone stands in theappropriate continuity relations. If more than one person does so, then thisis no cause for concern but a sign that one’s continuing existence should notbe understood in terms of a personal identity rather than an (appropriate)continuity relation. It is not identity that matters, but continuity. As long asat least one person exists in the future who will be related to P1 in the way inwhich P1 is related to her past selves, P1 continues to exist. Fission representsno particular difficulty to this approach because, unlike the normal one-to-one case, in such a scenario the person continues to exist “twice over” andshe has no reason not to value the existence of her successors as much as herown existence before fission took place. Since it is not identity that mattersin survival, no violation of transitivity of the identity relation takes place insuch fission scenarios (Parfit 1984, pp. 245–80).The bottom line of the discussion so far is that no complex account of

personal identity is wholly satisfactory. One way to deal with this result is toassume that our intuitive beliefs concerning personal identity are somehowconfused or even inconsistent. Another possibility is to assume that thethought experiments lead us astray because they do not represent realpossibilities in our world (Wilkes 1988, pp. 8–18). A third possibility is topropose an alternative metaphysical framework which can account for thesepuzzle cases. So far we have considered only a three-dimensional metaphys-ical framework. It seems, however, that four-dimensionalism fares betterwith the problems discussed than its three-dimensionalist alternative.

four -d im en s i ona l i sm

Theorists of personal identity generally hold that a human person exists atdifferent times. How she persists, however, is a matter ofmetaphysical dispute.Three-dimensionalism claims that a human person, like any other materialobject, is wholly present at each time that she exists and has no temporal parts.Four-dimensionalism, instead, holds that human persons not only have spatialparts, but successive temporal parts as well. They are temporally extendedcomposites filling up regions of space-time. Ted Sider writes:

My spatial parts extend through time like I do. We call them spatial parts becausethey are smaller than I, spatially speaking; they are “cut out of” me along a spatialdimension. Reverse time and space in this description and we obtain a description

Introduction 11

of my temporal parts, which extend through space like I do but are smaller than I,temporally speaking; they are what you get by slicing me along a temporaldimension. (Sider 2001, p. 2)

According to this approach, persons have temporal parts located atdifferent times, in a way that is analogous to having spatial parts locatedat different places. A person who exists throughout the decade between 1990and 2000 does so by having connected temporal parts at every time between1990 and 2000. She extends along not three but four dimensions, being“spread out” over a region of space-time. Persons, like material objects, areconceived as four-dimensional space-time worms whose single parts areunited by relations which – depending on the account one favors – can bespecified in various ways. John Perry (2008, pp. 7–12), for instance, under-lines that the unity relation is of crucial importance for understandingpersonal identity because it determines which parts belong to the sameentity. The identity relation, instead, indicates whether one or two (ormore) distinct entities are present. Thus, it has to be kept distinct fromidentity over time even though both relations are themselves closely related(see e.g. Perry 2008, pp. 7–12). Perry clarifies the distinction between thesetwo relations with the following example: If we want to learn about abaseball game we need to know when events in the game are parts of onesingle baseball game. If we just knew about the parts of the game, but notabout when they belong to one single game rather than to two differentones, we would not possess the concept of a baseball game. Thus, we arelooking for the unity relation between the single events of the one baseballgame. Unity is a relation between the events of one game specifying whichevents belong to the same game. When different events or parts stand in theunity relation, they are events or parts of one single entity. Identity, bycontrast, specifies whether game A is numerically the same as game B.

The unity relation can be understood in various ways. You could havepsychological continuity, bodily continuity or whatever unifies spatio-temporal parts into one single person. Depending on one’s preferredaccount of personal identity, variants of bodily theories of personal identityare as correct as variants of psychological ones. Different accounts ofpersonal identity over time simply define persons differently: whereas onetheory employs bodily connections between temporal parts to pick out“bodily” persons, the other theory employs psychological connections topick out “psychological” persons. Facts about personal identity over timeare up to us to determine, since it is up to us to decide which kind of“person” to refer to. Those who employ different unity relations may not


disagree about the correct approach to personal identity at all. Rather, theymay refer to different kinds of persons − persons being composed of person-stages related to one another by bodily continuity and persons beingcomposed of person-stages related to one another by psychologicalcontinuity.With this ontological framework, four-dimensionalism gets a grip on the

problems of fission and graduality (see e.g. Perry 1972). As noted, fissionviolates the transitivity of identity. Not so the unity relation, which is nottransitive. If x and y are parts of A and y and z are parts of B, then this does notimply that x must be a part of B and z be a part of A. Different objects canshare some but not all of their parts. The same holds for temporal parts ofpersons: if two persons share temporal parts P1 and P2 at t1 and t2, then thisdoes not imply that they share their temporal parts P3 and P4 at t3 as well:imagine there is one person composed of the temporal parts P1, P2 and P3, andanother composed of P1, P2 and P4, but there is no person composed of P2, P3and P4. So there is one single person with the temporal parts P1, P2 and P3, andanother with the temporal parts P1, P2 and P4. These two persons arenumerically distinct: they share temporal parts P1 and P2, and they thereforecannot be distinguished during the times t1 and t2. Fission does not pose aparticular problem to a four-dimensionalist account, because there is not onesingle person dividing into two but two persons parting ways after they havecoincided for some time.Four-dimensionalism suggests a solution to the problem of graduality,

too (Noonan 1989, pp. 140–8). Graduality is understood as the result ofsemantic underdetermination and not a result of metaphysical puzzles.Once it is agreed that all there is are spatio-temporal parts, then there isno open ontological question. We only have to specify our concept of thehuman person: that is, we have to specify the unity relation we pick out withthis concept, and then we can state precisely which spatio-temporal partsbelong to the same (kind of) person.The upshot of this discussion is that four-dimensionalism’s liberal ontol-

ogy of objects provides a powerful metaphysical framework for solving theproblems that plague complex approaches within a three-dimensionalistframework. Of course, not everybody sees this liberal ontology as a virtue.Some philosophers consider its cost to be prohibitive. There is no space todwell on this discussion here. It suffices to note that, by construing any filledregion of space-time as an object, for some philosophers four-dimensionalism countenances far too many objects which are not robustenough to make sense of our ordinary understanding of an object’s persis-tence (see e.g. Baker 2007d, pp. 199–217).

Introduction 13

the s imp l e v i ew

Some philosophers reject all the accounts of personal identity presented sofar. They consider them fatally flawed and as a consequence take themselvesto be justified in assuming that there are no informative non-circularconditions for personal identity. According to them, personal identity issomething so fundamentally basic that it cannot be accounted for in morebasic terms such as biological and psychological relations or temporal parts.E. J. Lowe, for instance, puts this account as follows: “The persistence of atleast some sorts of things must . . . be primitive or ungrounded, in that it canconsist neither in relationships between non-persisting things nor in thepersistence of other sorts of things” (Lowe 1988, pp. 77–8).

As a simple relation, personal identity does not have other relations asproper parts, and what has no proper parts cannot obtain partly, as thiswould presuppose that only some parts exist and others do not.Accordingly, personal identity either obtains or does not obtain, and assuch it does not admit of degree.

Taking personal identity to be ontologically basic does not imply anyform of skepticism about the reality of persons and their identity over time.It seeks merely to distinguish between epistemic criteria and conditions foridentity over time. Biological and psychological relations are epistemiccriteria for justifying the assumption of personal identity, but they are nottruth conditions for its obtaining. Even if one knew everything about aperson’s psychological and bodily relations, the question of personal iden-tity would remain open. According to the simple view, it is metaphysicallypossible for there to be two worlds which are identical in their physical andpsychological details, except for the distribution of personal identity. Ofcourse, under ordinary circumstances psychological and bodily relations candoubtlessly be considered to be reliable signs for the obtaining of personalidentity. Nevertheless this does not justify conflating them with conditionsfor the obtaining of identity over time because in the light of the simple viewthere is exactly one such a truth condition, which is the obtaining ofpersonal identity itself.

The claim that personal identity is simple implies that it cannot beanalyzed because it is not possible to appeal to other entities of a suitablekind for formulating in a non-circular way what personal identity consistsin. This does not imply that the simple view is entirely uninformative. Thediscovery − if it is a discovery − that personal identity is not analyzable is akind of progress in understanding it. In addition, it should be clear by nowwhy the simple view is mainly a negative thesis in light of which the most


one can realistically hope to show is that its rivals come with prohibitivecosts of their own.Arguments for the simple view may be grouped into arguments from

intuition, epistemic arguments and ontological arguments. Arguments fromintuition appeal to the intuitive force of the simple view by considering thevarious counterintuitive problems, such as fission and graduality, that burdenthe complex view (Chisholm 1976, pp. 110–12; Swinburne 1984, pp. 13–19).The simple view provides a clear answer to the problem of fission: sincepersonal identity is a simple fact, either it obtains or it does not. In the fissionscenario the simple view offers three possibilities: P1 is identical with P2 andnot with P3, P1 is identical with P3 and not with P2, or P1 ceases to exist and isidentical with neither P2 nor P3. Still, however, there is a fact of the matterabout which possibility is realized, and therefore there is no need for thesimple view to draw on conventionalism to decide whether identity obtains.The same is true in the case of graduality: there must be a definite borderlinethat demarcates identity from non-identity, because identity, being a simplefact, cannot obtain partly. One peculiar implication of the simple view is thatwe may well be ignorant about how identity is actually distributed. It may bethat we simply cannot tell for certain whether P1 is identical with P2, with P3,or with neither. Though this might be unfortunate, it is an epistemologicaland not an ontological problem. Human beings are not omniscient, and factsabout personal identity in fission and gradual cases might be among thethings we do not know.The second type of argument for the simple view appeals to the idea that

knowing everything about bodily and psychological properties and theirrelations would still leave open the question of personal identity, becausethis question must be answered from a first-person perspective. This type ofargument starts from the claim that personal identity is conceivable in theabsence of psychological and bodily relations, and moves on to its meta-physical possibility (e.g. Madell 1981, pp. 78–106). Accordingly, there aretwo conceivable possibilities: first, I might have had a totally different life:for example, living in a different century with a different body and adifferent psychological makeup. I might, for instance, have lived inseventeenth-century France, been born of different parents and had differ-ent memories, intentions, desires and so forth. Nevertheless, so the argu-ment goes, from my first-person perspective it is still conceivable that it isme that lives that life. A similar argument points to changes of body andpsychology (see e.g. Swinburne 1984, pp. 22–3). I can conceive of myself ashaving your body and psychology and of you as having mine. I could alsoimagine that I might not have existed, but that instead someone else exists

Introduction 15

with the same life and body that I actually have. If these scenarios really aremetaphysical possibilities, then the obtaining of psychological or of bodilyrelations is neither necessary nor sufficient for personal identity: there is apossible world in which I exist without the bodily and psychologicalproperties that I actually have, and another in which the bodily andpsychological properties I actually have belong to another person.

It should be noted that this is an argument against the claim thatsynchronic personal identity is analyzable. However, synchronic and dia-chronic identity are not different in kind. Rather, they are instances of thesame relation as it obtains either at one time or at different times.Accordingly, if synchronic personal identity cannot be analyzed, then wehave every reason to assume that diachronic personal identity cannot beanalyzed either.

There is another epistemic argument closely related to this one. It arguesthat whether one has a conceptual grasp of personal identity over time hasnothing to do with whether one has knowledge of psychological and bodilyrelations. In other words, we understand clearly what it means for a personat t1 to be identical with a person at t2 without knowing any bodily andpsychological criteria for personal identity (see e.g. Nida-Rümelin 2010).This becomes clear on consideration of epistemically underdeterminedcases such as fission scenarios: if there is perfect symmetry between P1 andP2 as well as between P1 and P3, then there are no bodily or psychologicalcriteria available to decide which of the two successors at t2, P2 or P3, isidentical with P1 at t1. Nevertheless, one can clearly conceive of the differ-ence between the case in which P2 is identical with P1 and the case in whichP3 is identical with P1. In the former case P2 would have P1’s first-personperspective, whereas in the latter case P3 would have it. There is, accordingto this argument, a conceptual difference between bodily and psychologicalrelations and identity. As a human person, each one of us is able to take theperspective of P1 and from this viewpoint she knows, independently ofconsiderations about the bodily and psychological relations obtainingbetween P1 and P2 and P1 and P3, what must be the case for herself as P1to be identical to P3 in contrast to P2 (or vice versa). What is claimed is thatone has a clear conceptual grasp of personal identity independently of one’sknowledge of psychological and bodily relations.

Nida-Rümelin corroborates her conceptual reflections by arguing thatone cannot reasonably give up this understanding of personal identity, sinceit is deeply rooted in our capacity for first-person thought. She argues thatthere is good reason to interpret the findings of this conceptual analysis in astraightforwardly realist sense – as long as there are no strong reasons against


it. There is a difference between bodily and psychological criteria forpersonal identity and personal identity itself.Although these epistemic arguments resemble each other, they are dis-

tinct: the first claims that it is metaphysically possible for a person to have atotally different psychological and bodily composition and still be herself(and the other way around). The second argument says that, even inepistemically underdetermined cases, we have a clear conception of whatmust be the case for there to be a definite identity relation.A third type of argument for the simple view is ontological (Lowe 1988;

Lowe 1994). It aims to show that persons are simple individual substances,and that as such they do not have proper parts. Entities without properparts cannot have non-circular identity conditions, for then the latterwould have to refer to the identity conditions of those entities constitutingthe entity in question. This would imply, however, that the entity hasparts and as such would not be simple. Without delving into the details ofthe argument for the simplicity of persons (e.g. Lowe 2000, pp. 15–21;Barnett 2010, pp. 161–74), it suffices to note for present purposes that,once one accepts that persons have no parts, it seems plausible to assumethat their identity conditions cannot be spelled out in terms different frompersonal identity itself.

conc lu s i on

The section on the simple view completes our presentation of the contem-porary discussion about personal identity over time, a relation for whichmanifold approaches try to account. It is mostly taken for granted thatpersonal identity consists in simpler relations of whatever kind, and that themajor task is to specify in detail when personal identity obtains and when itdoes not. Against the background of this discussion, this volume tackles themore fundamental question first − whether personal identity is analyzableinto simpler relations at all. Thus Personal Identity: Complex or Simple?contributes to a better understanding of what it means for human personsto be identical over time.

Introduction 17

part i

Framing the question

chapter 1

Chitchat on personal identityDavid Barnett

“Kid looks like you,” says Jitney.“Kid looks like you,” says Cletus.Jitney and her grown twin brother, Cletus, are cleaning out their mother’s

attic. Cletus has found a photograph of a child with a squirrel in one hand, ameatball in the other, and a nametag that reads “Kid.” Cletus and Jitneymull over the photo from the comfort of two ragtag armchairs.“I vaguely remember one of us being called ‘Kid’,” says Jitney.“I might be Kid,” says Cletus.“Kid’s body was different from yours, Cletus. Kid’s body was small and

youthful, with flawless skin. Your body is big and old, with spots.”“I might be Kid.”“Kid’s mind was youthful and innocent. Kid lived in the moment, with

thoughts of slingshots and bubblegum. You live in the past and future.”“I might be Kid, Jitney.”“You might be, Cletus. After all, you’re just that kind of a guy.”“What do you mean by that?”“You’re the sort of thing that could survive significant physical and

mental changes.”“So are you, Jitney.”“Thanks, Cletus.”“Just how significant do you reckon these changes could be?”“That depends on just what sort of thing you are, Cletus.”“How about one of them rapid, complete, body swaps, like in the movie

Freaky Friday? Could I survive one of them?”“Not if you are your body. And not if you are one of its organs – say, its

brain. If you’re either of those sorts of things, you could not survive havingone living body, with one set of organs, one moment, and a numericallydistinct living body, with an entirely different set of organs, the next.”

21

“So what sort of thing am I?”“That’s no easy question, Cletus.”“I’ve got all day and plenty of gin. I’ll think real hard.”“A full day of thinking by youmight not be enough. In fact, an eternity of

thinking by someone even smarter than you might not do.”“Why’s that?”“Because the question of your nature might not be the sort of question

that can be answered by thinking alone. Take one candidate answer: thatyou are a human brain. Thinking alone can’t even verify that there existsuch things as brains, so it surely can’t verify that you are a brain. If you’re abrain, the only way you can know that you’re a brain is to get out of thatchair and take a look in that thick skull of yours.”

“Already done that.”“And?”“Doctors say there’s a brain.”“What else did they say?”“Doctors say what the brain does seems a pretty good sign of what I’m

about to do, what I’m thinking and what I’m feeling.”“Well, then, maybe you are that brain.”“Maybe I’m just a prisoner of that brain, made to be how the brain makes

me be.”“Did the doctors find anything else in that skull?”“Nope.”“Then maybe you’re that brain.”“Maybe I’m invisible.”“Right. And maybe you’re a squad of cheerleaders.”“I’m not that kind of a guy, Jitney.”“I was being sarcastic. Is it any less absurd that you might be invisible

than that you might be a squad of cheerleaders?”“My being invisible may seem unlikely, but it doesn’t seem impossible.

My being a squad of cheerleaders seems unlikely because it seemsimpossible.”

“Look at you, Cletus, a real philosopher, making distinctions.”“And look at us, Jitney: sittin’ here, sippin’ on gin, thinkin’ hard, and

we’re two steps closer to knowing what I am.”“How’s that?”“As you said, I’m a real philosopher, a distinction maker, a thinking

thing. That’s one. And by your own sarcastic tongue, whatever the doctorssay, I’m no squad of cheerleaders. That’s two.”

22 david barnett

“Fine. You’re a thinker and you’re no squad of cheerleaders. But that leavesmore than a few options open, including your being a brain – and you’re notgoing to settle whether you’re a brain without listening to the doctors.”“Maybe the doctors are wrong. Maybe that oversized camera of theirs is

broke, and inside my head is something other than a brain.”“Right. And maybe there are no such things as doctors. Maybe your

senses have been deceiving you your entire life. Maybe you are, and alwayshave been, dreaming. Maybe I’m just a figment of your imagination.”“Good point, Jitney. You might be that kind of a gal. And I just might be

that kind of a guy.”“Sarcasm, Cletus. While it may be less likely that your senses have been

deceiving you your entire life than that your doctors are deeply mistakenabout what’s in your head, neither idea is worth taking seriously – both areabsurd.”“You mean impossible?”“No. I mean very unlikely.”“Then we’ll throw ’em together with the idea of my being invisible; all

three seem possible, but unlikely. And we’ll keep ’em separate from the idea ofmy being a squad of cheerleaders, which seems impossible and thus unlikely.Now, from these chairs, with this gin, let’s get thinking. Let’s take everypossibility seriously. Let’s not trust the doctors or even our sensations, past orpresent. We’ll forget about what’s actually going on out there and think hardon what is and isn’t possible. Then maybe – just maybe – we’ll narrow downthe range of things I might be, so that we know somethin’ worth knowingabout the sorts of changes I could survive.”“But Cletus, just as things might not be as they seem with your sensory

experience, they might not be as they seem with your intellect: what seemsto you to be possible might be impossible, and what seems to you to beimpossible might be possible.”“Fair enough.”“So why trust your intellectual seemings but not your sensory

experience?”“Because my sensory experience takes a back seat to my intellectual seem-

ings. No sensory experience is going to convince me that I am a bunch ofcheerleaders. Why? Because it seems impossible, in the first place, for me to bea bunch of cheerleaders. So what seems possible or impossible limits what, bythe lights of my sensory experience, seems actual.”“Are you claiming that what you can learn from your armchair, through

hard thinking about what’s possible, limits what you can learn from yoursensory experience about what is actual, but not vice versa?”

Chitchat on personal identity 23

“I am.”“And what if the best explanation of all your sensory experience, past

and present, is that your seemings about what is and isn’t possible areinaccurate?”

“Then I’d be in a bit of a pickle.”“Then we’re back to where we started, Cletus. Thinking alone may not

be enough to answer the question of what you are. Thinking alone maydeliver plenty of seemings about what’s possible. But, first of all, what’spossible, from the perspective of that armchair, regarding your nature, maybe too broad to bear in any meaningful way on what sort of thing you areand what sorts of changes you could survive – so you may need to call onyour sensory experience to narrow down the possibilities. Second, yourarmchair seemings about what’s possible may ultimately be underminedby past, present and future sensory experience. So before you draw any hardconclusions, you’d better get out of that ragtag chair and check in with yourfive senses.”

“But if we agree to pay heed to our sensory experience, then we oughtto do a whole lot of poking around in the world – something I ain’t up fortoday.”

“Me neither.”“So let’s have a gin-and-thinkin’ day today, and see where it gets us.

Tomorrow we’ll start checking our results against as much sensory experi-ence as we can gather up.”

“Deal. Today we’ll stick with the following question: if things are as theyseem to you, from a perspective in which you’ve set aside all knowledge thatdepends on the past or present accuracy of your sensory experience, regard-ing what is and isn’t possible, thenwhat sort of thing might you be and whatsorts of changes could you survive?”

“Bingo. I’ll start by saying that I’m not just any old thinking thing; I’m aconscious being, the sort that has sensory experience, including the taste ofsweet gin. I’m that kind of a guy.”

“Not so fast, Cletus. You agreed to set aside the deliverances of yoursensory experience. So on what grounds do you know that you have sensoryexperience?”

“I agreed to not trust the accuracy of my sensory experience. I didn’t agreeto ignore the fact that I’m having sensory experience. That much I knowfrom the chair.”

“Fine. I’ll grant that you’re a thinking, conscious, being. Now, let’sreturn to the idea of a rapid, complete, body swap – Freaky Friday style.What’s your feeling on it?”

24 david barnett

“Seems possible. Seems that a conscious being – a being of my stripe –could have one living body, with one set of organs, one moment, and anumerically distinct living body, with an entirely different set of organs, thenext. So I’ll go with yes, it’s possible for me to survive such a swap. I’m thatkind of a guy.”“Slow down, Cletus. It seems to you that you’re a conscious being, right?”“Check.”“And it seems to you that a conscious being could survive a rapid,

complete, body swap?”“Check.”“Could conscious beings come in more than one stripe?”“I reckon so.”“Could one stripe – say, an embodied angel – be capable of surviving

such a swap, while another – say, a conscious brain – be incapable?”“I reckon so.”“Have you established, from that chair, which stripe you are?”“I reckon not.”“Then you better slow down. Granted that your armchair seemings are

accurate, all you know so far is that it’s possible for there to be a sort ofconscious being who could survive a rapid, complete, body swap. You don’tknow that you are that sort of being.”“Point taken, Jitney.”“Let’s turn next to brain transplants. Imagine there are two conscious

beings, Elrod and Cooter. Each has a relatively normal human body. Elrodhas two green eyes. Cooter has one brown eye and one blue. The doctorsknock Elrod and Cooter clean unconscious, scoop their brains out, andswap each one back into the other’s skull. Elrod’s original brain is now partof a body withmixed eyes; Cooter’s original brain is now part of a body withgreen eyes. Cooter wakes up and looks in a mirror. What color eyes doesCooter have?”“Green.”“Why do you say that?”“In my experience, people tend not to stray far from their brains.”“You mean in your sensory experience?”“I do.”“So, based on the deliverances of your sensory experiences, you would say

that it’s very likely that Cooter goes with his brain?”“Yep.”“And, for present purposes, you have agreed to set aside the deliverances

of your sensory experiences?”


“Yep.”“What’s wrong with you, Cletus?”“I’m drunk.”“Get it together. I’ll forget your first answer and let you try again. For all I’ve

said, might Cooter be the sort of thing that could wake up with green eyes?”“Yep.”“And might Cooter be the sort of thing that could wake up with mixed

eyes?”“Yep. As I said already, it’s possible for there to be a conscious being that

could survive a rapid, complete, body swap. Surely, then, it’s possible forthere to be a conscious being that could survive a rapid, partial, body swap –a mere brain swap. For all you’ve said, Cooter might be that kind of a guy.”

“That’s better, Cletus. Now let’s talk about the celebrated brain chop fora moment. What are your feelings on it?”

“Tell me more.”“Earlene is a conscious being with a relatively normal red-headed human

body. Earlene’s not feelin’ herself today. Doctors say she’s got a stormybrain, so they chop it in two, right down the middle. One half they put intoa new blond-headed body; the other into a new brunette-headed body.Someone wakes up with blond hair; someone wakes up with brunette hair.Which one’s Earlene?”

“Which one acts like Earlene?”“Both.”“Which one claims to remember things that happened to Earlene?”“Both.”“And what about from the inside? Which one seems to remember things

that happened to Earlene? And which one feels like Earlene?”“Both.”“Then it’s impossible to know.”“Agreed.”“So I’ll go with neither : Earlene got split into oblivion.”“Where’d that come from, Cletus? Suppose I’m about to flip a fair coin.

Will it land heads or tails?”“Impossible to know.”“And by your lights, it follows that it won’t land heads or tails. But of

course it will.”“Good point, Jitney. Let’s back up. All I know is that, before the chop,

Earlene was a conscious being with a relatively normal human body. I don’tknow her exact relation to that body. Could be she was that body. Could beshe was its brain. Could be she was merely a prisoner of that body and its

26 david barnett

brain, something like an angel with a human body. If she was merely aprisoner, then, after the chop, she could get the blond’s brain; she could getthe brunette’s brain; she could get some other brain; she could be liberatedfrom the material world and sent to Heaven; she could go entirely out ofexistence. For all you’ve said, there’s some sort of conscious being for whichany of these outcomes is possible.”“Sounds right to me. From these chairs, we are in no position to say that

Earlene has the blond’s brain; we’re in no position to say that she has thebrunette’s brain; and we’re in no position to say that she has neither.”“The sweet taste of progress.”“Progress? So far it seems that just about anything goes when it comes to

mixing and matching conscious beings with bodies and their parts – atleast if we leave open the proper type of conscious being that’s up formixing and matching. How this counts as progress toward figuring outwhat sort of thing you are, and what sorts of changes you could survive,I don’t know.”“Feels like progress to me, Jitney. Let’s not lose our momentum.

What’s next?”“Teletransportation. You got a feeling?”“Star Trek style?”“Close. Suppose that Rubyjane is a conscious being with a normal

human body. She steps into the entry module of a teletransporter. Themodule surveys the complete physical state of Rubyjane’s body, destroys thebody, and sends the information at the speed of light to an exit module,where a perfect physical duplicate of the body is constructed out of newmaterial. Out walks a girl. Who is it?”“The girl looks just like Rubyjane from the outside, and she feels just like

Rubyjane from the inside?”“Right.”“Is teletransportation a normal happening where Rubyjane’s from?”“Let’s suppose so.”“Then I suspect her folk would say the girl who walks out is Rubyjane.

And I suspect the girl who walks out would say she’s Rubyjane.”“I suspect so.”“And I suspect you suspect I’ll conclude that the girl isRubyjane. But I’m

reformed, Jitney. I’m not about to make that mistake. For all you’ve said,Rubyjane could be something like an angel with a human body and brain.In which case, when her original body is destroyed, any of a number ofthings is possible: she could pop out of existence; she could continue to existin a disembodied state; or she could take on a new body, including the body


that just stepped out of the exit module. So it’s possible that Rubyjane is thegirl who walks out of the exit module, and it’s possible that she’s not.”

“Seems right to me.”“Could be this gin, but I’m pretty sure I taste another wave of progress.”“Speak for yourself, Cletus. I’m about to call it quits. From these arm-

chairs, it seems that just about anything is possible. There could be a sort ofconscious being who could survive a rapid, complete, body swap; a sort whocould survive a rapid, partial, body swap; a sort who could survive a brainchop; and a sort who could survive teletransportation. But all this isconsistent with there being a sort of conscious being who could not surviveany of these events. And we’re no closer to knowing which sort you are, sowe’re no closer to answering your question.”

“Knowin’ what we can’t know is knowin’ something, Jitney. We knowthat, from the armchair, we can’t know who goes where in a brain swap, abrain chop, or a teletransportation, at least not without knowin’more aboutthe sorts of beings involved in such antics. And we know that we can’t, fromthe chair, know that rapid, complete, body swaps are beyond my repertoire.That’s a lot of knowin’, in my books.”

“Whatever tickles your fancy. Let’s shift gears and talk about mentalchanges. What’s your take on complete amnesia?”

“Do I know you?”“Let’s say that Trixie-Lynn has a normal human body. She swings a

baseball bat at a basketball. The bat bounces back into her face, knockingher clean out. Someone wakes up dizzy, with a purple knot between her eyes.She doesn’t know where she is or how she got there. She has no memory ofanything that ever happened to Trixie-Lynn. Who is she?”

“If I had to bet, I’d say she’s Trixie-Lynn. But this bet would be based,once more, on the fact that folks don’t usually wander far from their brains –a fact I know bymy past sensory experience, and I’mnot about to make thatmistake again. So all bets are off. From the chair, given only what you’vesaid, all I can say is that the purple-headed galmight be Trixie-Lynn, and shemight be someone else – some lucky lady who happened into Trixie-Lynn’sformer body.”

“And what if I told you that this gal acts nothing like Trixie-Lynn actedbefore the accident? This gal is polite and generous. Trixie-Lynn was rudeand selfish. This gal is witty, creative, lively, and engaging. Trixie-Lynn wasdull, boring, and aloof. This gal excels at math and baseball. Trixie-Lynncouldn’t add two and two, and didn’t know the difference between abaseball and basketball. In fact, Trixie-Lynn’s friends and family say that,ever since the accident, Trixie-Lynn is a totally different person.”

28 david barnett

“Are they drunk?”“No.”“And yet they think that Trixie-Lynn was one person before the accident

and a different person after the accident?”“Be charitable, Cletus. They think that Trixie-Lynn is, and always has

been, identical to one and only one person: namely, herself. However, theybelieve that, qualitatively speaking, Trixie-Lynn is radically different sincethe accident: her memories, personality and abilities are nothing like whatthey were before the accident. This is upsetting to her friends and family, allof whom formed close relationships with Trixie-Lynn based largely on theirappreciation of her personality before the accident.”“You mean their appreciation of who Trixie-Lynn was?”“Loosely speaking – or, rather, qualitatively speaking – yes: I mean their

appreciation of the family of mental characteristics by which they – andTrixie-Lynn herself – distinguished Trixie-Lynn from other people inmeaningful ways, and on the basis of which they formed intimate relation-ships with Trixie-Lynn.”“For all practical purposes, Trixie-Lynn is dead to them.”“That’s right. All that they valued in their connections with Trixie-Lynn

is gone. And all that Trixie-Lynn valued in her connections to herself atearlier times is gone. For all intents and purposes, Trixie-Lynn is dead.”“God bless.”“God bless, Cletus.”“She didn’t have to die.”“What?”“Sorry, I meant to say that she didn’t have to go changing like that.”“Sounds like you agree with her friends and family that, strictly speaking,

Trixie-Lynn survived the accident with her original body, and that it’s notthe case that a numerically distinct conscious being has taken over her body.Numerically speaking, the person after the accident is the same as before.Qualitatively, the person is very different. Is that your view?”“I suppose that would be my view if I were a member of the scenario who

witnessed the events, or if I were considering the scenario with the aid ofmy past sensory experience. But I am no such person. To repeat: from thischair, given all you’ve said, it seems possible that the purple-headed gal isTrixie-Lynn, and it seems possible that she’s not.Without the aid of sensoryexperience, the facts of the story don’t tip the scale one way or the other.”“Then consider a different story, Cletus. This one starts like the last:

Trixie-Lynn whops herself unconscious with a bat, and thereafter the personwith Trixie-Lynn’s original body has a very different personality from the one


Trixie-Lynn had prior to the whopping. What’s more, at the time of thewhopping, Trixie-Lynn’s mother is across town, in a wicker rocker, on herfront porch, in her pajamas, snoring like a walrus.”

“Doesn’t change my view.”“There’s more. Some kids are playing rock-ball.”“Like baseball with rocks?”“Yep. One smacks an egg-sized rock straight between the mother’s eyes.

The snoring stops; the rocking stops; the dreaming stops; the breathingstops – everything stops. A couple of minutes pass. Then, just as Trixie-Lynn’s body comes back to life, so too does the mother’s. Only, mentallyspeaking, the person with the mother’s body doesn’t resemble the motherprior to the event. Her personality is totally different. And yet it’s distinctivelyrecognizable to her friends and family. Guess what they recognize it as?”

“No!”“Yep. Trixie-Lynn’s personality, prior to her accident.”“Freaky.”“The gal with the mother’s body is now rude, selfish, dull, boring, aloof,

mathematically inept and unable to distinguish a baseball from a basketball.Did I mention that the gal with Trixie-Lynn’s body has the personality themother had prior to the accident? And we’re not even to the freaky part yet.”

“Let’s get our freak on!”“Alright, Cletus: the gal with the mother’s body appears to remember all

sorts of events that Trixie-Lynn experienced prior to her accident. And theseappearances ain’t but a stone’s throw from the truth.”

“No!”“They’re as accurate as anyone’s.”“And the gal in Trixie-Lynn’s body? Does she seem to remember what

the mother experienced prior to her accident?”“Yep. Last thing she seems to remember from before the accident is

falling asleep in a rocking chair on a porch.”“Who do these gals think they are?”“Trixie-Lynn and her mother. That’s who. They think they’ve swapped

bodies, Freaky Friday style, only their apparent swap happened on a Sunday,not a Friday.”

“Freaky Sunday. And what do their friends and family think?”“Most are body swappists. But a small group of dissenters maintain that

body swapping is impossible and unscientific.”“How do they explain the fact that each gal seems to remember things

that really did happen to the person with the other body prior to theaccidents?”

30 david barnett

“Coincidence. Remarkable coincidence.”“So they deny that these are genuine memories?”“Yep. They say that each gal appears to remember experiencing all sorts

of events that she never experienced. By remarkable coincidence, for eachapparent memory, the other gal did in fact experience an event quite like theone apparently remembered. Trixie-Lynn seems to remember falling asleepin a rocking chair just before the accident, even though she never experi-enced this. By remarkable coincidence, her mother did experience an eventof just this sort.”“Seems these dissenters have some explaining to do.”“They do. But so do the body swappists. How in the world – the natural

world – do two people just swap bodies like that?”“Maybe the people are like angels – non-physical souls – with human

bodies. Then there wouldn’t be much preventing such a swap, exceptmaybe some angel-physical laws of nature.”“Or maybe they are like restaurants. Restaurants can swap buildings

without a hitch. One day Denny’s is here and Taco Bell is there; the nextTaco Bell is here and Denny’s is there. No big deal. No souls. No mystery.Everything is natural. Except the meat.”“Mmm . . . meat . . .”“Focus, Cletus. Are you a swappist on this one?”“If I were a member of the scenario who witnessed the events, or if I were

considering the scenario with the aid of my past sensory experience, thenthere’s a good chance I’d be a swappist, as the swappin’ hypothesis wouldexplain a lot. On the other hand, I’d be open to a remarkable coincidencehypothesis, with a no-angels clause, as it sure would simplify things. I mighteven want to hear more about your people are like restaurants idea, since itmight make for a simple story too.”“But you’re not a member of the scenario, and you’re not considering the

scenario with the aid of your past sensory experience. You’re a chair-bound,gin-soaked, amateur thinker. So what’s it gonna be?”“Same as usual: from this chair, given all you’ve said, it seems possible

that the purple-headed gal is Trixie-Lynn; it seems possible that she’sTrixie-Lynn’s mother; and it seems possible that she’s neither. The samegoes for the gal with the older body: could be the mother; could beTrixie-Lynn; could be neither.”“Love this progress.”“Feels good, doesn’t it?”“Sarcasm, Cletus. You’re no closer to knowing what you are and what

sorts of changes you could survive than when we started. The most you


know, even granting the accuracy of your armchair seemings, is that therecould be a conscious being who could survive a rapid, complete, body swap;one who could survive a rapid, partial, body swap; one who could survive abrain chop; one who could survive a teletransportation; and one who couldsurvive a rapid, complete, personality change. But all this is consistent withthere being a sort of conscious being who could not survive these sorts ofevents. And we’re no closer to knowing which sort you are, so we’re no closerto answering your question.”

“Slow down, Jitney. What sort of conscious being could not survive arapid, complete, personality change?”

“Suppose Trixie-Lynn is nothing but a collection of her beliefs, sensa-tions, memories and personality traits. If that’s all she is, then what sense isthere to the idea of her surviving a rapid, complete, change in beliefs,sensations, memories and personality traits?”

“But you said Trixie-Lynn was a conscious being.”“I did.”“How in the world could a collection of features be a conscious being?

How could a collection of sensations itself experience the sensation of gettinghit in the face with a baseball bat?”

“Just playing Devil’s advocate, Cletus. I agree that the idea seems ratherpreposterous on its surface, but maybe it’s a possibility nonetheless, inwhich case, you might be that sort of a guy. And if you are that sort of aguy, then perhaps you’re not the sort who could survive a rapid, complete,personality change.”

“I’m startin’ to see what you mean about the lack of real progress. Weneed to narrow the possibilities, Jitney.”

“You ready to get out of that chair and see what you’re made of?”“No, ma’am.”“Then how are we going to narrow the possibilities?”“Let’s quit focusing on the positive. It’s overrated. Let’s focus on the

negative. Rather than asking what’s possible, for some conscious being orother, let’s ask what’s impossible, for any conscious being.”

“Sounds like six of one, half-dozen of the other, to me.”“What I mean, Jitney, is that we should stop assuming that, when it comes

to conscious beings, anything goes: it’s possible for them to be human bodies,human brains, non-physical souls, restaurant-like things, or even collectionsof features. Maybe not anything goes. Maybe it’s impossible for them to berestaurant-like things, souls or collections of features.”

“Instead of asking what’s possible regarding the changes a conscious beingcould survive, on the assumption that just about anything is possible

32 david barnett

regarding the nature of conscious beings, you want to back up and questionthat assumption, by asking what sorts of things no conscious being could be.Do I have it right, Cletus?”“Bingo. If it turns out to be impossible for any conscious being to be, say,

a restaurant-like thing, that’s one less possibility for this conscious being.”“So what sorts of things could no conscious being be?”“Let’s start with a squad of cheerleaders. No conscious being could be

one of those. I don’t need to know what stripe of conscious being I am toknow that I’m no squad of cheerleaders. From this chair, by pure thinking,I know I’m not that kind of a guy.”“You’re a true visionary.”“So you agree that it’s impossible for a conscious being to be a squad of

cheerleaders?”“I agree; I’m just not impressed.”“Let me tell you a story, Jitney. There’s a squad of cheerleaders. The end.”“Now I’m impressed.”“Given only what I’ve said, can you know, from your chair, without the

aid of any past or present sensory experience, that you’re not that squad ofcheerleaders?”“I can.”“But I haven’t even said what team they’re cheering for.”“Doesn’t matter, Cletus. I’m not their squad.”“I haven’t told you what color their uniforms are.”“Doesn’t matter.”“But there’s so much I haven’t told you, Jitney. Maybe you should hold

your judgment until you find out whether they have human bodies.”“Doesn’t matter. I’m not their squad.”“You don’t know their number, size or location.”“Doesn’t matter, Cletus.”“For all I’ve said, they could number in the billions.”“Fine.”“They could be really small.”“Fine.”“They could live in your skull.”“Fine.”“They could have spooky-shaped bodies.”“Fine.”“They could be caught up with one another in a twisted, complex, web of

relations . . .”“I don’t care, Cletus. I’m not their squad!”


“. . . while they’re dead!”“Now that’s different. Why didn’t you ask me in the first place whether

I might be a squad of billions of miniature, talented, zombie cheerleaders,living happily inside my skull?”

“If they’re happy, they’re not dead. And if they’re dead, they’re nothappy, Jitney.”

“You’re missing the point. There’s no need for you to fill in any details ofyour original story. You’re wasting your breath. All I need to know is thatthere is a squad of cheerleaders. I don’t need any more details to know thatI’m not that squad.”

“That’s progress.”“How so?”“If you can know that you’re not that squad of cheerleaders without

knowing the number, location, or skillset of the squad, without knowingwhether its members are dead or alive, miniature or gigantic, and withoutknowing what their bodies are like, then surely the fact that they arecheerleaders is irrelevant to what you know.”

“Seems right. Could be a squad of child soldiers, matchbox cars ordead ants for all I care. I’m no squad of anything, Cletus. I’m not that kindof a gal.”

“Then you’re no squad of neurons.”“And I’m no squad of sensations, beliefs or personality traits.”“And you’re no squad of cells.”“Or atoms.”“Or living organs.”“Or particles.”“That’s a lot of possibilities out the door.”“I do taste progress, Cletus.”“If a brain is nothin’ but a squad of neurons, caught up with one another

in a twisted, complex, web of relations, then I know from this chair that I’mno brain. And if a human body is nothin’ but a squad of organs or cells,caught up with one another in a twisted, complex, web of relations, thenI know from this chair that I’m no human body.”

“Slow down, Cletus. We agreed from the start that we might be brains orbodies. It certainly seemed possible. Sowhy think the results of our cheerleader-centered investigation are any better off on this score than how things seemedto us originally?”

“Maybe our original seemings were gotten from thinkin’ of brains as biglumps of whatnot, and forgetting the fact that lumps of whatnot are oftennothin’ but squads of tiny whatnots. If we think of brains as squads of tiny

34 david barnett

whatnots, caught up with one another in a twisted, complex, web of relations,those original seemings seem a lot less vivid. Ask yourself, Jitney: does it reallyseem possible that are you a squad of other things?”“It does not, Cletus.”“Hence the progress. Originally, something seemed possible. On reflec-

tion, it no longer does.”“Of course, without paying heed to our sensory experience, we can’t

know what human brains and bodies are made of, or whether they evenexist. What we can know is that, should they exist, there’s a good sense inwhich they are nothin’ but squads of things, caught up with one another in apossibly twisted, possibly complex, web of relations.”“And we know we’re not squads.”“So we know we’re not brains or bodies.”“Or restaurant-like things – squads whose membership can change

rapidly and radically.”“Still, Cletus, there’s a lot we don’t know we’re not.”“Like pure and partless angels. They ain’t squads of nothin’.”“That’s right. We might be partless angels.”“Simple souls.”“Well, some souls are quite fanciful, Cletus.”“I mean simple in composition, not quality.”“Like quarks. We might be quarks.”“What’s that, Jitney?”“A little bitty physical particle.”“Ain’t a squad of other particles?”“Nope. Pure and partless, simple and plain.”“You learned that in science?”“Yes sir.”“No sensory experience allowed.”“Look who’s drunk now. I retract. For all we know from these chairs,

quarks are simple, and for all we know, they’re squads.”“So we might be simple, non-physical, beings – like partless angels – and

we might be simple, physical, beings – like partless physical particles.”“Or so it seems from these chairs.”“Correct. Tomorrow we’ll help ourselves to as much sensory experience

as we can gather up.”“And the next day too.”“And the next.”“And if the best explanation of all that experience is that things just ain’t

as they seem, from these chairs, regarding what is and isn’t possible, then


we’ll begin the hard job of negotiating between what seems actual by oursenses and what seems possible by our intellects.”

“But maybe what seems possible by one line of intellectual thinkin’would seem impossible by another, and vice versa. Maybe if we thought anew line of thought, it would no longer seem impossible for a consciousbeing to be a squad.”

“Maybe, Cletus. But unless you’ve got a good reason for thinking so,I don’t have a good reason for taking this maybe seriously.”

“Think about it this way, Jitney. We’ve only thought about the questionof whether it’s possible for a conscious being to be a squad by one method:asking ourselves whether we might be squads. Our feeling on this questionis a firm no: no matter what sort of conscious being we turn out to be, weain’t no squads. But it’s just a feeling. There’s no argument.”

“Speaking for myself, it’s a very strong feeling.”“But maybe it’s worth hittin’ the question with a different method, from

the chair, to see whether we get a different answer. Instead of askingourselves directly whether we might be squads, let’s ask ourselves whetherthe sort of feature had by every conscious being is the sort of feature thatcould be had by a squad.”

“What sort of feature is that?”“Experiential. Like feeling sad.”“I see where you’re going, Cletus. You like French fries?”“Mmm . . .”“You like McDonald’s French fries?”“Mmm . . .”“You like their taste, don’t you Cletus?”“Mmm . . . taste . . .”“That’s an example of an experiential feature, right?”“Mmm . . . experiential feature . . .”“Wake up, Cletus!”“Right, Jitney. The question is whether a squad is the sort of thing that could

have that sort of feature. Could a squad experience the taste of French fries?”“How do squads get their features?”“From their members, I suppose.”“And how do squads get their features from their members?”“The old-fashioned way?”“Nope. Not by hard work. Squads don’t have to lift a finger to get their

features from their members. They get them automatically.”“The way McDonald’s fries get their golden brown color from that

boiling beef-flavored lard?”

36 david barnett

“Nope. That is automatic in a different sense. The boiling lard causes thefries to change color, without anyone intervening. The squad members’features do not cause the squad to have its features. Rather, they constitutethem. The location of a squad of cheerleaders is constituted by the locationof its members.”“And its weight is constituted by the weight of its members.”“And its shape is constituted by the shapes and spatial relations of its

members.”“If God wants to give a squad a feature, He can only do so by giving other

things – its members – certain features.”“What about the feature of being named ‘Ed’? Couldn’t God name the

squad ‘Ed’ without laying a finger on any of the squad’s members?”“Good point, Jitney. I’m talking about inner features, the ones that give

the squad its inner glow.”“Let’s call those intrinsic features, Cletus.”“Fine. If God wants to give a squad an intrinsic feature, He can only do so

by giving other things – its members – certain features.”“Agreed.”“But conscious experience is an intrinsic feature. And giving other things

features is no way to give someone the gift of conscious experience.”“Unless doing so causes that person to have the experience.”“Bingo. If you experience the taste ofMcDonald’s French fries, you do not

qualify as having this experience by virtue of the fact that other things in theworld have various features. This factmight cause you to have the experience –it might bring about the separate event of your having the experience – but itdoesn’t itself qualify you as having the experience. If you experience the tasteof McDonald’s French fries, you qualify as having the experience directly.”“Agreed, Cletus. Other things being the way they are may cause me to

have my experience, but it can’t constitute it. Other things having variousfeatures and standing in various relations is one thing . . .”“Your having an experience is another.”“Agreed.”“So this confirms our original feeling that no conscious being could be a

squad of anything. Squads earn their intrinsic features only indirectly: theremust exist a number of things, namely, the members of the squad, none ofwhich is individually identical to the squad, whose features jointly con-stitute the intrinsic features of the squad.”“You sure do sound smart, Cletus.”“By contrast, conscious beings have at least some of their intrinsic features,

namely, their experiential features, only directly: they have conscious


experience, not by virtue of other things having various features, but simplyby virtue of directly experiencing sensations, such as the taste of French fries.”

“If God wants to give me the taste of French fries, He can do it directly.”“No need for Him to dilly dally around with other things. To be sure, His

dilly-dallying around with other things could cause a distinct event, namely,your experiencing the taste of French fries. But it could not constitute it.”

“Or so it seems from the chair.”“Hallelujah, Jitney.”“Let’s recap our day’s fruit.”“We agreed to set aside the deliverances of our sensory experience, past

and present.”“But not the sensory experience itself.”“Correct. You then took me on a fantastical journey.”“I told you some stories, Cletus.”“Right. And, frommy chair, by hard thinking alone, given only what you

said, I asked myself what seemed possible and impossible.”“And what seemed possible?”“Seemed possible for there to be a conscious being who could survive a

rapid, complete, body swap; one who could survive a rapid, partial, bodyswap; one who could survive a brain chop; one who could survive ateletransportation; and one who could survive a rapid, complete, personalitychange.”

“I wasn’t impressed.”“Because you said this was consistent with there being a sort of conscious

being who could not survive these sorts of events, and that we were no closerto knowing which sort of conscious being I was, so we were no closer toanswering my original question: what sort of thing am I and what sorts ofchanges could I survive.”

“At this point, nearly anything seemed possible regarding the nature ofconscious beings. It seemed that a conscious being could come in just aboutany stripe.”

“A human body.”“A brain.”“A soul.”“A restaurant-like thing.”“Even a collection of features.”“Those were some strange stripes, Cletus.”“Then I had me an idea.”“I didn’t think you had it in you, Cletus. I really didn’t.”“We’d think real hard on what sorts of thing no conscious being could be.”

38 david barnett

“A squad of cheerleaders.”“A squad of dead cheerleaders.”“A squad of tiny dead cheerleaders.”“Caught up in a twisted, complex, web of relations with one another.”“And with the outside world.”“A squad of molecules.”“A squad of cells.”“A squad of features.”“A squad of organs.”“A squad of anything.”“A squad-like thing, like a human body.”“Or a brain.”“Or a restaurant-like thing.”“Then we surveyed the damage. Which possibilities were left, Jitney?”“A non-physical simple soul, like a partless angel.”“Or a physical simple particle, like maybe a quark.”“Rednecks are simple.”“And snobs from Boulder.”“All conscious beings are simple.”“I’m simple, Jitney! I’m that kind of a guy.”“Then we confirmed this chair-bound feeling of simplicity with an

argument.”“Squads get their intrinsic features indirectly, from the features of other

things.”“Conscious beings get at least some of their intrinsic features directly.”“They experience their sensations directly.”“No other way to experience a sensation.”“So, I’m simple. But am I physical?”“Can’t know from the chair.”“I suppose not, Jitney.”“And what sorts of changes could you survive?”“If I’m not physical, it’s clear that pretty much anything goes: I could

survive a rapid, complete, body swap; a rapid, partial, body swap; a brainchop; a teletransportation; and a rapid, complete, personality change. Atleast this is so if we set aside any physical to non-physical laws of nature.”“You’re an impressive guy, Cletus. And what if you are physical?”“Then I suppose we’d need to pause to think about what it would be to

have a body in the first place.”“If you’re not physical, then you have a particular body just in case you’re

connected with that body in the right sort of intimate way.”


“It just feels like my body.”“Perhaps.”“It feels under my authority.”“Perhaps.”“And yet, I feel like its prisoner: damage to the body is damage to my

mind, whether I like it or not.”“Perhaps, Cletus. It’s hard to say what exactly that intimate connection is.”“And if I’m a simple physical particle?”“Then, in some sense, you are your body; you’re a tiny, simple, physical,

body of one.”“That’s doubletalk, Jitney. Even if I’m a simple physical particle – a

physical body of one – I could still have a larger body – a body of many – inthe sense we just spoke about: I could have that special intimate connectionwith a large body of matter – say, a human body.”

“And would you be a component of that body?”“Maybe, maybe not. Either way, I could have that intimate connection.”“Seems you could, Cletus.”“And since I wouldn’t be identical to that larger body, I’d be capable of

having such intimate connections with other bodies, too, one after another.”“You’re that kind of a guy, Cletus.”“Whatever my stripe – simple physical or simple non-physical – I could

survive a rapid, complete, body swap; a rapid, partial, body swap; a brainchop; a teletransportation; and a rapid, complete, personality change.”

“Or so it seems from the chair.”“So it does, Jitney.”“Tomorrow we’ll take our philosopher hats off and put our mosquito

suits on.”“We’ll get our hands dirty.”“We’ll look under rocks.”“And scalps.”“And if we find that our sensations don’t jibe with today’s result . . .”“If their best explanation is that things ain’t what they seem from the

chair . . .”“Then we’ll start negotiating between our sensory and our intellectual

experience.”“And if those negotiations end with the verdict that conscious beings like

you andme are in fact squads of tiny, dead, strangely bodied cheerleaders . . .”“Or squads of tiny, strangely bodied squads of even tinier more-strangely

bodied things . . .”“Or even squads of anything . . .”

40 david barnett

“Then I’ll be darned, Cletus.”“And I’ll be lookin’ for a new theory of the sorts of changes I could

survive.”“You may not need to look too far. If negotiations end with the verdict

that you’re a human body, then you’ll know that you could survive what-ever changes a human body could survive. And if they end with the verdictthat you’re a human brain, then you’ll know that you could survive what-ever changes a human brain could survive.”“But maybe negotiations will take a turn for the bizarre. Maybe respect

for our sensory experience will lead us to conclude that we’re squads, andrespect for our intellectual experience will lead us to maintain that it’spossible for us to survive rapid complete body swaps, rapid partial bodyswaps, brain chops and teletransportations.”“What sort of a squad could survive those things?”“Funny you ask, Jitney. A four-dimensional, choppy, space-time worm,

for one.”“I don’t even need to know what that is to know that negotiations

shouldn’t take that turn. Think about it, Cletus. If respect for our sensoryexperience leads us to conclude that we’re squads, then it ought to lead us toconclude that things ain’t how they seem from the chair, regarding what isand isn’t possible. And if they lead us there, then we ought not have anyrespect for our positive intellectual feelings on the possibilities of bodyswaps, brain chops and teletransportation.”“Sounds like you never learned the art of negotiation, Jitney. It’s all about

give and take. If respect for our sensory experience leads us to conclude thatwe’re squads, then we ought not have any respect for those intellectual feelingsthat don’t fit with our being squads, such as the feeling that it’s impossiblefor any conscious being to be a squad of tiny, dead cheerleaders. But there’sno reason to throw the baby out with the bathwater. We could still respectthose intellectual feelings that do fit with our being squads, such as thefeeling that it’s possible for us to survive body swaps, brain chops andteletransportation.”“Sounds like you never learned the art of housekeeping, Cletus. If the

baby is a perennial source of dirty water, there’s only one way to clean thewater: throw out the baby.”“Sounds cruel.”“The baby is the conception that you have of yourself and all other

conscious beings from the armchair, prior to testing it against your sensoryexperience. Capiche?”“Mmm . . . capiche . . .”


“That conception is of a simple thing, something that is not composed ofother things. That conception is what generated your feeling, in the firstplace, that no conscious being is a squad of anything, and, in the second place,that you’re the sort of thing that could survive a rapid complete body swap, arapid partial body swap, a brain chop, teletransportation, or a radical, com-plete, personality change.”

“Agreed.”“So any reason you have to doubt that conception is a reason to doubt

those feelings generated by it. If you get out of that chair and find that yoursensory experience provides you with a strong enough reason to reject theidea that you’re simple, then you’re obligated to stop listening to any of yourfeelings based on your prior acceptance of that idea, including your feelingsthat you could survive these changes. Capiche?”

“Yes, please.”“If you end up throwing out the dirty bathwater – the idea that no

conscious being is a squad of anything – then you’d better throw out thefilthy baby that’s dirtying the water – namely, your conception of consciousbeings as simple – and all the dirty water that baby is generating, includingyour feelings that you could survive those changes. How’s that fornegotiation?”

“Respect, Jitney. If we ever get to the point of negotiating away ourarmchair conception of ourselves as simple, we’ll toss out all the feelingsgenerated by that conception. Which is not to say that we’ll positively rejectthe deliverances of those feelings, but only that we’ll not count thosefeelings as reasons to accept their deliverances.”

“If it comes to that point, and the question arises which squad of thingsare we, we’ll address that question without giving the slightest voice to thatdirty baby.”

“I don’t trust that baby.”“So far, this is all hypothetical, Cletus. So far we’ve not found a semblance

of a reason to turn our backs on the baby. We’re still in our chairs, settingaside sensory experience. It would take a heap of powerful and unexpectedsensory experience to bring us to the negotiating table in the first place. Andeven then, there’d be a ways to go before throwing out the baby.”

“It just doesn’t seem possible that we could get to that place, Jitney. Howin the world could I be a squad?”

“In some sense it seems possible; in some sense it doesn’t. Surely there’s achance our armchair seemings are off the mark, just as there’s a chanceyou’ve been dreaming your entire life. In that sense, it seems possible thatthe baby is dirty and ought to go. In another sense, I agree, it doesn’t seem

42 david barnett

possible. That’s the sense in which it seems that no conscious being, of anystripe, could be identical to a squad of things.”“I could go for a squad of golden French fries right now.”“Sounds like something Kid would have said.”“I might be Kid, Jitney.”“You might be, Cletus.”1

1 For helpful comments on an earlier draft of this chapter, I am grateful to Chris Heathwood, BradMonton, Michaela Markham McSweeney, Bob Pasnau, and Alex Skiles. I am also grateful to theparticipants of my Spring 2010 Graduate Seminar on Identity for helping me as I developed somecentral ideas for this chapter.


chapter 2

In search of the simple viewEric T. Olson

s im p l e and comp l e x v i ews

We tell our students that accounts of personal identity over time fall intotwo broad categories. First there are “complex” views according to whichpersonal identity consists in something else: when there is a fact about ouridentity or non-identity, there is some other, deeper fact that underlies andis responsible for it. There are familiar debates about what these deeper factsare – whether they have to do entirely with brute physical continuity, forexample, or with some sort of psychological continuity, perhaps with aphysical constraint. The second category, the “simple” view, rejects all this,and denies that our persistence or identity over time consists in anything.A certain past or future being is or is not you, and that is all there is to besaid. Nothing furthermakes it you or not you. The identity or non-identity ofan earlier person and a later one is already as deep as it gets. This taxonomy isshown in Figure 1.

Advocates of simple and complex views are supposed to disagree aboutsomething more fundamental than those who differ over what personalidentity consists in, namely, about whether it consists in anything at all.And because the simple view just denies what complex views assert, the twocategories must be both exclusive and exhaustive: any account of personalidentity over time must be either simple or complex and not both.

If this is right, it ought to be possible to say what the two sides disagreeabout.What proposition is it that friends of complex views accept and friendsof the simple view deny? What do you have to believe in order to accept acomplex view, and what belief (or lack of belief) characterizes the simple view?This question is surprisingly hard to answer. The trouble is not merely thatthere are hard cases – views of personal identity that resist classification aseither simple or complex – but that no answer even gets all the easy cases right.And those proposals that best approximate the traditional boundary betweensimple and complex lack the interest and importance that the boundary is

44

thought to have. If there is such a doctrine as the simple view, it is certainlynot what we thought it was. The debate over personal identity has beensystematically misdescribed. After explaining why I believe this, I will saysomething about how the debate might be better characterized.

p r e l im i na r i e s

First some brief points of clarification. There are three assumptions com-monly taken for granted in discussions of personal identity that I am notgoing to make, rendering some of my formulations unusual. One is that ifthe identity over time of any person consists in something, then necessarilythe identity of every person consists in something. A second is that theidentity of all people whose identity consists in anything must consist inthe same thing. I want to leave open the possibility that our own identityconsists in something but the identity of certain other rational, intelligentbeings consists in something else, or in nothing at all. (It may be, forinstance, that we are biological organisms that persist by virtue of somesort of brute physical continuity but there is also an immaterial, personalgod for whom this does not hold.) For this reason I will speak of the identityof human people, not of people generally. Third, I will not assume that aperson must always persist as a person. I leave open the possibility that eachof us starts out as an unthinking embryo and may end up in a vegetativestate, and that we do not count as people at those times. So I will speakof the identity or non-identity of a human person existing at one time toa thing – person or not – existing at another time.1 If these traditional

Views of personal identityover time

Complex views The simple view

Psychological-continuity views

Brute physicalcontinuity views

Fig. 1: A taxonomy of views about personal identity over time.

1 For further discussion of this point see Olson (1997b, pp. 22–7).

In search of the simple view 45

assumptions turn out to be true, my formulations will be equivalent to theusual ones and the difference will do no harm.

A terminological note: although I have been speaking of “the simpleview,” there may be more than one view about personal identity that is notcomplex. In that case we can take “the simple view” to name the propositionthat some simple view or other is true. Likewise for “the complex view.”

I take the simple view and its rivals to be views about personal identityover time. There are other views about personal identity broadly speakingthat have certain affinities with this one. For instance, Wiggins says that wecannot give a complete account of what it is to be a person: personhood is, atleast to some extent, primitive and indefinable (Wiggins 1980, p. 171). Orsomeone might deny that there are any non-trivial conditions for our“transworld identity” – that is, for how we could have been. As these claimsappear to be independent of the simple view of our identity over time (theyare answers to different questions), I will say nothing about them.

There are also claims associated with the simple/complex debate thatreally are about our identity over time, but not about whether it consists insomething other than itself. One is that our identity over time must alwaysbe determinate: there could never be a being existing at another time thatwas neither definitely you nor definitely not you. Another is that the facts ofpersonal identity over time are somehow up to us to decide rather than todiscover. To my knowledge, all those said to be advocates of the simple viewdeny both that our identity can be indeterminate and that it can be up to usto decide – though some philosophers said to hold complex views say thesame.2 As I have been unable to discover any important connection amongthese issues, I will set them aside.

ground ing and cr i t e r i a

Complex views are supposed to imply that if a person existing at one timeis (or is not) numerically identical with a being existing at another time,something must make this the case – something beyond the mere fact thatthey are (or are not) the same.

There seem to be two things that this might amount to.One has to do withhow our identity over time relates to other things: advocates of the complexview think it consists in or depends on or holds by virtue of something other

2 Noonan (2003, ch. 6) accepts that identity statements can have an indeterminate truth-value, butdenies that there can be indeterminacy of identity itself. Chisholm (1976, pp. 108–11), Olson (1997a),and Merricks (2001b) argue against “identity voluntarism.”

46 eric t. olson

than itself. Not that a future being’s having inherited my mental states in acertain way (for example) causes him to be me, in the way that stubbing mytoe causes me pain. The dependence is logical or metaphysical. So:

Whenever a human person existing at one time is identical (or non-identical) to abeing existing at another time, this consists in something else.

Call this the grounding claim. Alternatively, personal identity might notconsist in anything else: it might be “brute” or “primitive.” An intermediateposition is also possible: that when a human person existing at one time isidentical or non-identical to a being existing at another time, this sometimesconsists in something else and sometimes does not. There is no consensusabout whether this would count as simple or complex.The second thought is that according to complex views there is a criterion

of identity for human people – not an evidential but a “constitutive” criterion,giving conditions individually necessary and jointly sufficient for a humanperson to persist from one time to another (“persistence conditions”).Consider this formula:

Necessarily, if x is a human person at time t and y exists at another time t*, x = y ifand only if . . .,

where “=” expresses numerical identity. The thought is that there is a truecompletion of the formula that is not trivial or otherwise degenerate. Callthis criterialism. Again, intermediate positions are possible. One is that thereis a criterion of identity for some kinds of people and none for others.Another is that there are certain conditions necessary for a person to persist,and perhaps even certain sufficient conditions, but no set of conditionsindividually necessary and jointly sufficient (Merricks 2001a, pp. 195–6).3 Infact every supposed advocate of the simple view that I know of accepts somenecessary conditions for personal identity: for instance that a thing existingat another time can be me only if it is not then a stone. If there are anymodal truths at all, one of them is surely that it is impossible for somethingto be a human person at one time and a piece of basalt at another time. (Orif one of us really could come to be made entirely of stone, none of us couldbecome abstract like the number seven.) But this sort of thing does nothingto suggest any non-trivial sufficient condition for our identity over time, orany way of completing the formula.

3 Noonan (2003, p. 97) calls such views complex. If I am right to say that everyone accepts some suchnecessary conditions, this implies that no one holds a simple view.


What is it for such a criterion to be non-trivial? Here are some that aretrivial:

Necessarily, if x is a human person at t and y exists at t*, x = y iff x = y.Necessarily, if x is a human person at t and y exists at t*, x = y iff anomniscient being would believe that x = y.

Necessarily, if x is a human person at t and y exists at t*, x = y iff whateveris true of x is true of y and whatever is true of y is true of x.

Necessarily, if x is a human person at t and y exists at t*, x = y iff x and yhave the same individual essence (where a thing’s individual essence isa property that it has essentially and that no other thing could possiblyhave).

The truth of these claims should not vindicate criterialism, and no friend ofthe simple view would deny them. They are not “proper” criteria of personalidentity.

Why not? Well, they are entirely uncontroversial. And they are true notbecause of the nature of personal identity, but because of the nature ofomniscience or individual essences or identity in general. Analogous prin-ciples apply to all objects: the result of replacing the term “human person”with any other count noun would be equally true and uncontentious. Sothe criteria tell us nothing about personal identity as such. That seems toaccount for their triviality.

They are also uninformative, in that we could not know whether theconditions were satisfied without already knowing whether identity holds.If Buggins suffers severe and irreversible brain damage of a certain sort, wecould not discover whether the resulting being lying insensible on thehospital bed had Buggins’ individual essence and use this to work outwhether it was Buggins. Of course, an informant might tell us whether ithad Buggins’ individual essence; but the original source of the informationabout essences could not have obtained it before knowing who was who.

What is more, the examples fail to support the connection we expect tofind between a criterion of identity and an account of what grounds it.Obviously x cannot be y because x is y. Nor can it be the case because anomniscient being believes it. The dependence is the other way round:whether an omniscient being believes something depends on whether it isthe case. Likewise, whether everything true of x is true of y depends onwhether x is y, as one of the things true of x is whether it is identical to y.Similar remarks go for the other examples. These criteria cannot tell us whatmakes x and y identical or non identical because they appeal to conditionsthat depend on whether this is so (or, in the case of the first example, to thevery fact to be accounted for).

48 eric t. olson

There is more to be said about what makes a criterion of identity over timesubstantive and non-trivial. Rather than exploring this further, I propose thata proper criterion is one that completes this formula:

Necessarily, if x is a human person at t and y exists at t*, x = y iff and because . . .

where the word “because” expresses the logical ormetaphysical dependence ofthe grounding claim. With any luck, this will avoid triviality and uninforma-tiveness. Criterialism, then, is the claim that some such criterion is true.How the grounding claim and criterialism are related is a nice question.

Although criterialism as stated entails the grounding claim, the converseentailment is less clear. You might think that if our identity consists insomething else, it must be possible to articulate that something else in theform of a proper criterion of identity. But I do not know how to argue forthis. Perhaps our identity over time could consist in something that is notexpressible in terms of necessary and sufficient conditions. (Or perhapsthere could be infinitely many such conditions, or even an uncountableinfinity. I leave open whether this would count as a criterion of identity.)

ant i - c r i t e r i a l i sm

Let us turn now to attempts to state the simple and complex views.The most obvious proposal is that the complex view is criterialism and

the simple view is its negation, anti-criterialism (that there is no propercriterion of personal identity). Anti-criterialism is probably the best candidatefor being the simple view. Yet it is such a strong claim that few if any supposedfriends of the simple view would accept it. Consider this thesis:

Necessarily, if x is a human person at t and y exists at t*, x = y iff and because thething that is x’s soul at t is the thing that is y’s soul at t*,

where a soul is an immaterial thinking substance, and a soul belongs to aparticular being at a time just if that being has its mental properties then byvirtue of the soul’s having them. Call this Cartesianism. Some philosophersaccept it, namely those who think that we have souls as parts but are notourselves souls – that a living human person is composed of a soul and abody (Swinburne 1984, p. 21). In that case, they say, a person existing now isidentical to a being existing at another time if and only if and because thatbeing has the person’s soul then. (If people are their souls, by contrast,personal identity and soul identity are the same, and one cannot depend onthe other, contrary to Cartesianism.)


Cartesianism is far from trivial: it tells us something about human peoplein particular, and is not true of concrete objects in general. It is alsoinformative, in that one need not know whether x is y before knowingwhether x has y’s soul. There may of course be practical obstacles to findingout whether we have the same soul before finding out whether we have thesame person, owing to the fact that we cannot observe souls. But theseobstacles appear to be only contingent.

So Cartesianism is a proper criterion of personal identity. It is a versionof criterialism, and thus, on the current proposal, a complex view. Yet it isconsidered a paradigm case of a simple view (Swinburne himself calls itsimple: 1984, p. 19).

Cartesianism also entails the grounding claim: it implies that personalidentity consists in something else, namely sameness of soul. (That is unsur-prising, given that criterialism entails the grounding claim.) This showsthat both criterialism and the grounding claim are too weak to capturewhat distinguishes complex from simple views – assuming, at least, that thephilosophical community is right about where the boundary lies.

Accepting anti-criterialism would mean rejecting all proper criteria ofpersonal identity. That would appear to rule out our being wholly materialthings, as in that case there could hardly fail to be a criterion in terms of atomsor matter, even if not a specific or satisfying one – something like the atomiccriterion of p. 59 (Zimmerman 1998).4 It would also rule out our being partlymaterial and partly immaterial, as that would presumably imply Cartesianism.It looks as if anti-criterialists must say that we are wholly immaterial souls.

They would also have to reject all non-trivial criteria for a soul existing atone time to be identical to a soul existing at another time. For instance,there could be no criterion in terms of causal relations, no matter howvague. Even the causal-dependence criterion of p. 58 would be false. Itwould have to be metaphysically possible that you are not the soul who readthe previous sentence – that it ceased to exist just now and you came intobeing in its place – even though all the causal relations are exactly as theyactually are (see Merricks 2001a, p. 196). If we asked why that soul perished,or why another appeared, there would be no answer. It would not bebecause of the intervention of some deity, or a local disturbance in theectoplasmic field, or anything else. There would be possible situations thatdiffered only in the facts about our identity over time.

Well, someone could say that. But I do not know whether any realphilosopher ever has. It would be quite a surprise if this were the only possible

4 Merricks (1998b) denies this, though I believe he would accept the atomic criterion.

50 eric t. olson

simple view – if some versions of substance dualism were simple and others,including those most commonly held, were complex, and the differenceturned on such arcane matters as whether we are souls as opposed to soul–body compounds and whether any sort of causal dependence is necessary andsufficient for a soul to persist. If that were the simple view, it would hardly beworth including in the undergraduate curriculum. Whether there are humansouls is an important fact about our identity over time, but whether there isany criterion of identity for them is at best a footnote.

ana l y z a b i l i t y

Can we do better? The simple view is often said to assert that personalidentity is “unanalyzable,” and the question of what it consists in is some-times put by asking for an analysis of the concept of personal identity.5 Thissuggests that the issue at stake is whether any proper criterion of personalidentity could count as an analysis. Analytic criterialism, then, would be theview that some such criterion is analytic – that is, true by virtue of theconcepts that figure in it, and the negation of which can be transformed intoan explicit contradiction by a process of analysis. Analytic anti-criterialismwould be the view that no such criterion is analytic. Might these be thecomplex and simple views?Swinburne’s main objection to complex views is in fact an objection to

analytic criterialism (Swinburne 1984, pp. 18–20). It is a version of Moore’s“open-question argument.” Consider any criterion of the form: necessarily,if x is a human person at t and y exists at t*, x = y iff and because x bears R at tto y as it is at t*, where R might be some sort of psychological or brutephysical continuity. Now imagine that some future being really does bear Rto you. Given this fact, Swinburne says, “mere logic” – or logic togetherwith the meanings of the relevant expressions – cannot determine whetherthat being would be you. Neither those who say yes nor those who say nowould thereby contradict themselves or commit any other logical or con-ceptual error. It is, as Moore would say, an open question. Thus, Swinburneconcludes, the proposed criterion cannot be an analytic truth. And thisholds for any proper criterion.Whatever the merits of this argument may be, it does not rule out such a

criterion’s being true, but at most its being analytic. So if this is an argumentfor the simple view, the simple view must be analytic anti-criterialism.

5 Noonan, for instance, says that according to the simple view “personal identity is an ultimateunanalyzable fact, which resists definition in other terms” (Noonan 2003, p. 95).


And maybe Cartesianism, if true, would not be analytic. That is certainlywhat the open-question argument suggests: given that each of us is acompound of a soul and a body, is it not an open question whether abeing who has my soul at some other time would have to be me? (Lockefamously said no.) Suppose my soul had its mental contents “erased” butwas otherwise unharmed, and those contents were then replicated inanother soul. No amount of conceptual analysis can determine whether Ishould then be the person with my original soul, the person with the newsoul and my replicated mental contents, or neither.6

But whereas ordinary anti-criterialism was too strict to be the simpleview, analytic anti-criterialism is too lenient: it is consistent with manyviews considered paradigms of complexity. Consider this one:

Necessarily, if x is a human person at t and y exists at t*, x = y iff and because x’sbiological life at t = y’s biological life at t*,

where a biological life is a self-sustaining chemical event that always coincideswith a living organism (Van Inwagen 1990b, pp. 83–90). If this were true, ouridentity over time would consist in something other than itself, namely inthe continuation of a biological life – a sort of brute physical continuity(unless we are ourselves lives – but no philosopher I know of believes that).This is a complex view if anything is. But it can hardly be analytic. Reflectingon the relevant concepts will never enable us to work out that we persist justas long as our biological lives continue. If nothing else, analytic truths areknowable a priori, and we cannot know a priori that human people even havebiological lives, never mind whether they must always go where their lives go.So the life criterion is perfectly consistent with analytic anti-criterialism (andimmune to Swinburne’s open-question argument). Many other proposedcriteria that are considered complex are no more analytic: Unger’s “physicallybased approach,” for example (Unger 1990, ch. 5).

advoca t e s o f ana l y t i c cr i t e r i a l i sm

Would any view of personal identity count as complex if the simple viewwereanalytic anti-criterialism? Does anyone think that what it takes for us topersist is analytic? This would apparently mean that the concept of a humanperson had the persistence conditions of human people built into it: that

6 In fact Swinburne takes Cartesianism to be analytic (personal communication). If the simple view isanalytic anti-criterialism, that would make Cartesianism a complex view.

52 eric t. olson

you could work out a complete set of necessary and sufficient conditions fora human person to persist just by analyzing the concept of a human person.You might suppose that it is not the concept of a person, but the concept ofpersonal identity over time that has the criterion of personal identity builtinto it. But what is the concept of personal identity, if not the concept ofidentity applied to people? And the concept of identity (or of identity overtime) by itself provides no substantive criterion of identity. So if the criterionof personal identity is analytic, it must derive from the concept of a person.I know of two views that may have this consequence.7One is Shoemaker’s

theory that mental properties, by their very nature, give their bearers causalpowers that fix their persistence conditions (Shoemaker 1999). It is a familiarthought that mental states are by nature disposed to combine in certain wayswith other mental states to produce certain effects: this is the core of thefunctionalist theory of mind. Shoemaker adds to this the further claim that amental state must be disposed to combine in this way only with states of thesame subject, and to produce the relevant effects in that subject alone. Yourbeing hungry, for instance, must tend to cause you to eat if you believe there isfood before you, unless you have some competing goal – you and only you.Thus, any being that was caused, in this way, to eat by your state of hungermust be you. It follows that some sort of psychological continuity suffices foryou or any other psychological being to persist.8 If a person is by definition asort of psychological being, and if Shoemaker’s theory of mental properties isanalytic, then this account of personal identity will be analytic, and presum-ably a version of analytic criterialism.A view that pretty clearly makes the conditions of personal identity analytic

is the ontology of temporal parts or “four-dimensionalism.” It implies that forevery candidate for being the career or history of a human person, and everypart of such a candidate, there is a conscious, intelligent being whose history itis. So there is now a being sitting here that is just like me, both physically andmentally, except that it came into being at midnight last night and will ceaseto exist at midnight tonight. Another such being came into existence whenmy biological life began, some sixteen days after I was conceived, and willcease to exist when my life ends. Yet another came into being at conception,and will persist after my death as a corpse until my remains become dust.

7 Another possible candidate is Parfitian “reductionism,” but I do not understand it well enough toknow for sure. Shoemaker (1985) and Noonan (2003, pp. 97–100) are good discussions. Merricks(1998b, pp. 111–16) argues against analytic criterialism.

8 Though whether psychological continuity would be necessary for a person to persist is less clear; seeOlson (2002).


Many more such beings share my current stage. But according to most four-dimensionalists, only one of these beings (vagueness aside) is a person. Beingrational and self-conscious is therefore insufficient for personhood. In factthe vast majority of beings psychologically like ourselves are not people, butmerely share temporal parts with a person.

To say what it is to be a person, on this view, we have to say what it takesfor a person to persist. Thus Lewis (1976), for example, says that a person is amaximal aggregate of psychologically interconnected person-stages. A person-stage is a more or less momentary being with the mental properties thatcharacterize people: rationality and self-consciousness, perhaps. Two person-stages are psychologically connected when one “inherits” its psychologicalproperties from the other in a special direct way: when one has a memory ofan experience of the other’s, for instance. An aggregate of psychologicallyinterconnected person-stages is a being composed of person-stages, each ofwhich is psychologically connected with every other, and it is maximal whenit is not a part of any other such aggregate. This gives the conditions necessaryand sufficient for a person to persist. And because it is part of the definition of“person,” it is analytic if anything is.

But this can hardly be the frontier between simple and complex, with viewslike Lewis’s and Shoemaker’s on the complex side and all the rest, includingthose of Unger and Van Inwagen, counting as simple.

emp i r i c i s t theor i e s

Another issue sometimes discussed in connection with the simple/complexdebate is whether personal identity consists in any of the conditions we useas evidence for judgments about it. Swinburne complains that some phi-losophers fail to distinguish between evidential and constitutive criteria:“their account of the logically necessary and sufficient conditions of per-sonal identity [is] in terms of the evidence of observation and experiencewhich would establish or oppose claims of personal identity” (Swinburne1984, p. 3; for similar remarks see Chisholm 1976, pp. 111–13). He calls theseempiricist theories of personal identity.

Might complex views be empiricist theories? Is this the issue at stake inthe simple/complex debate – whether there are conditions individuallynecessary and jointly sufficient for personal identity that we use as evidencefor claims about who is who? Well, what do we use as evidence for theseclaims? What is the basis for my belief that the man emerging from theoffice next door is my colleague Nils, rather than another colleague or a

54 eric t. olson

complete stranger? Well, he looks like Nils did when I last saw him. And heacts like Nils: he greets me rather than staring blankly, and responds as Nilshas in the past when engaged in discussion about departmental matters.And I take it to be rare for someone to look and act so much like someoneelse as to fool his colleagues. Our primary evidence for claims about personalidentity seems to be physical and behavioral similarity.If an empiricist theory is one according to which there is a criterion of

personal identity in terms of our primary evidence for such claims, thenempiricists are those who hold that our identity over time consists in physicaland behavioral similarity. But no one thinks that. Even if we could specify therespect of similarity generously enough to allow for the fact that people’sappearance and behavior change radically between infancy and old age, noone thinks that mere similarity of any sort is sufficient for identity: there couldbe someone other thanNils whose appearance and behavior were as similar tohis as you like.This cannot be what Swinburnemeant by an empiricist theory.What did

he mean? Well, we take physical similarity to be evidence for some sort ofphysical continuity – causal dependence of later physical states on earlierones – and we take behavioral similarity to be evidence for psychologicalcontinuity, which is also a sort of causal dependence. And many philoso-phers think that physical or psychological continuity is not merely evidenceof personal identity, but what it consists in. So perhaps empiricist views arethose according to which there is some condition that we use as evidence –not necessarily as primary evidence – for judgments about personal identity,which figures in a proper criterion of personal identity.But this would make Cartesianism an empiricist theory. If it were true

(and we knew it), physical and behavioral similarity would be evidence forsameness of soul: I should be warranted in believing that the man next doorhas Nils’ soul because he looks and acts like Nils. At any rate this is consistentwith Cartesianism. (It would be bad news for Cartesians if it were not true:it would seem to deprive us of any evidence for claims about who is who.)Yet having the same soul would not only be evidence of personal identity,but what it consists in. You might suppose that Cartesianism is not anempiricist theory because the condition that it takes personal identity toconsist in is not observable: which human being has which soul is not“evidence of observation and experience.” But psychological continuity isnot observable either. Both have to be inferred (in others, if not oneself )from behavioral similarity.So it looks as if Cartesianism is an empiricist theory if psychological-

continuity views are. This is not the right boundary either.


b rut ene s s

What couldmake Cartesianism, but not physical- or psychological-continuityaccounts, a simple view? Perhaps this: even if personal identity consists inidentity of soul, identity of soul consists in nothing. Nor is there any criterionof identity for souls. Though personal identity may depend metaphysicallyon something else, that something else does not itself depend on anythingfurther. So this might be the simple view: personal identity either consists innothing (that was anti-criterialism, or near enough), or consists in the identityof something else which consists in nothing. Alternatively, if there is anyproper criterion of identity over time for human people, it appeals to theidentity of something for which there is no such criterion. (For presentpurposes we can ignore the difference between these two variants.) Call thisproposal bruteness.

But some Cartesians deny bruteness. Swinburne, for instance, used to saythat souls, like all substances, persist by virtue of sameness of form andcontinuity of stuff, so that their identity is not brute (Swinburne 1984,p. 27; 1997, pp. 153–4).9 To make Cartesianism compatible with bruteness,you would have to combine it with bruteness about souls (a view brieflyexplored on p. 50). It would be pretty surprising if this were the frontierbetween simple and complex views. There are also views everyone wouldclassify as complex that are consistent with bruteness. Someone who endorsedthe life criterion on p. 52, but believed that identity of lives was brute, wouldaccept bruteness. Or imagine a Humean who believed that people are, as thegreat man said, “nothing but bundles of impressions,” and who endorsed thefollowing criterion: if x is a human person at t and y exists at t*, x = y iff andbecause y is composed of impressions at t* and a certain proportion of thoseimpressions are also parts of x at t. Combined with the view that there is nocriterion of identity for individual impressions, this would entail bruteness.

noonan ’ s p ro po s a l

Harold Noonan (2011) has recently proposed a subtle refinement of brute-ness. The simple view, he says, is that the only necessary conditions onpersonal identity belong to a certain special sort. A condition will belong tothis sort if it follows from what it is to be a person (or a human person) ata particular moment. For instance, you cannot be a person at a particular

9 He has since revealed in personal communication that he no longer holds this view. But he does notsay that he once held a complex view and now holds a simple view.

56 eric t. olson

moment if you are made entirely of stone then. Supposing, as Noonan does,that whatever is a person at some time must be a person whenever it exists(which we can grant for the sake of argument), it follows that no past or futurebeing can be you if it is made entirely of stone at that time. But as everyoneaccepts this, it ought to be consistent with the simple view. Noonan calls suchconditions “synchronic.” Conditions that are trivial in something like thesense discussed on p. 48 are also compatible with the simple view. And soare conditions that are “identity-involving”: those that require the identityover time of something other than the person in question. For instance,Cartesianism implies that a past or future being can be you only if it thenhas the same soul as you have now, where a soul is not a person. This iscompatible with the simple view. The simple view, then, is that the onlynecessary conditions for personal identity are either synchronic, trivial oridentity-involving. Complex views assert further conditions.Like the previous proposal, Noonan’s implies that at least some versions

of Cartesianism are complex, depending on what conditions they imposeon the identity of souls. If the only conditions for soul identity are syn-chronic (if they follow from what is necessary for a thing to be a soul at aninstant), trivial or identity-involving, then Cartesianism may be a simpleview. But if there are other necessary conditions for a soul to persist – forinstance, if it requires a soul’s later states to depend causally in some way onits earlier ones – then these will also be conditions for personal identity, andthe Cartesian view will count as complex. Likewise, it appears to count someversions of the life criterion and the Humean view as simple.If I have understood it, Noonan’s proposal makes the simple view so

strong that no one actually holds it. No philosopher I know of would acceptthat a person could be wholly material at one time and wholly immaterial atanother. If you are now made up entirely of matter, then no being that is notmade up even partly of matter at some other time can be you. Likewise, if youare wholly immaterial now – if you are a Cartesian soul, say – then you couldnot come to bemade up entirely ofmatter. This is not a synchronic conditionfor personal identity: it does not follow fromwhat it is to be a person, or evena human person, at a particular moment. From the mere proposition that xis a human person at t, it does not follow either that x is material at t or that xis immaterial at t, never mind that x is material or immaterial at any othertime. So it seems, anyway. Nor is the condition trivial or identity-involving.If so, Noonan’s proposal implies that you hold a simple view only if you rejectthis condition: only if you think it is possible for one of us to change frombeing wholly material to being wholly immaterial or vice versa. But everysupposed advocate of the simple view accepts the condition.


s p e c i f i c and un s p e c i f i c

I can offer no better account of what makes a view simple or complex. As faras I can see, no principle divides views about personal identity in the rightplace, or has the importance that the simple/complex distinction is tradi-tionally ascribed. Let me finish by trying to say something positive. Supposethe traditional taxonomy of simple and complex views really is hopelesslywrongheaded. What ought to replace it? How should views of personalidentity be categorized? What are we to tell our students?

Well, we can still divide views into those that appeal only to brute physicalconditions, those that appeal to some sort of psychological continuity, andthose that appeal to neither. This is an important three-way distinction.

Someone might suggest that simple views are precisely those appealingneither to brute-physical nor to psychological conditions – thus makingthe simple/complex distinction irreducibly disjunctive. That would makemost versions of Cartesianism simple – the exception being those that givepsychological persistence conditions for souls (Foster 1991, pp. 240–61). Butthis would be to abandon all our original starting points. It would do awaywith the thought that complex views take our identity over time to consistin something while simple views deny this. There are views according towhich personal identity over time consists in something other than brute-physical or -psychological continuity – or, for that matter, identity of soul.Kantians, for instance, might say that it consists in something noumenalthat we can form no conception of. The disjunctive proposal would countthese as simple views. But should they not count as complex? It wouldalso give up the conviction that simple views differ from complex ones oversomething more fundamental than what personal identity consists in,instead putting their difference on a par with that between physical andpsychological views.

Here is something else we can tell our students. Criteria of personalidentity differ enormously in howmuch information they provide. This onetells us rather little:

Necessarily, if x is a human person at t and y exists at t*, x = y iff and because thetotal state of x at t depends causally in an appropriate way on the total state of y at t*or vice versa.

It says only that personal identity consists in some sort of causal depend-ence. That is not trivial. If it is true, it is because of the metaphysical natureof human people. It may not hold for all objects, or even all concreteobjects. It is not true of shadows and sunbeams, if there are such things:

58 eric t. olson

their earlier states do not cause their later ones. It has controversial impli-cations: it rules out our surviving without any sort of causal dependence onour earlier states, contrary to some views about life after death (e.g. Baker2005; Hick 1990, pp. 123–4). But beyond that it is useless. It says nothingabout what sort of causal dependence our identity consists in. Think ofcases where it is hard to know who is who, even if we know everything else.Suppose we ask whether reincarnation is possible, or what would happen tosomeone whose brain was transplanted into a different head. The causal-dependence criterion is consistent with any answer to these questions.Or consider this “atomic criterion” (adapted from Zimmerman 1998):

Necessarily, if x is a human person at t and y exists at t*, x = y iff and because thearrangement of x’s atoms at t and the arrangement of y’s atoms at t* both belong toan atomic history of the sort that necessarily any human person has, and which isnecessarily the history of a human person,

where “the arrangement of x’s atoms” includes both the intrinsic nature ofthe atoms composing x and their spatio-temporal and causal relations to oneanother and to other atoms, and someone’s “atomic history” is a completedescription of her atoms throughout her existence. This is not trivial either.If it is true, it is because of how it is with human people and not merelybecause of the nature of atomic histories and arrangements. It has con-troversial implications, ruling out our surviving without being composed ofany atoms at all. But it tells us no more about what happens in most of thepuzzle cases than the causal-dependence criterion does.10

These accounts are enormously unspecific about what it takes for us topersist. The life account, that a past or future being is you just if it then hasyour current biological life, is far more specific. It tells us who would be who inmost of the puzzle cases. The same goes for familiar psychological-continuityviews.Yet even these views are less specific than they might be. The precise sort of

causal dependence that psychological continuity consists in is elusive. Nomatter how much detail an account provides, there will be possible caseswhere a being existing at another time has mental states that depend causallyon yours now (or vice versa) in such a devious way as to leave it unclearwhether he or she is psychologically continuous with you. And even if wecould specify the sort of causal dependence, psychological continuity will be amatter of degree, and it is hard, if not impossible, to say precisely what degree

10 It may be that these are somehow not “proper” criteria of identity, and thus ought to be consistentwith anti-criterialism, supporting the claim that anti-criterialism is the simple view. Those with thissuspicion will want to think about what makes them improper.


is needed. I have never seen a psychological-continuity account specificenough to tell us, in conjunction with the relevant underlying facts, inevery case whether the resulting being would be you, or not you, or whetherit would be indeterminate whether it was you.

You might think this is only because no one has taken the trouble to writeout a completely specific psychological-continuity view, or because no oneknows enough to do it. More generally, it may be that if some less specificaccount of personal identity is true, it must be because it follows from somemore specific account – perhaps even a maximally specific account that wouldentail, in conjunction with the underlying conditions, all the facts aboutpersonal identity over time. I have no idea whether this is the case. There iscertainly reason for doubt. Shoemaker, who has thought about these matters asdeeply as anyone, confesses that it seems impossible to specify the sort of“appropriateness” that figures in the causal-dependence criterion withoutalready knowing the precise conditions under which human people persist(Shoemaker 1979, p. 337). You might think it is impossible only for finiteminds: God must be able to do it. But who knows? It is worth noting thatMerricks, who presumably accepts the atomic criterion, explicitly denies thatthere is any more specific criterion of personal identity (Merricks 1998b).

So here is an important difference among criteria of personal identity:some give us quite specific information about what our identity over timeconsists in, while others say little. This bears some affinity to the traditionalsimple/complex distinction: those views called complex often appear morespecific than those called simple.

But the simple/complex distinction is not that between unspecific andspecific views. For one thing, the difference betweenmore and less specific isa matter of degree, while the simple/complex distinction is supposed to beabsolute: we tell our students that some views of personal identity are simpleand others are not, not that some are more simple than others.

And the two distinctions divide the territory in different ways. The causal-dependence criterion is normally taken to be a complex view: Shoemakerexplicitly contrasts it with that of Butler, Reid and Chisholm that “the cross-temporal identity of genuine continuants [cannot] be said to consist in theholding of any relations other than the relation of identity itself” (1979,p. 323). Yet it is consistent with the absence of any more specific criterion.Nor is it clear whether the soul criterion (p. 49) is any less specific, eventhough it is supposed to be a simple view. Think of reincarnation again.Could someone born after my death be me? The causal-dependence criteriongives an ambivalent answer: only if that being’s state then depended causallyin the appropriate way on mine before my death. But it says nothing about

60 eric t. olson

what way is appropriate, or whether it could be appropriate in this case. Theatomic criterion tells us even less. The soul criterion gives what sounds like amore specific answer: such a being would be me if and only if it had myimmaterial soul. If you were a deity wanting to reincarnate human people,this (if true) would be far more useful information. Thus, some views tradi-tionally taken to be complex appear, if anything, to be less specific than viewsconsidered simple.And the distinction between less andmore specificmay be not only amatter

of degree, but context-relative. The soul criterion might be less specific thanthe causal-dependence criterion relative to some contexts and more specificrelative to others – as in the case of our imaginary deity. But I do not thinkanyone ever took the simple/complex distinction to be context-relative.

e x p l ana tor y demand s

The distinction between specific and unspecific criteria of identity is closelyrelated to a difference in expectations. Philosophers discussing personalidentity often make explanatory demands. Suppose once again that we areconsidering the possibility of reincarnation. Here is an almost irresistiblequestion: what could make it the case that some infant born after I am deadwas me? Out of all those future newborns – many billions of them – whatcould make this one me, rather than some other one? To put it the other wayround, what could make it the case that a certain infant born after my deathwas me rather than someone else – you, say, or Socrates, or someone who hadnever existed before? Or if reincarnation is impossible, we shall want to knowwhat makes it impossible. What is it about the way someone born after mydeath would have to relate to me that would necessarily rule out its being me?Philosophers disagree about when such explanatory demands are legiti-

mate: about what must have an explanation in matters of personal identity,and where explanation comes to an end andwe face brute facts. They disagreeabout which questions must have answers. They also disagree about what sortof answers we can expect. At one extreme, someone might say that an infantborn after my death could be me but nothing would have to make it me. Itwould not have to be me because it had my soul, or because it bore anypsychological relation to me, or because its states then depended causally in acertain way onmine now, or because God willed that it be so. It could simplybe me, and that would be that.11 At the other extreme, someone might saythat such an infant could not possibly be me because its total state then

11 This may be the view of Mavrodes (1977).


would not bear to my total state now the sort of causal dependence that isboth necessary and sufficient for personal identity; and she might be able todescribe this relation precisely. (Certain versions of the life criterion might belike this.) Nearly all the answers actually proposed – both those consideredsimple and those considered complex – fall between these extremes.

When it comes to personal identity, some philosophers expect more facts tobe explained than others do, and there is probably some correlation betweenthis and whether one holds a view classified as complex or one classified assimple. But the correlation is far from perfect. Nor can I see any way oftransforming this difference of expectations into a difference of doctrine.

The simple view remains elusive.12

12 For discussions on the topic of this chapter I thank Andrew Jones, Harold Noonan, RichardSwinburne and the audience at the 2010 Obergurgl conference.

62 eric t. olson

chapter 3

Personal identity, indeterminacy and obligationRyan Wasserman

It is the foundation of all rights and obligations, and of all account-ableness; and the notion of it is fixed and precise.

(Thomas Reid 1975 [1785])

p e r sona l i d ent i t y and inde t e rm inac y

In Reasons and Persons (1984), Derek Parfit defends the complex view1 ofpersonal identity, which he formulates as follows:

The complex view: Facts about personal identity consist in other, impersonal facts.2

More specifically, Parfit defends the view that facts about personal identityreduce to facts about psychological continuity or connectedness (Parfit1984, p. 216). He then goes on to argue that this view implies the possibilityof indeterminacy:

The Indeterminacy of Personal Identity: It is possible for questions about personalidentity to lack determinate answers.

Parfit explains his reasoning as follows:

We can describe cases where, between me now and some future person, the physicaland psychological connections hold only to reduced degrees. If I imagine myself insuch a case, I can always ask, “Am I about to die? Will the resulting person be me?”On the Reductionist View, in some cases there would be no answer. My questionwould be empty. The claim that I was about to die would be neither true nor false.(Parfit 1984, p. 214)

Many people find this conclusion incredible, and not in a good way.Chisholm, for example, writes that:

1 In Reasons and Persons, Parfit’s focus is on what he calls reductionism, but it is clear that he takesreductionism to be equivalent to the complex view. See, for example, Parfit (1982, pp. 227–8).

2 Where “impersonal facts” are those which do not presuppose the existence of a continuing person.

63

When we use “the same person” in [the] strict way . . . although cases may well arisein which we have no way of deciding whether the person x is the same person as theperson y, nevertheless the question “Is x the same person as y?” will have an answerand that answer will be either “yes” or “no.” If we know that x is a person and if wealso know that y is a person, then it is not possible to imagine circumstances underwhich the question “Is x the same person as y?” is a borderline question – a questionadmitting only of a “yes and no” answer. (Chisholm 1970a, p. 171)

This way of thinking suggests the following argument against the complexview: if the complex view is correct, then personal identity can be indeter-minate. But personal identity cannot be indeterminate. Hence, the complexview is wrong.

This simple argument is a common one, having been suggested byChisholm (1970a), Swinburne (1973−4), and many others.3 But the centralpremise of the argument is seldom defended:why think that personal identitycannot be indeterminate?4

This omission is surely due to the inherent plausibility of the premise.When I imaginemyself in some of the cases described by Parfit and I considerthe question “Will I survive?” I do not know how to answer – but somethingdeep inside of me says that there must be an answer. Theremust be an answer,which is just to say the Indeterminacy of Personal Identity must be false.

The strength of this conviction is undeniable, but its source is unclear.The goal of this chapter is to identify the source of our anti-indeterminacyintuition and thus provide a better understanding of the standard argumentagainst the complex view.5

i nd e t e rm ina c y and obl i g a t i on

Imagine a machine that can make ever so subtle changes to your body –including your brain – and is therefore capable of making equally smallchanges to your psychology. There are approximately fifty trillion cells inthe human body, so let us imagine that the machine has a keypad that allowsus to enter any number from zero to fifty trillion. If we enter “0” and youenter the machine, nothing will happen – you will emerge from the machinein almost exactly the same state that you entered. If we enter the number “1,”

3 Harold Noonan suggests that this line of thinking is “historically one of the central components of theSimple View” (1989, p. 226).

4 Noonan (1989) bemoans the fact that this “question is seldom asked by the proponents of the SimpleView” (see also Thomson (2008, pp. 173–4)). For some exceptions to the rule, see Williams (1970) andHossack (2006).

5 More carefully: the goal of this chapter is to identify at least one source of this conviction. Ouropposition to indeterminacy could be overdetermined.

64 ryan wasserman

however, the machine will destroy a randomly selected cell from your bodyand replace it with a new cell – or perhaps several cells – which qualitativelyduplicates the corresponding part in the body of Albert Einstein (circa 1950).If we enter the number “2,” the machine will replace two cells in the sameway. And so on. If we enter “50,000,000,000,000,” every cell of your bodywill be destroyed and replaced so that the person who emerges from themachine will be a qualitative duplicate of Einstein himself.Now, let us suppose that Frank is a convicted killer who has been sentenced

to die. Suppose further that we place Frank into the matter-replacementmachine and enter the number “20,000,000,000,000,” so that twenty trillionFrank cells are replaced with approximately twenty trillion Einstein cells.The man who emerges from the machine will be, so to speak, 60 percentFrank and 40 percent Einstein. Let us call this creature “Frank−Einstein” or,more simply, Frankenstein.Outwardly, Frankenstein looks a bit like Frank and a bit like Einstein.

Inwardly, the transformation is similar. Frankenstein has many of thesame beliefs, desires and psychological dispositions that Frank had. But healso has some of the psychological features of Einstein. And some of hispsychological features are a “blend” of the two. Frank – before entering themachine – gave zero credence to the proposition that he had won a NobelPrize; Frankenstein is not so sure. Frank had never been that fond of ridingbicycles; Frankenstein is rather keen on the idea. Perhaps, most impor-tantly, Frank clearly remembered killing a man in cold blood; Frankensteinseems to remember this . . . but it is all a bit hazy.This kind of case raises two questions – one metaphysical and one moral.The metaphysical question is this: is Frankenstein the same person as Frank?

If the complex view is correct, then Frank has survived the procedure just incase there is sufficient bodily continuity or sufficient psychological continuityor sufficient continuity in both respects. This disjunction indicates that thereare various versions of the complex view. One could focus on physical factsor psychological facts or both. One could focus on continuity or connected-ness or both. And with respect to psychology, one could focus on memoriesor beliefs or desires or emotional states. The point is that there are many,many versions of the complex view. But every reasonable version willinvoke continuity or connectedness in some way or other. Continuity andconnectedness are a matter of similarity, and similarity comes in degrees. So,continuity and connectedness come in degrees as well. How continuous doesa series need to be to be continuous? That is like the question: how tall doyou have to be to be tall? Neither question has a determinate answer. Sincethere will be unsettled questions of continuity and connectedness, there will

Personal identity, indeterminacy and obligation 65

be unsettled cases of personal identity. I will assume that the Frankensteincase is a case like this, so I will assume that – on the complex view – it isindeterminate whether or not Frank is Frankenstein.

Now for the moral question: what is to be done with Frankenstein? Let usassume that Frank deserved to die for his crimes. Or, to put it in deonticterms, executing Frank would have been the right thing to do; letting him gowould have been wrong. What about Frankenstein? If Frank is Frankenstein,then we are obligated to kill Frankenstein, for Frankenstein is a killersentenced to die. If not, not. In other words, we are obligated to killFrankenstein if and only if Frankenstein is Frank. And we are obligated tonot kill Frankenstein just in case Frankenstein is not Frank. If it is indeter-minate whether or not Frankenstein is Frank, it is indeterminate whether ornot we ought to carry out the punishment. It is also indeterminate whether ornot we ought not to carry out the punishment. More formally:

O(Punish) ↔ Guilty O(¬Punish) ↔ ¬Guilty∇Guilty ∇¬Guilty

∴ ∇O(Punish) ∴ ∇O(¬Punish)6

The upshot is that the question “What is to be done with Frankenstein?” isone more “empty question.”

But that is crazy! Or at least it seems crazy. How can a question of life anddeath be an empty question?

To illustrate the oddity, imagine being juror in this case. You have todecide what is to be done with Frankenstein. And while you and the otherjurors are debating, God himself comes down to join in the conversation.You now have the ultimate advisor before you – perfect in knowledgeand wisdom. So you ask: “Should we sentence Frankenstein to death?” Inresponse, God shrugs. “Should we let our prisoner go?” God shrugs again.“But God,” you persist, “What are we supposed to do?”TowhichGod replies,“Don’t look at me!”

I find that picture unsettling. It is one thing to say that we humans do notknow what to do in a case like this. That is unsurprising. But if there is nofact of the matter, then there is nothing to know – not even for God. Onecan know all the facts without knowing what to do. And that does seemshocking.

The same seems true when we turn from morality to rationality. Toillustrate the point, let us quickly consider a second story involving thematter-replacement machine.

6 Where O is the obligatory operator and ∇ is the indeterminacy operator.

66 ryan wasserman

Suppose you are about to be forced into the Einsteinmachine. The machineis set to “20,000,000,000,000.”You knowwhat is about to happen. But beforethe machine is turned on, you are approached by a car insurance salesman. Thesalesman has it on good authority that the person who emerges from themachine – You* – will be in a car accident tomorrow. There will be $10,000of damage done – damage that You* will be responsible for. However, thesalesman offers to sell you a limited-term insurance policy. The policy will costyou $1,000 and it will cover all of the damages that will be incurred by You*.This is clearly a bad deal for the salesman. But is it a good deal for you?

That is, are you rationally obligated to buy the insurance? This is not so clear.For the sake of simplicity, let us suppose that you care only about money –

yourmoney. If you are going to be the person in the accident tomorrow, thenyour expected utility of buying the insurance is $9,000. If you are not goingto be in the accident, your expected utility is −$1,000. So, rationality dictatesthat you ought to buy the insurance if and only if you are going to be in theaccident.7 And you ought not to buy the insurance just in case you are notgoing to be in the accident. If it is indeterminate whether or not you are You*,it is indeterminate whether or not you are going to be in the accident. So, it isindeterminate whether you are rationally obligated to buy the insurance or toreject the deal.Once again, this is a strange result. Once again, the image of God

shrugging helps to emphasize the point. You have to decide whether ornot to buy the insurance. God is there to serve as your financial advisor. Heknows everything about the machine. He knows everything about theupcoming accident. He knows everything about your goals and utilitiesand your financial situation. All you want is a straight answer: “Should I buythe insurance?” And the best that God can do is . . . shrug.That seems wrong. That seems deeply wrong. Questions about what is to

be done – whether rational or moral – must have answers. In other words,the following principle must be false:

The Indeterminacy of Obligation: It is possible for questions about obligations tolack determinate answers.

If this principle follows from the Indeterminacy of Personal Identity, thenthat principle must be rejected as well.

7 Objection: you are either going to die when you step into the matter-replacement machine or not(though it may be indeterminate which). If you are going to live – and therefore be the person in the caraccident – you should buy the insurance. Moreover, there is no benefit to saving money if you are aboutto die. Thus, there is no reason not to buy the insurance. Response: let us suppose that you will have theopportunity to spend the $1,000 on pleasure-producing items if you do not spend it on insurance.


i nd e t e rm in ac y and mora l d i l emma s

I have suggested that our intuitive resistance to the Indeterminacy of PersonalIdentity flows from a more basic opposition to the Indeterminacy ofObligation. This is not a new idea. In his Essays on the Intellectual Powers ofMan, Thomas Reid writes that:

The identity . . . which we ascribe to bodies, whether natural or artificial, is notperfect identity; it is rather something which, for the convenience of speech, we callidentity. It admits of great change of the subject, providing the change be gradual . . .But identity, when applied to persons, has no ambiguity, and admits not of degrees,or of more and less. It is the foundation of all rights and obligations, and of allaccountableness; and the notion of it is fixed and precise. (2002 [1785], p. 113)

As I read Reid, he is suggesting an argument against the Indeterminacy ofPersonal Identity. This argument has an unstated premise – namely, thatfacts about rights, obligations and responsibility are, in his words, “fixed andprecise.” So the argument, made explicit, is this: morality must be deter-minate, and morality is founded on personal identity, so personal identitymust be determinate as well.

This old argument suggests a new question: why think that moralitycannot be indeterminate?

Several answers to this question will be suggested.8 But first, a point ofclarification: when Reid writes of morality, he mentions rights, obligationsand accountableness (or responsibility). I have chosen to focus on obligations,and one might object to this decision. Suppose, for example, that one isattracted to satisficing consequentialism – the thesis that an action is morallypermissible if and only if it is “good enough.”9 If the satisficing view is correct,then one is obligated not to perform an action if and only if it is not “goodenough.”What is it for an act to be “good enough”? On the picture presentedby the satisficer, all of one’s available actions can be ranked in terms of whichare better than others (where there may be many ties in the ranking). Thesatisficer then says that an action is “good enough” just in case it appears“high enough” in the rankings. Presumably, the line between those optionswhich are “high enough” and those which are not is unsettled, so it can beindeterminate whether or not one is obligated to not perform a certain action.This result is not obviously absurd. Compare: every human being can beordered in terms of who is taller than whom (where there may bemany ties inthe rankings). Someone counts as “tall” (relative to a context) just in case he

8 See below. See also section “Objections to epistemicism” (p. 74) and Wasserman (2004b).9 On satisficing, see Slote (1985, pp. 45–7).

68 ryan wasserman

or she appears “high enough” in the ranking. This threshold is determined bycontext but, presumably, most contexts fail to determine a precise point inthe rankings.10

What would seem absurd, for the satisficer, would be indeterminacy in therankings. That is, it would be absurd to think that there could be two availableacts, A and B, where there is no fact of the matter as to whether A is betterthan, worse than, or equal to B. (Once again, the image of God shrugginghelps to reinforce this feeling.) Given this, the arguments of the previoussection can be easily adapted for the satisficer. If personal identity can beindeterminate, then it can be indeterminate whether one option is better than(or worse than or equal to) another. For example, it will be indeterminatewhether it would be better to punish Frankenstein or to let him go. But that isimpossible. So the Indeterminacy of Personal Identity is unacceptable.The lesson is that the Indeterminacy of Personal Identity will imply

indeterminacy in many moral matters, and one might focus on differentmatters, depending on one’s moral theory. That being said, I will now setsatisficing aside and continue to focus on the Indeterminacy of Obligation.Let us return to our question: why think that morality – and moral

obligations in particular – cannot be indeterminate? One reason has to dowith the possibility (or impossibility) of moral dilemmas. A moral dilemmais commonly defined as a situation in which one is obligated to performeach of two mutually inconsistent actions (not prima facie obligated, butobligated obligated). In a genuine moral dilemma, one is guaranteed to failin one’s moral duty.A common example is the story of Sophie’s Choice, in which a Polish

woman is arrested by the Nazis and sent to Auschwitz. On arrival, she is“honored” for not being Jewish by being given a choice: one of her childrenwill be spared the gas chamber, but she has to choose which one. In theagony of indecision, as both children are being taken away, she chooses herson (who, she reasons, is older and stronger, and therefore more likely tosurvive). Tragically, she loses track of her son and never learns his fate.The suggestion is that, in this case, Sophie is obligated to save her son and

obligated to save her daughter, despite the fact that she cannot do both.Of course, many people reject this description of the case, since they deny

the possibility of genuine moral dilemmas. There are at least three reasonsfor thinking that such cases are impossible. First, moral dilemmas wouldviolate the “ought-implies-can” principle, at least if we accept the standarddistribution principle:

10 For more on gradable predicates and comparatives, see Bolinger (1972).


Distribution: (OA & OB) ↔ O(A & B)

Second, the existence of moral dilemmas is inconsistent with two otherstandard principles of deontic logic:

Principle of Deontic Consistency: OA → ¬ O¬ APrinciple of Deontic Entailment: (A → B) → (OA → OB)

The third reason for rejecting the possibility of moral dilemmas begins withthe thought that morality must be action-guiding – and not just action-guiding, but uniquely action-guiding. A set of rules that tells you to do bothA and B (where these are incompatible options) tells you too much andthus does not tell you enough: it does not tell you whether you should doA rather than B or B rather than A. And that is exactly what one wants toknow in a case like Sophie’s Choice.

This third reason for rejecting the possibility of moral dilemmas is also areason for rejecting the possibility of moral indeterminacy – at least the kindof indeterminacy that concerns Parfit. According to Parfit, there is no fact ofthe matter about what ought to be done in the Frankenstein case. It is notobligatory to kill Frankenstein, nor is it obligatory to let him go. Neitheraction is permissible; neither is impermissible. In short, we have a failure ofmoral guidance.

In fact, one might go on to argue that the Frankenstein case is itself amoral dilemma – at least for those who accept indeterminacy. Let ussuppose that, generally speaking, the punishment ought to fit the crime.In particular, let us suppose that (all else being equal) it would be wrong todeterminately punish someone for a crime he has not determinately com-mitted. Also, all else being equal, it would be wrong to determinatelyabsolve someone who is not determinately innocent. More formally:

Principle 1: (¬ Δ Guilty) → (O¬Δ Punish)Principle 2: (¬Δ¬ Guilty) → (O¬Δ¬ Punish)11

Frankenstein is not determinately guilty and not determinately not guilty.Given the two principles, we are obligated to not determinately punish himand to not determinately not punish him. In other words, we are obligatedto make it the case that it is indeterminate whether or not Frankenstein ispunished. But that might not be within our power. Perhaps we could throwFrankenstein back into the matter-replacement machine. In that case,justice would be served since it would be indeterminate whether or notwe carried out the death sentence. But if we do not have something like

11 Where Δ is the determinacy operator.

70 ryan wasserman

this available to us, we are stuck. We must either determinately punishor determinately not punish. But we are obligated to refrain from both.Hence, we have a moral dilemma.

i nd e t e rm in ac y and e p i s t em i c i sm

Having identified the source of our anti-indeterminacy conviction, we cannow return to the original argument against the complex view. Given thepreceding, we can formulate the argument as follows:

The Indeterminacy Argument(1) If the complex view is true, then the Indeterminacy of Personal Identity

is true.(2) If the Indeterminacy of Personal Identity is true, then the Indeterminacy

of Obligation is true.(3) The Indeterminacy of Obligation is false. So,(4) The complex view is false.

We will consider two objections to this argument, the first of which is theepistemicist response.Epistemicism is a view on vagueness often associated with Timothy

Williamson (1994).12 In a slogan, the view says that vagueness is ignorance.Williamson believes, for example, that there is a precise line between thebald and the non-bald – we just do not know where that line is. A borderlinecase of baldness, then, is either a case of unknowable baldness or a case ofunknowable non-baldness.Slightly more carefully, Williamson identifies vagueness with a specific

kind of ignorance: ignorance due to semantic plasticity.13 I will illustrate theidea of plasticity and then explain its connection to ignorance.Let us focus on the predicate “bald.”The extension of this predicate is fixed by its use in accordance with

the semantic laws that link use and meaning. Of course, there are manycandidate properties to serve as the semantic value of “bald” – there is theproperty of having fewer than 1,000 hairs on one’s scalp, the property of havingfewer than 1,001 hairs on one’s scalp, and so on.14 Let us suppose, for the sakeof argument, that the overall pattern of use determines that “bald” expressesthe first property on our list, so that someone is bald just in case he has fewer

12 For an alternative version of epistemicism, see Sorenson (2001).13 I owe this phrase to Hawthorne (2006).14 These are all “candidates” in the sense that none of them is obviously not the semantic value of

“bald” – they are all in the running, so far as we can tell.


than 1,000 hairs on his head. That is how things are, but things could havebeen slightly different – a few more people could have been a little lessinclined to apply the term “bald” to individuals with 1,000 hairs. If usagehad differed enough, the semantic laws would have drawn slightly differentlines. So, it very easily could have been the case that “bald” picked out theproperty of having fewer than 999 hairs on one’s scalp. It is in this sense thatthe predicate “bald” is semantically plastic – it could have very easily been“bent” toward a different semantic value.

Now, suppose I believe a particular man to be bald, just on the basis ofobservation. And suppose that the man in question has exactly 999 hairs onhis head. In that case, my belief will be true. But,Williamson says, my beliefwill not count as knowledge. His explanation begins with the thesis thatknowledge requires safety. If someone forms a true belief, but could just aseasily have formed a false belief, then he is not being safe – epistemicallyspeaking – and thus does not know. For example: suppose that I look at abig tree and guess that it has 1,872 leaves on it. That might turn out to betrue, but we would not say that I knew the precise number of leaves on thetree. That is because my vision is not sufficiently sensitive to small varia-tions. So, I would have made the same estimate, even if there had been oneleaf more or one leaf less. So, I could have easily been mistaken. So, I do notbelieve safely. So, I do not know.

Williamson says that it is the same thing when it comes to my belief thatMr. 999 hairs is bald. In this case, my belief-forming mechanism is notsufficiently sensitive to small changes in use. I would still have believed thatMr. 999 is bald – I would still have assented to the sentence “Mr. 999 isbald” – even if the dispositions of other language-speakers had been slightlydifferent. However, if use had been different, I would have manifested afalse belief by assenting to the sentence “Mr. 999 is bald.” Hence, I couldeasily have been mistaken. Hence, I do not know.15

This is how semantic plasticity leads to ignorance in borderline cases.And that is the kind of ignorance that is characteristic of vagueness.

Now that I have explained epistemicism, I can explain the epistemicistresponse.

When it comes to the Indeterminacy Argument, the crucial question forthe epistemicist is: how are we to understand talk of “indeterminacy”? Somephilosophers (in at least some contexts) treat “indeterminate” as synonymouswith “indefinite” and thus understand talk of “indeterminacy” in terms ofvagueness (see, for example, Evans 1978 and Wright 1987, section 5). Others

15 For a more complete presentation of this argument, see Williamson (1994, ch. 8).

72 ryan wasserman

treat “indeterminate” as synonymous with “metaphysically unsettled,” so tosay that something is “indeterminate” is to say that there is simply no fact ofthe matter (or, perhaps, no determinate fact of the matter) (see, for example,Field 2003 and Barnett 2009). Since there are two readings of “indetermi-nacy,” there are two ways of understanding the Indeterminacy Argument.On each interpretation, the epistemicist will deny a different premise.Let us take the second reading first, since this is what Parfit means when he

says that personal identity admits of indeterminacy (Parfit 1984, pp. 213–15).And let us focus on the first premise of the Indeterminacy Argument.On the current reading, the Indeterminacy of Personal Identity amounts

to the following claim: it is possible for personal identity to be metaphysi-cally unsettled. Thus it is possible for there to be “yes or no” questions aboutpersonal identity where we cannot answer “yes” or “no,” since there is nofact of the matter either way. The epistemicist will deny that this conclusionfollows from the complex view. In other words, she will deny the firstpremise of the Indeterminacy Argument.It is not difficult to see why. Parfit’s combined spectrum presents us with

a traditional sorites argument. If the matter-replacement machine replacesone of Frank’s cells, he clearly survives. If we replace all of Frank’s cells, heclearly does not survive. In between, there are many borderline cases – likethe one where we replace 40 percent of Frank’s cells. In that case, it isunclear whether or not Frank survives. But, for the epistemicist, this lack ofclarity is simply the result of our epistemic limitations. There is still a factof the matter as to whether or not Frank survives, so our question in thiscase will not be empty. That is just to say that there is no indeterminacy, inParfit’s sense of the word.What about the other sense? Let us suppose that “indeterminate” simply

means “vague.” And let us turn our attention to the third premise of theIndeterminacy Argument.On this reading, the Indeterminacy of Obligation amounts to the follow-

ing claim: it is possible for questions about obligations to lack non-vagueanswers. For the epistemicist, this is true just in case it is possible for there tobe moral questions that we cannot know the answer to (where this igno-rance is due to semantic plasticity). I think the epistemicist will be happy toendorse that claim and, thus, to deny premise (3).The crucial point is that, when we are thinking like epistemicists, the

Indeterminacy of Obligations does not seem that bad. That is becauseindeterminacy, on this reading, is consistent with there being some fact ofthe matter. Go back to the car insurance case and the thought experimentinvolving God as your financial advisor. You are about to get into the


matter-replacement machine and you must decide whether or not to buythe insurance for tomorrow’s crash. Since this is a borderline case, you donot know whether you are going to be the person in the crash. But Goddoes! God knows where the line is between sufficient and insufficientcontinuity, just as he knows where the line is between the bald and non-bald individuals. The good news is that God is not going to shrug. He isgoing to give you an answer. Thus, there is no embarrassment in embracingthe Indeterminacy of Obligation.

ob j e c t i on s to e p i s t em i c i sm

I have suggested that the epistemicist has a simple reply to the IndeterminacyArgument: If “indeterminate”means “metaphysically unsettled,” then premise(1) is false. If “indeterminate” means “vague,” then premise (3) is false. Eitherway, the argument is unsound.

There are at least three reasons why one might reject this response.First, one might think that epistemicism is simply crazy.16 It is insane to

think that one hair could make a difference between being bald and beingnot, or that one penny could make a difference between being rich andbeing not, or that one cell could make a difference between being the sameperson and being someone else.

I agree that epistemicism is crazy, but that does not mean that it is false.Here, it is important to remind ourselves that the sorites paradox is a genuineparadox – the argument is intuitively valid, the premises are intuitively true,and the conclusion is intuitively false. Every solution will thus be counter-intuitive (at least initially), in which case the oddity of epistemicism need notcount against it.17

Second, one might have a special reason to dislike epistemicism when itcomes to the case of personal identity. Recall that (Williamson’s version of)epistemicism requires many candidates. In the case of “bald,” there was theproperty of having fewer than 999 hairs, the property of having fewer than1,000 hairs, and so on. The predicate “bald” expresses one of these proper-ties, but it would have expressed a different candidate if use had been a bitdifferent. This raises the question: what are the many candidates in the caseof personal identity?

16 Field, for example, describes Williamson’s view as “beyond belief” (2003, p. 458).17 Of course one might, on reflection, judge epistemicism to be more counterintuitive than its

competitors, and that might be a mark against the view. My point is that the view should not berejected just on the grounds that it is counterintuitive.

74 ryan wasserman

Let us focus on the proper name “Frank.” It is clear that this name refersto the person before the operation, but it is vague whether or not it refers toFrankenstein – the person who exists after the operation. If it is vague, then,for Williamson, there must be multiple candidates. The name “Frank”refers to one of these objects, but it would have referred to a different objectif use had been a bit different. What are the objects in question? The mostnatural suggestion is: different sums of instantaneous temporal parts.To illustrate the idea, let t1 be the first moment of Frank’s existence and

let t2 be the last moment before the matter-replacement machine is acti-vated. Let t3 be the moment at which the machine is activated18 and let t4 bethe final moment of Frankenstein’s existence (see Figure 2).According to the doctrine of temporal parts, persistence through time is

like extension across space. Just as an extended line is made up of smallerline segments (and, ultimately, point-sized parts), a persisting object is madeup of shorter-lived objects (and, ultimately, instantaneous temporal parts).19

Thus, the temporal parts theorist will say that, in the case described, there isa distinct person-stage – PS1, PS2, etc. – for each time in the story. There is

A

A B

PS3PS2PS1

Space

t1 t2Time

t3 t4

PS4

B+

Fig. 2: Different sums of person-stages.

18 To simplify matters, let us suppose that the machine works instantaneously, so that the moment ofactivation is also the moment at which the machine completes its matter-replacement work.

19 For a more careful characterization of the temporal parts view, see Sider (2001, ch. 3) and Wasserman(2004a).


also the sum of all the pre-operation person-stages – A – and the sum of allthe post-operation stages – B. Finally, there is the sum of all the person-stages from t1 through t4 – A⊕B. Now, suppose for the sake of argumentthat there is sufficient continuity between PS2 and PS3, so that the personafter the operation is the same person that we had before. In that case, thename “Frank” will refer to the sum A⊕B. However, if use had been slightlydifferent, then the predicate “is sufficiently continuous with” would nothave applied to PS2 and PS3, in which case the predicate “is the same personas” would not have applied to the pre-op person and the post-op person. Inthat scenario, the name “Frank”would refer to A. This would be yet anotherexample of semantic plasticity.20 This gives rise to ignorance and that,according to the epistemicist, is what gives us vagueness.

Of course, to offer this explanation, the epistemicist must first accept theexistence of temporal parts, as well as multiple cross-time fusions of thoseparts. Many philosophers do not accept the existence of such objects, sothey must reject the epistemicist response.

Here is a third and final worry for the epistemicist. Suppose, once again,that there is sufficient continuity between PS2 and PS3 (Figure 2). We areassuming this is a borderline case, which means that there is just barelyenough continuity. But still, as things stand, Frankenstein – the person afterthe operation – is the same person as Frank – the man from before theoperation. So, Frankenstein can be held morally responsible for Frank’scrimes. So it is permissible for us – and perhaps even obligatory for us – tocarry out the relevant punishment. But here is the problem: if we had usedcertain words – words like “person” and “similar” and “same” – just a littlebit differently, then that same sentence would have expressed a falsehood.The sentence “We ought to carry out the punishment” is true, but couldeasily have been false. The worry is not that that sentence could have easilybeen false because the non-linguistic facts could have easily been different.That is true, but unobjectionable. Rather, the worry is that the sentence inquestion could have easily been made false by our dispositions to use certainwords in certain ways.

To clarify: I am not claiming that we have the power to determine moralmatters with our words. The objection is rather this: if one is an epistemiciston moral matters, for example, one requires many candidate meanings forthe relevant moral terms. Moreover, the moral terms must be semanticallyplastic with respect to those candidates – for example, one must say that

20 For more on the relation between vagueness, personal identity and temporal parts, see Sider (2001,pp. 120–39) and Sider (2000).

76 ryan wasserman

“is morally obligatory” has one of the candidate semantic values but couldvery easily have had one of the others. And that seems wrong. It seems as ifour moral terms should be rigid in a way that many non-moral predicatesare not.I suggest that we understand rigidity, in this context, in terms of reference

magnetism. Start with David Lewis’s theory of meaning, according to whichsemantic content is determined by fit and eligibility (Lewis 1983, 1984; seealso Sider 2001 pp. xxi−xxiv and, especially, Sider 2012). Fit is a function ofuse – this is our contribution to content determination. Eligibility, on theother hand, is a contribution from the world. On Lewis’s picture, propertiesare ranked in terms of naturalness. Naturalness, like fit, comes in degrees:some properties are perfectly natural (being negatively charged, perhaps),some are less than perfectly natural (like being blue) and some (like beinggrue) are highly gerrymandered and unnatural. The more natural a prop-erty, the more eligible it is to serve as the semantic value of a predicate.21

Thus, Lewis claims that the content of a term is whatever does the bestoverall job of matching use and naturalness.22

On this “best candidate” theory, a natural property acts as a “referencemagnet.” To illustrate: take the predicate “gold,” which expresses a naturalproperty (not perfectly natural, but pretty natural). We could have differedslightly in our use of that term – we could have, for example, been somewhatmore disposed to call fool’s gold “gold.” In that case, some other candidatesemantic value – like the property of being gold or fool’s gold –might have beena better fit with use. But the actual semantic value – being gold – would stillhave fitted fairly well, and it would have done a much better job with respectto naturalness. Hence, it would still have done the best overall job. Hence,“gold” would still have meant gold, despite the difference in use.It is a very different story when it comes to non-natural properties. Take,

for example, the case of “bald.” Once again, there are many candidatemeanings for this predicate – there is the property of having fewer than1,000 hairs on one’s scalp, the property of having fewer than 1,001 hairs on one’sscalp, and so on. All of these properties are equally natural, so the world doesno work – the semantic content of “bald” is determined solely by use, whichis why that predicate is so pliable.

21 The issue of what qualifies a property as perfectly natural is a contentious one; for further discussionsee Lewis (1986, pp. 59–69) and Sider (1996).

22 It is no part of the theory that exactly one candidate will score highest in both fit and eligibility. Theremay be ties. There may be cases where one property scores highest in fit and another scores highest ineligibility. What the theory says is that the reference relation picks out the property with the highestoverall score.


We can now put the final worry as follows: on the epistemicist picture,moral predicates are more like “bald” than “gold.” And I think that a moralrealist – a seriousmoral realist – should resist this conclusion. For the seriousmoral realist, moral terms pick outmetaphysically important properties.23Theproperty of being morally obligatory, for example, is more like the property ofbeing gold, than being bald. So, the semantics of moral terms should be morelike the semantics of natural kind terms. So, moral terms are not semanti-cally plastic. So, the epistemicist response must be denied.

i nd e t e rm in ac y and sub j e c t i v i sm

I will close by considering a second response to the Indeterminacy Argument.Let us begin with the following case, due to Frances Howard-Snyder:

Agnes’s brakes fail. Should she continue straight into the busy intersection orshould she swerve into a field? Add to the story what Agnes does not and cannotknow, that the following two counterfactuals are true. If she were to enter theintersection, other cars would swerve and narrowly miss her car, no one would behurt and she would stop on the slight hill across the intersection. If she were toswerve into the field, she would hit a man hidden asleep in the grass and kill him.(2005, p. 265)

The question is: should Agnes swerve? Some say “yes.” Some say “no.” Andsome say “yes and no.” In one sense, Agnes ought to swerve, since that wouldmaximize expected utility. (It is the best option from her point of view.) Inanother sense, Agnes ought to go straight, since that would maximize actualutility. (It is the best option from the fully informed point of view.) Thosewho give the third response distinguish between the subjective “ought” and theobjective “ought.” If “ought” is ambiguous in this way, that will have signifi-cant implications for our understanding of the Indeterminacy Argument.

Consider again the Indeterminacy of Obligation. One is obligated to dowhat one ought to do, so if “ought” is ambiguous, we have two differentreadings of that principle. There is:

The Indeterminacy of Objective Obligation: It is possible for questions aboutobjective obligations to lack determinate answers.

And:

The Indeterminacy of Subjective Obligation: It is possible for questions aboutsubjective obligations to lack determinate answers.

23 This way of thinking about moral realism is hinted at in Wasserman (2004).

78 ryan wasserman

Now return to the car insurance case. Remember, you have the option ofbuying car insurance for $1,000. The policy will cover the $10,000 of damagesthat will result when You* gets into an accident tomorrow. But it is indeter-minate whether or not you are You* and thus indeterminate whether or notyou will be the person in tomorrow’s car accident. Let us suppose that we areworking with a non-epistemic reading of indeterminacy, so there really is nofact of the matter. Our question is: are you rationally obligated to purchase thecar insurance? Or should you turn the salesman down?Objectively speaking, you ought to purchase it if and only if you will be

in the car accident and you ought to decline it if and only if you will not bein the accident. Since there is no fact of the matter about whether or not youwill be in the accident, there is no fact of the matter about what is to bedone, objectively speaking.But that does not mean that there is no fact of the matter about what is to

be done, subjectively speaking. In the subjective sense of “ought,” youought to maximize expected utility. In order to calculate expected utility,you must first assign credences and utilities to the relevant outcomes. Andthis brings up a very interesting question: what credence should you give toP when P is indeterminate? There are various ways of thinking about thisquestion. One could treat “indeterminate” as synonymous with “indefinite”and then think about the question as an epistemicist. But the question ismost gripping when one employs a semantic or metaphysical reading of“indeterminate”: how should you distribute your credence across P and ¬Pwhen there is no fact of the matter as to whether P or ¬P?I do not have a general answer to this question. But I will make a few

specific suggestions about a few specific cases.First, imagine that your car has been placed in a car crusher and that you

cannot free it until tomorrow. Unfortunately, the machine has just beenturned on. Lucky for you, the machine happens to be an indeterministicsystem – a signal has been sent from the start button, but there is a 40 percentchance it will fizzle out before the crushing mechanism is engaged. (Ofcourse, there is also a 60 percent chance that the signal will not fizzle, inwhich case your car will soon be crushed.) More good news: you happen tohave an insurance salesman with you. He offers you a limited-term policy for$1,000 which will replace your $10,000 car in the event that it is crushed.Should you buy the insurance? Intuitively, the answer is yes. Of course

you should buy the insurance! (At least if your utilities go with dollars andmoney is the only relevant consideration.)That is the intuitive reaction. Here is the theoretical support. First, we

have the Principal Principle: subjective credence ought to match known


objective chance.24 So, in this case, you should assign .6 to the propositionthat your car will be crushed and .4 to the proposition that it will not.Second, we have decision theory: once credences and utilities are set, thequestion of what you ought to do is just a matter of number crunching (seeTable 1). When I say, with the decision theorist, that you ought to buy theinsurance, I am using the subjective “ought.” Buying the insurance is whatmakes sense, given what you know. Objectively speaking, there is no fact ofthe matter, since what you ought to do depends on what the machine isgoing to do, and there is no fact of the matter about that.

Now go back to the original car insurance case, where you are being putinto a machine and where 40 percent of the molecules in your body aregoing to be swapped for Einstein cells. I suggest that these two cases arerelevantly similar, so you should approach the cases in the same way. Inboth cases, it is indeterminate whether or not you are going to be stuck witha crushed car. In one case, that is because it is indeterminate whether or notthe car will be crushed. In the other case, it is because it is indeterminatewhether or not it will be you with the crushed car. But still, it is indeter-minate in both cases. So, you should approach the cases in the same way. Inparticular, you should assign the same credences (.6 that you will survive; .4that you will not) and use the same decision theory matrix.

Now go back to the case of Frank and Frankenstein. Once again, it isindeterminate whether or not Frankenstein is the same person as Frank.And, once again, we are operating with a non-epistemic reading of indeter-minacy. Our question is: ought we to punish Frankenstein? Or are weobligated to let him go?

Objectively speaking, we ought to punish Frankenstein just in case he isguilty and we ought to let him go just in case he is not guilty. There is nofact of the matter as to whether or not he is the same person as Frank, so it isindeterminate whether or not he is guilty. So it is indeterminate what is tobe done, objectively speaking.

Table 1

Crushed Not-crushed Expected utility

Buy .6 × −$1,000 + .4 × −$1,000 = −$1,000Do not buy .6 × −$10,000 + .4 × $0 = −$6,000

24 For a more careful statement (and defense) of the Principal Principle, see Lewis (1980).

80 ryan wasserman

Subjectively speaking, matters are more complicated. Note, first, it isdifficult to assign utilities in this case. It is bad to punish the innocent and itis bad to let the guilty go unpunished. Many will say that the first is worsethan the second, but it difficult to put dollar amounts on these outcomes. Itis also difficult to assign credences in this case.25 You might think that, since60 percent of Frank’s cells remain, we should be .6 on the proposition thatFrank is Frankenstein (and .4 on the proposition that he is not). But that istoo simplistic. Credences should not map cleanly onto cell replacement, oreven onto degrees of continuity – after all, a .99 degree of continuity ispresumably sufficient for identity over time, and determinately so.Despite these complications, there may well be some fact of the matter

about what is to be done, subjectively speaking. For my own part, I amapproximately .5 on the claim that Frank is Frankenstein. And I think it isquite a bit worse to punish an innocent person than to let a guilty party go.So I say, let him go. You might disagree with that verdict, but the mainpoint is that there could still be a fact about what you should do, sub-jectively speaking, even if there is no objective fact in this case.To tie all of this back to the Indeterminacy Argument: I have suggested

that there are two ways of understanding the Indeterminacy of Obligation,so premises (2) and (3) are both ambiguous. On the subjective reading, (3) istrue, but (2) is false. As I have just argued, the Indeterminacy of PersonalIdentity is consistent with there being some determinate fact about whatought to be done, from the subject’s perspective. On the objective reading,(2) is true, but (3) is false. To deny (3) is to accept the Indeterminacy ofObjective Obligation, but that seems like an acceptable consequence.Crucially, objective indeterminacy does not entail divine shrugging, forGod can still offer advice on what He would do, were He in your shoes.That is just to say that there might still be some fact about what is to bedone, subjectively speaking.If personal identity is genuinely indeterminate, then that is all we can ask

for. Perhaps that is all we need.26

25 This second worry also applies in the previous car accident case.26 I thank the audience at the 2010 conference on personal identity in Obergurgl, Austria. Special thanks

to the conference organizers at the University of Innsbruck. I also thank Berry Crawford, DanHoward-Snyder, Frances Howard-Snyder, Hud Hudson, Ned Markosian and Dennis Whitcomb –all of whom read this chapter and provided me with helpful comments.


chapter 4

Personal identity and its perplexitiesHarold W. Noonan

i n t roduct i on

What is the problem of personal identity? What is the distinction betweenthe simple and complex views of personal identity? What distinguishes thesortal concept of a person from other sortal concepts? How can we accountfor indeterminacy in personal identity?

First, I characterize the problem of personal identity in a way that con-forms to Lewis’s dictum (1986) that there are no problems about identity.Second, I build on this to draw the distinction between simple and complexviews of personal identity. Third, I argue that the concept of a person isindexical and appeal to this contention to defend the neo-Lockean psycho-logical continuity account against the animalist’s “too many minds” objec-tion. Fourth, I argue that we cannot account for indeterminacy in personalidentity in terms of indeterminacy in identity and that complex theoristsshould accept that it requires a multiplicity of thinkers.

the prob l em of p e r sona l i d ent i t y

Lewis denies that there are problems about identity: “We do state . . .genuine problems in terms of identity. But we needn’t state them so.Therefore they are not problems about identity” (Lewis 1986). If whatLewis says is correct, the traditional problem of personal identity shouldbe reformulated as a question that is not about identity. In what followsI offer such a reformulation.

Forget about personal identity for a moment and think about anotherfamiliar puzzle case: that of the statue and the clay. Statues can survivechanges pieces of clay cannot and vice versa. Hence, a statue (Goliath) andthe piece of clay from which it is made (Lumpl) may be actually at all timescoincident but possibly divergent. Our concept of a statue plausibly implies

82

that no statue can survive radical reshaping. Our concept of a piece of clayplausibly implies that any piece of clay must survive radical reshaping in whichall its matter is preserved in one coherent whole. These propositions specifypersistence conditions for statues and pieces of clay and, as they illustrate,these persistence conditions are of two types.The proposition that no statue can undergo radical reshaping can be

expressed as:

Necessarily, if x is a statue then if the matter that constitutes x at t is radicallyreshaped at t, then x ceases to exist.

This specifies a “passing-away” condition for statues (this terminologycomes from Penelope Mackie, personal communication).The proposition that any piece of clay must survive radical reshaping in

which all its matter is preserved in one coherent whole can be expressed as:

Necessarily, if x is a piece of clay then if the matter that constitutes x at t is radicallyreshaped at t but preserved in one coherent mass, x survives.

This specifies a “preservation” condition for pieces of clay.Sortal concepts for persisting things are governed by such conditions

because they constrain the histories of the things they apply to, and suchconstraints can always be expressed in the form:

Necessarily, if x is an S then if x exists at t and t* then Rxtt*,

or in the form:

Necessarily, if x is an S then if Rxtt* x exists at t and t*.

The “passing-away” condition for statues is entailed by a principle of the firstform (stating that a statue cannot have radically different shapes at differenttimes) and the “preservation condition” for pieces of clay is entailed by aprinciple of the second form (stating that if the matter composing a piece ofclay is in one coherent mass at both of two times, whatever shape it is in, thepiece of clay exists at both times).These principles state necessary conditions for being a thing of sort

S. What distinguishes sortal concepts under which persisting things fallfrom non-sortal concepts under which persisting things fall (even ones thatnecessarily apply to a thing at any time it exists, like permanent bachelor) isthat they are governed by such de dictomodal principles (Mackie also arguesthat sortals provide (de dicto) passing-away conditions whereas concepts likepermanent bachelor do not, personal communication).

Personal identity and its perplexities 83

Questions about personal identity over time can similarly be rephrased asquestions about necessary conditions of membership of the kind person.One question we can ask, without mention of identity, is:

(Q1) What conditions R satisfy the following schema: (P1) Necessarily, if x is aperson then if x exists at t and t*, Rxtt*?

Another is:

(Q2) What conditions R satisfy the following schema: (P2) Necessarily, if x is aperson then if Rxtt* x exists at t and t*?

Solutions to the problem of diachronic personal identity will take the form:“all and only the following conditions satisfy (P1): . . . and all and only thefollowing conditions satisfy (P2): . . .”

Of course, in answering these questions we only specify some of thenecessary conditions of personhood. One other type of necessary conditioncan be thought of as a synchronic constraint, capturable in the form:

Necessarily, if x is a person then if x exists at t, Fxt,

where F represents an expression for a non-historical property, i.e. oneentailing nothing about the past or future.

In addition the general problem of the criterion of identity for personscan be formulated by asking for a specification of the relation R satisfyingthe condition:

(1) Necessarily, if x is a person and y is a person then (x = y iff Rxy).

(1) is equivalent to the conjunction of:

(1a) Necessarily, if x is a person then Rxx

and

(1b) Necessarily, if x is a person then if y is a person and Rxy then (x = y).

Together these imply that necessarily any person is (identical with) theperson R-related to it. They thus give a sufficient condition for identity withany person by giving a necessary condition of personhood (to say that anyperson is the person R-related to it is to say something of the form “anyperson is . . .” which is the form of a statement of a necessary condition).Thus all intelligible questions about personal identity are equivalent toquestions about the necessary conditions for being a person.

84 harold w. noonan

the d i s t i nc t i on b e tween the s imp l e and thecomp l e x v i ew o f d i a chron i c p e r s ona l i d ent i t y

I now offer an account of the distinction between the simple and the complexviews of personal identity which conforms to the Lewisean dictum.1

Traditional defenders of the complex view are Locke and Hume, defend-ers of the simple view Butler and Reid. We also think of Chisholm andSwinburne as defenders of the simple view and Shoemaker, Parfit, andLewis as defenders of the complex view. But how is the distinction to becharacterized?One difference is that defenders of the simple view emphasize the

distinction between diachronic personal identity and the identity of otherobjects; they insist that in the case of the other familiar things that figure inphilosophical enquiries about identity – ships, plants and so on – the correctview is the complex one. We therefore need an account of the distinctionwhich allows us to speak generally of “the complex/simple view of thediachronic identity of things of sort S” where S is a sortal term.Recall two of the types of constraint on personhood just distinguished.

There are synchronic constraints capturable in the form:

Necessarily, if x is a person, then if x exists at t, Fxt,

where F represents a term for a non-historical property. And there arediachronic constraints capturable in the form:

Necessarily, if x is a person, then if x exists at t and t*, Rxtt*.

At first pass, the complex view is that there are diachronic constraints of thisform not logically equivalent to or logically implied by the constraint “xexists at t and t*” or “x is a person and x exists at t and t*” and not entailed bythe totality of synchronic constraints. If we call a diachronic constraint onpersonhood logically equivalent to or logically implied by one of the twojust mentioned “trivial,” and call a diachronic constraint entailed by thetotality of synchronic constraints “redundant,” we can put this by sayingthat the complex view is the view that there are non-trivial, non-redundantdiachronic constraints on personhood. The simple view is that there are none.This fits well with the classification of the complex theorists listed above.

1 Parfit (1984) proposes that the simple view of personal identity is the view that personal identity doesnot consist in any “further fact.” Arguably this trivializes it: for any x, the fact that x = x does not consistin any further fact, so since for any x and y, if x = y the fact that x = y is the fact that x = x, for any x and y,if x = y the fact that x = y consists in no further fact.


For Locke, for example, personal identity consists in the relation of co-consciousness. So if a person x exists at times t and t*, x will be co-consciousat t with x at t*, a constraint which is non-trivial and non-redundant in thesense explained. Again, the familiar modern-day psychological continuityaccounts satisfy this definition of a complex view, as do the competing non-psychological, or bodily, accounts.

But this characterization of the complex view is inadequate. Oneversion of the simple view is that persons have bodily parts so they arenot souls, but personal identity is constituted by identity of soul(Swinburne and Shoemaker 1984). This implies a non-trivial, non-redundant diachronic constraint on personhood: if x is a person whoexists at t and t* then x has at t the same soul that x has at t*. This isnon-trivial, since it is not logically equivalent to “x (is a person and) existsat t and t*” (even if it is metaphysically equivalent in some sense). It isnon-redundant since even if a person needs a soul whenever he exists hemay have different souls at different times. Nevertheless, the view thatpersonal identity is constituted by soul identity (if nothing more can besaid about the latter) must be classified as a version of the simple view ofdiachronic personal identity.

To accommodate this point call a diachronic constraint on personhood,Rxtt*, “identity-involving” if its satisfaction requires that something whichis not a person exists at times t and t*. That x has at t and t* the same soul isan identity-involving diachronic constraint. We may now say that thecomplex view is that there are non-trivial, non-redundant, non-identity-involving diachronic constraints on personhood. The simple view is that theonly non-trivial, non-redundant diachronic constraints on personhood areidentity-involving.2

This characterization of the distinction accords with the widespreadidea that the simple view is that the diachronic identity of persons cannotconsist in the holding of any relations other than the relation of identityitself.

But I have to consider an objection. Swinburne suggests an account ofpersonal identity according to which personal identity is constituted by soulidentity and the criterion of identity of souls, as (Aristotelian) substances, is

2 More precisely, the simple view is that the concept of a person is the concept of a sort of persistingobject not governed by non-trivial, non-redundant, non-identity-involving diachronic constraints.The complex view is that the concept of a person is the concept of a sort of persisting object which isgoverned by such constraints. A third view is that the concept of a person is not a sortal concept at all.The proponent of the simple view needs to explain what distinguishes the concept of a person fromnon-sortal concepts (so does the proponent of the complex view, but for him it is easy).

86 harold w. noonan

identity of form and continuity of immaterial stuff (Swinburne 1984, p. 27). Onthis proposal there is a non-trivial, non-redundant, non-identity-involvingdiachronic constraint on personhood: if person x exists at t and t* the soulwhich x possesses at t must have continuity of immaterial stuff with the soulwhich x possesses at t*. This is not an identity-involving constraint, sincecontinuity of stuff does not require identity of stuff, and by itself, withoutidentity of form, does not entail identity of substance, i.e. soul.I think the thing to say is that this “Aristotelian” account is, indeed,

contrary to Swinburne’s intention, a version of the complex view.Why not?It is a dualist view. But this does not rule out complexity. Locke’s account ofpersonal identity is dualist: it explains personal identity as consisting in co-consciousness of immaterial thinking things that “think in” persons. Butit is paradigmatically complex. A feature of the “Aristotelian” account notpossessed by Locke’s is that it entails that if a person exists at two times,some other entity – his soul – which determines his identity, exists at thosetimes. But this is also a feature of the brain account of personal identityaccording to which, although persons are not brains, a person persists just incase his brain persists (whether or not it carries psychological continuitywith it), which is paradigmatically a version of the complex view (brainsare paradigms of material objects whose persistence does not require persis-tence of matter but requires, but is not entailed by, material continuity).Swinburne’s “Aristotelian” theory, according to which souls can undergoa complete replacement in their psychologies, but, though only parts ofpersons, determine personal identity, is a sort of dualist version of the brainaccount.I conclude that we can define the complex view of personal identity as the

view that there are non-trivial, non-redundant, non-identity-involvingdiachronic constraints on personhood.So defined it comes in two varieties. According to the first, if a person

exists at times t and t* there exist at those times some entities distinct fromthe person, the obtaining of a certain relation between which suffices for theperson’s existence at both times. Locke’s view is of this variety. If a personexists at t and t* then there are thinking substances S and S existing at t andt*, such that S at t is co-conscious with S* at t*. The Lewisean psychologicalcontinuity account has the same structure: if a person exists at t and t* thereare person-stages existing at t and t* which are psychologically continuous(R-related). Complex accounts of diachronic personal identity of this varietyare “two-level”; they ground the identity of persons in a relation (otherthan identity) between non-persons. According to the second variety ofthe complex view, the existence of a person at two times does not require the


existence at those two times of any entities distinct from the person (whichare not adjectival on, or states of, the person) related in a way that suffices forthe existence of a person at both times. Shoemaker’s account of personalidentity is of this kind, as also, perhaps, is Van Inwagen’s.3 Since complexaccounts of this second type are possible, denying that personal identity isgrounded in a relation between other things does not entail acceptance ofthe simple view.

I finish this section with a comment on the link between the simple viewand the possibility of indeterminacy in diachronic personal identity.

Standard puzzle cases of diachronic personal identity are often describedas cases where, though everything else is clear, diachronic personal identityis unclear. If the Brown/Brownson case (Swinburne and Shoemaker 1984),in which Brown’s brain is transplanted into Robinson’s body with cons-equent transference of psychological traits, is thought of as such, forexample, then the situation must be one in which it is perfectly determinatethat there is exactly one person where Brown is before the transplant,perfectly determinate that there is exactly one person where Brownson isafter the transplant, but indeterminate whether Brownson is Brown, henceindeterminate whether there is someone, exactly one person, where Brownis before the transplant and where Brownson is afterwards.

According to either variety of the complex view, as I have explained it, thisindeterminacy can be due to indeterminacy whether a non-trivial, non-redundant, non-identity-involving diachronic constraint on personhood,Rxtt*, is satisfied, where t and t* are the two times pre- and post-transplant.But according to the simple view there are no non-trivial, non-redundant,non-identity-involving diachronic constraints on personhood. So, giventhat everything but the fact of diachronic identity is clear, the simpletheorist can account for the indeterminacy only by saying that it is indetermi-nate whether some identity-involving constraint on personal identity, Rxtt*, issatisfied, e.g. that it is indeterminate whether the soul x possesses at t is the soulx possesses at t*. In this case there cannot be any indeterminacy in the terms“the soul x possesses at t ” and “the soul x possesses at t*,” for everything is clearexcept the fact of diachronic personal identity. Indeterminacy in diachronic

3 Shoemaker’s neo-Lockean psychological continuity account of personal identity is neutral betweenmaterialism and dualism. So the existence of a person at two times does not require, on this account,the existence at the times of brains or material particles. Shoemaker is also opposed to the four-dimensionalist Lewisean position. Finally, in allowing the possibility of dualism he is not allowing thepossibility of a Lockean position according to which as well as persons there are thinking substanceswhich “do the thinking for” persons (Shoemaker 1985). Van Inwagen (1990b) thinks that humanpersons are human animals and that in the material world there are at most living organisms andsimples. But he does not claim that the existence of organisms requires the existence of simples.

88 harold w. noonan

personal identity according to the simple view as characterized above, then,must be accounted for in terms of indeterminacy in identity itself, in a way thatoffends against Evansian (Evans 1978) sensibilities.4

the i nd e x i c a l i t y o f the conce p t o f a p e r son

So far nothing I have said about the concept of a person requires it to befundamentally different from any other sortal concept. But it is differ-ent, because it is indexical. We can formulate the problem of personalidentity over time by asking: “What are the constraints on the history ofa person?” Or each of us can formulate it for himself by asking, “Whatare the constraints on my history?” The latter, indexical, formulation isthe basic one. Our interest in personal identity is, of course, funda-mentally an interest in our own identity (the problem of personalidentity is more illuminatingly titled “the problem of the self’s iden-tity”). Thus “person” as it occurs in standard formulations of theproblem just means “object of first-person reference,” i.e. that whicheach of us calls “I.”5

Thus a gap exists between the concept of a person and the concept of athinker of first-person thoughts. That a thinker of first-person thoughts isan object of first-person reference is not trivially analytic. The neo-Lockeanpsychological continuity theorists (Lewis, Parfit, Shoemaker et al.) canexploit this gap to respond to the “thinking animal” or “too many thinkers”objection.The problem of the thinking animal (Olson 1997b), briefly, is that it seems

indisputable that all normal healthy adult human beings are thinkers. But so,by definition, are persons. However, according to the psychological continu-ity theorist, persons are not human beings (they differ in their persistenceconditions). So the psychological continuity theory entails the existence oftoo many thinkers. Moreover, it creates an impossible epistemic problem:how do I know that I am the person and not the coincident human animal

4 Evans denies that statements of identity can be indeterminate when both singular terms are determi-nate in reference. This is what we think possible in the case of statements about friendship, forexample. But Evans takes to be ruled out what correspond to two prima facie possibilities in the case offriendship: that “friend of” is indeterminate in reference or that it is epistemically indeterminate in thesense that though it has a particular reference it does so unknowably because it could easily have hadanother. What remains is just the possibility that the relation of identity (the relation of friendship)itself is vague. Thus Evans sees himself as arguing against the possible ontic indeterminacy of identity.

5 Note that the move from the indexical to the non-indexical formulation embodies the substantialassumption that there is some kind of thing for which persistence conditions can be given to which allobjects of first-person reference belong.


thinking falsely that he is a person? Finally, if human animals, in additionto persons, are thinkers, they must be persons after all, since their thoughtshave whatever complexity and sophistication any definition of “person” couldrequire – they have the same thoughts, after all, as the persons with whom,on the psychological continuity account, they “cohabit.” So the neo-Lockean’s attempt to identify the persistence conditions for persons collapsesinto incoherence, since he must acknowledge different kinds of person withdifferent persistence conditions.

I think a defender of the neo-Lockean position can respond to the firstcharge against it – that it entails the existence of too many thinkers – byaccepting that his theory entails that in a case in which the line of psycho-logical continuity goes one way and the line of biological continuity goesanother, there are two temporarily coincident non-identical entities present –the person and the human animal – each of which is, say, thinking that it willrain. But this is nomore problematic, he can say, than accepting that there aretwo coincident non-identical entities present, each of which weighs 140pounds, is flesh-colored, is breathing and is walking,6 which is in turn nomore problematic than accepting that the temporarily coincident statue andpiece of clay are two non-identical entities each of which weighs five pounds,is brown and is being carried across the room. The same kind of situation canexist as far down the evolutionary tree as there are interesting psychologicaldifferences between members of the same species. Thus, if in the brainswap story Brown and Robinson are two dogs, there are two temporarilycoincident non-identical entities where Brown is before transplant, each ofwhich weighs 50 pounds, has a thick black coat, is breathing and walkingand thinking about a bone. The “too many thinkers” objection is no morepersuasive than a “too many walkers” objection.

To the charges that the neo-Lockean psychological theory creates in thespecific case of persons – objects of first-person reference – an impossibleepistemic problem, and collapses into incoherence since it entails theexistence of different kinds of person with different persistence conditions,the answer must be more complex.

I claim: (a) that the most sensible thing for the psychological continuitytheorist to do in response to these charges is to distinguish the concepts ofan object of first-person reference and a thinker of first-person thoughts and tosay that when an animal coincident with but distinct from a person is a

6 So if the psychological continuity theorist is confronted with the challenge of the thinking animal orthe problem of too many thinkers, he is also confronted with the challenge of the animate person orthe problem of too many animals.

90 harold w. noonan

thinker of a first-person thought, not he, but the coincident person, is theobject of the thought and (b) that this is not a reductio of the psychologicalcontinuity theory.My argument is as follows:

(1) All persons and only persons are objects of their own first-personreference.I take this to be trivially analytic, given that the concept of a person is an

indexical concept. (Sometimes, of course, we pretend that something whichis not a person is an object of first-person reference: I put a sign on my doorsaying “I am unlocked, please come in and wait.”)(2) All persons are psychological continuers.This is a formulation of the standard neo-Lockean view.

(3) Some normal healthy adult human animals are not psychologicalcontinuers.

This is undisputed by neo-Lockeans.(4) All normal healthy adult human animals are thinkers of true first-

person thoughts.This is what the animalist7 urges on the neo-Lockean as an evident truth.

Some neo-Lockeans, notably Shoemaker (2007), deny it. I explain belowwhyI disagree with Shoemaker.From (1)–(4) follows:

(5) Some normal healthy adult human animals have first-person thoughtswhich are not about themselves but about psychological continuerswith which they are not identical.

I claim that (5) is what the neo-Lockean, the defender of (2), must say ordeny a trivial analytic truth, (1), or a plain empirical fact, (3), or an evidenttruth, (4).But is this not a reductio of neo-Lockeanism? I argue not.Given the triviality of (1) the substantive neo-Lockean claim (2) becomes

the claim that objects of first-person reference are psychological continuers. Itfollows (given empirical facts and evident truth) that there are more thinkers

7 Animalists say that we are animals, understood as entities with purely biological persistence con-ditions. Much of the plausibility of animalism derives from its apparent consonance with thecommon-sense thoughts that we are animals and that animals are the subjects of psychologicalattributes. But it is not obvious that animals have purely biological persistence conditions. Whenthe brain or cerebrum of my dog, Brown, is transplanted, I think that my dog goes with his brain orcerebrum and acquires a new body. I do not deny that persons are animals in the same sense that dogsare. But neither are biological animals in the sense of entities with purely biological persistenceconditions. Nevertheless, I think that entities with purely biological persistence conditions are also thesubjects of psychological attributes and hence the psychological continuity theorist must accept thatthere are many coincident thinkers.


than common sense acknowledges, some of which are not persons. But itis hardly news that neo-Lockeanism leads to this unless, like Shoemaker’sversion, it denies what seems evidently true elsewhere. Locke himselfdistinguished between thinking substances and persons,8 and so acknow-ledged the existence of thinkers distinct from persons that were not objectsof first-person reference. And, of course, the four-dimensional version of thepsychological continuity account, defended by Lewis, endorses a multi-plicity of thinkers that are not objects of first-person reference (Lewis 1979,1981) and entails that some thinkers of first-person thoughts are not theirobjects.

But perhaps what is wrong with my response to the problem of too manythinkers is that it creates a mystery. How can a normal healthy adult humanbeing, capable of sophisticated first-person thought, not be capable ofreferring to himself in the first-person in such thought, if the psychologi-cally indistinguishable person is? How can the human animal lack thiscapacity which its psychologically indistinguishable twin has?

But to ask this is to misunderstand the proposal. The neo-Lockean claimsthat it can be established by conceptual analysis that the following is a dedicto necessary truth:

All objects of first-person reference are psychological continuers.Or:

Only psychological continuers are objects of first-person thought.

Consequently, if A is a human animal it is a de dicto necessary truth that:

If A is not a psychological continuer A’s first-person thoughts are not thoughtsabout A.

It does not follow, given that A is not a psychological continuer, that A isnecessarily or essentially something whose first-person thoughts are notabout itself – no de re necessity follows. This is so since the neo-Lockeanstory does not entail that A could not have been a psychological continuer.To say that A is not a psychological continuer is to say something about A’shistory, and, consistently with neo-Lockeanism, A could have had a differ-ent history. The sense in which the neo-Lockean account entails thatA cannot think “I”-thoughts about itself is just that, qua something whichis not a psychological continuer, its “I”-thoughts cannot be about itself, justas qua someone who is not Prime Minister, I cannot be correctly referred toby the Queen as “my first Minister.”

8 Locke did not have to confront the thinking animal problem since he did not ascribe thoughts to men.

92 harold w. noonan

Discomfort may remain. According to the neo-Lockean, something that isnot a psychological continuer cannot be an object of its own first-personthoughts, no matter how sophisticated its thoughts.9 It cannot think that ititself is F, though it can think of itself that it is F (as Perry’s shopper (1979)thinks of himself that he is making a mess, but does not think that he himselfis making a mess).10 But why should this be so? However, the explanation issimple: to be a psychological continuer is to have a certain kind of history. Itis like being a past and future Prime Minister. But whether something is athinker now and the level of sophistication of its present thoughts cannotdepend upon its indefinitely long-distant future or its indefinitely long-distant past. So no matter how sophisticated a creature’s present thoughts,they cannot insure that it is a psychological continuer. Consequently, giventhe neo-Lockean thesis (2), they cannot insure that it is an object of its ownfirst-person thoughts.As a psychological continuer theorist I think that human animals think

“I”-thoughts whose referents are coincident but non-identical personsbecause I deny that whether something is a thinker now, and the level ofsophistication of its present thoughts, can depend upon its long-distant pastor future. On this point I disagree with Shoemaker. Shoemaker thinks that(biological) human animals are not thinkers at all, given the difference intheir histories, despite having, in the sense the animalist emphasizes, “whatit takes.”11 This is perfectly coherent. Many properties are had by things attimes in virtue (partly) of what happens at other times. This holds of theproperty being a widow. It also holds of the property being a statue. Whethersomething is a statue depends on its originating in some maker’s intentions,so the matter now constituting a statue cannot count as a statue just becauseit coincides with one. It is equally implausible that the matter momentarily

9 This is also trivially true on Shoemaker’s account, since on that account it is not possible forsomething that is not a psychological continuer to think (first-person thoughts) at all.

10 X thinks that he himself is F in the relevant sense if he thinks a first-person thought and in doing sothinks about himself. Perry’s shopper thinks about himself but not in the first-person way; thesophisticated animal thinks in the first-person way but not about itself. Geach (1976) suggests that weexplain “X thinks that he himself is F” in a Fregeanmanner as “for some way of thinking α, α = ego(X)and X thinks that α is F.”Here ego(X) stands for the way of thinking X employs when he thinks in thefirst-person way. In the case of the sophisticated animal the way of thinking it employs when it thinksin the first-person way is not a mode of presentation of itself. Similarly, when a Lewisean person-stageemploys a first-person way of thinking in thinking about the future, its thought is not a thought aboutitself but about the perduring person of which it is a part.

11 If we imagine hypothetical beings that are otherwise identical but differ from psychological continuersonly in their futures, e.g. because they come into existence fully formed, Shoemaker would still denythat they were thinkers, because of their differing futures. In fact, he would deny that human beingspermanently coincident with persons were thinking things (just as he would deny that Goliath wasLumpl).


coinciding with a person or an animal thereby qualifies as a person, an animalor a thinker. It is thus perfectly consistent to say both that I, a thinkingbeing, currently coincide with a human animal distinct from me, and thatthat animal is not a thinker, because like the matter which currently con-stitutes both of us, it is not thinking my thoughts or any thoughts. This isShoemaker’s position. But the difficulty is this. I and the animal with whichI currently coincide have been and will be coincident for many years. So ifI am thinking it is raining now and the coincident human animal is not, whyis this so? The difference must either be ungrounded or grounded in some farfuture difference between us (but there need be none) or grounded in somelong-past difference between us (e.g. such as the fact that I was never a fetusand the human animal was) or grounded in some difference in unactualizedpotentialities. But neither the thought that the difference between a thinkerand a non-thinker may be ungrounded, nor the thought that that differencecan be grounded merely in a difference in unactualized potentialities seemsacceptable. And though properties can be possessed at times in virtue of whathappens at other times, such long-past differences as are present in this caseseem irrelevant. Of course, there are properties closely related to the propertyof thinking it is raining that I now possess and the human animal does not ifit is not a person: for example, the property being a person thinking it israining. And it may be suggested that the property we in fact attribute withthe predicate “is now thinking it is raining” is a property of the form is aperson who now has property P. But although this is a possible response tothe thinking animal objection to the psychological approach it implausiblyrequires that far-distant past or future differences are relevant to whethersomething now satisfies the predicate “is a thinker.” Although it is absurd tosay that the matter now momentarily coinciding with me is thereby thinkingit is raining, it seems far from absurd to say that the (biological) humananimal with which, on Shoemaker’s view, I have been and will be so longcoincident, is now thinking it is raining. Surely it “has what it takes,” in termsof present equipment, and sufficiently appropriate past and future? Thinking,like breathing and walking, takes some time, which is why it is absurd todescribe the momentarily coincident parcel of matter as doing any of thesethings. Whether something qualifies as breathing or walking or thinking at atime depends on its properties in an extended period including that time.12

But the period does not have to be years long. So Lewisean day-long

12 A defender of Sider’s stage ontology (2001) would rather say “depends on properties possessed at othertimes by its temporal counterparts.”

94 harold w. noonan

person-stages can count as walking and thinking (Lewis 1981, p. 76),13 as canthe animal which is coincident with me for so much longer than a single day.So I accept that human (and other) animals are thinking things, accept themultiplicity of thinkers and resist the animalist’s attempted reductio of thepsychological continuity account at another point.

the comp l e x v i ew and inde t e rm in ac y

All complex theorists accept the possibility of indeterminacy in personalidentity. It is part of the complex view that personal identity is not radicallydifferent from the identity of other familiar types of thing that figure inphilosophical puzzle cases about identity – ships, statues, plants and so on.But the possibility of indeterminacy in these cases is undeniable. Moreover,to deny the possibility of indeterminacy in personal identity the complextheorist must make a large difference – the difference between life anddeath – depend on a trivially significant one, a fraction more or less brain-matter, one more or less insignificant memory, etc.I want to argue, first, that complex theorists cannot regard the source of

this indeterminacy as indeterminacy in identity itself – even if identity can beindeterminate. In fact no one can (and so, unsurprisingly, the simple theorist,who must do so, must deny the possibility of indeterminacy in personalidentity).There are two familiar types of case in which according to complex

theorists vagueness is possible for persons.First, a person’s spatio-temporal boundaries may be vague. In this respect

people are like clouds or mountains or (other) animals.Second, it may be vague whether a person existing at one time is a person

existing at another. If the brain transplant is botched it may be indeterminatewhether Brown, the person in room 100 before the operation, is Brownson,the person in room 101 after the operation. In this case it is clear that there isexactly one person at a certain time with a certain property, clear that there isexactly one person at a certain time with another property, but indeterminatewhether the person with the first property at the first time is the person withthe second property at the second time.It is again uncontroversial that we can construct puzzle cases about the

identity of entities other than persons with this structure. Familiar examples

13 Lewis adds “a stage cannot do everything that a person can do, for it cannot do those things a persondoes over a longish interval.”


in the literature are Shoemaker’s case of the four-centuries-old bridge ofSanta Trinita in Florence, Parfit’s example of the reconstituted club andShoemaker’s spatial example of Alpha Hall and Beta Hall and the linkingwalkway (Shoemaker and Swinburne 1984).

Now in all these cases we can formulate statements involving the conceptof personal identity which are vague. It will be vague in cases of the first sortwhether the person in place P1 at time t1 is identical with anyone alive atanother time t2. And in cases of the second type it will be vague whetherthe person in place P1 at time t1 is identical with the person in place P2 attime t2. But there will also be indeterminate statements not involving theconcept of personal identity. In both cases it will be indeterminate whethersome person is at place P1 at time t1 and at place P2 at time t2 (though in thesecond case it will also be determinate that there is some person (exactlyone) in place P1 at time t1 and determinate that there is some person (exactlyone) in place P2 at time t2). Other statements not involving personal identitywill similarly be indeterminate, e.g. whether someone is fat at one time andthin at another.

These statements can be trivially reformulated to bring in the concept ofidentity. But it does not follow that their indeterminacy can be explainedby appeal to the indeterminacy of identity. Suppose it is indeterminatewhether the cleverest boy in the school is tall. Now this statement is logicallyequivalent to one in which the concept of identity is employed – “the cleverestboy in the school is identical with someone tall” – but indeterminacy inidentity can be no part of the explanation of its indeterminacy. It may beindeterminate because “the cleverest boy in the school” is indeterminate inreference between a tall referent (who, say, excels at math) and a short referent(who excels at English). Or it may be that, in accordance with the epistemicview, “the cleverest boy in the school” refers to a single (tall) person butunknowably so because there is another (short) candidate which it would havereferred to if our linguistic practices had been marginally different (if we hadcounted ability at math as slightly less important) but we cannot know thatour linguistic practices are not thus marginally different. Or it may be that“the cleverest boy in the school” refers to a single vague object with a ratherlarge region of spatial (vertical) indeterminacy.14 Or it may be that “thecleverest boy in the school” determinately and knowably refers to a perfectly

14 Tomake sense of this last possibility consider this situation. Amouse has a partially severed tail, whichis no longer clearly part of it. Some philosophers suppose that there is a single vague object which isdeterminately the reference of “the mouse” but is indeterminate in spatial extent (so it is indetermi-nate whether it is 5 inches or 12 inches long).

96 harold w. noonan

precise object, a boy who is manifestly intellectually superior to every otherboy in the school but is of middling to tallish height and that the predicate “istall” is vague so that the statement is indeterminate. What is clear is that theexplanation of the indeterminacy of the statement that the cleverest boy inthe school is tall cannot be that the relational expression “is identical with” orthe concept of identity or the relation of identity is in some way vague (in away that it might plausibly be thought that either the relational expression“is a friend of” or the concept of friendship or the relation of friendship is insome way vague), since this relational expression does not occur in thestatement. If a statement is indeterminate some expression occurring in itmust be the source of the indeterminacy (by what it refers to being referen-tially or epistemically indeterminate or by determinately referring to an entitywhich is in itself indeterminate in its nature); it can be neither sufficientnor necessary that some expression not occurring in it is characterized byindeterminacy.15

It is the same with the statements given above, not involving any expres-sion for identity, which are indeterminate in the puzzle cases of personalidentity. In these cases the crucial concept involved is that of a person (we cansuppose that the application of all the other concepts is clear).16 Once againthe source of the statement’s indeterminacy in truth-value may be indeter-minacy in reference (indeterminacy in the reference of “person”), epistemicindeterminacy (ignorance about the reference of “person”) or ontic indeter-minacy (the existence of vague objects which are indeterminate examples ofpersonhood). But its source cannot be any indeterminacy in identity since thestatements do not contain any expression for identity.That this is so, of course, follows from the thesis argued above, that questions

about personal identity are not questions about identity but about the con-ditions of personhood. If so, indeterminacy in their answers cannot be dueto indeterminacy in identity – but I have here argued for this consequenceindependently.

15 There are counter-examples to this claim, but ones clearly irrelevant to our concerns. If my shirt isborderline blue “My shirt is in the extension of ‘blue’” is indeterminate but contains no vagueexpression (it contains the quotation name of the word “blue” not the word “blue”).

16 Alternatively, the indeterminacy of “some person exists at t and t*”may be located in the vagueness ofthe relational expression “exists at,” which may be thought to be vague in the way “friend of” is vague.But in the botched brain transplant it is also determinately true that there is exactly one person whoexists at t and exactly one person who exists at t*. However, vagueness in “exists at” may be appealedto to explain vagueness in explicit statements of spatio-temporal boundaries – the idea being thatone’s extent is vague in the way that the extent of one’s circle of friends is vague. I put this implausibleidea aside in what follows.


What I wish to argue now is that whatever account of indeterminacy heoffers any complex theorist must accept the existence of a multiplicity ofcoincident or nearly coincident thinkers in cases of indeterminacy inpersonal identity or allow that tiny differences in spatial extent or far-offfuture or past differences between some pairs of temporarily (nearly)coincident entities can disqualify one of them from being a current thinker.

Let us think about the first kind of vagueness noted above, vagueness inspatio-temporal boundaries. If persons have vague boundaries then on anyaccount of their vagueness that locates it in indeterminacy of reference thisis a matter of the existence of a multiplicity of candidates, largely coincidentin spatio-temporal extent, between which our singular terms are indetermi-nate in reference.17 On an epistemic story there must likewise be a multi-plicity of such candidates in nearly the same place and time. These will allnow have what it takes to be thinkers unless it is allowed that the tinydifferences in their spatial extents or far-off past or future temporal boun-daries matter to whether they count as thinkers now. To deny the multi-plicity of thinkers it seems that the complex theorist who does not want toallow this must insist that there is in fact in any such case only one personcandidate present – an object with ontically indeterminate boundaries. Butcan this view be maintained?

Let us call the (putatively unique) person candidate present in such a case“T.” The view is that “T” determinately refers to a single vague object, aperson, an object whose boundaries are either or both spatially and tempo-rally fuzzy. There may indeed be many spatio-temporally overlappingprecise objects in the vicinity of T, but none of them is T, since they areall determinately not persons, but at most person-constituters. Persons arevague objects and there is only one such object where T is.

An alternative view, still in conformity with the existence of vague objects,is: “T” is a singular term of indeterminate reference. There is a multitude ofspatio-temporally imprecise, massively overlapping vague objects betweenwhich its reference is indeterminate (none is determinately a person, sincethey are all determinately distinct from one another and there is determi-nately only one person where T is). Only vague objects can qualify as persons:hence any precise objects in the vicinity – if any – are at most person-constituters, not persons. But this is not enough to secure “T” a determinatereference, since there are many massively overlapping vague objects roughly

17 In accordance with the previous footnote I assume that it is not the relational expression “exists at”which is indeterminate in reference. I leave it as an exercise for the reader to see what effect discardingthis assumption has on the following arguments.

98 harold w. noonan

where T is and hence “T” is indeterminate in reference between them. Onthis alternative the answer to the question whether “T” or T is vague is: both.Is the view that there must be a unique vague object in what we think of as

the vicinity of T defensible? It is hard to see how the possibility of massivelyoverlapping vague objects, like the possibility of massively overlapping preciseobjects, can be denied. If persons and other animals are vague objects,conjoined twins provide an example. Consider also this variant on the familiarexample of Tibbles and Tib, the “tail complement” of Tibbles, i.e. that objectwhich is all of Tibbles without his tail.18 Suppose Tibbles has been in a scrap,and both his ears are partially severed so it is indeterminate whether they areparts of Tibbles any more. According to the vague object theorist, Tibbles is avague object. So is Tib. So Tib and Tibbles are two massively overlappingvague objects. Consider now the left-ear complement of Tibbles and theright-ear complement of Tibbles. These are also vague objects (it is indeter-minate whether the right ear is part of the left-ear complement) and they aredistinct from Tibbles and each other since parts that are determinately notparts of, for example, the left-ear complement are indeterminately parts ofTibbles and the right-ear complement. So they are massively overlappingvague objects, differing only with respect to the region of indeterminacy.(Examples of temporally overlapping vague objects can be similarly imag-ined.) Returning to the case of T, we can argue as follows: if there is a vagueobject where T is, there could be an exactly similar vague object elsewhere. Ifso, there could instead be a vague object elsewhere very, but not exactly,similar, differing perhaps only in the region of its indeterminacy (as the left-ear complement of Tibbles differs from the right-ear complement of Tibbles).But if there could be such a very similar vague object elsewhere there could bea vague object exactly similar to it where T is, with a location massivelyoverlapping the location of T.Moreover, suppose, with the first, that is, the unique vague object, view, that

there is only one vague object in the vicinity of T, a person to which “T”determinately refers. There is a counterfactual community using singular andgeneral terms similar to our terms “T” and “person” – let these be “T*” and“person*” – which they apply very largely as we do “T” and “person” but withsome slight differences. Some few particles we say are indeterminately parts of Tthey say are determinately parts or not parts of T* and/or some few particles wesay are determinately parts/not parts of T they say are indeterminately parts ofT*. If the many vague objects view is correct we can charitably interpret their

18 Some philosophers deny the existence of Tib and indeed tail. For them the example of Tibbles willnot convince. The other arguments to be given do not depend upon its acceptance.


term “T*” as referring to something other than our term “T” and do not have toascribe mistaken views to them. But if the unique vague object view is correct,their term “T*” is either without reference or refers to a precise object or is co-referential with our term “T,” in which case if we are speaking the truth, theyare making false statements about T. But their disagreement with us may befaultless; there need be no cognitive or perceptual failing from which they, orwe, suffer. The same is true of the innumerable other counterfactual linguisticcommunities we can imagine differing from each other and from us as we differfrom the first counterfactual community just imagined. Only one of thesecommunities can be right about the boundaries of T if the unique vague objectview is correct. So if this view is correct it is overwhelmingly likely that we arejust wrong about the boundaries of T (according to this view there is just onevague object in the vicinity, which is the determinate reference of our term “T,”an object about the boundaries of which there is a matter of fact, but this viewdoes not require that our opinions about these boundaries are correct). ButT can be any vague object which is a thinker and according to the vague objectview persons are typically such vague objects, so it implausibly implies that allthe time we are very probably making incorrect ascriptions of (indeterminate)temporal and spatial boundaries to persons.

Consider now the second kind of vagueness noted above − vaguenesswhether a person existing at one time, Brown, is the person existing atanother time, Brownson. In this case it is indeterminate whether there is(exactly) one person in place P1 at time t1 and at place P2 at time t2, though it isdeterminately true that there is exactly one person in P1 at t1 and determi-nately true that there is exactly one person in P2 at t2. This is similar to the wayin which, in Shoemaker’s case of AlphaHall and BetaHall, it is indeterminatewhether there is (exactly) one building in which Smith, who is located inAlphaHall, and Jones, who is located in Beta Hall, are both lecturing, thoughit is definitely true that Smith is located in just one building and Jones islocated in just one building. On any account according to which there aremultiple person candidates present (one existing both at t1 and at t2, oneexisting at t1 and not existing or only indeterminately existing at t2 and oneexisting at t2 and not existing or only indeterminately existing at t1) it is nottrue that there is only one thinker existing at t1 (t2) unless far-off future or pastdifferences between a pair of temporarily coincident entities can disqualifyone of the pair from being a current thinker. However, because of thetemporal symmetry of the situation (comparable to the spatial symmetry ofthe Alpha Hall/Beta Hall example) one vague object cannot suffice to explainthe indeterminacy. It is definitely true that a person exists at t1 and definitelytrue that a person exists at t2. So to postulate a single vague object which is

100 harold w. noonan

indeterminately a person is inconsistent with the description given of thesituation as one in which it is definitely true that a person exists at t1 anddefinitely true that a person exists at t2. And if there is a single vague objectwhich is determinately a person but which is such that it is indeterminatewhether it exists at the earlier time and indeterminate whether it exists at thelater time it is false that, as specified, it is definitely true that a person exists atthe earlier time and definitely true that a person exists at the later time. So toaccept this kind of indeterminacy in personal identity the complex theoristmust accept that in this kind of case there are many thinkers present at t1 andat t2 or else allow that far-off future or past differences between some pair ofcoincident entities can disqualify one of any pair from being a currentthinker. If, as I have argued above, this is implausible, the complex theorist,who has to allow this kind of indeterminacy in personal identity, must acceptthat any such case involves a multiplicity of thinkers.


part i i

Arguments for and against simplicity

chapter 5

How to determine which is the truetheory of personal identity

Richard Swinburne

the p rob l em

The simple view of diachronic personal identity holds that personal identityis not constituted by continuities of mental or physical properties or of thephysical stuff (that is, the bodily matter) of which they are made, but is aseparate feature of the world from any of the former, although of course it iscompatible with personal identity being caused by such continuities. Onthe simple view, as I shall understand it, a person P2 at t2 can be the sameperson as a person P1, at an earlier time t1, whatever the physical or mentalproperties and whatever the body possessed by each person. P2may not at t2remember1 anything done or experienced by P1 at t1 or earlier, and may havean entirely different character and a totally different body (including brain)from P1 at t1. The main arguments in favor of the simple view consist inadducing thought experiments in which persons undergo radical changes ofmental life and bodily constitution, and in which – it is claimed – it is“possible” that they continue to exist; from which it follows that continu-ities of the kind mentioned are not necessary for personal identity.I begin with one example (of very many which have been put forward) to

indicate the role of thought experiments in supporting the simple view.Suppose I have a severe brain disease affecting the right brain hemisphere.The only way to keep my body functioning is to replace this hemisphere. Sothe doctors remove my current right hemisphere and replace it by a righthemisphere taken from a clone of me. The new right hemisphere, let ussuppose, contains the brain correlates of (that is the neurons, the states ofwhich are the immediate causes of ) similar but slightly different memoriesand character traits from mine. The resulting person would presumably to

1 Ordinary language sometimes assumes that only true beliefs are correctly called “memories.” Thus itassumes that if I am correctly said to have a “memory” that I did so-and-so, I really did so-and-so. Ishall not follow that usage here, but shall understand by “memory” what on that usage would only bean apparent memory: it seeming to the subject that he “remembered” so-and-so.

105

some extent behave like me and remember having done what I did and alsoto some extent behave like my clone and remember having done what myclone did (when what I did was different from what my clone did). Nowsuppose that the disease spreads to the left hemisphere, and that too – twoyears later – is replaced by a left hemisphere taken from a different clone ofme, again containing the brain correlates of similar but slightly differentmemories and character from mine. Then my body would be directed by abrain made of totally different matter and sustaining rather different mem-ories and character from those I had two years previously. Yet presumably tosome extent, but to a lesser extent than after the first operation, the resultingperson would still behave like me and remember having done what I did.

But would the resulting person be me? That person would be a personlargely continuous with the earlier me two years ago, apart from having hadtwo large brain operations. One might think that the continuity of manymental and physical properties over this period has the consequence that thesame person continues to exist. Yet the resulting person would have none ofthe brain matter and only some of the memories and character which werepreviously mine. I suggest that it is totally unobvious whether in thissituation the resulting person would or would not be me. Yet the question“Would the resulting person be me?” is logically equivalent to the question“Would I survive the operations?” and so have the (pleasurable or unpleas-ant) experiences of the resulting person. And surely no one about to have aserious operation can think that the question of whether he will “survive” abrain operation is simply a question requiring an arbitrary decision aboutwhich of two senses we should give to the word “survive.” We (or at leastmost of us) seem to understand the alternatives as mutually incompatiblefactual alternatives – that I survive, or that I do not survive – in one clear andnatural sense of “survive.” Yet it is totally unobvious what is the answer. Ifyou think that – one way or other – the answer is obvious, it is easy to alterthe thought experiment in such a way that the answer is no longer obvious.If you think it is obvious that the continuity of at least half the brain matterover each of the operations two years apart insures that I continue to exist,suppose the second operation to be performed after only one year or sixmonths. If you think it obvious that when half my brain matter is removedin one operation I no longer exist, suppose a series of operations replacingonly a tenth of the matter each time.

In such a situation, which I call an ambiguous situation, it does seempossible that I have survived (i.e. continued to exist), and possible that I havenot survived; and yet that we do not know (and have no further means offinding out) whether I have or have not survived. If what seems possible is

106 richard swinburne

indeed possible, my survival does not require any particular degree of strongphysical and mental continuities2 which make it obvious that I do survive. Itthen follows that the difference between the situations of different degrees ofcontinuity consists in the strength of the evidence that I continue to exist.Under normal conditions of very strong continuity of body (and in particularof the brain, the physical sustainer of mental life), memory (of what happenedto a person with that body) and character, it is enormously probable that Icontinue to exist; it becomes less and less probable until we reach theambiguous situation where it is as probable as not that I continue to exist.Why it is enormously probable that under those normal conditions I

continue to exist is first because it is a fundamental epistemological principlethat (apparent) memory beliefs are probably true (in the absence of counter-evidence), and my personal memories (that is memories “from the inside”about what I did and experienced) concern the actions and experiences ofthe person who had a brain strongly continuous with my present brain.Unless memories as such (in the absence of counter-evidence) are probablytrue (and so do not require to be rendered probable by evidence of someother kind in order to be probably true), we would know very little aboutthe world. For we depend on memory for all the knowledge which webelieve that we have acquired from others about history and geography, etc.;and while mymemory of these things may coincide with yours, at any time Idepend on my own memory of what others have told me for my belief thatour memories do coincide, and so the personal memory of each of us mustas such be probably true if we are to have virtually any knowledge at all. Andthe second reason why it is enormously probable that under those normalconditions I continue to exist is that the simplest, and so most probable,hypothesis supported by the strong continuity of memory and charactersustained by the same brain is that these are mental properties belonging tothe same person. It would be less simple, and so less probably true, tosuppose that the memory and character strongly continuous with theprevious memory and character sustained by a brain having strong con-tinuity with the previous brain are the memory and character of a differentperson. So being the same person does not entail strong continuity of brain,character and memory; but the latter is good evidence of the former. This isthe simple view.

2 By brain, memory or character being “strongly continuous” with a previous brain, memory orcharacter, I understand that there exist between them both what Derek Parfit (1984, ch. 10) callsstrong “connectedness” (that is strong similarity) and what he calls strong “continuity” (that isoverlapping chains of strong connectedness), the continuity of memory and character being causallysustained by the strong continuity of the brain.

The true theory of personal identity 107

Some philosophers hold that personal identity, like the identity ofartifacts, can be a matter of degree. On this view a later person can beonly partly identical to some earlier person, and so the result of suchoperations as I have described might be that the resultant person was onlypartly me. I do not myself think that it is logically possible that some personbe partly me. But even if this were a possible result of the operations, it doesnot seem to be a necessary truth that the operations will have this result,because the history of all the physical bits and all the mental propertiesassociated with them seems compatible with the subsequent person notbeing partly me. It still seems possible that, just as the resulting person isfully me if I have both heart and liver replaced, so after the half-braintransplants the resulting person is still fully me; and it is also possible that itis not me at all. Yet if we include the subsequent person being partly me as apossible result of the operations, we would now be ignorant about whichout of three (rather than two) possible results of the operations had in factoccurred. If what “seems possible” is possible, that I survive the operationsnot merely in part but wholly (or alternatively not at all), partial survival iscompatible with the simple view.

The alternative to the simple view, the complex view, claims that personalidentity is constituted (not merely evidenced) by a certain particular degree ofcontinuity of physical and mental properties and of the matter which forms aperson’s body. The main arguments in favor of this view are that theparadigm examples of personal identity are all ones in which there is con-tinuity of these kinds, and that the simple view leads to contradictions. Thereare innumerable varieties of the complex view according to which degrees ofcontinuity ensure the identity of a later person with an earlier person. Onevariety is the view that the concept of personal identity has no applicationoutside normal situations of strong physical andmental continuities. Anothervariety of the complex view holds that necessarily (not merely possibly, as inthe version of the simple view) personal identity is a matter of degree. Theweaker the continuities, the lesser the degree to which the later person is thesame as the earlier person. Again there is an issue about this variety of thecomplex view, as about the similar variety of the simple view, as to whetherthe notion of partial identity of persons makes any sense.

log i c a l po s s i b i l i t y

Which of these views is correct depends on what is possible, and so I come toa central topic of this chapter: what is it for some situation to be “possible,”and how can we know whether it is or not? We can have no discussable


knowledge of possibilities (or necessities or impossibilities) which cannot beexpressed in sentences, and so I will discuss the question of which situationsare “possible” (or whatever) by discussing which sentences describing situa-tions describe “possible” (or whatever) situations, and I shall call suchsentences “possible” (or whatever) sentences. In talking about possibility inthis kind of context, we are talking about the widest or weakest kind ofpossibility there can be, which I shall followmost contemporary philosophersin calling “metaphysical possibility.” Some state of affairs may not be practi-cally or physically possible, but it may still be metaphysically possible. Hencemetaphysical necessity and metaphysical impossibility are the narrowest orstrongest kinds of necessity or impossibility. Among metaphysical possibil-ities, etc., are ones discoverable a priori: that is, discoverable by reflection onwhat is involved in the claim made by the sentence. I shall call these logicalpossibilities, etc. No sentence could be more strongly impossible than a self-contradictory sentence. It claims both that something is so and also that it isnot so (and is normally expressed by a sentence of the form “s and not-s”). Forsuch a sentence could only be true if that something is so, and the sentenceasserts that it is not so. But any sentence which entails a self-contradictionwould be as strongly impossible as a self-contradiction. For similar reasons nosentence could be more strongly necessary than one whose negation entails aself-contradiction. Such necessities and impossibilities are logical ones, sincethey are discoverable by deriving the relevant contradiction by means of therules of the language.I see no reason to suppose that there are any other a priori impossibil-

ities as strong as those which entail a contradiction, or any other a priorinecessities as strong as those whose negation entails a contradiction. Tobegin with the case of impossibilities: what is asserted could only be apriori impossible if the impossibility is detectable merely by under-standing what is involved in what is actually said. To be impossible asentence must have the form of a declarative sentence in which thecomponent words already have a sense in the language. It will be asubject−predicate sentence, an existential generalization, or some otherone of many recognized forms of declarative sentences. It will – to put thepoint loosely – assert something about some substance or property orevent or whatever that it has or does not have some property or relation tosome other substance, property, etc.; or that there are or are not certainsubstances, properties or whatever. Words have a sense insofar as it is clearwhat are the criteria for an object, property or whatever picked out by theword being that object, property or whatever. They therefore delimit aboundary to the sort of object or property it can be or the sort of properties


it can have. Hence it will be inconsistent to affirm that an object pickedout by some word is of a kind ruled out by the very criteria for being thatobject. And the form of any sentence will exclude some alternative; and soit will be inconsistent to affirm the sentence together with that alternative.If a sentence is not inconsistent in these ways (or does not entail one thatis), we would have no reason to suppose that that sentence is a prioriimpossible. A similar argument shows that we would have no reason tobelieve that any sentence whose negation is not inconsistent in these waysis a priori necessary. So I shall assume that all logically impossible senten-ces entail a contradiction, and all logically necessary sentences are suchthat their negations entail a contradiction;3 and so assume that all declar-ative sentences which do not entail a contradiction are logically possible.

But what determines the rules of a language, and so the truth conditionsof sentences, and so what entails what? Sentences of a language mean whatmost of its speakers (or some group of expert speakers) mean by them. Eachof us learns the meanings of certain sentences by being shown manyparadigm observable conditions under which those sentences are regardedas true or false, and by being told of other sentences to which a speaker isregarded as committed by uttering those sentences, and other sentenceswhich are such that someone who utters them is regarded as committed tothe former sentences. We learn the meaning of a word by being taught thedifference to the meaning of a sentence made by that word playing a certainrole in the sentence. By being taught the meanings of individual words andof sentences of various forms, we may then come to an understanding of themeaning of a sentence in which those words are arranged in a certain way,even if we have not been shown observable conditions under which thesentence is regarded as true or as false. Showing a language learner “observ-able conditions”may involve pointing to them or describing them in termsalready introduced. We need to observe many different paradigm examplesof observable conditions under which a sentence containing a certain wordin various roles is regarded as true or false, and of the commitments speakerswho use sentences containing that word in various roles are regarded ashaving; and this allows us to acquire an understanding of the conditionsunder which some new sentence containing that word would be regarded astrue or false. We extrapolate, that is, from a stock of supposedly paradigmexamples (of observable conditions and relations of commitment) to an

3 Robert Adams (1987, pp. 213–14) argues (in effect) that there are logical necessities whose negations donot entail a contradiction. In Swinburne (2010, pp. 318–19) I argue that the example of a sentence bywhich he seeks to show this does not support his view.


understanding that the sentence would be regarded as true (or false, as thecase may be) under conditions sufficiently similar in certain respects to mostof the paradigm examples.This process normally leads, within limits of vagueness and minor

idiosyncrasies of use, to words (and longer expressions) and sentenceforms, and so sentences having a “correct” use. It leads, that is, to publicagreement about what in general are the circumstances in which a givensentence would be true and the circumstances under which it would befalse; and so to the commitments of sentences to other sentences. We maycall a rule for what one is objectively committed to by a sentence a rule ofmini-entailment. s1 mini-entails s2 if and only if anyone who asserts s1 isthereby (in virtue of the rules for the correct use of language) committed tos2. s1 entails sn if they can be joined by a chain of mini-entailments, such thats1 mini-entails some s2, s2 mini-entails some s3, and so on until we reach asentence which mini-entails sn.Given this agreement, we are then in theory in a position to determine the

logical necessity, possibility or impossibility of sentences. To show somesentence s to be logically impossible we need to find an agreed chain ofmini-entailments from s to a contradiction; and to show s to be logicallynecessary we need to find such a chain from not-s to a contradiction. Gettingagreement that such a chain has been found is, however, often a difficultmatter. An opponent of the claim that s entails a contradiction may challengesome suggested link in the chain − say the suggestion that pmini-entails q, byclaiming that q is not something to which anyone is committed when using pin the correct sense. This disagreement may be overcome if the proponent ofthe claim that s entails a contradiction can get his opponent to recognize somer such that pmini-entails r and rmini-entails q. Or the disagreement may bebypassed if the proponent can find a different chain of mini-entailments froms to a contradiction which an opponent will recognize as such.A sentence is logically possible if it does not entail a contradiction. Of

course any logically necessary sentence is logically possible. But to showsome other sentence to be logically possible (and so logically contingent)may be an even more difficult matter than to show a sentence to be logicallyimpossible or necessary. Sometimes it is very obvious that some sentencedoes not entail a contradiction, and so is logically possible. A true sentenceentails no contradiction, and if it is obvious that some sentence (e.g. “mydesk is brown”) is true, then it is obvious that it is logically possible.Sometimes too it is very obvious that some sentence, which may be false,entails no contradiction (e.g. “my desk is red”). And more generally it issometimes very obvious that some description of a world very different from


our world entails no contradiction. To show some disputed sentence s to belogically possible requires showing that it is entailed by a sentence r whichdisputants agree to be logically possible without this needing to be shown byargument. The argument then consists in showing that r entails s by asequence of agreed mini-entailments. For if r does not entail a contra-diction, neither does any sentence entailed by r. For example, someonemay try to show that “there are two spaces” − a space being a system ofplaces each of which is at some distance in some direction from each otherplace of the system and from no other places – is logically possible, bydescribing in detail a situation under which it would be true.4 That is, theyclaim that the latter description entails that there are two spaces; and thatsince the latter description is logically possible, “there are two spaces” is alsologically possible.

However, the use of these procedures to determine logical possibilitypresupposes that it is clear what are the truth and falsity conditions ofsentences, and which sentences mini-entail other sentences. But thelanguage-learning process, which normally produces very similar under-standings of meanings in members of a language group, sometimes prod-uces somewhat different understandings of these conditions andentailments in different sub-groups. This may occur because differentlearners learn meanings from somewhat different paradigm examples; andwhen this is recognized, language users can acknowledge that the same wordor sentence has more than one meaning. But it may also occur even whenboth sub-groups acknowledge the same paradigm examples of observableconditions and commitments. And then it sometimes happens that one oftwo sub-groups objects that the sense in which some word (or longerexpression) derived from the same paradigm examples (of observable con-ditions and mini-entailments) used by the other sub-group is not a real orlegitimate sense of that word, in that its use in that sense entails contra-dictions. Or one of the sub-groups may object that the sense in which theword (or longer expression) is used by the other sub-group is not the senseimplicit in some of the paradigm examples. It is objections of these twokinds that produce the disputes about the meaning of “personal identity.”

Most of us have been taught the meaning of the expression “is the sameperson as the person who” or its more natural equivalent “is the personwho” by many observable paradigm examples of the same kind (e.g. “this isthe same person as the person you saw last week,” “you are the person whohad a headache only thirty seconds ago”) and many similar paradigm

4 For an argument of this form in favor of two spaces being a logical possibility, see Quinton (1962).


mini-entailments (e.g. “A is the same person as B” and “B is the same personas C”mini-entail “A is the same person as C”). We will all recognize most ofthese observable paradigm examples of “same persons” as examples ofpersons with strongly continuous bodies, memories and character. Sosome philosophers provide an analysis of the meaning of “P2 is the sameperson as P1” in terms of P2 having a body, memory and character stronglycontinuous with those of P1. That is, they advocate the complex theory ofpersonal identity as a conceptual truth, in my sense a logically necessarytruth. But others of us (including myself) think that it is not the normalsense of “P2 is the same person as P1”; and that in the normal sense thisexpression designates a continuing identity of a different kind whichnormally underlies the strong physical and mental continuities but is notconstituted by them and can occur without them. That is, we claim, thesimple theory of personal identity is a conceptual truth, in my sense alogically necessary truth.The only way to resolve this disagreement is by persistent continuing use of

the methods described earlier. Advocates of the complex theory as a con-ceptual truth try to get us to recognize the logical impossibility of a personalidentity independent of strongmental and physical continuities. They do thisby trying to show that some sentence using the expression “is the same personas” in a different sense from theirs entails a contradiction which would notarise if the expression were used in their sense. For example, they may claimthat “Socrates is the same person as the mayor of Queenborough, but hasnone of the same brain, memory or character as the mayor,” together withwhat they may claim to be a necessary truth “no one should be punished forany act which they cannot remember doing” entails “both {the mayor shouldbe punished for any immoral acts of Socrates} and not-{the mayor should bepunished for any immoral acts of Socrates}.”5 If they can get us to recognizethis in one case, then they may get us to recognize that other sentences wherethe expression “is the same person as” is used in a sense other than their sensewill have the same consequences, and so to see that any such sense of theexpression is not a legitimate one. We who claim that the simple theory is aconceptual truth are of course likely to deny, with respect to the example justdiscussed, either that the first conjunct of the purported entailment (“themayor should be punished for any immoral acts of Socrates”) is indeedentailed, or that “no one should be punished for any act which they cannotremember doing” is a necessary truth.

5 This example is of course a formalized version of Locke’s argument for the necessity of same memory(which he calls “same consciousness”) for personal identity (Locke 1975 [1690], ii, xxvii, 19).


We advocates of the second sense are happy to acknowledge that the sensein which our opponents understand “is the same person as” − that is, assomething like “has a body, memory and character strongly continuouswith” − is a perfectly legitimate sense of an expression, but claim that it isnot the normal sense of “is the same person as.” This is because – contrary tothe claims of the advocates of the complex theory – it is not compatible withthe sense implicit in some of the paradigm examples of personal identity bywhich the expression has been introduced into language. Some of theseexamples concern our opponents’ own identity. They must recognize thatthey themselves often have streams of overlapping experiences. For example,the second half of the experience of a pain during a “specious present” mayoverlap with the first half of an experience of some noise; this noise maycontinue for a short while (during several overlapping periods of speciouspresent), and overlap with a certain tactual experience, and so on. It is aparadigm example of personal identity that two overlapping conscious eventsare experiences of the same person, from which it follows that any stream ofsuch events are also experiences of the same person.6 Then they must recog-nize some very recent past experiences which they remember so vividly that itis obvious that they occurred (e.g. “you are the person who had a headachethirty seconds ago”); and it is obvious that – as Reid (2002 [1785], iii.4) put it –“mymemory testifies not only that [a certain past action] was done, but that itwas done byme who now remembers it.”We thus point out to any opponentthat some of the paradigm examples of personal identity that he mustrecognize are ones in which he has a direct awareness of personal identity;and what the awareness is of is not continuity of body, memory and character,but something which can only be described as an awareness of himself as acontinuing subject of experience. Once we have focused on the paradigmexamples of our opponent’s own personal identity over time,which give rise tothe understanding of himself as a continuing subject of experience, we mustget our opponent to recognize that, as with any experiences, what he is awareof (the continuity of his mental life) could occur without his subsequentlyremembering it. And so, more generally, we must get him to see that thiscontinuity could occur without any of his criteria of personal identity being inanyway satisfied.To do this we need to describe some situation inmuch detailby a sentence (normally consisting of a long conjunction, such as a thoughtexperiment in which someone is described as surviving events of a kind

6 It was John Foster who drew attention to the phenomenon of a stream of overlapping experiences asthe foundation of our understanding of personal identity. (See Foster 1979, 176: “it is in the identity ofa stream that we primarily discern the identity of a subject.” See also Foster 1991, 246–50.)


described at the beginning of this chapter) which our opponents are preparedto recognize as logically possible; and then find a route of mini-entailmentsfrom it to a sentence that claims that some later person is the same person as anearlier personwithout our opponents’ criteria being satisfied. Given that this islogically possible, our opponents’ theory which claims that it is not logicallyimpossible must be false.My feeling about how this debate goes these days is that we are getting

our message across. If it is admitted that it is physically possible, and so afortiori logically possible, that a series of operations such as those describedat the beginning of the chapter could occur and that it is logically possiblethat the person before the operations could subsequently have the experi-ences of the person after the operations, it then follows that being the sameperson as a previous person does not entail having the same brain or stronglycontinuous memory and character. Our opponents may, however, insist onsome residual physical continuity, e.g. that the replacement of brain matterdoes not occur all at once. But someone softened up by physically possiblestories of the kind described at the beginning of the chapter may then beginto acknowledge the logical possibility of a person acquiring a new body all atonce without gradual replacement of parts; and so come to acknowledgethat it is logically possible that a person could be the same person as a personat a later time without there being any continuity of body (including brain),memory or character between them. So, our opponent should recognize thesecond sense as the normal sense of personal identity.The same arguments that will show that there is no contradiction in an

unnamed person continuing to exist and have experiences under thesecircumstances, are unaffected by whom one supposes the person to be. Sowe may conclude that it is logically possible for me or any other human tosurvive total replacement of body, memory and character. Logical possibi-lity is the kind of metaphysical possibility which can shown to be sucha priori. But to determine whether it is metaphysically possible for me orany other human to continue to exist without any continuity of body,memory or character, we need to show that this is also a posteriori meta-physically possible. That, someone may claim, will depend on what sort ofpersons we humans are − that is, what is the essence of a human person −and that that is not something to be determined a priori.

a po s t e r i o r i me t a phy s i c a l po s s i b i l i t y

I will begin my discussion of a posteriori modal claims in terms of what it isfor a sentence to be a posteriori metaphysically necessary; the application to


a posteriori metaphysical impossibility and possibility will then becomeapparent. A posteriori metaphysical necessity is supposed to be a necessityas hard as logical necessity, yet discernible only a posteriori. No different typeof necessity could be as hard as logical necessity, and so a posteriori necessitymust be in some way reducible to logical necessity. And, as far as I can seefrom all the plausible examples of a posteriori metaphysically necessarysentences that have been adduced, the way “a posteriori” comes into it isthat we need tomake empirical enquiry to determinemore adequately what isthe substance or property or whatever about which the claim being made bythe sentence is being made. When we have determined that, if the claim ismetaphysically necessary, the necessity of the claim will be detectable a priori.

Sentences pick out the substances, properties or whatever with whichthey are concerned either by “rigid designators” (as defined by Kripke 1981,p. 48) − that is, expressions which (given that their meaning remains thesame) always refer to the same substance, property or whatever, howeverdifferent the world might be from how it is − or by non-rigid designatorswhich may pick out different substances or whatever if the world is differ-ent. “Green,” for example, is a rigid designator because it always refers to thecolor green, whereas “Amanda’s favorite color” is a non-rigid designatorbecause it would refer to a different color if Amanda had different colorpreferences from her present ones. I will call a rigid designator φ an“informative designator” if we can (when favorably positioned, faculties inworking order, and not subject to illusion) recognize when something is φand when it is not, merely in virtue of knowing what the word φ means. Iwould not understand the word “green” unless (when the stated conditionsare satisfied) I could recognize when an object is green and when it is not.When our referring expressions are informative designators (or can bedefined in terms of informative designators), we know the necessary andsufficient conditions for the things referred to to be what they are; and so, Ishall say, we know the essence of what is being designated. When all thedesignators in a sentence are informative (or can be defined in terms ofinformative designators), it is a pure a priori exercise to determine whetherthe sentence is logically necessary or whatever. When we know what we aretalking about, mere thought can showwhat that involves. Mere thought, forexample, I suggest, can show that “all trilateral figures are triangular” (a“trilateral” figure is a closed surface bounded by three straight lines; atriangle is a closed surface bounded by straight lines and having threeinterior angles) or “no surface can be both green and red all over” arelogically necessary; and we can now see that this is because “surface,”“straight,” “green,” “red,” etc. are all informative designators, and so


understanding the sentences involves knowing the essences of what is beingreferred to and so comprehending fully what the sentences are claiming.But there are rigid designators which refer to substances or whatever, such

that a speaker can understand to what they are referring on some occasionswhen the thing exhibits certain non-essential features, without knowing theessence of what is referred to. Clearly one will not understand what some rigiddesignator means (what its role is in the language) unless one knows how touse it on some occasions, and understands the kind of thing to which it is usedto refer and so the kind of criteria by which to distinguish one thing of thatkind from other things of that kind; but language users may not be able inpractice to use these criteria to determine to which thing of that kind it isreferring. But in that case one would not be able to recognize that thing onoccasions other than the ones on which it exhibits those non-essential features(its “stereotype”). I shall call such designators for which the criteria of“informative designators” are not satisfied (and cannot be defined in termsof such designators), “uninformative designators.”7 A sentence is thena posteriori metaphysically necessary if it would be logically necessary whenwe substitute informative designators (or expressions definable in terms ofthese) for its uninformative designators.Thus – to use the example discussed by Putnam (1975) – the word

“water” as used in the eighteenth century was an “uninformative designa-tor.” This is because although people used “water” as a designator of a stuff,and so knew that to be the same stuff something would have to have thesame chemical essence, they picked out a volume of stuff as water in virtueof its superficial contingent properties (being liquid, in our rivers and seas,etc.), yet – in ignorance of what that chemical essence was − they would notbe able to recognize it on occasions when it did not have those superficialcontingent properties. So they were unable to say whether or not sometimesstuff found elsewhere than in our rivers and seas was water or not. Whenpeople discovered that chemical essence ( H2O), they could then recognizewhether stuff not in our rivers and seas was water or not. Hence, since theclaim being made about water is a claim about H2O, we can substitute“H2O” for “water” in “water is H2O” (as used in the eighteenth century)and the sentence then reduces to a logically necessary truth. (I assume herethat “H2O” can be defined in terms of informative designators: that is, in

7 Similar distinctions to my distinction between “informative” and “uninformative” designators arethose made by Chalmers (1996) between expressions with “primary intensions” and ones with“secondary intensions,” and by Bealer (1996) between “semantically stable” and “semantically unsta-ble” expressions.


terms of such expressions as “mass,” “volume,” “smaller by 10–1 than” and soon.) Somewhat similar is the sentence used by Kripke (1981, p. 100) toillustrate a posteriori metaphysical necessity, “Everest is Gaurisanker,”where – Kripke supposes – “Everest” was used by early explorers todesignate a mountain having a certain shape when seen in the distancefrom Tibet and “Gaurisanker” was used to designate a mountain having acertain shape when seen in the distance from Nepal. The explorers under-stood the sentence “Everest is Gaurisanker” because they understood whatit would be like for the two mountains to be the same – it would consist intheir being made of the same chunk of rock. The rock of which eachmountain was made constituted its essence. But they did not know whetherthe two mountains were the same (whether they had the same essence), andit required empirical investigation to discover that they did. Once theyknew what each mountain essentially was, they knew that the claim beingmade by the sentence was necessarily true with a necessity as hard as that of“all squares have four sides.” Thus – in my terminology – they could knowthat there is an informative designator of the form “mountain made of suchand such rock” which can be substituted for both “Everest” and“Gaurisanker,” so that the sentence has the form of an identity sentence“a is a” and so is logically necessary.

The application of my account of a posteriori metaphysical necessity toa posteriori metaphysical impossibility and possibility should now beevident. The metaphysical modal status of a sentence is its logical statuswhen informative designators are substituted for uninformative ones. Evenif we cannot find out what is the essence of some substance or whatever, ourunderstanding of how to use the designator may give us enough knowledgeof the kind of essence involved to enable us to see or deduce the modalstatus of some sentence using it. For example, even if we do not know theessence of water, we can see that “water is the number 42” is impossible. Butonly if all designators in a sentence are informative (or can be analyzedin terms of informative designators) is it guaranteed that mere a priorireasoning can determine its logical status.

Many of the words by which we pick out properties are informativedesignators (“green,” “square,” “has a length of one meter,” etc.). And manywords by which we pick out properties, which are not themselves informativedesignators, can be analyzed in terms of them – e.g. “has a length of 10–18

meters” can be defined in terms of the informatively designated property “hasa length of one meter” and eighteen applications of the informatively des-ignated relation of “being shorter by one-tenth than.”However it seems thatthere are at present some substances which we can only pick out by


uninformative designators. For example, it is an unresolved issue (French2006) whether some fundamental particles, such as quarks and electrons, arethe particles they are merely in virtue of their properties (such as mass andcharge, and causal relations to other particles), or whether they are what theyare partly in virtue of the particular matter of which they are made. For thisreason we do not know what are the necessary and sufficient conditions forsome fundamental particles to be the ones they are (that is, we do not knowtheir essence) and have to pick them out by uninformative designators.Now what sort of designator is “I,” or “Richard Swinburne,” as used by

me? These seem to be informative designators. If I know how to use thesewords, I cannot be mistaken about whether or not they apply to a certainperson – given that I am favorably positioned (e.g. his body is my body),with faculties in working order, and not subject to illusion. And when I amconsidering applying these words to a person in virtue of his having someconscious event, these conditions will be maximally satisfied and no mistakeis possible. I am, in Shoemaker’s (1994, p. 82) phrase, “immune to errorthough misidentification.” I cannot know how to use the word “I,” recog-nize that someone is having some conscious event (e.g. a pain) and stillwonder whether it is I who am having that event or someone else, in the waythat an early explorer could know how to use the word “Everest,” and yetwonder whether the mountain at which he is looking fromNepal is Everest.My knowledge of how to use “I,” like my knowledge of how to use “green”and “square,”means that, in the sense analyzed above, I know the essence ofwhat I am talking about when I use the words. Hence if “I will existtomorrow with a new brain” or “I will exist without any memory of myprevious existence” are logically possible, they are also metaphysicallypossible. I claim therefore that the considerations which should lead peopleto conclude that such sentences express logical possibilities should also leadthem to conclude that they express metaphysical possibilities. And since Ican know this merely in virtue of knowing to what my use of the word “I”refers, other people can know the same about themselves. Each of us, wemay properly conclude, can continue to exist without any continuity ofbrain, memory or character.Of course I can still misremember what I did in the past, and indeed

misremember how I used the word “I” in the past. But this kind of problemarises with every claim whatsoever about the past. “Green” is an informativedesignator of a property, but I may still misremember which things weregreen and even what I meant by “green” in the past. The difference betweeninformative and uninformative designators is that (when my faculties are inworking order, I am favorably positioned and not subject to illusion) I can


recognize which objects are correctly picked out at the present time byinformative designators, but not necessarily when they are picked out byuninformative designators (in the absence of further information). And so Iknow what a claim made now about the past or future amounts to when it ismade by informative designators (but not when it is made by uninformativedesignators) whether or not I have any reason for supposing it to be true. Forto claim that some informatively designated object a will exist or have aninformatively designated property φ tomorrow is just to claim that somethingwhich I can understand (a existing or being φ today) will hold tomorrow. Itfollows that I can understand what it is for me to exist tomorrow or yesterdayand to have such and such experiences. Not so with Everest or water, whenthese things can be picked out only by uninformative designators. I do notknow what would constitute a past or future substance being water or Everestif I ammerely in the position of the “water” user in the eighteenth century, orthe early explorers using “Everest” in the way described.

I conclude that given that each of us can come to see that it is logicallypossible that they can survive without any continuity of brain, body orcharacter, in the crucial sense inwhich they subsequently have the experiencesof the surviving persons, and so come to see that this is metaphysicallypossible and so come to see that the simple theory of personal identity is true.

the human sou l

I stated the simple view as the view that personal identity is not constitutedby continuities of mental or physical properties or of the physical stuff (thatis, matter) of which persons are made, but is a separate feature of the worldfrom any of the former. But this leaves open the possibility that personalidentity might be constituted by a non-physical part, a “soul.” Substancedualism holds that each human on earth consists of two parts, a body and asoul – the soul being the essential part, and the body a contingent part. Onthis theory, while any physical stuff of which the body is composed, and anyparticular physical and mental properties, are not necessary for the contin-ued existence of a person, the continued existence of his soul is necessary. Isuggest that this further step is forced upon us if we admit the logicalpossibility of a certain thought experiment, and the high plausibility of acertain principle about what it is for any substance to continue to exist.

The thought experiment is this. Suppose that there exists instead of ouractual world W1, a world W2 which is exactly the same as W1 except in thatinstead of a certain person S1who lives a certain life inW1 there is a person S2who has the same body and the same mental and physical properties


throughout his life as S1, but is not S1. And surely our world could bedifferent in the sole respect that the person who lived my life was not me.For it is not entailed by the full description of the world in its physicalaspects and in respect of which mental properties are associated with whichbodies that the person who has any particular body and mental and physicalproperties should be me. We can see this if we imagine that before thisworld exists we are shown a film of what is going to happen in it; and thatthe film has some device for showing us what will be the mental lives of thepeople in the world. Each of us would still not know whether we were goingto live one of the lives in this world. And if so, which one. So because W2

can be seen, I suggest, when we reflect on it, to be logically possible; andbecause – as before – persons can be picked out by themselves by theinformative designator “I,” W2 is a metaphysically possible world.But a substance at a particular time is the substance it is in virtue of its

parts, what they are made of and their properties, and the relations the partshave to each other. For example, if there is a substance composed of certainfundamental particles with certain properties (including certain relations topast particles) related to each other by certain causal and spatio-temporalrelations, there could not be instead of it a different substance composed ofall the same particles with the same properties and relations to each other.Andre Gallois (1998, p. 251) has called the view that there could be anothersuch substance “strong haecceitism.” He writes:

Strong haecceitism seems to me incredible. Consider a car on a parking lot. It is notat all incredible to suppose that a qualitative duplicative of the car in question mighthave existed even if there is no qualitative difference at any place or time as a result.It is incredible to suppose that throughout history all of the atoms that actually existmight have been configurated at each time in exactly the way they are actuallyconfigurated without the car on the parking lot existing.

I suggest that it follows from our very understanding of what a substance isthat what Gallois describes as “incredible” is false; and in particular itfollows from our understanding of what a person is that two personscould not be different if all their parts and all their properties were thesame. (This does not commit me to the identity of indiscernibles, whichholds that two substances are the same if they have all the same properties(in the sense of universals). It allows that two substances may be differenteven if they have all the same properties, so long as their parts are different.)It follows that in the thought experiment described above, S1 and S2must

have different parts; and since all their physical parts are the same, thedifference must arise from each having different non-physical parts, that is


different souls. My earlier thought experiments suggested that a person cancontinue to exist with a different body and different mental properties. Thethought experiment just described in this section suggests that a personneeds a particular soul in order to continue to exist. It is, however, compat-ible with substance dualism to hold that to exist a person needs a body, butnot any particular body – though I believe that further thought experimentsshow that a person can exist without a body and so without any physicalproperties at all. So the only essential properties necessary for a person toexist are the essential properties of any soul, which – I suggest – are simplythe one property of having (in some sense) a capacity to be conscious.8

8 In my book The Evolution of the Soul (Swinburne 1997, pp. 153–4, and pp. 327–32), I followedSt. Bonaventure in analyzing a soul as a form (a collection of essential properties) instantiated in somemental stuff, soul-stuff. But it now seems to me that any stuff must be capable of being divided intosmaller chunks of the same stuff; and given my view that humans (and so their souls) cannot bedivided, the soul cannot be made of any stuff. It is an “immaterial particular.”


chapter 6

Against simplicitySydney Shoemaker

i

Proponents of simple views of personal identity often contrast personalidentity with the identity of other sorts of things. Sometimes, as in the caseof Bishop Butler and Thomas Reid, this takes the form of a view which takesidentity of persons to be identity in the “strict and philosophical sense” andconstrues the identity we ascribe to other sorts of things, especially artifacts,as identity in only a “loose and popular sense,” something which, as Reidput it, we call identity “for conveniency of speech.” Whether or not theidentity ascribed to other sorts of things is held to be “fictitious,” not reallyidentity, what is thought to distinguish it from personal identity is its being“analyzable,” its consisting in facts closely related to the facts that serve asour evidence for ascribing it. There are held to be “constitutive criteria” forthe identity of ships and trees, and it is denied that there are such criteria forthe identity of persons. The tendency to regard personal identity as specialin this way often – but not always – goes with acceptance of some form ofmind–body dualism.But if the rejection of constitutive criteria for the identity of persons is the

mark of the simple view about personal identity, then on one understandingof constitutive criteria there are plenty of philosophers who hold a simpleview about the identity of ships, trees, rivers, etc. For there are many thatreject “reductionism” about the identity of such things. There are, ofcourse, different understandings of what it is for an account of the identityof so-and-sos to be “reductive,” and what it is for the identity of so-and-sos tobe “analyzable.” But it is worth asking whether there are any grounds fordenying that the identity of persons is not analyzable that are not equallygrounds for denying that the identity of other sorts of things is not analyzable.I will discuss this later. But first I want to present a case in favor of a complexview about the identity of things other than persons – I will focus on theidentity of trees – and then argue that the case applies equally to persons.

123

i i

When a thing persists over an interval of time there is a series of events andproperty instances that constitutes the career of the thing during that interval.I will speak of this simply as a series of property instances, since the events inthe series can be thought of as causally connected series of property instances.But the causal connections in the series are of course not limited to thoseholding among property instances that make up individual events; it willbe true generally that what property instances occur later in the series willbe causally due, in part, to what property instances occurred in it earlier.A special case of this will be the retention of a property over time;myweddingring’s having a certain shape at a time is due in large part to its having had thatshape earlier. But there are also cases in which there is systematic change inwhich later states of a thing grow out of earlier states of it and are causally andcounterfactually dependent on them – as when a tree gradually changes itsshape and size in playing out its biological nature.

What a career is really a series of are not individual property instances butsets of simultaneous property instances. These sets will consist of instancesbelonging to a single thing – instances standing to one another in a syn-chronic unity relation. These can be called “thing-stages.” If the career is thecareer of a tree, each property instance will be a state of affairs consisting ofa tree’s having a certain property at a time, and each tree-stage will be a stateof affairs consisting of a single tree having a certain multiplicity of propertiesat a time. The question of whether a series of tree-stages is the career of a treeis the question of whether it is one and the same tree that is involved inthese different states of affairs. There is of course a question as to whatthe synchronic unity of a thing-stage consists in – e.g. what it is for twosimultaneous tree-states to be states of one and the same tree. But my concernhere is with diachronic unity.

Perhaps some proponents of “perdurance” views of persistence over timewould identify persisting things with their careers. But most of us wouldnot. And one does not have to hold that view in order to hold that theexistence of a career entails the existence of a thing whose career it is. Thatis a claim on which proponents of the simple view and proponents of thecomplex view can agree. Where they cannot agree is on what it takes for aseries of thing-stages to be a career.

On a complex view, what makes a series of thing-stages a career is therelation between the properties instantiated in it at different times, includingthe nature of the causal relations between different property instantiations inthe series. These make the series a career; and in so doing they constitute the

124 sydney shoemaker

persistence of the object over the interval occupied by the career. This hasthe consequence that if a series of thing-stages duplicates the series of propertyinstances that makes up the career of some thing of kind K, in the sensethat in the duplicate series the same properties are instantiated in the sametemporal order and the property instances are causally related in exactly theway their counterparts are in the career, then the duplicate series is itself acareer of something of kind K, and there is a thing of kind K to which theproperty instances in the series belong.1

Notice that such an account of the persistence of things through timemakes free use of the concepts of the kinds of things said to persist. For itsthing-stages are states of affairs involving such things – e.g. a tree-stage is astate of affairs consisting in a tree having certain properties at a time. One cansay if one likes that tree-stages and tree-property instances “get their identity”from the trees they involve. But it would be a confusion to say that this makesthe account of the identity of trees across time circular. The notion of a treeenters into the characterization of the series of tree-stages, but the notion ofsameness of trees over time does not – nowhere in the characterization is itsaid that it is the same tree that is involved in the different tree-stages.A similar charge of circularity will be addressed later on.A proponent of a simple view about the trans-identity of things of a certain

kind can certainly agree that in general the occurrence of a series of thing-stages that duplicates in this way a series that makes up a career of a thing ofthat kind will itself make up the career of a thing of that kind. But can he holdthat the existence of such a series constitutes the career of a thing of that kind?If he does, he will be holding that the existence of such a series constitutesthe persistence over time of a thing of that kind. But that is what he issupposed to be denying in holding a simple view about the persistence ofthings of that kind. It would seem that hemust hold that it can at best be onlya contingent truth, not a necessary one, that duplicating the series that is thecareer of a persisting thing of a kind gives us the career of a thing of that kind.Of course, if we are allowed to include in the specification of the series

that makes up a career that the different thing-stages in the series stand toone another in an unanalyzable relation of diachronic unity, a relation that

1 This is true only with a qualification that does not affect the point I am making. To use a well-knownexample ofDonaldDavidson’s, we can suppose that a highly unlikely consequence of lightning striking aswamp might be the coming into being of a creature, a “swamp man,” that is a physical duplicate of anactual man. We could also have swamp tigers and swamp trees. On some views of species individuationthese swamp creatures would not belong to the biological kinds their real world physical duplicatesbelong to – although arguably they would persist through time in the way their duplicates do. We canallow for this by modifying the claim in the text so that it concerns cases in which the duplicate seriesoccurs in a setting, including a pre-history, like that of the series it duplicates.

Against simplicity 125

necessarily holds just in case the relata belong to the very same thing, thenthe advocate of the simple view can happily allow that duplicating the seriesmaking up a career will itself make up a career. On the simple view, thisrelation of diachronic unity will have to be something over and above thecausal relations that hold among the property instances. And apparently itwill be a relation that could be there whether or not the causal relationshold. The proponent of the simple view is apparently committed to holdingthat a series that duplicates the career of a persisting thing with respect towhat property instances it contains and how these are ordered and causallyrelated could fail to be the career of a persisting thing of that sort, namelythrough the members failing to stand in the unanalyzable diachronic unityrelation to one another. And he also seems committed to holding that aseries of property instances whose members are causally unrelated couldnevertheless be the career of a persisting object, because its members standto one another in the unanalyzable diachronic unity relation despite theirlack of causal connectedness.

These are, to put itmildly, highly counterintuitive consequences. Accordingto the first, what we can establish empirically about a series of thing-stages isinsufficient to establish that it is the career of a persisting object. I observeover a period of time what I take to be a tree growing in my yard, and whataccounts for what I observe is the impact on my sensory system of a series ofinstantiations of tree properties whose members stand in precisely the sorts ofcausal relations that obtain among the property instances in the career of agrowing tree.2 But on the simple view, apparently, it is compatible with whatI seem to observe in such a case that the series of property instances is notthe career of a tree. What more could I observe, or otherwise establish, thatwould show me that there really was a single tree persisting throughout thatinterval? It is hard to see how on this view we could have any justified beliefsabout the persistence of trees through time.

According to the second consequence of the simple view, there could be aseries of tree-property instantiations that constitute the career of a tree eventhough there are no causal relations among its members. So there are norelations of counterfactual dependence among its members – it is not true ofour tree that if a branch had been lopped off yesterday it would have hadfewer branches today. Of course, its having the branches it has today should

2 I am not saying that what I observe is a series of tree-stages. What my perceptual states represent is atree persisting over time, not a succession of tree-stages. But my having perceptual states that representthis is due to my sensory system registering the impact on it of there occurring, withinmy field of view,a succession of tree-stages, i.e. a succession of tree-property instances.


have a cause. But it would seem that on the simple view it is not requiredthat this cause should lie at least in part in its past career. Moreover, if thediachronic unity relation among property instances cannot consist partly incausal relations, it is hard to see how it could consist partly in spatio-temporalrelations. And then it is not clear why a series that we would take to be thesum of parts of the careers of different trees – a part of a different tree careerfor each day of the week – could not be the career of a single tree.It may be suggested that these embarrassing consequences vanish when

we see that the properties instantiated at different times in a career includesortal properties, like the property of being an oak tree. The instantiation ofsuch a property at a time and place, it will be suggested, imposes constraintson what can be true at temporally and spatially adjacent times and places.A tree cannot exist for just an instant; if it exists at a time then it must existthroughout an interval that includes that time. The existence of a tree at aplace at a time requires its existence at nearby places at nearby times. Anddifferent trees cannot be at the same place at the same time. So if there isa temporally and spatially continuous series of instantiations of the propertyof being a tree, there must be a single tree that is the subject of all of theseinstantiations. Temporally proximate instantiations of the property mustbelong to the same tree, since otherwise we would have different treesoccupying the same place at the same time. And by the transitivity of identity,this implies that it is the same tree throughout. This will be true, the suggestionis, whether or not the successive property instances are causally related in theway required by the complex view. And we can know it to be true withoutknowing that they are so causally related.This may seem prima facie plausible. But it is not clear how the simple

theorist can be in a position to say it. Crucial to the argument is the claim that,as I have put it vaguely, the existence of a tree at a place and time requiresits existence at nearby places at nearby times. This implies that the persistenceof trees requires spatio-temporal continuity. (I think that it is arguable thatthe requirement of spatio-temporal continuity stems from the fact thatpersistence requires causal relations and that causal chains must be spatio-temporally continuous.3 But I need not rely on this here.) Considerationsabout causality aside, it is unclear how a proponent of a simple view about theidentity of a certain kind of objects can hold that the careers of such objectsmust be spatio-temporally continuous. For one cannot hold this withoutholding that the persistence of such things partly consists in spatio-temporal

3 If it is allowed that there can be causal paths that are not spatio-temporally continuous, it is a plausibleconsequence that there can be careers that are not spatio-temporally continuous.


continuity – and that seems to be something that a proponent of the simpleview cannot hold.

We should also notice that while it is true that the property instances thatmake up the career of a persisting thing must include instances of sortalproperties, the instantiation of sortal properties supervenes on the instan-tiation of other properties. It is because of other properties a thing has thatit counts as being a tree. The property of being a tree will be realized inphysical properties, and, ultimately, in microphysical states of affairs thatrealize instances of these physical properties. And it will be the causalprofiles of these more basic physical properties and microphysical states ofaffairs that make them realizers of the property of being a tree. While theinstantiation of the property of being a tree cannot be instantaneous, theinstantiations of the physical properties in virtue of which something is atree can be. It will be series of instantaneous physical property instancesthat realize instances of the sortal property of being a tree. So at one level ofdescription, the series of property instances that constitutes the career of thetree will be the series of the physical realizers of the property of being a tree.4

They are realizers of that property because their causal profiles include itscausal profile. The reason why the instantiation of the property of being atree at a time constrains what is true at other times is that this property ispartly individuated by causal features that contribute to determining what istrue at subsequent times. When this property is instantiated one or anotherof its realizer properties will be instantiated, and its causal features will affectwhat is true at subsequent times. The effects will include the instantiation ofproperties that are themselves realizers of the property of being a tree. So thephysical properties of a tree at a time, properties that realize the property ofbeing a tree, will be such as to generate, or contribute to the generation of,slightly different physical properties at a later time, and these will also realizethe property of being a tree. And because of the way the latter propertiesgrow out of the former, the two sets of properties will belong to the sametree. The point that the existence of trees cannot be instantaneous rests onthe fact that trees necessarily have careers that exhibit these kinds of causalconnections. So it can hardly be used against the claim that the persistenceof trees consists in the existence of such careers.

4 If realizers must be sufficient for what they realize, then an instantaneous property instance cannot be arealizer of a non-instantaneous property instance. What can be a realizer of the latter is a series ofinstantaneous property instances. In such a case let us speak of the instantaneous property instances inthe series as quasi-realizers of the non-instantaneous one. So what I should say here is that whatconstitutes the career of the tree is the series of quasi-realizers of the property of being a tree.


Further, if we focus on the microphysical states of affairs that are theultimate realizers of tree property instances, we see that those that realize theproperty instances in the career of a tree form a series that can be describedwithout any use of the concept of a tree. There is a good sense in which thisseries constitutes the career of the tree, and so constitutes the persistence of thetree over the interval in question. If this series is described in physical terms, itwill not be a priori that a series of this description constitutes the career of atree. But that it does so will be an a posteriori necessary truth – one that holdspartly in virtue of the empirical fact that the causal features of themicrophysicalstates of affairs are such as to make them realizers of the tree-property instancesin the career, and partly because of the fact – which I think will be a priori –that these, and the relations among them, are collectively such as to constitutethe career of a tree. This gives us a set of facts whose specification does notrequire the use of the concept of a tree but which constitute a fact about treeidentity and tree persistence. That seems enough to show that tree identityconsists in something, contrary to what a simple view about it would claim.I have claimed that on the simple view there could be a series of property

instances that is the career of a persisting object despite there being no causalrelations, and no relations of counterfactual dependence, between itssuccessive members. It might be replied that all that the simple theorist iscommitted to is that there being such relations cannot be, or be part of, asufficient condition for a series of events being the career of a persistingthing, and that this does not prevent him from holding that it is a necessarycondition for this. The same might be said of the requirement that thecareer of a persisting thing must be spatio-temporally continuous. But it isnot clear how these could be necessary conditions without being parts of asufficient condition. If we deny that a series of property instances is thecareer of a persisting object of a certain sort on the grounds that it is notspatio-temporally continuous, or that its members are causally unrelated,we are saying that it does not have what it takes to be a career of an object ofthat sort. And that surely implies that there is something that constitutesthe persistence of objects of that sort, something that is lacking if there isno spatio-temporal continuity or no causal connectedness. The diachronicunity relation cannot be simple and unanalyzable if its obtaining requiresspatio-temporal continuity or causal connectedness.

i i i

Most simple theorists about identity over time are not simple theoristsabout the identity of trees – or of rivers, automobiles or statues. So they


could accept what I have said against the simple view about tree identity.Their claim concerns the identity of persons.

But it would seem that what I have said about tree identity should carryover to the identity of persons. Consider the career of a person, and considera series of person-states (person-property instances) that duplicates thiscareer with respect to what properties are instantiated and in what order,and with respect to how the instantiations are related causally. The denialthat this duplicate series would have to be itself the career of a person is, tosay the least, extremely implausible. It seems to lead to an extreme sort ofskepticism about third-person identity judgments and, indeed, about judg-ments about other minds. It is hard to see how, on this view, the phenom-ena we take to be evidence of the persistence of a person over an interval oftime could really be so. And it is also hard to see how, on this view, thephenomena we take to establish the existence of mental states in otherpersons could really do so – for the behavioral evidence for the existence ofmental states consists in phenomena occurring over a period of time anddepends on successive stages of the behavior being manifested by one andthe same person. I think, therefore, that the simple view is just as implau-sible in the case of persons as it is in the case of trees, rivers and the like.

What partly underlies the feeling that there is a difference here is thememory access persons have to their own pasts. First-person memory judg-ments are not themselves identity judgments, but in making them onecommits oneself to the truth of trans-temporal identity judgments aboutoneself. If I say that I remember having eggs for breakfast this morning,I commit myself to being the same as the person I remember eating eggs thismorning. And that identity judgment appears to be one I know to be truewithout the use of any criterion of identity. My knowledge of it seems to bedirect and not grounded on evidence of identity of any sort. We do not havethis sort of knowledge of the truth of trans-temporal identity judgments aboutother things (including persons other than ourselves). And this can make itseem that where we do base identity judgments on evidence, the status of thisevidence is different when the judgment concerns the identity of a person thanwhen it concerns the identity of a ship or a tree or a river. In the former casethe evidence can have the status of being a symptom of identity – somethingthat is known to be evidence of an identity fact because it has been foundempirically, using one’s memory access to one’s past, to be correlated withfacts of that sort. The symptoms would include things like similarity ofappearance, similarity of fingerprints and similarity of DNA – things thatnearly everyone would regard as symptoms rather than criteria. But theywould also include, on this view, spatio-temporal continuity, continuity


with respect to properties of one sort or another (psychological properties,biological properties, etc.), and causal connections between successive states.The discovery of the correlation of course would involve the access each of ushas in memory to facts about his or her own identity, an access that gives usknowledge of identity that is direct and not grounded on evidence. The use ofthis access to ground symptom-based judgments about the identity of personsother than ourselves would of course have to involve something like theargument from analogy – a projection onto others of correlations we havediscovered to hold in our own case. Where our identity judgment concerns aship or a tree or a river, we must have evidence concerning such things thatdoes not consist of symptoms in this sense. Sometimes, to be sure, ourjudgments about such things would be based on symptoms, but the discoveryof the correlations that tell us that these are symptoms would have to involveevidence that does not consist of symptoms – it cannot be symptoms all theway down. And it may seem that this non-symptomatic evidence mustconstitute criteria of identity – and that these criteria must be constitutive,and not merely epistemic. Thus reflection on the special access we have to ourown pasts in memory, and on the absence of a similar access to the pasts ofother sorts of things, can lead to the view that personal identity is alone inbeing simple – where its being simple is equated with there being noconstitutive criteria for it.But the fact that we have this special memory access to facts about our

identity does not at all imply that there are not constitutive criteria of personalidentity. We do not use any criteria of identity in making memory-basedfirst-person identity judgments, so of course we do not use as a criterion thefact that the memory involved stands in certain causal relations to the pastaction or experience that it represents. But this does not mean that this factcannot be a constitutive criterion of personal identity that can be used as anepistemic criterion by persons other than the subject of the memory. Theseries of property instances that make up the career of a person will includeepisodes of making memory judgments, where these are made spontaneouslyand on the basis of no evidence, and in most cases these will correspond incontent to, and will be appropriately causally connected to, events earlierin the career. On a psychological view of personal identity this can be one ofthe facts about such a series that makes it the career of a person; and it iscompatible with this that first-person memory knowledge of personal iden-tity is not grounded on this – or any other – criterion of personal identity. Itis also compatible with non-psychological accounts of personal identity,e.g. animalism, that no criterion, psychological or otherwise, is used inmaking memory judgments about one’s own past. It can be held that it is


just a contingent fact about human beings that they have in memory a directaccess to facts about their own pasts. More generally, it is compatible with thecomplex view, the view that there are constitutive criteria of personal identity,that persons are capable of having direct knowledge of their own pasts that isnot grounded on such criteria.

But it is often urged against complex accounts of personal identity thatthey are inescapably circular. This is especially urged against psychologicalaccounts. It is an old objection to memory accounts of personal identity thatthe notion of memory implicitly involves the notion of personal identity, itbeing a necessary condition of a state’s being a memory of a past event thatits subject be identical with a past subject or witness of that event. In previouswork I have countered this objection by arguing that the work done bythe notion of memory in psychological accounts can be done by the notion of“quasi-memory,” which is like the notion of memory but does not includethe requirement that there be an identity between the person who quasi-remembers an event and a past subject or witness of that event – also by sayingthat much more than memory continuity goes into the psychological con-tinuity that constitutes personal identity (Shoemaker 1970). Since my presentpurpose is not to argue for a psychological account but to argue that somecomplex view must be true, I will not pursue this matter further here.

But E. J. Lowe has raised an objection to the complex view of personalidentity that, while explicitly directed against the psychological version ofthe complex view, would if cogent go against other versions as well. He saysthat any criterion of personal identity on Lockean lines “cannot be deemedinformative in the sense in which any adequate criterion of identity is requiredto be so. For conscious states – and, indeed, mental or psychological statesquite generally – cannot themselves be individuated or identified save interms which presuppose the identity of persons (or conscious subjects) whosestates they are” (Lowe 2009, p. 134).

Now there certainly is a good sense in which the individuation of con-scious states involves the identity of persons. A conscious state of a person isnecessarily a state of that person, so any situation in which it exists will be onein which that same person is its subject. But it is not only conscious states, orpsychological states, whose individuation involves the identity of persons inthis sense. Any state, psychological or not, necessarily belongs to the thing ofwhich it is a subject, and so gets its identity from the identity of its subject.And it is not only states of persons of which this is true. So if this sort ofinvolvement of the identity of persons makes a psychological account ofpersonal identity circular, any account of the identity of any sort of persistingthing in terms of relations between successive states would be circular. If this


were right, Lowe’s argument against the complex view of personal identitywould generalize into an argument against the complex view of any sort ofpersisting thing – which would be contrary to his intent, since he thinks thatin the case of entities other than persons there can be non-circular constitutivecriteria of identity.But as I noted earlier, in talking about the identity of trees, it is a confusion

to think that this sort of involvement of identity in the individuation of states,property instances, thing-stages, etc. makes circular accounts that hold thetrans-temporal identity of things to consist in relations between such items.For what is involved in such individuation is neither the diachronic identitynor the synchronic identity that is at issue in discussions of personal identity(or in other discussions of the identity conditions of the possessors of propertyinstances and states). In such discussions what will be at issue is what it is fordifferent states, either occurring at the same time or occurring at differenttimes, to belong to one and the same person. Each of the states will necessarilybelong to a person, and what is at issue is what it is for it to be the same personall of them belong to. That the identity of each of the states is fixed by theidentity of the person who has it is in noway incompatible with the claim thatit is relations between the states (no doubt involving relations to yet otherstates) that make it the case that they belong to one and the same person.It is not altogether clear why it is natural to use the term “identity” to make

the point that states, property instances, etc. necessarily have the subjects theyhave. Perhaps the identity here is identity across possible situations – across“possible worlds.” If a state in a given situation is the same as one in another,what has it in the one situation is the same as what has it in the other. This ofcourse has no bearing on questions of diachronic or synchronic identity.Lowe thinks, mistakenly, that accounts that frame criteria of personal

identity in terms of person-stages assume a four-dimensionalist view of persist-ing things, a view he repudiates.5 But there is no need to equate person-stages,as he does, with “time-slices” of persons – one can think of a person-stage asjust the total state of a person at a time. A state of a thing is not a part of it,and the fact that a series of person-stages has temporal parts does not implythat persons have temporal parts. Putting aside his rejection of four-dimensionalism, Lowe says that he can

see no prospect of our being able to introduce the notion of a “person-stage” in away that would make the individuation and identification of person-stages possiblewithout reference to the persons whose stages they would supposedly be. In

5 Lowe (2009, p. 137, fn 58) identifies me as “one leading advocate of the four-dimensionalist approach.”In fact, I have always been a staunch opponent of this approach.


consequence, I can see no prospect of our being able to formulate an informative,because non-circular, criterion of personal identity framed in terms of person-stages. (Lowe 2009, p. 137)

This is the same mistake as before. Person-stages are indeed individuated byreference to the persons of which they are stages. But that leaves it openwhether different person-stages are stages of one and the same person, andno circularity is involved in holding that whether they are is determined byrelations, including causal ones, between them. It is such an account thatI have offered.

i v

But there is a version of the circularity objection that has more substance thanthe ones I have considered. And it is an objection that applies as much tocomplex accounts of the trans-temporal identity of other sorts of things as itdoes to complex accounts of personal identity. This is an objection I raised ina paper written over thirty years ago (Shoemaker 1979). I did not think then,and do not think now, that it shows that there cannot be informativeaccounts of what the trans-temporal identity of various sorts of things consistsin. What I did claim is that it precludes a “reductive” account of the identityof such things. It remains to be considered what that claim amounts to.

The argument rests on the claim that there is an internal relation betweenthe persistence conditions for things of a given kind and the causal profiles ofproperties characteristic of things of that kind. Properties are individuated inpart by the effects their instantiation has on the future career of the things thathave them. This is most evident in the case of dispositional properties likefragility, malleability and elasticity. For something to be malleable is for it tobe such that, when subjected to certain forces, it – that same thing – acquiresa different shape. For something to be elastic is for it to be such that, whensubjected to certain forces, it – that same thing – will change in shape, andthat when that force is removed, it – that same thing –will revert to its formershape. But it is also true of the properties that ground such dispositions thattheir causal profiles include tendencies to affect the future careers of thethings that have them.Here the nature of the properties is explained in part interms of the trans-temporal identity of the things that have them. But on a“complex view” of the identity of such things, the identity of things over timeconsists in relations, e.g. relations of causal dependence, between instances ofthese properties occurring at different times in their careers. This seems topose a threat of circularity; the trans-temporal identity of things is explained


in terms of relations between instantiations of properties, yet this trans-temporal identity figures in the explanation of the nature of these properties.In fact, I think, what we have here is not circularity but a kind of holism.

The nature of the properties and the nature of the persistence conditions ofthe things that have them cannot be explained independently of one another.It is compatible with this that we can give an informative account of whatthe persistence of a given sort of thing consists in. The account will begrounded in an account of what sorts of properties things of a given sorthave and what properties are essential to them. Take, for example, the currentdebate between psychological (neo-Lockean) accounts of personal identity and“animalist” accounts that hold that personal identity consists in the sort ofbiological continuity that characterizes the career of a persisting organism.These are competing complex views. In differing about what personal identityconsists in they are at the same time differing about what is involved in thecausal profiles of the properties persons can have. To put it very roughly, onthe psychological view the causal profiles of mental properties are such that thesuccessor states their instances contribute to generating will be states of thesame person, even if the successor states of the biological properties possessedby that person earlier belong to someone other than the possessor of thosemental states, while on the animalist view the causal profiles of biologicalproperties are such that the successor states their instances contribute togenerating will be states of the same person, even if the successor states ofthe mental states possessed by that person earlier belong to someone otherthan the possessor of those biological states. These views give differentaccounts of what personal identity consists in, each of which if true wouldbe an informative account of it.It should be noted that in most cases our concept of a property is far from

including a full specification of its causal profile. So the internal relationhere is not between the concept of the trans-temporal identity of a kind ofthing and the concepts of the causal profiles of the properties things of thatsort can have – it is between the nature of the trans-temporal identity andthe nature of the causal profiles of the properties. This shows, I think, thatthere being this internal relation does not make accounts of trans-temporalidentity in terms of continuous series of property instances circular. It is truethat in some cases our concepts of properties do include quite a bit ofinformation about their causal profiles. On a functionalist view this is thecase with many of our concepts of psychological properties. To the extentthat this is so there may be a kind of circularity involved in psychologicalaccounts of personal identity. But it is a benign circularity, reflecting aconceptual holism. It may be compared with circularities we run into in


specifying the causal roles of mental states, e.g. that characterizing the causalrole of a belief requires mention of relations to desires, while characterizingthe causal role of a desire requires mention of relations to beliefs. What suchcircularities point to is the need for a “package deal” account, and that iswhat we need here.

Wemight take reductionism about trans-temporal identity to be the viewthat the persistence of the things through time consists in relations betweenentities that can exist independently of their involvement in the careers ofpersisting things. Whether the interdependence of trans-temporal identityand the causal profiles of properties stands in the way of such a reductionistaccount depends on whether properties have their causal features essen-tially. If they have them only contingently, instances of them could existwithout being involved in the careers of persisting things. Even this wouldnot mean that the actual instances of them could have existed without beingso involved; only in worlds governed by different laws could the propertieshave different causal profiles, and only there could instances of them occurwithout being involved in careers in the way instances of them are here. Inany case, my own view is that properties have their causal profiles essentially.So I reject reductionism on this interpretation. What I have been arguing isthat the rejection of reductionism is compatible with the complex view,taken as the view that there are informative criteria for the trans-temporalidentity of things, including persons.

v

One of the central claims of this chapter is that such difficulties as there are ingiving an informative and non-circular account of trans-temporal identity aredifficulties that arise no matter what sort of persisting thing we are dealingwith. None of them is such as to make the complex view about personalidentity problematic in a way complex views about the identity of trees,rivers, etc. are not. As I suggested earlier, a major source of the view thatpersonal identity is unique in being simple is the special access we have inmemory to our own identities. But, as I argued, this is not a good reason forthe view – complex views of personal identity can easily accommodate thisspecial access. Advocates of the simple view have often been dualists. But suchreasons as dualists may have for holding the simple view are not reasons forthe rest of us. And I think there are no good reasons for the rest of us.


chapter 7

The probable simplicity of personal identityE. J. Lowe

My aim in this chapter is to show that personal identity is, in all probability,“simple” rather than “complex,” in the sense that there is no informativeand non-circular criterion of personal identity. This will require me toexplain what I understand a criterion of identity, quite generally, to be.Some philosophers, of course, are skeptical about the very notion of acriterion of identity – but that, I think, is simply because they misunder-stand the notion, which is really perfectly straightforward and clearly hasapplication. Others would want me to be more specific in my character-ization of “simplicity,” insisting that what is strictly at issue is the existenceor non-existence of an informative and non-circular criterion of diachronicpersonal identity. However, in my view, this also is to misunderstand thenotion of a criterion of identity. Where persisting things of a certain kind Kare concerned, an adequate criterion of identity for Ks should certainlyprovide an account of their identity across time, but can do so only inconjunction with an account of their identity at a time. It is an error tosuppose that “diachronic” and “synchronic” identity are different sorts ofidentity and so demand different identity criteria. Identity itself is univocaland unanalyzable. It applies to any kind of entity whatever, whether itpersists through time or not.My procedure will be as follows. First of all, I shall try to motivate the

quest for a criterion of personal identity, partly in order to subvert theattacks of those philosophers who are skeptical about the very notion of anidentity criterion. Then I shall say something about identity itself, from alogical point of view. Following that, I shall explain what I take a criterion ofidentity to be, and how, in my view, identity criteria are properly formu-lated. Subsequently, I shall focus on the specific case of persons, concen-trating my discussion on Locke’s conception of personal identity andmodern developments of that approach – often called the “psychological”approach. I shall argue that this approach offers little prospect for a satis-factory “complex” account of personal identity. Then, more briefly, I shalldiscuss other approaches which favor a “complex” account, but dismiss

137

them too, leaving a “simple” account as the most plausible option. I shallalso try to explain why we should not be surprised by this finding. Finally,I shall address an important objection to my principal reason for rejecting acomplex account.

why should we s e e k a cr i t e r i on o fp e r sona l i d ent i t y ?

Persons, I take it, are prime examples of minded beings. However, it mightbe supposed that, since the notion of identity is a universal one, there can benothing special to say about personal identity as such, beyond saying that itinvolves the application of this notion to minded beings of a certain kind.Some philosophers would undoubtedly agree with this view. They wouldurge that the theory of identity is exhausted by an account of the logicalproperties of the identity relation, which reduce to the fact that it is areflexive relation governed by Leibniz’s law: that is, by the principle of theindiscernibility of identicals. This is just the principle, taken to be a necessarytruth, that things that are identical share all their properties, or that what-ever is true of something is true of anything identical with that thing. If thatview were correct, then there would be nothing to be said about personalidentity beyond the banality that persons, like anything else, can be said tobe identical only if they are indiscernible from one another.

On this view, there is nothing more to be said regarding the hypothesisthat I am identical with, say, Julius Caesar than that it is true only if I differfrom Julius Caesar in no discernible way. It might be supposed that this isthen the end of the matter, since it is surely just obvious that there areindeed discernible differences between Julius Caesar and me, such as that heconquered Gaul but I did not. But why should anyone be so confident thatI did not conquer Gaul? The reply may be offered: because I obviously didnot even exist at the time of that conquest. But why should anyone be soconfident of that? It can only be because something is being presupposedabout the nature of persons which constrains the possibilities of identifyingone person with “another,” such as that I cannot be identical with a personnone of whose experiences I can remember having, or with a person whosebody was destroyed before the body that I have now was created. It ispresuppositions like these that make it seem “obvious” that I cannot beidentical with Julius Caesar: but they have nothing to do with Leibniz’s lawas such, since they relate specifically to the presumed nature of persons, asopposed to things of various other kinds.

138 e. j . lowe

This suggests, then, that much more needs to be said about personalidentity than can be captured simply by applying the logical properties ofthe identity relation to the particular case of persons. Specifically, whatseems to be called for is a principled account of the identity conditions ofpersons – or, to use John Locke’s helpful phrase, an account of what theiridentity “consists in.” In modern parlance, what we might hope to establishis a criterion of identity for persons. However, before we can examine anyparticular proposal concerning the identity conditions of persons, we needto look more closely at identity and identity criteria considered quitegenerally. Furthermore, we should bear in mind that we cannot just assumein advance that our quest for a criterion of personal identity is bound to besuccessful, if pursued hard enough. Indeed, as I indicated at the outset ofthis chapter, one of my main aims in what follows is to show that such aquest is very probably doomed to failure, while at the same time urging thatthis is no reason to regard the concept of a person with suspicion.

i d ent i t y f rom a log i c a l po in t o f v i ew

Now, clearly, the expression “is identical with” – symbolized in logic andmathematics by the equality sign, = – is a relational expression and hencemay naturally be supposed to denote a certain relation in which things canstand to one another. However, if so, then it is a very peculiar relation, inthat it can never literally hold between one thing and another thing, but onlybetween a thing and itself. Other relations can, of course, hold between athing and itself, such as the relation of loving: someone can obviously lovehim- or herself. But this relation can also hold between different things, aswhen Peter loves Jane. Identity is peculiar as a relation in that it necessarilyholds only between a thing and itself, and, indeed, this has led somephilosophers to deny that it is “really” a relation at all. However we classifyit, though, it can certainly seem strange. Since, of necessity, everything isidentical with itself and with no other thing, one might wonder how facts ofidentity can fail to be utterly trivial and uninteresting.Part of the solution to this puzzle is provided by distinguishing, as we

must do in any case, between identity and identification. Identification, inthe sense of the term that I now have in mind, is a cognitive act and a far fromtrivial or easy one. One and the same object may often be identified indifferent ways, even by the same thinker, and it may not be evident to such athinker that, indeed, he or she has identified the same object in two suchways. To be able to identify an object is, typically, to be in possession ofsome descriptive information which applies uniquely to that object. But, as

The probable simplicity of personal identity 139

Frege (1960 [1892]) pointed out, a thinker can be in possession of two suchpieces of information without necessarily thereby knowing that they applyto the same object. To cite his famous example, it was an astronomicaldiscovery of considerable magnitude that the Evening Star (Hesperus) is theMorning Star (Phosphorus). Similarly, it would be a remarkable discoveryto find out that the conqueror of Gaul (Julius Caesar) is the author of thischapter (E. J. Lowe). One function of identity criteria is to impose certainconstraints on what can count as an intelligible act of identification. Beforewe discuss criteria of identity, however, we need to say more about identityitself.

Identity, as I have already remarked, is a reflexive relation – a relationwhich, of necessity, holds between every thing and itself. We can formalizethis as follows:

∀xð Þ x ¼ xð Þ:As I also remarked earlier, identity is subject to Leibniz’s law, which for ourpurposes may be formalized in the following way:

∀xð Þ ∀yð Þ x ¼ y→ ∀F Þ Fx ↔ FyÞÞ:ðððHere F stands for any condition that may hold true of an object, so that theabove formula effectively affirms that, for any things x and y, if x is identicalwith y, then anything true of x is also true of y, and vice versa. From theforegoing two principles, it is easy to derive two other logical properties ofthe identity relation – its symmetry and its transitivity, expressible by thefollowing two formulas:

∀xð Þ ∀yð Þ x ¼ y→ y ¼ xÞ:ð∀xð Þ ∀yð Þ ∀zð Þ x ¼ y &y ¼ zð Þ→ x ¼ zÞ:ð

Together, these four formulas exhaust the properties of the identityrelation, from a purely logical point of view. They pin that relationdown uniquely, as being not only an equivalence relation – reflexive,symmetrical and transitive – but also, more specifically, as being theonly such relation all of whose equivalence classes are necessarily single-membered, with each such member being an ordered pair of a thing anditself, of the form ⟨x, x⟩. To make this latter point clearer, observe that eachequivalence class of the same height relation is the class of all those pairs ofobjects that share a certain height and, clearly, while it might happen to bethe case that only one object has a certain height, it is also possible formore

140 e. j . lowe

than one object to have the same height. Hence, some of these equivalenceclasses may contain many ordered pairs of different objects, such as ⟨Peter,Jane⟩, ⟨Jane, Mary⟩ and ⟨Peter, Mary⟩, assuming that Peter, Jane andMary all have the same height. But the equivalence classes of the identityrelation are all perfectly uniform, each having a unique member such as⟨Peter, Peter⟩ or ⟨Jane, Jane⟩ – because, obviously enough, Peter isidentical only with Peter, Jane only with Jane, and so on. These ratheraustere logical points are not made idly here, since they have a directbearing on what can qualify as a satisfactory criterion of identity for thingsof a given kind. We may sum up the situation by saying that while anequivalence relation such as the same height relation may be described asbeing an exact similarity relation, the identity relation is necessarily stricterthan that, in that it can fail to hold even between objects which are exactlysimilar in every qualitative respect.

cr i t e r i a o f i d ent i t y : th e i r f o rm s andadequac y cond i t i on s

As I understand it, a criterion of identity is a principle which specifies, in asubstantive way, a logically necessary and sufficient condition for theidentity of objects of a given sort or kind, K. The qualification “in asubstantive way” is needed to exclude principles that are trivial, uninforma-tive or circular. Such a principle may take one or other of two differentforms and, depending on which it takes, it may be described as being eithera “one-level” or a “two-level” identity criterion (Lowe 1989c, 1997). One-level criteria take the following form:

∀xð Þ ∀yð Þ Kx &KyÞ→ x ¼ y↔RK xyÞÞ:ðððHere, RK denotes what we may call the criterial relation for objects of kindK. And note that such a relation must, of course, be an equivalence relation –reflexive, symmetrical and transitive – because identity itself is an equiv-alence relation and RK has to hold between Ks just in case they are identical.The best-known example of such a one-level identity criterion is the axiomof extensionality of set theory, which tells us that if x and y are sets, then x isidentical with y if and only if x and y have the same members – so that in thiscase having the same members is the relevant criterial relation. Observe thatthis relation is, as required, an equivalence relation.However, Frege – who founded the formal theory of identity criteria –

favored two-level identity criteria, which may be written in the form:


∀xð Þ ∀yð Þ f K xð Þ ¼ f K yð Þ↔RK xyÞ:ð

Here, fK denotes what might aptly be called the K-function. The best wayto illustrate this is by means of Frege’s own famous example of such anidentity criterion, his criterion of identity for directions (Frege 1953 [1884],p. 74). A direction (in the geometrical sense of the word) is always adirection of something, namely, a line. And Frege’s criterion of identityfor directions is just this: the direction of line x is identical with the directionof line y if and only if line x and line y are parallel. So, in this case, theK-function is the direction of function and the criterial relation for directionsis parallelism between lines. Observe that the relation of parallelism betweenlines is again, as required, an equivalence relation.

It should be easy to see why the two different forms of identity criteriareceive their respective names. A two-level criterion specifies the identityconditions of things of a kind K in terms of an equivalence relation holdingbetween things of another kind – thus, in the case of Frege’s criterion, itspecifies the identity conditions of directions in terms of an equivalencerelation holding between lines. In contrast, a one-level criterion specifies theidentity conditions of things of a kind K in terms of an equivalence relationholding between the very things of kind K whose identity is at issue – thus, inthe case of the axiom of extensionality, it specifies the identity conditions ofsets in terms of an equivalence relation holding between the very sets whoseidentity is at issue. We shall see that this difference between the two forms ofidentity criteria is significant in the context of a search for an adequatecriterion of personal identity. For a two-level criterion of personal identitywill be appropriate only if we can and should think of persons as beingentities of a “functional” kind, in the sense that directions are.

Something now needs to be said about the requirement that a criterion ofidentity be substantive and hence non-circular. Clearly, it would be bla-tantly circular to allow the criterial relation in a one-level criterion ofidentity for Ks simply to be the relation of identity itself. It is true, butjust trivially so, that if x and y are Ks, then x is identical with y if and only ifx and y are identical. But sometimes a putative identity criterion can becircular in a less obvious way: for example, the putative identity criterion forsets which states that if x and y are sets, then x is identical with y if and only ifx and y include exactly the same sets. It is indeed logically necessary andsufficient for the identity of sets x and y that x and y include exactly the samesets (bearing in mind that every set includes itself), but since what we areseeking is an informative way of specifying the identity conditions of sets, it

142 e. j . lowe

is clearly unsatisfactory to do so by appealing to a criterial relation – in thiscase, the relation of including the same sets – which is itself defined at leastpartly in terms of sameness (that is, identity) between sets.Another example of such circularity is provided by Donald Davidson’s

well-known proposal regarding the identity conditions of events, namely,that events x and y are identical if and only if x and y have the same causes andeffects (Davidson 1980 [1969]). For, since he takes causes and effects them-selves to be events, this proposal amounts to the circular claim that eventsx and y are identical if and only if the same events cause both x and y and thesame events are caused by both x and y (Lowe 1989a). A criterion of identityfor Ks should never appeal to or rely upon, in its formulation of the criterialrelation for Ks, sameness (that is, identity) between Ks. Unfortunately,circularity of this kind in a putative identity criterion is not always easy todiscern and sometimes needs considerable work to discover. This is aproblem that afflicts certain well-known attempts to formulate an adequatecriterion of personal identity, as we shall see.One final point should be made about identity criteria in general. This is

that they are here understood to be metaphysical principles, not merelyepistemic or heuristic ones. Thus, for example, while it is true in the case ofhuman persons that having the same fingerprints provides strong empiricalevidence for identity between such persons, it certainly is not true thathuman personal identity consists in having the same fingerprints – for, quiteapart from anything else, a human person can obviously survive the loss ofhis or her fingerprints (by losing his or her fingers) and indeed can even, inthese days of modern medicine, acquire someone else’s fingerprints (as aresult of a hand transplant). So it cannot be true, quite generally and ofnecessity, that human persons x and y are identical if and only if x and y havethe same fingerprints.

what i s a p e r son ?

Locke very wisely observed that “This being premised to find whereinpersonal identity consists, we must consider what Person stands for” (1975[1690], ii, xxvii, 9). We cannot hope to formulate an adequate criterion ofidentity for objects of a kind K unless we have an adequate idea of what Ksare. But what exactly are we asking when we ask a question of the form“What are Ks?”? The correct answer, I consider, is that we are enquiring intothe nature or essence of Ks. As for what the word “essence” means in thiscontext, we again do well to cite Locke, who said that “in the proper originalsignification” of the word “it denotes the very being of any thing, whereby it


is, what it is” (1975 [1690], iii, iii, 15). From this we may gather that, at thevery least, we do not know what a K is unless we know to what ontologicalcategory Ks belong. Unfortunately, in the case of persons this immediatelygives rise to a difficulty: namely, that different philosophers over the agesand across cultures have had very different views as to what, in this sense,persons are. Some have held that persons are essentially immaterial sub-stances (“spirits” or “souls”), some that they are “combinations” of such asubstance with a material one (a “body”), some that they are purely materialsubstances (such as living animals), some that they are “phases” of suchsubstances (rather as caterpillars and butterflies are different “phases” of thesame kind of insect), some that they are non-substances (such as “bundles”of experiences, or “functional roles” that substances can occupy), some thatthey are not even individual entities of any kind but rather universals of acertain type, some that they are “transcendental” entities which cannot beidentified with items of any kind that are located in the world of space andtime, and some that they are literally non-entities having a purely “fictional”status.

What is the source of this remarkably wide difference of opinion concern-ing the nature or essence of persons? Perhaps this: the key ingredient inanyone’s conception of a person seems to be the conviction that he or sheherself is a person. Thus, possession of the first-person perspective is at the heartof anyone’s conception of a person, whatever else may also be part of it. Aperson is, first and foremost, something that conceives of itself as thinking,feeling or doing various things (Lowe 1996, ch. 1). Such a conception isone that requires the deployment of the first-person pronoun, “I,” or someexpression equivalent to that, for its articulation. But the peculiar feature ofthis pronoun, from a semantic point of view, is that its competent useapparently does not require of the user any very specific conception of whatkind of thing it designates. This is why Descartes (1984 [1641], ii) couldfamously claim to be certain of the truth of the cogito – I think – and therebycertain of his own existence, while still professing uncertainty as to what hewas. In the end, of course, he concludes that he is essentially a thinking thing, asubstance whose essence is thought (in the broadest sense of that term) andwhich excludes any physical property such as shape or mass.

Locke is less prescriptive concerning the nature or essence of persons,saying only that a person is “a thinking intelligent Being, that has reason andreflection, and can consider it self as it self, the same thinking thing indifferent times and places” (1975 [1690], ii, xxvii, 9). This definition ofpersonhood certainly builds in the notion that a person is a self-awaresubject of thought and experience, but we should not take Locke to be

144 e. j . lowe

implying, by his use of the capitalized word “Being,” that persons aresubstances, much less that they are essentially immaterial substances. Infact, it appears that Locke held human persons to be, strictly speaking,non-substances, their ontological status being that of highly complex modes(“mode” being Locke’s preferred term for an individualized property, orwhat metaphysicians nowadays call a “trope”). This is because, while hebelieved that thoughts and other mental “modes” have to be borne bysubstances and that these substances are in all probability “spiritual” ratherthan “material” in nature, he held that an individual human person who hassuch thoughts cannot be identified with any such substance, since thathuman person could in principle survive a change with respect to thesubstance bearing its thoughts at different times in its life. This, of course,is connected with Locke’s own theory of personal identity and his preferredcriterion of personal identity, to which we shall return shortly.So the problem is that, while practically everyone might agree that,

whatever else a person is, a person is something that is, or at least is capableof being, aware of itself as having thoughts, this formulation apparentlyleaves it almost entirely open what kind of thing this “something” is. In fact,it even seems to leave open the possibility that there need be no one kind ofthing that a person could be. If that is the case, however, then it wouldappear to be misguided to search for a criterion of personal identity as such,since persons of different kinds could be expected to comply with theidentity criteria, whatever they might be, associated with the kinds inquestion. For example, if it is held that human persons – as opposed, say,to android persons of science fiction lore – are animals of a certain kindand thus that I, as a human person, am identical with such an animal(a biological organism of the species Homo sapiens), then it should beconcluded that my identity conditions are just those of one such animal –that I began to exist when it did and will cease to exist when it does. Thisview – known as animalism – is currently fairly popular amongst metaphy-sicians (Olson 1997a), perhaps on account of its thoroughly naturalisticflavor and perhaps, too, because it effectively does away with many of thetraditional problems of personal identity of the sort that Locke’s accountgenerates.On the other hand, the idea that persons are not really a single kind of

thing and thus that things of radically different kinds, with quite differentidentity conditions, could all qualify as persons is prima facie counter-intuitive and even rather alarming in its apparent moral implications. AsLocke so aptly put it, “person” is “a Forensick Term appropriating Actionsand their Merit” (1975 [1690], ii, xxvii, 26): it is indispensable for our moral


and legal practices of apportioning praise and blame and offering rewardsand punishments. One’s natural presumption is that each person has andshould have a moral concern for his or her own future and, more generally,for the futures of all other persons. But if there is no unified conception ofwhat would count as “the future” of a person as such, because persons ofdifferent kinds can have quite different identity conditions, it may be hardto see what exactly could be the basis of such a universal moral concern.Indeed, if animalism were true regarding human persons such as you and I,why, after all, should I have anymoral concern for your or my future as such,given that the animals that you and I supposedly are have identity con-ditions which do not entail that those futures are ones in which you or I existas persons at all?

Reflections such as these suggest that it is strongly built into thecommon-sense conception of a person that all persons are essentially per-sons, so that my ceasing to be a person would entail my ceasing to existaltogether. Locke’s definition of personhood, whatever its defects, is clearlyintended by him to have this consequence and to that extent seems to bemore in tune with common sense than a view like animalism is. This, in anycase, is a good point at which to look more closely at Locke’s own proposedcriterion of personal identity, which has remained highly influential. Even ifin no other regard, I certainly want to follow Locke in maintaining that“person” denotes a specific kind of thing, governed by a single criterion ofidentity if it has such a criterion at all.

lock e ’ s c r i t e r i on o f p e r sona l i d ent i t y

According to Locke,

[S]ince consciousness always accompanies thinking, and ’tis that, that makes everyone to be, what he calls self; and thereby distinguishes himself from all otherthinking things, in this alone consists personal Identity, i.e. the sameness of arational Being. And as far as this consciousness can be extended backwards toany past Action or Thought, so far reaches the Identity of that Person; it is the sameself now as it was then; and ’tis by the same self with the present one that nowreflects on it, that that Action was done. (1975 [1690], ii, xxvii, 9)

It is a matter of dispute amongst Locke scholars exactly how this passageshould be interpreted (Lowe 1995, ch. 5; Lowe 2009, ch. 8), but mostcommentators take it to be expressing a memory-based criterion of personalidentity, on the understanding that the kind of memory that we arehere concerned with is what is sometimes called “autobiographical” or

146 e. j . lowe

“experiential” memory (for example, remembering seeing a certain filmsome years ago), as opposed to the mere memory of impersonal facts (such asremembering that the Battle of Waterloo was fought in 1815).Here is one way in which one might attempt to frame Locke’s proposed

criterion in the form of a “one-level” identity criterion, as such criteria wereformulated earlier:

∀xÞ ∀yÞ x is a person& y is a personð Þ→ðððx ¼ y↔ ∀t1Þ ∀t2Þ ∀eÞ x has e at t1→ y remembers e at t2Þ&ððððððy has e at t1→x remembers e at t2ÞÞÞÞ;ð

where t1 and t2 are any two times at which both x and y exist (with t1 beingearlier than t2) and e is a variable ranging over individual conscious experi-ences, such as a conscious experience of having a particular thought orundertaking a particular action. Locke himself, it should be remarked,explicitly appeals to consciousness rather than to memory when formulatinghis criterion of personal identity in the passage just quoted above, thoughhis talk of “extending consciousness backwards” is very naturally interpret-able as referring to conscious memory.What the foregoing formula says, in plain English, is just this: if x and y

are persons, then they are the same person if and only if any consciousexperience had by x at any earlier time is remembered by y at any later time,and vice versa (restricting ourselves here to times at which both x and y exist,of course, since no person can experience or remember anything at a time atwhich he or she does not exist). This criterion entails, obviously, that aperson must always remember every conscious experience that he or sheever formerly had. That, however, is extremely implausible. Indeed, itsimplausibility was fairly soon exploited by Thomas Reid (1975 [1785]) toconstruct a refutation of Locke’s proposed criterion, by means of his famous“brave officer” example. As Reid points out, we can readily imagine therebeing an elderly general who remembers saving the regiment’s standardwhen in battle as a young officer and who, as a young officer, rememberedstealing apples as a boy. But it also seems quite conceivable that the elderlygeneral has entirely forgotten the boyhood episode. The nub of the prob-lem, it seems, is that identity is a transitive relation, whereas Locke’sproposed criterial relation for personal identity appears not to be transitive.However, according to many modern neo-Lockean theorists of personalidentity, all that we need to do to save Locke’s criterion of personal identityfrom Reid’s objection is to replace Locke’s proposed criterial relation by the


so-called ancestral of that relation, which is logically guaranteed to be atransitive relation. The elderly general will certainly stand in this relation tothe boy, provided that there is a finite chain of persons, beginning with theelderly general and ending with the boy, such that each person in this chainstands in Locke’s proposed criterial relation to the preceding member of thechain. It might be worried, though, that by thus quantifying over persons informulating a revised version of Locke’s criterial relation, we are introducingcircularity into the revised Lockean criterion of personal identity, since weare now appealing to the very class of entities for which we are attempting toprovide an identity criterion. But this is a worry that I shall set aside here,concerning though it is, as deeper problems for neo-Lockeanism are still tobe encountered.

Now, the fact that the ancestral of Locke’s proposed criterial relation is atransitive relation evidently does not guarantee that it is, as required, anequivalence relation, since to have that status it needs also to be reflexive andsymmetrical. That it is reflexive might seem to be relatively uncontroversial,but that it is symmetrical is certainly not, owing to the seeming possibility offission cases, in which two distinct persons, A and B, existing at a time t2 bothstand in this relation to a single person, C, existing at an earlier time t1. Ofcourse, it may be objected that the imagined cases of personal fission (andlikewise those of fusion) are purely fictional and not really possible. But thatis too big a debate to be entered into here. Suffice it to say that such casespresent at least a prima facie problem for the revised Lockean criterion.

Another prima facie problem – first raised by Joseph Butler (1975[1736]) – leads to the accusation that Locke’s criterion, both in its originaland in its revised form, is implicitly circular. The point here is that thecriterion appeals to the notion of a person, P, remembering some pastexperience, e. However, it is a logically necessary condition of P’s genuinelyremembering e (in the first-personal, autobiographical sense of “remember-ing”) that P him- or herself should actually have experienced e. That beingso, then – as Butler urged –memory presupposes personal identity and hencecannot be what constitutes it. The standard modern response to this objec-tion is to concede it, but then to revise the Lockean criterion still further byappealing instead to the notion of “quasi-memory,” where this is understoodto be a mental state with all the features of autobiographical memory exceptthat it is not a defining condition of the state that one can “quasi-remember”only experiences that one had oneself (see, for example, Parfit 19 84 , ch. 11). Itmust be acknowledged, however, that the notion of quasi-memory is far frombeing uncontroversial, since a good many philosophers doubt whether itreally makes any sense (see, for example, Wiggins 2001, ch. 7).

148 e. j . lowe

Suppose, however, that we set aside all such doubts. Does this mean thatsome suitably revised version of Locke’s criterion – call it a neo-Lockeancriterion – may be expected to satisfy all the requirements of an adequatecriterion of personal identity? Very arguably, the answer is “no,” no matterwhat further modifications to Locke’s original criterion are made. For, as weare about to see, any version of Locke’s criterion seems vulnerable toanother charge of implicit circularity of a quite different sort from Butler’s.

the f a t a l c i r cu l a r i t y i n an y neo - lock e ancr i t e r i on

Recall again that any version of the Lockean criterion appeals to the notionof a person, P, remembering – or, if one prefers, quasi-remembering – somepast experience, e. And it must be emphasized here that it is vital that theexperience, e, that is had by some person at an earlier time is the very sameexperience that is later remembered by P, if P is to be identified, inaccordance with the criterion, with the person existing at the earlier time.So the criterion certainly presupposes some account of the identity con-ditions of experiences. But that means that we now need to ask ourselveshow experiences themselves are individuated. Such items aremental states orevents. So what are their identity conditions? We can already rule out theDavidsonian criterion of identity for events as a way of settling this ques-tion, because we found it to be implicitly circular. It was so because it soughtto identify events on the basis of the sameness of their causes and effects,while also taking these causes and effects to be events themselves: so itdefined “sameness of events” in terms appealing to sameness among events –a blatantly circular procedure leaving us no clearer as to what the identity ofevents consists in. But any neo-Lockean criterion likewise appears to beimplicitly circular, albeit in a rather more roundabout way. For it is stronglyarguable that the only adequate criterion of identity for mental states andevents will be one which makes reference to their subjects – which, in thecase of personal experiences, will be the persons who have those experiences(compare Strawson 1959, ch. 3).Let us focus then on the specific case of experiences, although the same

reasoning will apply to mental states or events of any kind, so that thefollowing objection extends to any “psychological” account of personal iden-tity. On the view now being recommended, part of what makes an experienceof mine numerically distinct from a qualitatively indistinguishable experienceof yours is the very fact that it is mine as opposed to yours. The only otherpossible distinguishing feature seems to be the time at which an experience


occurs. In short, the following seems to be a very plausible criterion of identityfor personal experiences:

ð∀xÞð∀yÞððx is a personal experience & y is a personal experienceÞ→ðx ¼ y↔ ðx and y are qualitatively indistinguishable&

∃P1Þð∃P2Þð∃t 1Þð∃t 2ÞðP1 has x at t 1&P2 has y at t 2 &ðP1 ¼ P2 & t1 ¼ t2ÞÞÞÞ;

where P1 and P2 are variables ranging over persons and t1 and t2 are variablesranging over times. In plain English, what this formula says is just this: if xand y are personal experiences, then they are the same personal experience ifand only if x and y are qualitatively indistinguishable experiences had by thesame person at the same time. It is quite clear that the criterial relationinvoked by this criterion is, as required, an equivalence relation. But,equally, it is obvious that it appeals to the notion of sameness between personsand hence presupposes that notion. Accordingly if, as I strongly suspect is thecase, this is the only adequate criterion of identity for personal experiences,then any neo-Lockean criterion of personal identity is implicitly circular inthat it will need to rely on the foregoing criterion for a specification of theidentity conditions of the experiences to which it appeals for the purposes ofidentifying persons. Clearly, at any rate, we cannot both individuate personsin terms of their experiences (as any neo-Lockean criterion attempts to do)and individuate personal experiences in terms of the persons having them(as the foregoing criterion does). And to the extent that the foregoingcriterion of identity for personal experiences looks to be in good order, itis any form of neo-Lockean criterion that must be rejected as inadequate(compare Lowe 2009, ch. 8).

Note, incidentally, that I use the term “individuation” in this context inthe way that Davidson does in his paper “On the Individuation of Events”(Davidson 1980 [1969]), whereby a criterion of identity for Ks is taken to be aprinciple of their “individuation.” I do not mean to imply by this thatI favor this usage myself, but I adhere to it for present purposes owing to itscontinuing popularity and thus in order to avoid possible confusion.

Now, some philosophers will undoubtedly consider that my reasoning inthis section is fallacious, or at least suspect. Since this is likely to be acommon reaction, I shall return to the matter shortly, discussing a quitespecific objection to my line of argument. Before doing that, however,I want to broaden our perspective somewhat by considering, albeit rather

150 e. j . lowe

briefly, some non-Lockean approaches to the question of personal identity.I shall then introduce and try to motivate the simplicity thesis that there is infact no informative and non-circular criterion of identity for persons.

s ome non - lock e an a p p roache s and thes imp l i c i t y the s i s

I have focused so far on the neo-Lockean approach because it is, deservedly,by far the most prominent one in the modern literature on personalidentity, whether it is being endorsed or being attacked. But somethingshould now be said about some alternative approaches. First of all, we haveso far considered only the prospects for a one-level criterion of personalidentity. But on some views of what persons are, a two-level criterion mightseem more appropriate – for instance, if persons are taken to be functionalstates or roles that objects of appropriate kinds can occupy. Thus, one suchview would be that a person’s body, or a special part of that body, such as itsbrain, is the object that occupies the functional role in question. Supposethat being a person is a functional role of a brain (for example, it might betaken to be the role of being a producer of first-person thoughts). Then acriterion of personal identity could be expected to take something like thefollowing two-level form, where the variables of quantification range overbrains:

ð∀xÞð∀yÞðthe person of x ¼ the person of y↔ x and y are RP -relatedÞ;

where “RP” denotes a certain equivalence relation holding between brains.(In plain English: the person of brain x is identical with the person of brain yif and only if brain x and brain y are RP-related.) Indeed, on one view, thisequivalence relation would simply be identity itself. There would be nocircularity in the criterion on this account, since it would simply be definingpersonal identity in terms of brain identity, and persons and brains are herebeing taken to be items of quite different kinds. So this approach is by nomeans identifying a person with his or her brain. The brain-identitycriterion of personal identity just implies that a person’s identity tracksthat of the person’s brain – so that, for example, if a person A’s brain istransplanted into the evacuated head of another person B’s body, thenperson A acquires person B’s body; and if person A’s brain is switchedwith person B’s, then we have a body swap, with person A acquiring personB’s body and person B acquiring person A’s body. The scenario is really verysimilar to Locke’s famous imaginary example of the prince and the cobbler,


who supposedly undergo a body swap – although what Locke envisaged wasthat the soul of the prince entered the cobbler’s body and the soul of thecobbler entered the prince’s body (1975 [1690], ii, xxvii, 15). However,although Locke thought that this scenario was in principle possible, he didnot, of course, subscribe to a soul-identity criterion of personal identity,because he thought that the same person could in principle have differentsouls at different times and that the same soul could, at different times, bethe soul of different persons. Incidentally, an interesting modern defense ofa soul-identity criterion is offered by Richard Swinburne (1986, ch. 8).

I have nothing to say in recommendation of a two-level approach like thebrain-identity criterion, although it will clearly appeal to some philosophersand psychologists. Such an approach clearly seems inappropriate if weregard the term “person” as denoting a distinct kind of substantial being –an individual substance – rather than a certain kind of state or role that such asubstance can occupy. Certainly, common sense and ordinary languagestrongly suggest the former view. I feel myself to be some thing, withdistinctive properties such as thought and feeling, rather than my beingmerely some property or feature of some other thing, such as my brain. Buthow, then, can it be explained why a satisfactory one-level criterion ofpersonal identity that supports this conviction appears to be so very elusive?

My tentative suggestion in answer to this question is that personalidentity is just so basic in our ontological scheme that we should not reallyexpect to be able to formulate such a criterion. A crucial point here is that, aswe have seen, one-level criteria of identity for objects of a kind K alwaysappeal to entities of other kinds in specifying a criterial relation forK-identity. They must do so in order to avoid circularity. This, for example,is why the principle that if x and y are sets, then x is identical with y if andonly if x and y include exactly the same sets is defective as a criterion of identityfor sets, as indeed is Davidson’s criterion of identity for events. But ifpersons really are fundamental in our ontological scheme, as I very muchsuspect they are, then we simply should not expect to be able to appeal toother entities of suitable kinds in their case. Rather, persons will be amongthe things that criteria of identity for other less fundamental kinds of entityappeal to – entities such as mental states, for example. That being so, weshould probably conclude that personal identity is primitive and “simple,”in the sense that nothing more informative can be said about the identity ofpersons than that in some cases it just obtains and in others not (Lowe2009, ch. 8).

Certainly, it seems clear that not all entities can be provided withinformative and non-circular criteria of identity, once again on pain of

152 e. j . lowe

circularity – not necessarily circularity in any single criterion of identity,but at least circularity in the entire set of identity criteria putativelygoverning, between them, all the entities that there are. This kind ofcircularity is exhibited in the conflict, discussed earlier, between any neo-Lockean criterion of personal identity and the criterion of identity forpersonal experiences proposed in the previous section. That persons arewhat P. F. Strawson (1959) called basic particulars is in fact very plausible,for all the reasons he gave. It is something very close, if not identical, tothat proposal that I am recommending here. And it should be clear that,in maintaining for these reasons that there is no informative and non-circular criterion of personal identity, I am not in the least voicing anykind of skepticism about the reality of persons or their persistencethrough time. Obviously, it is very difficult to prove a negative thesislike the simplicity thesis, since this would require us to refute any con-ceivable “complex” account of personal identity. The most that one canrealistically hope to do is to show that leading complex accounts are fatallyflawed and that the simplicity thesis is well motivated, both of which tasksI hope to have carried substantially forward in this chapter, even if I havenot by any means said the last word on the matter.

r e p l y to an ob j e c t i on

I now return to an objection to my argument (pp. 149−50) that any neo-Lockean criterion of personal identity is subject to a fatal circularity.Commenting on an earlier presentation of this line of argument (in Lowe2009, ch. 8), Sydney Shoemaker (Chapter 6 of this volume) voices thefollowing complaint. He accepts, first of all, that “there certainly is a goodsense in which the individuation of conscious states involves the identity ofpersons.” But he points out that the same holds with regard to any state andthe thing that is its subject or bearer: “Any state, psychological or not,necessarily belongs to the thing [which is its subject], and so gets its identityfrom the identity of its subject.” He concludes:

So if this sort of involvement of the identity of persons makes psychologicalaccounts of personal identity circular, any account of the identity of any sort ofpersisting thing in terms of relations between successive states [of such things]would be circular. If this were right, Lowe’s argument against the complex view ofpersonal identity would generalise into an argument against the complex view ofany sort of persisting thing – which would be contrary to his intent, since he thinksthat in the case of entities other than persons there can be noncircular constitutivecriteria of identity.


Now, I agree with the first sentence in this passage, but not with the second.I do indeed think that “any account of the identity of any sort of persistingthing in terms of relations between successive states [of such things] wouldbe circular.” But the second sentence does not follow logically from the first.I hold that there can be non-circular “constitutive” criteria of identity forpersisting things of certain kinds which are not framed in terms of successivestates of the things in question. Instead, they can be framed, for instance, interms of the persisting parts of such things. For example, I am entirely happyto follow Locke in saying that if x and y are (what he called) masses or parcelsof matter, then x is identical with y if and only if x and y are composed of thesame “atoms.” As he puts it: “the Mass, consisting of the same Atoms, mustbe the sameMass, let the parts be never so differently jumbled: But if one ofthese Atoms be taken away, or one new one added, it is no longer the sameMass or the same Body” (Locke 1975 [1690], ii, xxvii, 3). As for thepersistence of atoms – or, more precisely, what physicists now call “funda-mental particles” – I hold that it is simple, like the persistence of persons.

But Shoemaker’s objection to my line of argument runs deeper than this.He considers that I have misunderstood what is really at issue in discussionsof personal identity (and indeed in discussions of the identity conditions ofthings of other kinds). As he puts it:

In such discussions what will be at issue is what it is for different states, eitheroccurring at the same time or occurring at different times, to belong to one and thesame person. Each of the states will necessarily belong to a person, and what is atissue is what it is for it to be the same person all of them belong to.

However, this suggests that Shoemaker is implicitly seeking what I earliercalled a two-level criterion of identity for persons, of the following form,where the variables of quantification range over person-states:

ð∀xÞð∀yÞðthe person of x ¼ the person of y↔ x and y are RP -relatedÞ:

In plain English: the person of person-state x is identical with the personof person-state y if and only if person-state x and person-state y are RP-related. Now, clearly, person-state x and person-state y will be, asShoemaker puts it, states belonging to the same person just in case the personto which person-state x belongs is identical with the person to whichperson-state y belongs, that is, just in case the person of person-state x isidentical with the person of person-state y – which is precisely what the left-hand side of the above biconditional says. Hence the above biconditionalwould certainly appear to express the kind of truth that Shoemaker is

154 e. j . lowe

looking for when he seeks to establish “what it is for it to be the same person[that different person-states] belong to.” This appears to confirm that whathe is implicitly seeking is a two-level criterion of personal identity of theabove form.I have already explained why I do not favor a two-level approach myself

in the case of persons: namely, because I do not regard persons as “func-tional” entities of any kind but rather as individual substances. But, in anycase, two-level identity criteria are no less subject to the requirement of non-circularity than are one-level identity criteria. If Frege had tried to formulatea two-level criterion of identity for lines in terms of some equivalencerelation holding between their directions, this would have vitiated his owntwo-level criterion of identity for directions in terms of the parallelism oftheir lines, because advocating both criteria together would have committedhim to a circularity of individuation whereby neither lines nor directions gettheir identity conditions properly determined. This is not to say that such apair of principles – if both are true – must be entirely uninformative in thesense of being just trivial, although they will indeed be uninformative as towhat the identity conditions of the entities in question actually are, preciselyon account of the circularity that they harbor. In any case, the importantpoint is that adequate identity criteria, whether considered singly ortogether, need to avoid circularity. Hence, if I am right that person-states,including the mental states of persons, have their identity conditionsdetermined at least in part by reference to the persons whose states theyare, then a two-level identity criterion for persons of the form indicatedabove will indeed introduce a fatal circularity into an account of the identityconditions of persons, no less than in the case of a one-level “neo-Lockean”criterion of the kind discussed earlier.


chapter 8

Reply to E. J. LoweSydney Shoemaker

The following is a response to what E. J. Lowe says in the last section ofChapter 7 in reply to criticisms I made of his view in Chapter 6.

Lowe thinks that holding that there is what he calls a two-level criterionof personal identity is incompatible with holding that persons are sub-stances. I think there is no good reason to believe this. It would be closer tothe truth to say that the status of persons as substances requires that therebe such a criterion – i.e. that there be a diachronic unity relation that holdsbetween the successive states of a person.

Lowe also thinks that the view that there is such a criterion violates therequirement that an account of the identity conditions be non-circular.This too I see no good reason to believe. What he thinks is a reason is thefact that states of persons get their identity from the persons that have them,it being a necessary condition of states being identical that the personshaving them are identical. But what the two-level criterion of personalidentity gives is an account of what the trans-temporal identity of a personconsists in, and the identity of persons that figures in the identity of states isnot trans-temporal identity.

To be sure, a full account of the identity conditions for persons shouldinclude an account of synchronic identity, i.e. of what it is for simultaneousstates to belong to one and the same person. Such an account might bealong functionalist lines – states are “co-personal” (to use Russell’s term) invirtue of being such that they are apt to have certain joint effects in certaincircumstances. This would have to be combined with an account ofdiachronic identity, since some of the effects would have to be future statesof the subject of the states that produced them. What the account offers isan account of what different states having the same possessor consist in, andit surely is not made circular by the fact that each individual state necessarilyhas the possessor it has.

156

chapter 9

The non-descriptive individual natureof conscious beingsMartine Nida-Rumelin

i n t roduct i on

According to an idea that has been around for centuries, the individualnature of individual things transcends any description.1 We cannot grasp,according to this intuition, what makes a particular thing that particularthing. Individuals do not have a “qualitative essence.” Every particular is,according to this idea, the individual thing it is in a primitive and non-reducible manner; it is not in virtue of any property or combination ofproperties that an individual is the one it is. According to this intuition it isimpossible to get “a cognitive hold” of an individual thing as such in any“descriptive manner.” These formulations are far from clear and they onlyhint at an intuition which, however, is not difficult to “see.” Let us use thefollowing formulation to express the idea: individual things have a non-descriptive individual nature.The proposal of the present chapter may be expressed as follows. Not all

individual things have a non-descriptive individual nature: tables, moun-tains and other concrete material objects do not. However, consciousindividuals do have that particular status. I will propose an account ofwhat it is for an individual to have a non-descriptive individual nature

1 Famously Duns Scotus argued for the view that the individuality of an individual is to be explained byits “haecceitas,” which he apparently took to be a non-qualitative property specific to the individual atissue (Cross 2010 [2003]; Gracia 1994). The idea that, in a sense, the individual in its individualitycannot be conceived of at least by the humanmind and can become the object of one’s thought only insome kind of “ostension” appears to occur, as far as I understand it, in the fourteenth century and inparticular in the work of Pierre Auriol (Suarez-Nani 2009). For contemporary discussions of the ideathat individual things have, in some sense, a non-descriptive nature, see, for example, Black (1952),Adams (1979) and Swinburne (1995). I am not claiming that all these people had the same underlyingclaim in mind; nor is the notion of a “perfect individual” introduced in the present chapter meant tocapture the underlying common idea. But I do believe that all these ideas have a common intuitivesource; to show this would, however, require a serious and cautious analysis of the relevant texts. Thenotion of a “perfect individual” here proposed might be helpful to capture an aspect of that commonintuitive source.

157

and I will motivate the claim that conscious beings, contrary to ordinarymaterial objects without consciousness, have a non-descriptive individualnature.2 I hope that I might at least succeed in rendering the intuitiveattraction of that metaphysical claim about the individual nature of con-scious beings clearly visible to the reader.

What I am going to say is closely related to the view about trans-temporalpersonal identity which has been called the “simple view” and which hasbeen defended by, among others, Bishop Butler, Thomas Reid, RoderickChisholm, Richard Swinburne and E. J. Lowe (Butler 1975 [1736]; Reid 1975[1785]; Chisholm 1970a, 1989; and the contributions of Swinburne andLowe to this volume). A difference is, however, that I will not restrict therelevant claim to the human domain. All reasons for the view that people areperfect individuals apply with the same force, or so I will argue, to any otherconscious being. I will not explicitly address the issue about trans-temporalidentity in what follows. It will, however, become quite obvious for thereader that the so-called simple view about trans-temporal personal identityand the claim that conscious beings have a non-descriptive individualnature share their intuitive core. Presupposing the truth of these viewsone may say that they are based on the same deeper insight.

the con s t i tu t i ona l b a s i s and met a phy s i c a lb a s e s o f an ind i v i dua l ’ s e x i s t enc e

During a walk on a beach near a small town called Kiola in Australia in Julyof this year, I found a beautiful small flat black stone and I took it with me toFribourg. It is now lying in front of me on the left-hand side of the keyboardI am writing on. Let us call that black stone “Nero.”

Suppose our planet had developed in a very different manner. Supposethat – due to some difference in volcanic activities or to a clash of the earthwith a huge meteor – the distribution of continents over the earth’s surfacehad evolved differently and quite different biological species had come intoexistence. Let us now consider the following question: what would have tobe the case under circumstances of that kind for it to be true that this

2 A similar view is defended in Nida-Rümelin (2006, 2010, 2012). My view has, however, changed intwo respects: in those earlier publications I defended the claim that our conception of ourselves andother subjects of experience as “perfect individuals” (in a sense similar to the one introduced here) hasits origin in special features of first-personal self-directed thought which carry over to the way we thinkof others. I now believe that it is a mistake to see these features of first-personal self-directed thought asfundamental. Furthermore, I have changed my view with respect to whether conscious beings have anon-trivial metaphysical basis (in the sense defined in this chapter). The denial that there is such a basisis no longer part of my view.

158 martine nida-rumelin

particular stone, Nero, had nonetheless come into existence in these cir-cumstances? To address this question is to ask for Nero’s individual nature,as I will use the term. The question “What is Nero’s individual nature?” canbe translated, without loss, into the following question: what are thefeatures of a situation such that a situation having those features is ipsofacto a situation in which Nero exists? To search for Nero’s individualnature is, in other words, to search for those conditions that make it the casethat a situation is a situation in which Nero exists. Let us call thoseconditions the constitutional basis of Nero’s existence. If C is the constitu-tional basis of Nero’s existence, then any situation fulfilling these conditionsis ipso facto a situation in which Nero exists. If C is the constitutional basisof Nero’s existence, then a description of C fully expresses what it takes forNero to exist under given factual or counterfactual circumstances. Any fulldescription of C then expresses what the fact that Nero exists consists in. InNero’s case, or so I suggest, the constitutional basis of Nero’s existence issomething like this: there is an object with roughly the same history,composed of roughly the same individual particles arranged in roughlythe same way, having roughly the same form. It is the existence of such astone in a given situation, of some stone having those commonalities withthe real stone Nero, which renders it true that Nero, this stone, exists in thatsituation. That there is such a stone in a given situation is what it takes forNero to exist in that situation, or so I suggest. (An argument for this claimwill follow.)I wish to distinguish another, similar notion from the notion of the

constitutional basis of an individual’s existence: the notion of ametaphysicalbasis of an individual’s existence. A condition is a metaphysical basis of, forexample, Nero’s existence if and only if its satisfaction in a given situationmetaphysically implies that Nero exists in that situation. In other words: ifM is a metaphysical basis of Nero’s existence then M metaphysicallynecessitates Nero’s existence. Circumstances where M is satisfied and yetNero does not exist are metaphysically impossible.3 Let us, for instance,consider a situation in which there is a perfect counterpart of Nero. A perfectcounterpart of Nero is a stone (existing in a counterfactual situation underconsideration) which came into existence in that situation in exactly thesame way as Nero did in the real world; it consists of the very same particles

3 The notion of metaphysical necessity used here is of course problematic and there is no consensusabout its precise meaning. I use the term here in roughly the following way: a situation is metaphysi-cally possible if it is compatible with the nature of the things, properties and relations at issue. To saythat pmetaphysically necessitates q is then to say that there is no situation compatible with the natureof the things, relations and properties at issue where p but not q is the case.

Consciousness and non-descriptive individuality 159

and these particles are arranged in that stone in exactly the same manner.The resulting stone looks like Nero and it has the same form. Thiscondition, or so I will argue, is a metaphysical basis for Nero’s existence.There is no possible situation in which there is a perfect counterpart of Neroand yet Nero does not exist.

The constitutional basis of an individual’s existence is, obviously, also ametaphysical basis for that individual’s existence. If C is what makes it thecase that a given real individual exists, then there are no metaphysicallypossible situations where C is satisfied and yet the individual does notexist.

non -d e s c r i p t i v e con s t i tu t i ona l b a s i sf o r an ind i v i dua l ’ s e x i s t ence

I propose to understand what it is for an individual to have a non-descriptiveindividual nature in the following way: a thing has a non-descriptiveindividual nature if and only if the constitutional basis for its existence isnon-descriptive. The constitutional basis of a thing’s existence is non-descriptive if and only if there is in principle no way to say in a non-circular manner what constitutes the existence of that particular thing. Ifthis is so, then any attempt at formulating a constitutional basis for thething’s existence fails in the following way: we can only find formulations ofsuch conditions that already contain in a trivial manner the assumption thatthis particular thing exists. I will use Nero’s example to make this notion ofwhat it is for an individual to have a non-descriptive individual natureintuitively more accessible.

There is always an easy way to formulate metaphysically sufficient con-ditions for a thing’s existence. For instance, every possible world in whichNero lies in a particular spot in Australia today is a world where Nero exists;that Nero lies in that spot today is a metaphysical basis for Nero’s existence.However, the name “Nero” is used in this description as a rigid designator:it is used to say of this particular stone that it is lying in a particular place inthe counterfactual situation; it follows trivially that this object in front of meexists in the circumstances so described.4We have succeeded in formulatinga metaphysical basis for Nero’s existence but the condition is trivially

4 For readers not familiar with the notion of rigid designators introduced in Kripke (1981) the followingrough explanation will make the notion clear enough for present purposes. A rigid designator is a termused to refer to an individual in the real world and to talk of that individual in the description ofcounterfactual circumstances.


sufficient for Nero’s existence.5 In a similar way we can formulate triviallysufficient conditions for Nero’s existence demonstratively referring to Nerosaying that this stone has certain properties in those conditions. In general,there is always an easy way to describe a metaphysical basis for an individ-ual’s existence: using a rigid designator D which refers to that individual thecounterfactual situation is described as one fulfilling the condition “Dexists” or “D has property P” (where P is a property the referent of D canhave). The use of a rigid designator for the thing at issue trivializes any suchproposal.Suppose now that, in a given case, we can only formulate a metaphysical

basis of the thing’s existence by rigidly referring to that thing in a way whichtrivializes the proposal. We then know that there is no non-circular way toformulate a constitutional basis for that thing’s existence either (since everyconstitutional basis is a metaphysical basis). In that case there is no way tosay in a non-circular manner what it takes for that thing to exist in counter-factual circumstances. The fact that this thing exists in a given situation isthen an irreducible simple fact which can only be expressed rigidly referringto that thing and saying of it (directly or indirectly) that it exists. In that casethe thing at issue has a non-descriptive individual nature. This notion cannow be defined as follows: an individual X has a non-descriptive individualnature (is a perfect individual) if and only if the constitutional basis of itsexistence can only be described using a rigid designator which refers to X.

the i nd i v i du a l na tur e o f a s tone

Is Nero’s individual nature non-descriptive? If so, then we must rigidly referto Nero in any description of the constitutional basis of Nero’s existence.There is no reason to think that this is the case, or so I will argue. Let us firstconsider situations where there is a perfect counterpart of Nero. In thosesituations there is a stone which looks just like Nero, consists of the samematerial (the same individual particles) arranged in the same way, and cameinto existence (in that counterfactual situation) on the basis of the samecausal processes. This could be so, for instance, if the situation we consideris only minimally different from the real one. Is it metaphysically possiblethat all these assumptions are satisfied in a counterfactual situation and stillNero does not exist in that situation? The answer, I suggest, must be “no.”Asituation in which there is a stone exactly like Nero in all respects (same

5 The triviality is due to the well-known semantic property of names: names are rigid designators, as hasbeen argued famously and convincingly by Kripke (1981).


form, same material, same particles arranged in the same manner, sameorigin) ipso facto is a situation in which Nero exists and the stone satisfyingall those criteria in the counterfactual situation is the one lying here next tome. Is this already the constitutional basis of Nero’s existence we aresearching for? Not yet, since the condition is too strong. A slightly differentstone (it has, for example, lost a few particles) would still be Nero. But wemay safely assume that the constitutional basis of Nero’s existence is someweaker condition than the one just considered: it is a condition implied bythe existence of a perfect counterpart of Nero which does not imply theexistence of a perfect counterpart of Nero. If we wish to decide whetherNero has a non-descriptive individual nature we have to decide whether anyformulation of that weakened condition requires rigid reference to Nero. Ifwe can see that the strong condition (that there is a perfect counterpart ofNero) can be formulated without rigid reference to Nero, then we knowthat the constitutional basis of Nero’s existence can be described withoutrigid reference to Nero as well. So if the condition that there is a perfectcounterpart of Nero can be formulated without rigid reference to Nero,then Nero does not have a non-descriptive individual nature. Let us there-fore have a closer look at the way that stronger condition can be described.

There is rigid reference involved in the description: we refer to Nero(using a name or some other rigid designator) when we say that, in thesituation we wish to describe, some object shares the relevant properties (itsform, the stuff it is made of and its origin) with Nero, the stone next to menow.6However, it is in principle possible to free that description from rigidreference to Nero: rigid reference to Nero does not occur essentially – as onemay put it – in that description. It is practically impossible for us tointroduce names for every single particle involved in Nero’s genesis and itis practically impossible for us to name every single particle making up thematerial that stone is constituted of. But it is clear that it would in principlebe possible to eliminate rigid reference (and indeed any reference) to “Nero”in the description of the relevant condition. The condition only concernsthe way a huge number of particles interacted with one another in thoseevents that led to Nero’s formation and it concerns the way a huge numberof particles were finally arranged. So, in principle, there is a way to formulatethe condition that there is a perfect counterpart of Nero without rigidly

6 Another way to see that rigid reference to Nero does not occur essentially in that description is this: todescribe a situation in which Nero has a perfect counterpart we only have to pick out Nero in the realworld. Therefore, it does not make a difference whether we use a rigid or a non-rigid designator forthat purpose.


referring to Nero.7 Rigid reference to Nero does not occur essentially in thedescription of that condition and we can conclude: Nero is not a perfectindividual.Is there any reasonable way to resist this conclusion? Can someone, with

convincing reasons, insist that Nero and other material things have a non-descriptive individual nature? In the present chapter I cannot try to addresstheoretical arguments a metaphysician might develop in favor of the viewthat material things have a non-descriptive individual nature. However, ifthe way in which I propose to understand what it is for an individual to havea non-descriptive nature is accepted, then the above reasoning seriouslyundermines that view.There are, in principle, two ways to avoid the conclusion. The first is to

doubt the explication proposed of what it is to have a non-descriptiveindividual nature. According to this response, an important sense of theunderlying idea has not been captured bymy explication. I need not excludethat this might be so in the present context.8My aim here is to render visiblea difference between ordinary material objects (without consciousness) andconscious beings, a difference which I take to be important and “deep.” Iwill try to express the difference I have in mind by motivating the view thatconscious beings but not ordinary material things have a non-descriptiveindividual nature in the sense here explicated. The success of this enterprisedoes not depend on whether or not the notion here proposed of what it is tohave a non-descriptive individual nature already fully captures the under-lying idea. So I can set this first response aside.9

A second response to the above argument is to doubt that there is a non-circular way to describe the constitutional basis of Nero’s existence. Anydescription of that constitutional basis would require, according to thatresponse, rigid reference to Nero. An opponent who wishes to defend thatview would have to argue that any non-trivial candidate for a constitutionalbasis of Nero’s existence fails to capture the particular feature of a situation

7 This is not to say that any rigid reference can be eliminated. We need rigid reference to parts of Nero,or to the stuff it is made of, or to the particles that constitute Nero. But we do not need rigid referenceto Nero.

8 For instance, someone may agree with the reasoning presented and yet insist that Nero has a non-descriptive individual nature in virtue of being composed of that particular stuff. I would respond thatthis is only a derivative way of “having a non-descriptive individual nature” which is not grounded inNero’s nature as such.

9 To free myself from the assumption that I have been able to fully capture the idea of an individual’snon-descriptive individual nature, I could abandon using this expression and just use the term “perfectindividuals” (where what it is to be a perfect individual is, by stipulation, fully captured by thedefinition proposed).


which constitutes that Nero exists in that situation. (A non-trivial candidatehere is a condition which can be described without rigid reference to Nero.)If this were so, however, then even the strongest condition one mightpropose for a non-trivial constitutional basis − the condition that there isa perfect counterpart of Nero −would fail to capture that feature.10 But whatcould that feature consist in? Do we have any positive understanding of thatsupposed feature? It is evident, or so I claim, that we do not have any positiveunderstanding of the feature that must be added to a world in which there isan object exactly like Nero, made up of the same material and having thesame origin, in order to make it a world where Nero is that particular object.It clearly looks as though we have described a world in which Nero existswhen we have described a world in which there is an object having all thoserelevant properties Nero has in the real world. It “clearly looks like that” fora simple reason: we have no conception of the feature that might still belacking in those conditions in order for Nero to exist and in order for Neroto be identical with the particular stone so described.

At this point the reader might suspect an illegitimate move from a claimabout positive conceivability to a substantial metaphysical thesis. It mightlook as if the argument can be put as follows: there is no positivelyconceivable feature that might still be lacking for Nero’s existencein situations fulfilling the relevant condition. Therefore, there is no suchfeature that can be lacking in a situation fulfilling that condition (“can” inthe sense of metaphysical possibility). This, however, is not the argument Ihave in mind. I am not trying to establish in a direct way that there is no such“hidden feature” underlying the existence of Nero which is not yet capturedin the assumption that there is a perfect counterpart of Nero. My claim isweaker: I am trying to convince the reader that we have no reason to assumethe existence of such a hidden additional feature. We have no conceptionwhatsoever of that supposed additional feature which might still be lackingfor the existence of Nero in a world in which there is a perfect counterpart ofNero. Therefore, we have no reason to assume that there is such a hiddenfeature; and, therefore, we have no reason to assume that Nero has a non-descriptive individual nature. The situation is different, or so I will argue, inthe case of conscious individuals. In the case of conscious individuals we dohave a positive conception of the relevant additional feature.

10 The reasoning here is this: if that strongest condition does capture that feature, then the opponentmust be wrong. Either he has to admit that this strongest condition is itself the constitutional basis orthat it implies the constitutional basis. In both cases the constitutional basis can be formulatedwithout rigid reference to Nero.


p e r f e c t count e r p ar t s o f con s c i ou s i nd i v i du a l s

If you, the reader, are a perfect individual, then there is no way to formulatethe constitutional basis of your existence without rigidly referring to you.To judge whether this is so, it will be helpful to consider situations in whichthere is a perfect counterpart of you: a child came into existence on the basisof exactly the same biological processes that led to your existence, it wasborn to the same parents under the very same conditions, and it is genet-ically just like you. The child grows into an adult who looks like you actuallylook, who has your actual character, preferences and convictions, and whohas chosen the same profession, the same friends and partners. Satisfying allthese conditions is to be a perfect counterpart of you. In the description ofthese conditions I have been referring to you. But that rigid reference doesnot occur essentially in that description. In order to say, for instance, thatthe child has the biological origin you have in reality, we certainly need torigidly refer to biological cells, but this is not to rigidly refer to you. Allreference to you in the description of the counterpart’s properties could inprinciple be eliminated.If the description of a situation where there is a perfect counterpart of you

fails to capture what it takes for you to exist in a counterfactual situation,then you have a non-descriptive individual nature. This can be seen by thefollowing consideration. Any non-trivial constitutional basis of your exis-tence is a condition which is weaker than the assumption that there is aperfect counterpart of you. Therefore, if the stronger assumption that thereis a perfect counterpart of you does not capture what it takes for you to existin a counterfactual situation, then no candidate for a non-trivial constitu-tional basis captures what it takes for you to exist. (What a strongercondition cannot capture cannot be captured by a weaker one.) It followsthat every formulation of the constitutional basis of your existence requiresrigid reference to you, which means, according to the definition proposed,that you have a non-descriptive individual nature.So does the condition that there is a perfect counterpart of you capture

what it takes for you to exist in a given factual or counterfactual situation?Here is the beginning of an argument for a negative answer to that question.For simplicity I suppose that, if you exist in a situation where there is aperfect counterpart of you, then you are that perfect counterpart. It followsthat you do not exist in such a situation if you are not that counterpart. Nowlet us try to make sense of the difference between the two followingsituations: S1 is a perfect counterpart of you and you are that counterpart,S2. There is a perfect counterpart of you, yet you do not exist. Here is a


way to understand the difference: in S1 you have the counterpart’s body, yousee that body as your body when that person stands in front of a mirror, youare the one who is active when that person acts, you feel pain when that bodyis hurt: in short, it must be the case that you are the one who lives thatperson’s life. In S2 you never “wake up” in order to discover the world, younever experience anything, you do not exist. Pointing out this differencebetween S1 and S2 we have singled out a potential feature which constitutesthe difference between S1 and S2. And we have a clear positive under-standing of that feature.11 I will come back to the way in which we conceiveof that feature, but before doing so I must clarify an important detail.

It will be objected that S2 is not a genuine possibility − that no metaphysi-cally possible situation fulfills that description. The opponent might, forinstance, believe in the necessity of origin, in the claim that any individualhas its specific origin essentially (no other individual could possibly comeinto existence on the basis of the processes that led to its existence). Theprinciple of the necessity of origin excludes the metaphysical possibility ofS2. It is important to note that I need not argue against that view; it is notmy aim to establish the metaphysical possibility of S2. My present aim is toshow that we have a clear positive understanding of a particular featurewhich would distinguish these two possibilities, if S2 were metaphysicallypossible. In other words: we have a clear positive understanding of what itwould take for S2 to be realized if it was metaphysically possible. We have aclear positive understanding of a feature such that its presence in S1 and itsabsence in S2 would constitute a genuine difference between S1 and S2 (if S2were possible). This claim is compatible with the thesis that S2 is metaphysi-cally impossible and it is sufficient for my argument.

Compare the situation with Nero’s case. In Nero’s case we may describetwo situations in an analogous manner: S10 is a perfect counterpart of Neroand Nero is that counterpart, S20. There is a perfect counterpart of Nero andNero does not exist. In this situation we do not have a clear positiveunderstanding of some feature such that its presence in S10 and its absencein S20 would constitute a genuine difference between S10 and S20 if S20 weremetaphysically possible. We cannot form any positive understanding of anadditional feature that is necessary for Nero’s existence and not yet men-tioned in the assumption that there is a stone just like Nero, with the sameorigin and constituted of the same concrete stuff.

11 Some readers will be reminded of the last paragraph of Nagel (1965). For comments on that citationand its relation to the view proposed here, see Nida-Rümelin (2012).


the c entr a l conce p tu a l d i s ana log y

In Nero’s case, as argued before, there is no reason to suppose that thisparticular stone has a non-descriptive individual nature. The reasoningwas this: if Nero had a non-descriptive individual nature, then there is afeature of possible situations which constitutes Nero’s existence in thosesituations, which, however, is not yet captured in the assumption thatthere is a stone just like Nero with respect to its origin, form and internalconstitution. However, upon reflection it is quite easy to see that we donot have any positive conception of what that feature might be. Since wedo not have any such conception, we do not have a reason to suppose thatthere is such a feature. Now let us return to your case. Let us consideragain a situation in which there is a person with the same biological originas you have in the real world, a person who lives a life exactly like the oneyou are actually living, a person who looks like you and is like you in allpsychological and physical respects. Is there anything required for yourexistence which is not yet mentioned in that description? The answer, orso I suggest, must be “yes”: the feature required and not yet mentioned is,quite simply, that you are the one who lives that counterpart’s life, that itwas you who came into existence when that counterpart emerged. Thedescription, by itself, does not exclude that you do not exist in thesituation so described. It allows for the situation to be one where thereis nothing at all from your perspective.12

We understand the description of a case where someone is a perfectcounterpart of you and still you never came into existence. We have, as onemight say, the conceptual resources to understand that description; we areable to positively conceive of a situation satisfying it. This is more than tosay that there is no contradiction involved in the description of that case.There is no difference in this respect (lack of contradiction in the descriptionof the case) between the example about Nero and the example about you.There is no contradiction involved in the description of a situation wheresome stone is a perfect counterpart of Nero and still Nero − this stone − doesnot exist. The description might well be said to be in conflict with plausibleontological claims about the nature of stones and their identity acrosspossible worlds. However, the description in itself is non-contradictory.The same applies to your case. So far, the two cases are analogous.

12 I here presuppose that not every metaphysical consequence of a description is part of the content ofthe description. The description might metaphysically exclude that you do not exist; still, what it saysabout the situation (its content) is silent about whether or not you exist.


The difference lies in the fact that in a specific waywe cannot make sense ofthe idea that there is a perfect counterpart of Nero and yet no Nero while wecan make sense of the parallel idea in a case concerning you. To see thisclearly it is helpful to go back and forth in one’s thought between consid-ering a situation where there is a perfect counterpart of Nero which isNeroand considering a situation where there is a perfect counterpart of Nerowhich is not Nero. Doing this one can realize that it is impossible to get agrip on what the difference might consist in. We do not appear to be able tograsp a genuine difference between the two cases. Of course we maydemonstratively refer to Nero and repeat “Well, the difference is that inthe first situation this one is there and that in the second situation this one isnot there.” But doing so does not help. We cannot thereby get a grip onwhat would make the case that, in the second situation, this one is not thatperfect counterpart. No difference in the perfect counterpart we can in anyway positively conceive of might constitute the difference between thecounterpart being Nero and the counterpart not being Nero.

The situation is quite different when we come back to the case concern-ing you. Here again, to see the difference, it is helpful to go back and forthbetween considering a situation in which there is a perfect counterpart ofyou and the counterpart is you and a situation where there is a perfectcounterpart of you and the counterpart is not you (you do not exist at all).What does the difference consist in? In the first situation you are the oneliving that person’s life, you are the one who experiences the world from thatperson’s perspective, you have that person’s body, and you are the one whoenjoys that person’s pleasures. In the second situation, you never come intoexistence, there is no world from your perspective: you do not exist. There isa feature we can positively conceive of which constitutes the difference.Considering a perfect counterpart of Nero we have, as pointed out earlier,no positive conception of any feature that counterpart could possibly lacksuch that it is not Nero, the stone here beside me. Considering a perfectcounterpart of you we have a clear positive conception of a feature thatcounterpart could possibly lack so that it would not be you. If the counter-part has that feature, then, if we were in that situation, we would see youwhen looking into that counterpart’s eyes; if the counterpart lacks thatfeature then there is no way of meeting you in that counterfactual situation(you do not exist, you never started discovering the world).

Some people immediately agree with the disanalogy claim presented inthis section: they find the difference I have been describing obvious. Others,however, hesitate or spontaneously deny that there is such a differencebetween the case of the black stone Nero and the case concerning you.


This disagreement, I believe, is not a genuine disagreement. It is based, or soI hope, on misunderstandings. I will address these possible misunderstand-ings in the next section.

c l a r i f i c a t i on s o f the conce p tu a ld i s ana log y c l a im

What I wish to render “visible” to the reader may be put like this: we have aclear positive understanding of the difference between a case where a perfectcounterpart of you in a counterfactual situation is you and a case where aperfect counterpart of you in a counterfactual situation is not you. Weunderstand that difference by referring to you, in the real world, andconceiving of a situation where you have the body of that counterpart andlive that counterpart’s life as opposed to a situation where “there is nothingfrom your perspective” and you do not exist.In a sense, when conceiving of that difference, we take “your perspec-

tive” in thought, and conceive of the difference by considering the differ-ence “from your perspective.” This description may help to attractattention to the feature at issue the absence or presence of which con-stitutes the difference between the two cases. However, it should be clearthat the sense in which “we take your perspective” is not psychological: itis not taking your perspective by imagining being like you in importantrespects or by imagining being in your situation. The kind of “taking yourperspective” at issue may itself be described as “non-descriptive.” Onemay put it like this: in taking your perspective in the conception of thedifference, we conceive of the difference “in the first-personal way,” “fromyour perspective,” or, as one might say, “as if we were you.” In a sense weimagine being you, but we do so without imaging being like you in anyrespect. We refer to you, that subject of experience, and conceive of thedifference between a case where this subject exists and a case where thissubject does not exist. This is a fundamental difference for that subject (foryou, in that case). We understand the difference by understanding thedifference it would make for you.One might call this a “first-personal way” of grasping the difference. But

it is not first-personal in the sense of involving reference to oneself orreflection on one’s own mental states: we are, after all, talking of you. It isfirst-personal in the sense that grasping the difference requires seeing whatthe difference consists in “from your perspective” in the sense commentedon above. To say that we consider what the difference consists in from yourperspective should not, however, be taken to involve the idea that the


difference exists, according to that conception, only for you. On the con-trary, according to our understanding, the difference is an objective differ-ence, a genuine factual difference – a difference which is not dependent onthe way anyone experiences or judges the situation. One might get confusedhere since the objective feature at issue concerns, in a sense, “the subjectiverealm.” It is a fact concerning your “access to the world,” concerning whatyou experience “from your perspective”; it involves subjective facts aboutyou. But this does not imply that the relevant feature itself is in any senseonly subjective. It would be a genuine fact about a perfect counterpart who isyou that he or she is you, and it would be a genuine fact about a situationwhere there is such a counterpart non-identical to you that you do not existin that situation.

One might say that we have in a sense a first-personal conception of thedifference between S1 and S2. The only way to grasp that difference is tograsp it in that specific first-personal way, in thinking of you by non-descriptively taking your perspective. That mental activity (non-descriptively taking perspectives) deserves its name: it does not involveimaging oneself as satisfying a certain description; it does not involveimagining being like you or being in your situation; it involves nothingdescriptive; it only involves reference to a subject in the real world andconsidering possibilities “from its perspective.” This is just a metaphor, buthopefully a helpful one. Non-descriptive taking of perspective is notrestricted to people or human beings. One may refer to a dolphin or to araven and ask the very same kind of question in the very same sense “fromits perspective.” We have a clear positive understanding of the difference,for instance, between a situation where a perfect counterpart of the dolphinDelfina is Delfina and a situation where an equally perfect counterpart ofDelfina is not Delfina: the conceptual resources used in that understandingare the same as when we engage in the analogous conception with respect toa case concerning you − or so I suggest.

A first misunderstanding of my disanalogy claim has already been men-tioned. The claim that we have a clear positive understanding of the differ-ence between S1 and S2 should not be confused with the metaphysical claimthat there is such a feature which constitutes the difference between genuinepossibilities S1 and S2. I do not claim that both situations are metaphysicallypossible: my claim does not presuppose the genuine possibility of S2. It maybe a deep modal fact a metaphysician might discover that the existence of aperfect counterpart of you necessitates your existence. It might, for instance,be a metaphysical truth that only one specific experiencing subject (and noother) can possibly come into existence on the basis of a given concrete


fusion of cells. The claim here proposed is compatible with this hypothesis.As already said in the preceding section, the suggestion may be put like this:we have a positive understanding of what the difference between S1 and S2would consist in (if both were genuine possibilities). This is why the claimthat the existence of a perfect counterpart of X necessitates the existence of Xis a substantial claim if you are the individual X at issue. It is no substantialclaim if X is the stone Nero. (It is rather trivial in Nero’s case that Nero’sexistence is metaphysically necessitated by the existence of a perfect coun-terpart. We have no reason to assume the contrary, not even a reason towithhold an opinion, since we have no positive conception of any additionalfeature a perfect counterpart might lack for being Nero.)I have been using the locution that “a feature constitutes a difference” or

that it “constitutes your existence.” This locution may invite a secondmisunderstanding. Someone might reason as follows. We are assumingthat there is no difference between the counterpart of you in the firstsituation (in which the counterpart is you) and the counterpart of you inthe second situation in which the counterpart is not you. The humanorganism at issue consists of the same stuff in both situations and it hasdeveloped in the very same way from the same biological origin, thecounterparts do not differ in any psychological or bodily property andthey are phenomenally alike. So there is no difference between the counter-parts in the two situations in virtue of which the one in the first situation isyou and the one in the second situation is not you. So we cannot possiblyhave any understanding of the difference in virtue of which you are thecounterpart in the first situation but not in the second.There is nothing to be said against this reasoning and I agree with its

result. There is no difference between the counterparts in the two situationsin virtue of which the counterpart is you in the first situation but not in thesecond. The reason is that the situations considered do not differ in anyother respect: they only differ with respect to the perfect counterpart’sidentity; they only differ with respect to who is the perfect counterpart.When directed against the disanalogy claim the above reasoning is based onthe following misunderstanding: the claim is that we have a clear positiveunderstanding of the feature which constitutes the difference; this is not to beconfused with the different (and trivially false) claim that we have a clearpositive understanding of a further difference in virtue of which the counter-parts are not the same person.As this discussion shows, to clarify the view I am proposing it is helpful to

clearly distinguish between what it is for a feature to constitute the differencebetween S1 and S2 and what it means for S1 and S2 to be different in virtue of


a certain feature. It is quite natural to understand the second locution (butnot the first) in the following way: when we search for a feature in virtue ofwhich S1 and S2 are different, we are looking for a describable qualitativedifference between the counterparts in virtue of which they are differentpeople; we are searching for some independently understandable feature, afeature that can be described without already including in its description theassumption of identity or non-identity with you. But there is no suchfeature as the case is described, and this was precisely the point. Theexample could not serve its purpose if there was such a feature included inor compatible with the description of the example. On the other hand, thefeature constituting the difference, as I understand that locution, is notsubmitted to those constraints. When we talk of the feature that constitutesthe difference, the feature we are looking for may be an irreducible factabout personal identity (in my view it is such an irreducible fact). Using thislocution we leave it open that the feature we are looking for is no otherdifference than the one we can only describe saying of you (rigidly referringto you) that the counterpart is you in the first situation and not you in thesecond.We are not here searching for a feature that can be describedwithoutalready including in its description the assumption of identity or non-identity with you.13

Since there is no difference between the two counterparts with respect toall their physical and psychological properties, with respect to their socialenvironment and with respect to the kind of life they are living, one mightbe tempted to conclude that the alleged feature which constitutes the differ-ence must be elusive − “thin” − and that there cannot be anything positiveto grasp (just as in the parallel case of Nero). Yet, it is interesting to note, thisis not so. According to the way we conceive of that difference, the differenceis not elusive or “thin” at all. There is something positive to grasp here whenwe consider the difference between the two cases. But how is this possible,one might ask, despite the fact that there is no descriptive differencebetween the two cases (no difference in virtue of which the counterpartsare different subjects)? The answer, I think, must be this: the difference canbe grasped in the first-personal way described earlier. We can grasp thedifference on the basis of non-descriptively taking your perspective. Ofcourse, from the perspective of the perfect counterpart in the first situationand from the perspective of the perfect counterpart in the second situationall is exactly alike. But rigidly referring to you and assuming that in the first

13 The two locutions are sometimes used interchangeably. The distinction made here need not describecommon usage.


situation the counterpart is you while in the second it is not, we becomeclearly aware of a relevant difference. The first situation is a situation whereyou are living a rich human life, while the second is a situation where younever start experiencing anything. We have a rich conception of that differ-ence. The difference, of course, does not lie in the counterpart’s rich life inthe two situations (which is exactly alike). The difference lies in nothing butwho is experiencing that life. In order to grasp this difference we mustengage in what I called first-personal thought applied to others. Doing so weattain a rich conception of that difference.

the con s t i tu t i ona l b a s i s o f thee x i s t enc e o f con s c i ou s i nd i v i du a l s

Considering cases of the kind discussed we can discover facts about theway we think. We can discover what we implicitly take to be essential forbeing a subject. According to our understanding, subjects are the kind ofindividuals with respect to which it is adequate to engage in non-descriptiveperspective-taking. They are the kind of individuals with respect to which it isadequate to think in a first-personal way. Thinking about subjects in thisparticular way we can furthermore discover how we think about the constitu-tional basis of their existence. We discover that, according to our notion of asubject, the constitutional basis of any subject’s existence is non-descriptive.One may summarize the result of that reflection as follows: according to ourunderstanding of what it is to be an experiencing subject, it is essential forbeing an experiencing subject that one’s individual nature is non-descriptive.This assumption is built into our concept of an experiencing subject.All these observations about our own concept of conscious beings do not

imply that conscious beings have a non-descriptive individual nature. Itcould be, after all, that false assumptions are built into our concepts. But ourconcept of a conscious being is a special case. It is formed on the basis ofbeing an experiencing subject. On the basis of being an experiencing subjectwe have access to what it is to be an experiencing subject. This is why, inthat particular case, discoveries about our own concept can lead to insightsabout the referent of the concept, or so I will argue.

i m p l i c i t conce p t s and ph i lo so ph i c a l i n s i ght s

My argument will rely on a certain view about concepts and their role in theacquisition of philosophical insights that cannot be developed here. But Ihave to mention a few elements of that view to render the following


argument more accessible and more convincing. These elements will bebriefly presented without any argument in favor of the view theycharacterize.

Many of those concepts that play a central role in our thinking and in theway we perceptually and emotionally experience the world, ourselves andothers are only implicit and have no linguistic counterpart. They organizeand structure our mental life and they are relevant to the way thingscognitively, perceptually and emotionally appear to us. These conceptsoften incorporate assumptions about the nature of the phenomena (thekind of thing, the property, the relation) they refer to. However, in somecases it is quite difficult to bring those implicit assumptions to the surface,and to make them consciously accessible by finding an adequate expressionin language. Having these conceptsmanifests itself in how things cognitivelyappear to be (intuitions) and in what we can or cannot positively conceiveof. In many cases we must make use of these implicit concepts in anyadequate description of how things appear in perception, agency andemotion.Ways to bring those implicit notions and the implicit assumptionsabout the nature of their referents that they incorporate up to the surface ofconscious understanding are, among others, the testing of intuitions withrespect to concrete and counterfactual cases and the “method” applied inthe present chapter: finding out, by reflection upon counterfactual cases,what we can positively conceive of and in what kind of conceptualizing weare thereby engaged. Another such way to bring them to the surface is tocarefully enquire as to the content of emotions, perceptions or the experi-ence of agency.

As the examples mentioned will already suggest, the strategies availableto us to access those central, fundamental but implicit notions and theassumptions about the nature of the referent they incorporate have littleto do with conceptual analysis in the traditional sense. They are far morerich and complex and they do not aim to discover analytical truths(truths in virtue of linguistic meaning). Obviously, this cannot be the aimwhen implicit concepts which have no counterpart in natural language areat stake.

When we discover that a certain assumption about the nature of a kind ofthing (about what it takes to belong to that kind, about what is essential formembership in that kind) is incorporated in a central implicit notion, thenwe cannot in general conclude that the things belonging to the kind thenotion refers to actually satisfy that assumption. The implicit notion mightbe inadequate. This can be so if the reference of the notion is somehow fixedindependently of the relevant implicit assumption. In that case it might


be that the members of the kind the notion refers to do not in fact satisfy thecondition which is, according to our notion of the kind, essential formembership in that kind. In some cases the implicit assumption aboutwhat is essential for membership in the kind referred to fixes the reference tothe relevant kind. In this case the assumption cannot turn out to be false,but we may find out that there are no things belonging to that kind.

acc e s s to the na tur e o f con sc i ou s b e i ng s onthe b a s i s o f b e i ng a con sc i ou s b e i ng

In the present case the implicit notion at issue is the notion of an experiencingsubject or a conscious being. The terms “experiencing subject,” “experiencingbeing” or “conscious being” are technical terms which have no clear counter-part in natural language. They are introduced as expressions for an implicitnotion which plays a central role in our thinking and in the way things appearto us (cognitively, emotionally, perceptually and in agentive experience). Wecan see, or so I claim, that the assumption that experiencing subjects areperfect individuals is incorporated in our implicit notion of a subject and isthus deeply incorporated in the way we think and experience, and in the waythings cognitively and experientially appear to be.One might try to argue that the incorporated assumption at issue plays

a crucial role in fixing the reference to the kind referred to (which is thecategory of conscious individuals). If this is so, then the denial thathumans and other animals are perfect individuals implies that they arenot conscious beings. This may sound like a reductio ad absurdum but thephilosophical opponent will not be impressed. He or she will rather denythat the “inbuilt” assumption about our individual nature plays a signifi-cant role in fixing the reference to the kind we are talking about. He mightsay that the reference is rather fixed by our linguistic practices in theapplication of mental vocabulary, or something along these lines. Itfollows that conscious beings might very well not be perfect individuals,although this assumption is incorporated into our notion of what it is to bean experiencing subject. To this one might answer that the illusion theopponent thereby attributes to our thoughts and experiences is so funda-mental and omnipresent that it undermines his or her view.14 A moredirect way to argue for the claim that conscious beings are perfect indi-viduals will be briefly sketched in what follows.

14 For a version and an elaboration of this kind of argument, see Nida-Rümelin (2006, sections 3.15–3.20and 5.2–5.3).


Our implicit notion of conscious individuals has its origin in what we areaware of in experience. In any experience we have an implicit awareness ofourselves as the subject of the experience, in any action we have an implicitawareness of ourselves as the one who is active in that action, and in anyfirst-person memory we have an implicit awareness of ourselves as persistingacross time. We should, or so I suggest, accept a thesis of revelation withrespect to what it is to be an experiencing subject: in a non-conceptual andimplicit manner we are aware of what it is to be an experiencing subject bybeing an experiencing subject; we thus have implicit knowledge of what it isto be an experiencing subject given that we are subjects of experience; thenature of what it is to be an experiencing subject is revealed to us by being asubject of experience.

Our implicit notion of what it is to be a conscious individual is based onthe special kind of access we have to what it is to be a conscious individual bybeing a conscious individual. This is the best and most direct access onecould possibly have to what it is to be a subject of experience. Our implicitnotion of what it is to be a subject of experience is formed on the basis ofthat particular access. We therefore have a right to suppose that the notion isadequate. A notion is adequate if the assumptions about the nature of thereferent that are built into the notion are true. Therefore, if our notion of asubject of experience is adequate, as we should, in my view, suppose, thenrevealing these assumptions by careful reflection is a way to reveal the natureof the kind the notion refers to. If so, then the ontological claim thatconscious beings are perfect individuals is justified by the insight that thisassumption is deeply incorporated into our concept of what it is to be anexperiencing subject.

This argument presupposes a plausible claim which, however, needs tobe carefully elaborated and argued for. In order to find out what it is to be anexperiencing subject, we need not wait for any scientific discovery: wesimply have to uncover what we implicitly knew all along.


part i i i

Reconsidering simplicity

chapter 10

Personal identity: a not-so-simple simple viewLynne Rudder Baker

A number of different issues travel under the banner of “the problem ofpersonal identity.”My interests, like those of many other philosophers, aremetaphysical. In the first instance, I am not concerned with what is called“narrative identity,” or with how we re-identify a person, or with psycho-logical aspects of personality, or with ascriptions of the word “person.” “Theproblem of personal identity over time,” as formulated by Harold Noonan,“is the problem of giving an account of the logically necessary and sufficientconditions for a person identified at one time being the same person as aperson identified at another” time (Noonan 2003).If, as I believe, you and I are essentially persons, then we are persons at

any moment that we exist. (We could not change into non-persons and stillexist.) This is not uncontroversial, but I have argued at length for this viewof persons elsewhere (Baker 2000, 2007d). In any case, a metaphysicalaccount of personal identity should be tied to an account of the nature ofpersons, and that is how I shall proceed.

what i s a s imp l e v i ew of p e r sona l i d ent i t y ?

Typically, accounts of personal identity over time are classified as beingeither simple or complex. A simple account takes personal identity to be“an ultimate unanalyzable fact, distinct from everything observable orexperienceable that might be evidence for it” (Noonan 2003, p. 16), or a“further fact” that “does not just consist in physical and/or psychologicalcontinuity” (Parfit 1984, p. 210; Noonan 2003, p. 16). If personal identityover time is, as simple views hold, unanalyzable, then there are noinformative or non-circular necessary and sufficient conditions for aperson identified at one time to be the same person as a person identifiedat another time. Complex views specify informative necessary and suffi-cient conditions for personal identity over time – for example, persistenceof body and brain, psychological continuity, etc. The difference betweensimple and complex views of personal identity depends on whether they

179

are informative or not. Informative conditions of personal identity can bestated without presupposing the existence of persons over time. (I shall saymore about informativeness later.) Here are my characterizations of“simple” and “complex” views of personal identity:

An account of personal identity is a simple view if and only if the account offers noinformative necessary and sufficient conditions for a person identified at one timeto be the same person as a person identified at another time.

An account of personal identity is a complex view if and only if the account doesoffer informative necessary and sufficient conditions for a person identified at onetime to be the same person as a person identified at another time.

Most mainstream views (e.g. those of Shoemaker, Parfit and Lewis) arecomplex. Complex views (as far as I know) take personal identity to consistin a relation between items – such as mental states or brains or bodies –construed in sub-personal or non-personal terms. (That is what makes theminformative.) By contrast, typically simple views (e.g. those of Foster andSwinburne and sometimes Chisholm) appeal to some kind of immaterial-ism, according to which a person can exist independently of any body.1

Here I shall offer a simple view, but of a different sort from the views ofthose who take persons to be immaterial substances. On the one hand, I amnot a reductionist about persons: I do not believe that personhood can beunderstood in sub-personal or non-personal terms, and hence I do notbelieve that there are non-circular informative necessary and sufficientconditions for a person identified at one time to be the same person as aperson identified at a different time. On the other hand, I also reject theview that persons are “separately existing entities” that can exist apart frombodies or that are immaterial minds or souls. I reject both standard views,and describe a third way.

a f i r s t - p e r sona l a p p ro ach

According to my third way, we are fundamentally persons,2 who arenecessarily embodied, but we do not necessarily have the bodies that we

1 Parfit makes this point by saying that simple viewers hold that a person is “a separately existing entityfrom brain and body” (Parfit 1984, p. 251). I think that “separately existing entity” is ambiguous: it maymean “not identical to brain and/or body” or it could mean “exists independently of brain and/orbody.” On my view, we are persons on the first reading, but not on the second. (I suspect that manyphilosophers would not distinguish between the two readings, but there certainly is a conceptualdistinction.)

2 Throughout this chapter, I am talking about what I call “non-derivative persons.” Your body is aperson derivatively as long as it constitutes you; but you are a person essentially and non-derivatively.

180 lynne rudder baker

in fact have. I take person to be a primary kind. On my view, every concreteindividual in the natural world is of some primary kind or other (Baker2007d, pp. 33–9). An entity x’s primary kind F is the answer to the question“What, most fundamentally, is x?” Everything is of its primary kindessentially. What makes us the kind of beings we are is not what we aremade of; rather, what makes us the kind of beings that we are is a particularconceptual ability: an ability to think about ourselves and what we are doingin the first person, without identifying ourselves by any third-persondevice – like a name or description or a third-person pronoun.Here is an example to show that the existence of a particular thing does

not depend on what it is made of. Artifacts are what they are in virtue oftheir intended functions (abilities), not in virtue of what they are made of.What makes something a watch is that it is produced in order to tell time.Its primary kind – watch – depends on its intended function, which is bothrelational and intentional.What it is made of is wholly irrelevant to its beinga watch. If it is a wristwatch, there is a constraint on what it is made of: itmust fit on a wrist.Similarly, a person may be made of something – silicon, organic material,

what have you. But what makes the thing a person is a conceptual ability,not what it is made of. But if the thing is a human person, there is aconstraint on what it is made of: it begins existence constituted by a humanorganism. Just as whatever the watch is made of should support the abilityto tell time, so whatever the human person is made of should support theability to think of oneself as oneself – i.e. a robust first-person perspective.3

Furthermore, just as the intentionality required for there to be manufac-tured items that can tell time may have come about by natural selection, sotoo may the mental mechanisms required for there to be natural objects thathave first-person perspectives have come about by natural selection. So thisaccount of persons should be congenial to Darwinians.We are essentially embodied, but our bodies can be made of anything as

long as they provide the mechanisms that support our person-level activitiesand states. The relation between us and our bodies is constitution – thesame relation that a painting has to the canvas and paint that constitutes it.We are constituted by our bodies, and the bodies that constitute us now areorganisms. With enough neural implants and prosthetic limbs, we maycome to be constituted by bodies that are partially or wholly non-organic.

3 I am speaking of what I call a “robust first-person perspective.” I am omitting complications aboutbabies and higher non-human animals, which I take to have rudimentary first-person perspectives. Fordetails, see Baker (2007d, ch. 4).

Personal identity: a not-so-simple simple view 181

Our uniqueness lies in the fact that our persistence conditions – whichwe have essentially – are first-personal. If someone says, “I wonder how I amgoing to die,” or “I wish I were a movie star,” her first-person perspective isexemplified – that is, she (that person) exists. Everything else besidespersons – animals, artifacts, artworks – has third-personal persistence con-ditions. Any view that takes a person’s persistence conditions to be bio-logical, or physical, or “somatic” leaves out – must leave out – what isdistinctive about persons: the first-person perspective. This is so because, asDarwin emphasized, the animal kingdom is a seamless whole, and thepersistence conditions of organisms are third-personal. What is uniqueabout a human person is an essential first-person perspective, which I taketo be non-Cartesian. (Indeed, my philosophy of mind is externalist: Baker2007b, 2007c.)

On my view, human persons are emergent – as are organisms and otherconstituted things.4 Your first-person perspective is a property instance thatcannot be divided or duplicated. So, a molecule-for-molecule replica of yourbody would not have your first-person perspective. Hence, fission problems(mercifully!) do not arise.

So, here is my not-so-simple simple view: a person is a being with a first-person perspective essentially and persists as long as her first-person perspec-tive is exemplified. To allow for the possibility that persons are temporallygappy, I should say: a person exists when and only when her first-personperspective is exemplified. To put this in terms of possible worlds: at any timet and in any possible world w, I exist at t in w if and only if my first-personperspective is exemplified at t in w (Baker 2007a, pp. 237–9).5

This condition for personal identity over time is not informative inas-much as reference to the person is made in the explicans: “her first-personperspective” or “my first-person perspective.” Hence, this is a simple view.However, there is also a necessary condition for exemplifying a robust first-person perspective: a person must be constituted by something with mech-anisms (e.g. neural mechanisms) adequate to support first-person referenceto oneself as oneself.6

4 The relation between persons and bodies is not a supervenience relation; nor is it a part−wholerelation. Constitution – not supervenience, not mereology – is the glue of the universe.

5 Harold Noonan notes that we can take the problem of diachronic personal identity to be “the problemof specifying the relation which body a existing at time tmust bear to body b existing at time t0, if theperson occupying a at t is to be identical with the person occupying b at t0” (Noonan 2003, p. 94). Since,on my view, human persons are necessarily embodied, my view may be expressed in Noonan’s terms:the person occupying a at t is identical to the person occupying b at t0 if (and only if) the personoccupying a at t and the person occupying b at t0 have the same first-person perspective.

6 I think that this is close to what Noonan calls “the unoccupied view” (Noonan 2003, p. 97).


why not - s o - s im p l e ?

I have called my view “not-so-simple.” Why? Although the view conformsto the characterization of simple accounts that I started with, it differs fromother (perhaps more standard) simple views. For example, John Foster usesthe term “simple view” to apply to accounts of persons as non-physicalsubjects: “Jones and the non-physical subject to which the pain is attributedin the philosophically fundamental account are one and the same” (Foster1991, p. 238). My not-so-simple simple view differs from any simple viewthat supposes that there are immaterial substances that can exist withoutbodies in the natural world.The not-so-simple simple view differs from other simple views in other

ways as well. Simple viewers who invoke immaterial substances like soulsoften take personal identity over time to be determinate. That is, for anyperson P and time t, P either definitely exists at t or definitely does not existat t. For example, Thomas Reid said that when identity is applied topersons, it “admits not of degrees or of more or less” (quoted in Noonan2003, p. 16). However, the not-so-simple simple view rejects Reidiandeterminacy about persons and other ordinary objects. Here is the reasonthat I reject Reidian determinacy.Because I take human persons to be natural objects, I take them to

come into existence gradually. I think this follows from the empirical factthat no concrete object of any kind – a natural object or an artifact –comes into existence instantaneously or goes out of existence instanta-neously. (Our solar system took eons to come into existence. Fertilizationof an egg by a sperm takes up to twenty-four hours.) Hence, we have goodreason to believe that human persons do not come into existence instanta-neously; like everything else, they come into existence gradually. But ifpersons, who are essentially persons, come into existence gradually, thereis a time at which their existence is indeterminate. However, as I haveargued in The Metaphysics of Everyday Life (Baker 2007d), indeterminateexistence (of anything) depends on the determinate existence of thatthing. It is indeterminate whether some person P exists at t only if thereis some other time, t0, such that it is determinate that P exists at t0. Let meexplain by an analogy.Suppose that you had a house built. Consider the time t at which the

foundation was laid and the frame was in place; did your new house existat t? The answer depends on what happens subsequently: suppose that yourhouse was completed at t0, at which time it existed determinately. Since yourhouse existed determinately at t0, it existed indeterminately at t when the


foundation had been laid and the frame was in place. But if a tornado hadtorn the structure down right after t, it would not have been the case thatthere had been a house that existed indeterminately at t. That is, its havingexisted indeterminately at one time depends on its existing determinately atsome other time.

Indeterminacy, so understood, is metaphysical and not just semantic.However, this construal of indeterminacy does not entail that there isindeterminate identity. For any x and y, if x and y are identical, “they” aredeterminately identical. If x and y have vague temporal and spatial bounda-ries, x and y are identical – determinately identical – only if they are bothvague in the same ways to the same degree at the same times. If H isindeterminately a house at t, and H* is determinately a house at t0, then H =H* only if there is a unique x such that x is indeterminately a house at t and xis determinately a house at t0. Identity is determinate, atemporal andnecessary; the existence of temporal beings like houses and persons is not(Baker 2007d, pp. 226–33). Identity is all or nothing; temporal existence –when an entity is coming into existence, or going out of existence – is not.

Since the first-person perspective depends on the proper functioning ofneural mechanisms, and the proper functioning of neural mechanisms issubject to indeterminacies, there are indeterminacies in the coming intoexistence and going out of existence of persons. (Even if a person does notgo out of existence at death, the earthly body – perhaps gradually – ceases toconstitute the person.)

However, some philosophers take it to be a merit of the standard versionsof the simple view that they are committed to the determinacy of persons(Noonan 2003, p. 16). The merit of determinacy is that it seems to avoid theduplication problem. One version of the duplication problem is generatedby the “split-brain” thought experiment, where a person’s brain is split andone half is put into one body (“Lefty”) and the other half is put into anotherexactly similar body (“Righty”). Both successor persons survive, but on painof contradiction, they cannot be identical. So, which is the original person,Lefty or Righty? Determinacy about persons seems to assure us that there isan answer. But without determinacy, Lefty has no more claim to being theoriginal person than Righty, and vice versa. Although I have little patiencewith such thought experiments, if there is a puzzle here, it can be solved bymy construal of indeterminacy. There are not just two possible answers, butthree possible answers to the question: which is the original person? Theanswers are either Lefty, Righty or neither, and the not-so-simple simpleview is compatible with all three answers. We do not know which is thecorrect answer, but the not-so-simple simple view implies that there is a fact


of the matter that depends on whether Lefty or Righty or neither has theoriginal person’s first-person perspective. So, we do not need Reidiandeterminacy in order to avoid the duplication problem.There is still another way that my not-so-simple simple view differs from

standard simple views. For example, Roderick Chisholm and many otherphilosophers take the word “simple” to mean “having no parts,”where havingparts entails conforming to a classical mereological theory of “the part−wholerelation.” Thus, Chisholm and others take persons to be simple in that theyhave no parts.

do p e r son s hav e p a r t s ?

Do persons have parts?7 I cannot give a one-word answer to this question,because I think that it is ill-formed. The word “parts” is ambiguous. Personshave what I shall call “ordinary parts,” which are parts of ordinary objectslike animals, clocks and people; persons and other ordinary objects do nothave what I shall call “mereological parts.” Mereological parts are parts ofmereological sums, and governed by axioms of classical mereological theo-ries (Leonard and Goodman 1940). Although I have no objection tomereological theories as formal analyses of a relation called “the part-relation,” I do not believe that mereological theories merely sharpen theordinary notion of “parts.” Rather, mereological theories introduce a differ-ent idea of parts, such that they do not apply to everyday objects or to theirordinary parts unless supplemented by constitution. The relation of mereo-logical parts to what they are parts of is different from the relation of a cow’stail to the cow that the tail is part of.In classical mereology, “part” is a primitive, and all objects with parts are

taken to be sums. What is defined is a “sum” or “fusion”:

(S) y is a sum of the xs = df every x is a part of y, and every part of y overlaps someof the xs,

where x overlaps y if x and y share a part. Mereologists believe that all objectswith parts are sums.8 But on a three-dimensionalist view of ordinary objects,no instance of (S) is identical to any ordinary object. All that (S) can yield are

7 Throughout this discussion, I am construing parts non-derivatively (in the vocabulary of Baker2007a). Also, I am assuming three-dimensionalism throughout. For a defense of three-dimensionalism, see Baker (2007d, pp. 199–217; Baker 2009, pp. 1–14).

8 A variant of my view would reject universal composition, relativize composition to time, construeconstitution as a many−one relation, and appeal to plural quantification. Then: the xs have a sum at t ifand only if there is a y such that the xs constitute the y at t.


sums: sums are aggregates, not ordinary individuals.9 To see that a sumof three-dimensional objects is not an ordinary individual, consider the sumof atoms that compose my lectern. Now look at (S): exactly the same sum ofatoms would be scattered about if the lectern were smashed. There is asingle sum of (3-D) atoms, but before the smashing there is a lectern andafterwards there is no lectern, but the sum still exists. On a 3-D view, thedefinition (S) is satisfied nomatter how the atoms are arranged. The relationbetween a particular atom in the lectern and the sum of atoms in the lecternis obviously different from the relation between that atom and the lectern:the sum exists whether the lectern does or not. (3-D mereologists may saythat the xs have a sum when arranged lecternwise, but not when scatteredabout. But arrangements are not in the ontology.) I shall use the term“mereological part” for the term “part” in classical mereology.

Consider another kind of example: in classical mereology, any things thatexist have a sum.10 So there is no metaphysical difference between the sumof your head and the moon, and the sum of your head, limbs and trunk:your head and the moon compose one sum, and your head, limbs and trunkcompose another sum. There is a single relation – composition – and eachof the sums is a relatum of an instantiation of that relation. But if your headis related to the sum of your head and the moon in the same way that yourhead is related to the sum of your head, trunk and limbs, that relation is notthe relation between ordinary objects and their parts. At this moment, therelation between your head and the sum of your head, trunk and limbs isquite a bit more intimate than is the relation between your head and thesum of your head and the moon. At this moment, the sum of your head,trunk and limbs constitutes you; but at no time does the sum of your headand the moon constitute anything. Composition (in classical mereology)does not distinguish arbitrary from non-arbitrary sums, but arbitrary sumsare not ordinary material objects.11

9 If mereology is interpreted to apply to all material objects, then it is a false theory. Uninterpreted, it isan abstract calculus that is neither true nor false (non-Euclidean geometries).

10 I agree with Theodore Sider that mereological universalism is true because there cannot be vaguecomposition (Sider 2001). But from determinacy of composition – a relation between sums and theirmereological parts – nothing follows about genuine objects. Composition is too weak and promis-cuous to support a metaphysics of genuine objects. What can be vague is constitution: whichmereological sum constitutes Mt. Everest at t?

11 Although Peter Simons takes mereology to be an important part of ontology, he does not think thatmereology alone can tell us “what collections of objects compose others” (Simons 2006, p. 613). Mysuggestion is that we should give up on composition as a relation in which new individuals come intoexistence. Sums (e.g. sums of microparticles) are constituters, but I do not take a sum to be identical toa concrete object – ever. New concrete objects come into existence by constitution, not bycomposition.


In classical mereology, if S1 and S2 are sums and there is any difference intheir mereological parts, then S1 ≠ S2. Three-dimensional classical mereo-logy lacks the resources to have objects that gain or lose parts. I am awarethat Peter van Inwagen has proposed a temporal mereology that allowschange of parts. But his view entails that many sums that have no parts incommon are, nevertheless, one and the same sum (Van Inwagen 2006). Forexample, on Van Inwagen’s account, sum A (the sum of your cells when youwere born) and sum B (the sum of your cells fifty years later) are the samesum even if not a single cell in sum A is also in sum B. But the idea of sums(unlike the idea of persons) is conceptually tied to the idea of parts. Sum Aand sum B constitute the same object – you – but they are not the samesum. (They have no parts in common.) To construe sums in such a way thatidentity of parts is irrelevant to the identity of sums seems to me to rip theword “sum” from the theoretical home that gave it meaning.In any case, I do not believe that the formal properties of “the” part-

relation will give us an account of ordinary objects or their ordinary parts(Simons 2006). However, if we appeal to constitution, we can find a bridgefrom mereological parts to ordinary parts for ordinary objects at times. Thefollowing is a necessary condition for being an ordinary part at a time.12

(P) If x is an ordinary part of y at t, then ∃z(x is a proper mereological part of z and zconstitutes y at t).

An ordinary part of an ordinary object at t – a chair leg – is a propermereological part of a sum that constitutes the chair now. Sums havemereological parts, and have them essentially.Now we can see how different sums are from ordinary objects: sums and

only sums are composed of mereological parts; no constituted object likeyou or me has mereological parts. Sums cannot change parts; ordinaryobjects routinely change parts and are as real as objects can be. So, I thinkthat Chisholm’s view is exactly backward. It is not that genuine objects areconstrained by mereological essentialism. Mereological essentialism appliesonly to mereological objects – i.e. sums defined by (S) – that are not genuine

12 In Baker (2007d, 2008), I offered a definition of ordinary parthood at a time in terms of mereologicalparthood. But I now think that the definition does not fully capture ordinary parthood. It wouldallow that an atom inmy chair is an ordinary part of my chair. So, I have demotedmy “definition” to anecessary condition.Here is an alternative construal. Define “mereological part” and “ordinary part” in terms of a

primitive term “part” in a generic sense:x is a mereological part of y = df y is an aggregate and x is an item in y; andx is an ordinary part of y at t = df ∃z(x is a mereological part of z and z constitutes y at t).On this construal, any part is either a mereological part or an ordinary part.


objects at all.13 No sum is identical to an ordinary object; rather, sums (oflegs and tabletop, say) constitute ordinary objects.

The answer to the question “Do persons have parts?” depends on how weare using the term “parts.” If we mean “ordinary parts,” yes, of coursepersons have parts: by (P), your head is an ordinary part of you in virtue ofbeing a mereological part of a sum that constitutes the body that constitutesyou.14 (The sum that constitutes your body is the sum of your head, trunkand limbs.) But persons (and other ordinary objects) do not have propermereological parts (parts constrained by axioms of any extant mereologicaltheory known to me). Only mereological sums (aggregates) have mereo-logical parts.15 The relation between an ordinary part and an ordinarything – like a person – is a different relation from the relation between amereological part and a sum. So, the relation between an ordinary thing andits ordinary parts at t is not composition.

So, persons (and other ordinary things) are simples if we mean that theyhave no mereological parts – parts constrained by a formal theory; butpersons are not simples if we mean that persons have no ordinary parts. Anot-so-simple simple view of personal identity has a not-so-simple answer tothe question “Are persons simples?”

why ther e a r e no in format i v e c r i t e r i ao f p e r sona l i d ent i t y

The first-personal view is a simple view because it provides no informativecriteria of personal identity.16 The not-so-simple simple view provides noinformative criteria of personal identity because it takes personhood to be abasic property, and not susceptible to a non-personal or sub-personalaccount. By a “basic property,” I mean one that either was exemplifiedfrom the beginning of the universe, or one whose exemplification is emergent.New kinds of phenomena, objects and properties emerge over time – some by

13 The distinction between ordinary parts and mereological parts also makes clear the distinctionbetween constitution and composition. Composition is a mereological relation; constitution is not.Sums have mereological parts; ordinary objects, including persons, do not. Composition is a(necessary) relation that is exemplified by a sum and its mereological parts; constitution is a(contingent) relation that is exemplified by things of different primary kinds. Constitution explainssameness of parts at t of x and y when x constitutes y at t; hence, constitution should not be defined interms of sameness of parts

14 On my view, constitution is transitive.15 Like E. J. Lowe, I take us to be mereological atoms; but, unlike Lowe, I take us to have ordinary parts.

Mereology has a different role in my view from the one it has in Lowe’s view. See Lowe (2001).16 Merricks (1998b, pp. 106–24) argues that there are no criteria of identity over time at all.


means of natural processes and some by means of human ingenuity. Theoriginal property base at the Big Bang is not itself sufficient for all the lateremergent items, and hence the emergents are not reducible to the originalbase. A full account of reality would include the original base together withthe emergent items. Likely candidates for being emergents are self-replicatingmolecules, qualia and persons. If the first-person perspective is an emergentproperty, then it cannot be understood in non-personal or sub-personalterms, and it belongs in a full account of reality.It is not surprising, then, if I am right, that there are no informative

criteria for diachronic personal identity. To be informative, the criterion of“having the same first-person perspective” at different times would have tobe specifiable without using the idea of a person. There seems to be noprospect of doing that.17

However, rather than being embarrassed by the absence of informativepersistence conditions for persons, we should insist on it. The only alter-native is to construe our diachronic personal identity in non-personal terms.So, if you agree with me that we are irreducibly persons, then you too willwant to avoid informative persistence conditions.

ob j e c t i on s and r e p l i e s

obj. 1: If persons are not identical to animals, how can they be part of thenatural order?

reply 1: Biologically speaking, I am a Darwinian: I believe that there is importantcontinuity between the most primitive organisms and us, that we have animalnatures, and that biology can uncover all there is to know about our animalnatures. When our biological ancestors developed robust first-person perspec-tives (along with grammatically complex first-person sentences), entities of anew kind – persons – came into being.18My speculation is that grammaticallycomplex first-person language and human persons came into existencetogether – both in the course of what anthropologists call the period of“cognitive inflation” of human organisms (Mithen 2004, p. 164). So, we cancertainly imagine how we – with our ability to conceive of ourselves in thefirst-person and hence to enjoy inner lives – could have come about in thenatural course of things.

obj. 2: Sure, persons have abilities not shared by other kinds of beings, but whysuppose that the difference between persons and other things is ontological?

17 See the detailed discussion of Thomas Metzinger’s view in Baker (2007b).18 E.g. “I wish that I had more food” or “I believe that I am getting sick” are grammatically complex

sentences that indicate a robust first-person perspective.


reply 2: Biology can teach us a lot about ourselves, but not everything. I do notbelieve that biological knowledge suffices for understanding our nature, allthings considered. Even if we are the products of natural selection, ontologicalemergentism opens logical space for there to be more to us than our animalnatures.

For one thing, we can exercise our first-person perspectives to do thingsthat conflict with our animal natures. As Steven Pinker, a well-knownevolutionary psychologist, writes, “A Darwinian would say that ultimatelyorganisms have only two [goals]: to survive and to reproduce” (Pinker 1977,p. 541). But he also points out that he himself is “voluntarily childless,” andcomments, “I am happy to be that way, and if my genes don’t like it, theycan go jump in the lake” (Pinker 1977, p. 52). These remarks suggest thatPinker has a first-person notion of himself as something more than hisanimal nature as revealed by Darwinians.

Moreover, we persons have abilities that have no discernible antecedentsin any other species. Let me enumerate some of our features that, I submit,cannot be understood as simply extensions of features of non-humananimals (Hauser 2009). These uniquely human features are emergentfrom features that we share with non-human animals:* We share with other species the ability to communicate with conspecifics,

but only we human persons can have a fully articulated language withnecessity and possibility. Only we worry about the paradox of the heap.

* We share with other species the trait of having a perspective on ourenvironments, but only we human persons can have rich inner lives,filled with counterfactuals (“if only. . .”).

* We share with other species methods of rational enquiry (Where’s thebone? I saw it being buried over there yesterday. So, I’ll look over there),but only we human persons can deliberate about what to do and canattempt to rank preferences and goals, and try to resolve conflicts amongthem (and thus be rational agents).

* We share with other species activities like self-grooming, but only wehuman persons can have self-narratives.

* We share with other species the ability to make things that we need (e.g.nests), but only we human persons can make things that we do not need(e.g. SUVs).

* We share with other species the property of having social organization,but only we human persons have war crimes, international courts andhuman rights.


I believe that these are just a few of the uniquely human features that add upto an ontological difference between human beings and animals. All thesedifferences rest on our having robust first-person perspectives. Robust first-person perspectives bring with them a cascade of new kinds of abilities: we cansee ourselves as individuals and as members of communities; we can plan forour futures; we can hold ourselves responsible for what we do; we deceiveourselves; we can write memoirs; we try to reform; we have rich inner lives.And on and on. With respect to the range of what we can do (from trying tocontrol our destinies to daydreaming) and with respect to the moral signifi-cance of what we can do (from assessing our goals to confessing our sins), it isobvious that beings with robust first-person perspectives are unique.And the uniqueness is ontological. It follows from the theory of con-

stitution that if x constitutes y at t, then y is on a higher ontological levelthan x. When philosophers speak of levels, they usually mean levels ofdescription. I do not. I mean levels of reality (Baker 2007d). On theconstitution view, sums of molecules constitute people, but sums of mol-ecules are fundamentally different kinds of things from people. So, thedifference in level between molecule-talk and people-talk is not just adifference in level of description; it is a difference in what is being talkedabout. The idea of constitution makes sense of this ontological differencebetween molecules and people.

conc lu s i on

First, unlike other simple views, the not-so-simple simple view does notappeal to any immaterial substances in the natural world that can existindependently of bodies.Second, the not-so-simple simple view holds that, although persons are

essentially persons, the onset of human personhood is gradual and notinstantaneous.Third, the not-so-simple simple view holds that persons have ordinary parts

that are not the mereological parts of standard algebraic mereological theories.Fourth, and finally, the not-so-simple simple view holds that what

persons fundamentally are cannot be understood in exclusively third-personal terms. It is non-reductive. It does not pretend to understandpersons or their identity over time in non-personal or sub-personal terms.To be a person is to have a first-person perspective essentially, and a personcontinues to exist as long as her first-person perspective is exemplified.19

19 Thanks are due to Phillip Bricker and Sam Cowling for comments on a draft of this chapter.


chapter 1 1

Is “person” a sortal term?Christian Kanzian

i n t roductor y r emark s

The leading questions of my contribution are (a) whether “person” is a sortalterm in a technical sense, and (b) why this is relevant for the discussion ofpersonal identity, especially regarding the opposition between simplicityand complexity.

In the first parts of the chapter I discuss (standard) no answers: “person”is not sortal; then yes answers: “person” is a sortal in a strict and technicalsense, exactly like “car,” “sheep” or “Homo sapiens”; finally, views whichseem to be inconsistent concerning the semantic character of “person.”Then I try to present an alternative: “person” is a semantically unique term.This uniqueness consists in a kind of incompleteness, which I intend tospell out. I discuss this view in the context of a theory of personal identity.There I start with considerations about persons as a kind, and continue byfocusing on individual persons. In my center of interest lie specific personalindividual forms as the founding instances of personal identity. In con-clusion I borrow an argument from the philosophy of mind to strengthenmy thesis of personal individual forms as simple units.

At the beginning let me bring in a methodological remark on the generaltopic of our volume: diachronic personal identity, simple or complex? Totake something as simple means in general to regard the fact in question asnot subject to analysis and thus to a certain extent primitive. Conversely,complexes can be analyzed, at least as being composed of their constituents.It is such an analysis that offers the possibility of reduction: it might be thecase that the complex is nothing but the sum of its constituents. Somethingnon-analyzable does not fulfill this (necessary) condition for reducibility.Epistemologically considered, the non-analyzable cannot be explained, atleast not in an informative way. The simple is what it is.

Such statements of non-explainability sound suspect. The assertions ofsimplicity, unanalyzability and non-reducibility are often interpreted as

192

“not knowing what the fact in question really is” or “being unable to providea profound, ‘real scientific’ explanation”; in short: “being metaphysical orideological.” And thus we should stick to a complex view.My purpose is to express my conviction that such a conclusion against

simplicity is problematic. That can be made clear if we consider that everycomplexity is not complex in an absolute sense, but only in a relative one.Someone can take a complex view relative to diachronic personal identity.But the supporter of such a complex view concerning personal identity mustsay what personal identity consists in. And the explanation expressing suchconvictions must include references to some kind of:(a) basic entities like ontological “atoms”,(b) relations combining the atoms to complexes, and(c) a procedure for reconstructing personal unities from the bottom up

from (a) and (b) elements.This might be perfectly informative. But what about these relations them-selves? And – we must not forget – what about the atoms themselves? Evenif you try to give informative answers about the nature of the relations, as faras atoms are concerned, the end of analysis is near. Since Locke (see Locke1975 [1690], ii, xxvii, 3) we have known that atoms either are what they are,or they are not atoms at all. The same problems can be referred to process-ontological reconstructions of diachronic identity, and to psychologicalreductions of personal persistence in a Humean–Parfitean style. All ofthem might start with a complex view, but end as supporters of simple,non-analyzable units, whatever they are.An absolute complex view is impossible.This is why I think that the supposed epistemological advantage of

complex views is a relative matter. In other words: simplicity is unavoidable(cf. Benovski 2010). Every view has its unexplainable starting points – its“metaphysics,” if you like.1 This does not decide the complex–simplediscussion on personal identity, but deprives the followers of a complexview of one line of argumentation and takes away the burden of proof fromthe supporters of the simple view.However, the aim of this chapter is not to deal with these fundamental

questions concerning the complex–simple debate. My starting point is thesemantics of “person,” the question whether it is a sortal term, and how we

1 Even David Lewis seems to accept this point as a methodological unavoidable angle: “A system thattakes certain . . . facts as primitive, as unanalyzed, cannot be accused of failing to make a place forthem” (Lewis 1997, p. 198).

Is “person” a sortal term? 193

can understand the relevance of the question for an ontology of person-hood and personal identity.

“ p e r son ” i s not a sort a l t e rm

Let me come to the “no answer” position concerning the status of “person”as a sortal term. How can we understand such a position?

Before we seek an answer, we must give at least a provisional account ofwhat a sortal term, technically considered, is. Without being able to discussalternatives at this occasion, I suggest defining sortal terms such that they:(a) denote those (and only those) species which are called species infimae;(b) provide a maximally precise apparatus for the identification of individuals,

not only synchronic, but also diachronic, which means throughout theirwhole existence (since species infimae determine the identity of individualsat each moment, but also their persistence, in a maximally specific way);

(c) are (necessarily and sufficiently) connected with principles for countingand enumerating.

Condition (b) excludes terms denoting accidents (“being tall”), as well asso-called “phase sortals” (“caterpillar,” “juvenile”); condition (a) excludesterms for higher species, genera, and categories (“mammal,” “livingbeing,” “substance”), as well as for classes within the species infima(“woman”); and condition (c) also excludes mass terms (“gold,” “orangejuice”). Standard no positions urge that “person” lacks one of thesesemantic characteristics. The most promising way to argue in favor ofsuch a no position is to stick to a (b) procedure: “person” has either anaccidental character (radical no answers) or can be classified as phase sortal(weak no answers). The latter has the advantage of allowing one to link“person” with a remarkable identification apparatus, at least for thesynchronic identification of some entities.

The first (to my knowledge) to whom such an “accidental” understand-ing was imputed is Gilbert of Poitiers (1080–1155), who primarily intendedto oppose Boethius’s substance-view and thus suggested taking personhoodas something which is added (Lat.: extrinsecus affixa predicamenta) to thenatural constitution of human beings (Gilbert of Poitiers 1966). TheCouncil of Reims (1148) took him to task for this account, on the groundsthat it is committed to an interpretation of personhood as accidental. Icannot discuss whether Gilbert really wished to become the ancestor of all(more or less radical) no answers to our question on the sortal character of“person,” but I think he actually is.

194 christian kanzian

However, systematically considered, I think that the above-mentionedno answers are completely untenable. Weak no answers offer a sufficientargument against the radical ones: their identification function is essentialfor “person,” whereas terms for accidents have no identification function.Thus “person” is no accidental general term.Weak positions holding that “person” is a phase sortal term like

“juvenile,” “caterpillar” and so forth, assert the identification functionof the term in question, but actually restrict it conceptually or a priori to aperiod of time. This seems problematic, since standard uses of “person”do not provide such a priori restrictions;2 thus I think that “person”cannot be a phase sortal.To concede that “person” is a general term whose identification function

is not conceptually restricted to a specific period of time means we mustadmit that “person” is not a phase sortal at all. The supporter of such a thesisactually gives up the phase sortal theory, and in consequence the no answerposition.The problems with standard no answers (accidental, as well as phase

sortal accounts) are the best reasons to explore a yes answer position:“Person” should be accepted as a general term standing for a species infima,providing a complete and maximally precise apparatus for the identificationof individuals, and as connected with principles for counting andenumerating.

“ p e r son” i s a so r t a l t e rm

Boethius (480–526) can be considered to be the ancestor of this position.Boethius does not speak about general terms in a manner which can becompared with our dealing with sortal terms. Nevertheless, his effort tocombine personhood with substantiality establishes persons as a substance-kind, precisely to avoid the accidental character of persons; and, primarily,to explicate the Christological dogma of Chalzedon: one substance – oneperson and two natures within his categorical framework (Boethius 1988).And precisely this ontological view seems to go hand in hand with the thesisof “person” as a sortal term in the technical sense introduced above. Theontological and the semantic thesis are committed to each another.3

2 Another argument could be that the phase sortal conception concerning “person” would make“person” useless for theological contexts a priori. But I think it is also problematic to restrict humanpersonhood to a specific period, for conceptual reasons.

3 Here I can neglect the questions whether there are also sortals for the category of events and for thegenus of non-substantial things.


Although the substance-character of persons sounds harmless, I fear thatit is not. On the contrary, it provides one of the most significant problems ofthe yes view: “person” is a sortal. Were persons substances in their ownright, they would be things “like other things in the closed cosmos ofsubstances” (Lutz-Bachmann 1983, p. 66; original language of the quotationis German, translated by author). Persons would exist just as amebas andsheep do. I have no problems with amebas and sheep – they occur as such,but persons do not. There are no persons simpliciter. What obviously occurin everyday life are human persons, organisms of the species Homo sapiens,who we like to regard as persons. If persons as such were substances, wewould have to choose between neglecting the substantial character oforganisms (in which case we could defend the view that human personsare one and only one substance: persons) or conceding that we, humanpersons, consist of two substances.4 I think the first alternative leads veryquickly to a dead end, since the denial of the substancehood of organismsopposes all reasonable concepts of substance. The latter alternative would bea rather strong version of dualism. There is no point in rehearsing theproblems of substance dualism.

Before I start to ask for a third way between standard no positionsconcerning the sortal character of “person” and the yes view, I want todraw attention to another possible position, which we can call the “don’tcare” view.

Don’t care views can be accused of inconsistency concerning the logic of“person,” which does not bother them and is what provides reason forcalling them “don’t care-ists.”

the don ’ t c ar e v i ew

A paradigmatic “don’t care-ist” is Peter Singer in his Practical Ethics (Singer1978). On the one hand, he is committed to the accidental character ofpersonhood: according to Singer every (living) being that is able to gain the

4 A third alternative could be to take persons as being (nothing but) organisms of a specific kind. I amnot sure if this is compatible with a yes position: “person” is a sortal term in its own right. I ratherwould understand such a view to be either reductionistic or an accidental view. Obviouslycompatible with a yes position is a “constitutional view” on persons, as Lynne Rudder Bakerpresented it in her Persons and Bodies (Baker 2000) and in her The Metaphysics of Everyday Life(Baker 2007d). In the latter she explicitly states: “Person – like statue – is a primary kind, one ofmany irreducible ontological kinds” (Baker 2007d, p. 67). I concede that Baker’s view does notcommit to dualism. But I think that the premises of her constitutional view are themselves soproblematic that her theory cannot be regarded as a really successful yes theory. I ask for under-standing for not being able to go into details here.


status of person can lose it (which happens, say, when it is deprived of theintellectual capacity which is decisive for this status) and perhaps – underlucky circumstances – regain it. This is typical for an accidental account ofpersonhood or a (radical) non-sortal view on “person.”My interpretation issupported by examples given by Singer to point out the irrelevance ofpossible personhood for the (ethical) status of its bearer. For example: justas Prince Charles has the potential to become the king of England, but is notthe king of England now, so a being having the potential to become a personis not a person now, and thus need not be treated as such (Singer 1978,p. 120). Without discussing the example in detail, it seems clear that itsexplanatory power depends on the assimilation of “person” to “king,” and“king” is obviously a general term in the category of accidental properties.On the other hand we find passages in Singer’s Practical Ethics in which

he seems committed to the substantial character of personhood. A person ischaracterized as a being with the preference to exist in future as the sameperson (Singer 1978, p. 78). This presupposes that a person must be able toexist (as the same person) through time, and to identify herself as the sameperson diachronically. The latter would be impossible if “person” were notcombined with a sortal-like apparatus for identification. “Person” must beconsidered as a sortal, to make Singer´s characterization of persons sound.And sortals stand for substance species.5

We can conclude that Singer’s position is inconsistent with respect to thesemantic role of his concept “person.”

“ p e r son ” i s a s emant i c a l l y un i qu e t e rm

I am sure that we can criticize Singer’s theses in his Practical Ethics by asemantic analysis of his use of the concept “person.” But I do not intend tolead a debate on Singer here. I would prefer to return to my main point andask for an alternative between standard no positions and the simple yesanswer to the question of the sortal character of “person.”How, on the onehand, can we avoid the shortcomings of the no positions, yet on the otherhand also keep away from the problems of the yes position?The logical nature of “person” seems to require that we use it to identify

something, or rather someone: we pick out units from the manifold empiricalimpressions by marking them as persons. We mean to have reasons for

5 Since persons have the preference to exist in future as the same without conceptual or a priori restrictionto a specific period of time, “person” must really be considered as a sortal; the phase sortal account ofpersonhood cannot protect Singer from inconsistency.


assuming that the units are real units, not only at the actual moment ofidentification, but also through time. These reasons are just that the unitsfall under the concept “person.” Furthermore, we use “person” to countthese units, which we are able to identify and re-identify by applying theterm. What could be wrong with asserting that there are thirty-two personsin a room? Nothing, I concede, but we should nevertheless be careful not tojump to the conclusion that “person” is a sortal term in a strict and technicalsense. In order to make my caution plausible, let us examine in detail theidentification function of terms in general, and of our term in particular.

What in general is the reason why we can use an F-term for theidentification of x’s? It is because we have knowledge of some criteriawhich we connect with the F-term and which provide reasons for assumingthat something is an F-unit, both at one particular moment and throughtime.6 For example: why are we able to identify a living being with “sheep”?Because “sheep” provides clear and informative criteria for the decision whythe tail, for instance, belongs to the sheep but that the grass on which it islying does not; why there are two, and not only one sheep in the pen, andwhy we should accept that it is the same sheep there yesterday and today:maybe it is the sum of specific, sheep-like life functions, in which the partsof the organisms in question are integrated.

What about “person”? Is “person” connected with informative criteria forthe identification of units falling under it, like “sheep”? I do not think so.7

“Person” is not connected with corporal criteria for the identification ofbeings. This becomes especially clear if we consider reidentification: i.e.identification through time. We can change corporal parts without ceasingto be one and the same person. Nor are biological criteria satisfactory.Without being able to go into the details here, I want to recall JohnLocke’s famous thought experiment of the body change of two persons(Locke 1975 [1690], ii, xxvii, 27, 15), which can strengthen our anti-biologistic intuitions. Psychological criteria are problematic, because everyperson can change her psychological character radically. Moreover, it is hardto see which psychological criteria we should employ and which we shouldnot. Is there an isolated psychological criterion? I do not think so.

That neither physics, biology nor psychology can solve the problem ofidentification criteria for “person” is due to the impossibility of translating

6 I am of the opinion that we cannot identify something without taking it to be an F-unit.Without sort,there is no unit. But this is not at stake here.

7 I borrow this view from Lowe (1989b, ch. 7.3 and 2003). Other parts of my theory (the incompletenessof “person”) distinguish me from him.


“person” into a pure physical or physiological, biological or psychologicalconcept. If “person” were, for instance, a biological concept, biologywould be able to teach us sufficiently what units persons are – but this isnot the case.But, because we have no empirical criteria connected with “person,”

I think we must give up the idea of informative criteria for identificationconnected with this general term. We cannot explain why we identify unitswith “person.”Nevertheless we do actually identify units with “person,” and we have the

strong intuition that there should be reasons why we do so. How can weresolve this seemingly paradoxical situation? I think the answer is that we (canonly) use “person” for the identification of units, because we combine“person” with another general term, which indeed provides us with clearand informative criteria for the identification of entities. One of the bestcandidates for such a term is “human being” or “Homo sapiens,” which standsas a general term for a biological species.8 “Human being” is a sortal term in astrict and technical sense. It provides an apparatus for the identification ofindividuals, not only synchronic, but also diachronic, throughout their wholeexistence, and is connected with principles for counting and enumerating.Thus, what we do when we identify and count persons is to identify and tocount human persons. This sounds unspectacular, but it is not, semanticallyconsidered: it makes clear that “person” is an incomplete general term, withrespect to its identification function.Does that mean that we fall back into one of the standard no answer

positions, taking “person” as an accidental general term or as a phase sortal?No, because we stick to the identification capacity of “person,” which is notconceptually restricted to a specific period, but add that “person” per se isnot connected with informative criteria for the identification of anything.“Person” depends on the identification apparatus of another general term,of a real sortal term. If we take the informativeness of such criteria asnecessary for a term to be a count term, it becomes clear that “person” assuch cannot fulfill condition (c) for sortal terms, which also distinguishesour theory from yes positions.I want to conclude that “person” should be understood as a general term

with a unique semantic status: it is an incomplete term, which means that itdepends on other terms to yield a function to which its “logical nature” isdedicated: to identify something without conceptual or a priori temporal

8 NB: I do not pretend that there are logical reasons why “human being” should be the only candidate.


restrictions. All wrong accounts in the semantics of “person” start byneglecting this unique status and by putting “person” with other termsinto the same logical category of general terms.

In the next section I shall apply the semantic idea of the incompletenessof “person” to the kind of persons, and finally to an analysis of the“individual form” of persons.

the k i nd o f p e r son s

The first step can be taken with an enquiry into whether the kind person is aspecies infima in the sense introduced above: a species that determines theidentity of the individuals belonging to it in a very specific way, at eachmoment, but also throughout the whole time of their existence.

If my considerations concerning the term “person” are adequate, theanswer must be “no,” if only for conceptual reasons: I defined species infimaeas exactly those kinds for which sortal terms stand; “person” is not a sortal;thus person cannot be a species infima.

I hold this to be true, but nevertheless, I confess it sounds tricky, andneeds further argumentation. To meet this demand, let me explain in moredetail just what the determination of the identity of an individual by its kindconsists in (i), whether the kind person can provide such a determiningfunction or not (ii), and (iii) what we can conclude from (i) and (ii)concerning the ontological status of the kind person.

(i) The first aspect of the sortal determination of an entity is basic. I admitthat it is easier to grasp if you are guided by some Aristotelian premises inyour ontology: according to the Aristotelian tradition no entity (especiallyno substance, and we may restrict our considerations to this category here)can exist without a specific individual form, without a means by which (a“how”) its individual material is construed to make up the entity in ques-tion. No specific individual form can be imagined without being specifiedby a species infima; thus no entity, especially not in the category of sub-stances, can exist without a species infima.

Beneath this basic or existential aspect, we can state more concreteaspects of the determination of an individual by its kind: so entities aredetermined by their species infima in their continuity, spatially as well astemporally. That Susan is a sheep determines her spatial extension and thepeculiar characteristics of it. The same holds for her temporal career: herpersistence conditions are determined by her sheephood: general ones (e.g.that her diachronic identity cannot be interrupted) as well as special ones,


pertaining to the specific principles by which sheep-like organisms develop(Hirsch 1982, pp. 37–8; Oderberg 1993, pp. 22–3). Another concrete aspectof sortal determination is mereological and material composition: whichparts normally or essentially belong to a unity, of which matter these partscan or must be composed. All that is determined by the kind to which theentity belongs.(ii) However, the decisive question for our topic is whether the kind

person can provide such a full identity-determining function. Attemptingto answer this question lands us again in a paradoxical situation. There arestrong intuitions, as well as ontological reasons, supporting the view thatthe identity of a person is determined by her being a person. If we take theontological status of personhood seriously, we cannot give up the idea thatit has a determining influence on the identity of those individuals thatbelong to it. Personhood is not accidental. The ontological nature of thekind person seems to determine the identity of “its” individuals. But,on the other hand (and this yields the paradoxical situation), we cannotexplain how the kind person determines the identity of persons, as it isdemanded of a species infima. The first reason may be that, to determinethe identity of an entity in a maximally specific way, we need informativeconditions for the identity of that entity. But the kind person does notcontain any informative conditions for how the entity’s spatial and tem-poral career or its mereological and material composition are determined.For the kind “person” is not connected with material or biological fea-tures, or with other candidates for informative conditions. The argumentfor this can be borrowed from our considerations concerning the generalconcept “person.”(iii) The only way out of this paradoxical situation, as far as I can see, is to

assume that the kind person depends on other kinds, such asHomo sapiens, toget its “natural” identity-determining function. The kind person is depend-ent, and thus incomplete, and moreover unique in this incompleteness. Sinceonly complete kinds can be classified as species infimae, the kind person cannotbe one. This view is supported by the assumption that there are no entitiesthat are purely and simply persons (which is a rejection of extreme dualism orof spiritism). The identities of entities that are persons are not completely (butincompletely, which means in dependence on another kind) determined bytheir personhood. What we do when we provide conditions for the existence,continuity and composition of persons is to appeal to those conditions thatare connected with the species infima to which persons belong: e.g. humans.We determine the identity of persons by determining their identity as humanpersons.


This point of view is not dualistic, but is not its price reductionismconcerning the kind person? No: that a kind F depends in a central way on akind G does not imply that F is reducible to G, since there might remainfunctions at F that are not covered byG. Actually my argumentation (to sayit cautiously) excludes neither such functions nor their possible relevancefor an ontology of human persons.

Does this approach amount to “sortal relativity” concerning the identityof human persons? No: sortal relativity means that the identity of an entityis determined by two ontologically equivalent kinds or sorts, which is notthe case with the kinds of persons and of humans determining the identityof human persons.

Although there might be other remarks worth considering, I want tocome to my next point, which is an analysis of individual persons, especiallythe individual form of persons.

the i nd i v i dua l form of p e r son s

My result will be that the personal individual form is a simple and incom-plete unit, which depends in its function as form on other individual forms.But let me go step be step. From an explanation of what individual forms ingeneral are and how we can grasp their ontological functions, I will proceedto a characterization of personal individual forms in particular, and argue forthe above-mentioned thesis.

So, what is an individual form? Recall from the previous section that wecan understand the form of a substance as the how aspect of its ontologicalconstitution. The need for such a how aspect becomes clear when weconsider that there is no thing (in a technical sense) that is not formedmatter. (Unstructured material parcels are not things in this sense. The firstmay be the material aspect of the latter.) And the individual form is theconcrete formation of the material composing a thing. The concrete way inwhich Susan’s organism is arranged, structured and related to her environ-ment, its possibilities of developing and reproducing, is her individual form.That the individual form of a substance is determined by its kind was alsomentioned in the previous section. The concrete arrangement of Susan’sorganism is given by her being a sheep.

Not explicitly mentioned before, but easy to deduce from the assumptionsmade so far, is that the individual form is the instance of a substance whichfounds that substance’s identity. Susan’s identity is not founded, neithersynchronically nor diachronically considered, on a specific material part oron the sum of her parts, but on the life functions which are given by her


individual form.9 That her tail belongs to her but that the grass on which sheis standing does not is granted, because her tail, in contrast to the grass, isintegrated into her life functions. That she is the same sheep as she wasyesterday is not founded on the sameness of her cells, but on the continuity oflife, for which her individual form is the ontological principle, so to speak.An individual form is the identity-founding how aspect of a substance,

which is determined by the kind to which the substance belongs.How can we apply these general features to the specific personal indi-

vidual form? What characteristics must an individual form have, deter-mined by a kind, which is incomplete and thus dependent on other kindsin ontologically significant respects, and which is not connected withinformative conditions for the existence, continuity and composition ofindividuals belonging to its scope? I think the only possible answer is thatthese individual forms must also not be informatively explainable andanalyzable, i.e. simple in their ontological constitution. The same holdsfor personal identity insofar as it is founded in a simple individual form: it isnot complex, which means that it is not analyzable into other relations, likecontinuity (for instance of life functions). And the personal individual formmust be considered as dependent on another form in its ontological nature.As the kind person cannot carry out the ontological functions which areessential to kinds without other kinds (real species infimae), the personalindividual form cannot fulfill the function of a form without other indi-vidual forms: for instance, an organic one. This becomes clear when weconsider that the (primary) ontological function of a form is the composi-tion of a whole substance, which means the integration of its material aspectinto a substantial unity. The integration of complex matter cannot be doneby a simple and primitive form, but only by a complex one, paradigmati-cally by an organic form.We can conclude that the personal individual form is a simple and

incomplete unit, unique in this incompleteness. Human persons own twoindividual forms, one organic and one personal. “Diachronic personal iden-tity: simple or complex?” is the leading question of our volume. It must besimple, if my considerations are sound; but dependent in a unique way onsomething which is not simple: for example, on human identity. It is incom-plete in nature, because of its essential dependence on something complex.10

9 The thesis that life is the principle of the identity of organisms finds support from Locke’s famousChapter xxvii in Locke (1975) up to Van Inwagen’s Material Beings (1990b).

10 For more details of this thesis see Kanzian (2009, part iii).


the un i t y - o f - con s c i ou sn e s s a rgument

Individual personal forms are simple and incomplete. I have tried to findsemantic and also ontological arguments for my thesis, which primarilyrefer to an analysis of kinds in general and of the kind person in particular.In the last section of my article I want to point out that we can also findadditional support for my assumptions from considerations in the philos-ophy of mind, especially of consciousness.

Before I come to that, let me remind the reader of my remarks at thebeginning. There I argued that the supposed epistemological advantage ofcomplex views is a relative matter. Simplicity is unavoidable. I concededthat this does not decide the complex–simple discussion on personalidentity, but takes away the burden of proof from the supporters of thesimple view. I think that is important, but does not release us from thenecessity of looking for further direct reasons for accepting the simplicity ofpersonal individual forms.

The so-called unity-of-consciousness argument seems to be such anadditional argumentative support.

I do not want to bring up the manifold classical versions of the argu-ment,11 but restrict myself to one contemporary version: that of WilliamHasker. I refer to his book The Emergent Self (Hasker 2001), not becauseI think that all of Hasker’s theses are adequate (“emergentism” itself seemsto be more an explanandum than a solution of any problem), but because inone of his preliminary chapters he makes the decisive point discussable inthe terminological context of the present simple–complex debate on per-sonal identity – which is the only issue here. I start with a quotation of apassage in Hasker’s book:

The point is simply that . . . awareness . . . is essentially unitary, and it makes nosense to suggest that it may be “parcelled out” to entities each of which does nothave the awareness. A person’s being aware of a complex fact cannot consist of parts ofthe person being aware of the fact. A conjunction of partial awareness does not add upto a total awareness. (Hasker 2001, p. 128)

How can we understand Hasker’s point? Hasker starts by considering avisual field. According to standard biologistic interpretations of such a stateof consciousness (as well as others like perceptions and the content ofthoughts in general), the subject or the bearer of it is something bodily:

11 Descartes, Sixth Meditation; but most importantly Kant, Critique of Pure Reason, A 652. Peter vanInwagen has an interesting modern version of the argument (1990b, p. 118).


the brain. The brain is aware. The brain perceives. The brain thinks. Thedecisive point for the present discussion is that something complex is aware,perceives, thinks. Hasker is arguing against such a point of view: let usimagine the smallest part of the brain (Hasker calls it “V”), which containsthe modeling of all the information, for example, of the visual field. Is itpossible to say that this smallest part is aware of the complex visual field?

Perhaps so, but we need to consider the composition of V . . . we can say that V is awhole composed of physical parts. Many of these parts model information fromvarious parts of the visual field. But no proper part of V models all of thisinformation, so it is not possible for any of these parts to be aware of the entirevisual field. But if V is a whole composed of parts each of which is not aware of thevisual field, how can V itself be aware of it? [Hasker illustrates his points:] . . . thiswould be like saying that each student in a class knows the answer to one questionon an examination, and that in virtue of this the entire class knows the materialperfectly! It is true that the members of the class are able, working together, toreproduce all of the information, but there may in fact be no one at all who knows oris aware of all of it. (Hasker 2001, p. 128)

According to Hasker we may assert that it is false to say that the brain or apart of it is aware of something, perceives and thinks. States of awareness(or, more clearly, states of consciousness) are essentially unitary. Somethingessentially unitary cannot inhere in something non-unitary or in somethingcomplex, as the brain is. Consciousness – I would restrict myself to sayinghuman consciousness –must be grounded in something non-complex: thatmeans in a simple unity.The result of Hasker’s analysis is that the bearer of our states of con-

sciousness must be a simple unanalyzable unit. If that is the case, which iswhat I actually assume, we have the alternative of taking whole substances assuch simple units, which would lead us to a rather strong version of dualism,or – and this is what I would suggest – of sticking to individual forms as thebearer of specific personal acts of consciousness. The unity-of-consciousnessargument as such does not commit to substance dualism. Thus, it can betaken as the supporter of my thesis of simple personal individual forms.Perhaps we can formulate this in a stronger way: if the unity-of-consciousness argument is sound, and we want to avoid substance dualism,we have to accept simple individual forms as the only possible bearer ofspecific personal intellectual capacities.I think this is good news for friends of the simple view, even if personal

identity remains – let us not forget – essentially incomplete.


chapter 1 2

Materialism, dualism, and “simple” theoriesof personal identity

Dean Zimmerman

“ comp l e x ” and “ s im p l e ” theor i e s o fp e r sona l i d ent i t y

Derek Parfit introduced “the complex view” and “the simple view” as namesfor contrasting theories about the nature of personal identity. He detects a“reductionist tradition,” typified by Hume and Locke, and continuing insuch twentieth-century philosophers as Grice, Ayer, Quinton, Mackie,John Perry, David Lewis and Parfit himself. According to the reductionists,“the fact of personal identity over time just consists in the holding of certainother facts. It consists in various kinds of psychological continuity, ofmemory, character, intention, and the like, which in turn rest upon bodilycontinuity.” The complex view comprises “[t]he central claims of thereductionist tradition” (Parfit 1982, p. 227).

The complex view about the nature of personal identity is a forerunner towhat he later calls “reductionism.” A reductionist is anyone who believes

(1) that the fact of a person’s identity over time just consists in the holding ofcertain more particular facts,and(2) that these facts can be described without either presupposing the identity ofthis person, or explicitly claiming that the experiences in this person’s life are hadby this person, or even explicitly claiming that this person exists. (Parfit 1984,p. 210)

Take the fact that someone remembers that she, herself, witnessed a certainevent at an earlier time. When described in those terms, it “presupposes” or“explicitly claims” that the same person is involved in both the episode ofwitnessing and of remembering. Purging the psychological facts of all thosethat immediately imply the cross-temporal identity of a person will leaveplenty of grist for the mills of psychological theories of persistence con-ditions. Although veridically remembering may be ruled out, closely relatedpsychological states – and causal connections between them – can be

206

included. As has frequently been noted, reductionists can replace remem-bering with a relation of “quasi-remembering” (Parfit 1984, pp. 220–2;Shoemaker 1984, pp. 82–6; see also Shoemaker 1970). Personal identity isnot explicitly invoked by describing the experiencing of a certain event by aperson at one time, and the occurrence of an apparentmemory of just suchan event by a person later on, which recollection is caused, indirectly, by theoriginal experience in virtue of the normal processes that go on in ourbrains – whatever those are, exactly. I follow Parfit in calling psychologicalfacts that can be so described “impersonal.” Questions loom about theidentity conditions for facts, and about just how fully an impersonaldescription must characterize a fact in order for that fact to qualify asdescribable without presupposing or explicitly making claims about per-sonal identity. But I shall not pursue these questions here.The complex view favored by reductionists is contrasted with the simple

view of an opposing “non-reductionist tradition.” According to non-reductionists, “personal identity does not just consist in these [psychologicaland physical] continuities, but is a quite separate ‘further fact’” (Parfit 1982,p. 227; see also Parfit 1984, p. 210). To be a non-reductionist, one must“reject either or both of ” (1) and (2) (Parfit 1984, p. 210), though theyarguably stand or fall together.1

According to Parfit, non-reductionists come in two varieties. Somemaintain that “[a] person is a separately existing entity, distinct from hisbrain and body, and his experiences,” such as “a Cartesian Pure Ego, orspiritual substance,” or perhaps “a separately existing physical entity, of akind that is not yet recognized in the theories of contemporary physics.”Other non-reductionists accept a “further fact view”: they hold that,“though we are not separately existing entities, personal identity is a furtherfact, which does not just consist in physical and/or psychological continu-ity” (Parfit 1984, p. 210).The non-reductionists include, he says, “the great majority of those

who think about the question” (Parfit 1982, p. 227); but this claim standsin need of qualification. He mentions “Butler, Reid, Chisholm, Geach,Swinburne” (Parfit 1982, p. 227), but one wonders who else he has inmind. I should have thought that the dominant philosophical views aboutpersons would all satisfy Parfit’s definition of reductionism – at least, solong as attention is confined to English-speaking philosophers publishing

1 (2) entails (1), since (2) presupposes the existence of the “more particular facts” in which personalidentity is said to consist, according to (1). And, as Marco Dees has pointed out to me, the relation ofpartially consisting in should be transitive and irreflexive, in which case (1) entails (2) as well.

Materialism, dualism, and “simple” theories 207

articles or books since, say, the 1930s. Very few of these philosophersbelieve in anything like “a Cartesian Pure Ego” or “spiritual substance.”Almost all are materialists, committed to the view that the identity ofpersons, over time, must be determined by facts about psychological orphysical continuities with no need to appeal to souls or any new kind ofphysical entities “not yet recognized in . . . contemporary physics.” Theymay disagree about whether biological or psychological continuities arethe basis for the persistence of persons, but they typically attempt toformulate persistence conditions in their favored terms that leave noroom for sameness of persons as a “further fact.”

A closer look at his definition of reductionism can make one wonderwhether anyone really qualifies as a non-reductionist. Would someone whobelieved persons are immaterial substances, or a new kind of physicalparticle, really deny reductionism, given Parfit’s official definition? Well,can an immaterial substance or a new kind of physical particle, andthe conditions under which they persist over time, be described without“presupposing” or “explicitly claiming” anything about the identity ofpersons? At least some proponents of views along these lines might think so.

Chisholm is one of Parfit’s paradigmatic non-reductionists, identified asa defender of the simple view. The only hypotheses about the nature ofhuman persons that Chisholm took seriously were these two: (i) that each ofus is identical with “something of a microscopic nature, and presumablysomething located within the brain” (Chisholm 1986, p. 75); and (ii) thateach of us is “a monad” – an immaterial, indivisible substance (Chisholm1986, pp. 73, 77). Parfit explicitly consigns both sorts of hypothesis to thecategory of non-reductionism. But it is far from clear that either one isinconsistent with (1) and (2).

By Chisholm’s lights, the proper way for me to determine the nature ofmy persistence conditions is for me first to try to figure out what kind ofthing actually has the psychological properties with which I am directlyacquainted. Is it a special kind of physical particle or bit of matter? ThenI am that particle or piece of matter, and I have the persistence conditionsappropriate to particles or pieces of that sort. Is the bearer of my psycho-logical properties an immaterial substance? Then I am an immaterial sub-stance, with whatever persistence conditions are appropriate to such things.2

If there are facts – even brute facts – about the identity over time of bits of

2 Chisholm detects the same approach to personal identity in Bayle’s remarks about the rational self-concern of a group of atoms that radically change their form – remarks he cites with approval: seeChisholm (1976, p. 113).

208 dean zimmerman

immaterial substance or of some special physical particles that can havepsychological properties, then the discovery that I am a simple immaterialsubstance or a special physical particle should lead me to conclude – withParfit’s reductionists – that my “identity over time just consists in theholding of more particular facts,” ones that can be described without“explicitly” making claims about identity of persons. To further probethis question, I turn to some theses about the relations between persistenceand immanent causation.Chisholm was convinced that, whether we turn out to be physical or

immaterial, we must be fundamental substances – not mere “modes” ofsomething else, or “entia successiva” constituted by different things at differ-ent times (Chisholm 1986, pp. 66–71; 1976, pp. 104–8). For fundamentalsubstances (ones that are not made out of, or dependent upon, some otherkind or kinds of thing), persistence through time plausibly requires the sortof causal dependence of later stages upon earlier stages that is often calledimmanent causation. Elsewhere, I offer a theory about what such causaldependence should look like. The theory is based upon the idea that theearlier intrinsic features of a thing must be causally relevant to its laterintrinsic features (though not in just any old way) (Zimmerman 1997).Some philosophers – call them “immanent-causal reductionists” – will

think that, for fundamental things at least, the facts about persistenceconsist in nothing more than the facts about immanent-causal connectionsbetween stages of the things – i.e. that facts of the first kind are completelydetermined by, or supervene upon, facts of the second kind. (As Shoemakerpoints out, one can talk of “stages” in this context while remaining neutralwith respect to the doctrine of temporal parts – though many immanent-causal reductionists will believe in temporal parts (Shoemaker 1984, p. 75).)The immanent-causal reductionist will have to say that, whenever there areforking paths of immanent-causally connected stages, one must give thesame verdict about whether the original thing persists through the apparentfission. There are choices about what the verdict should be: for example, afundamental thing ceases to be, replaced by two new ones; or there were twothings all along; or a thing becomes multiply located; or . . . But a uniformanswer to all cases of branching paths is required by the immanent-causalreductionist’s commitment to the idea that persistence, for these funda-mental things, is fully determined by the facts about paths of immanentcausation. (Similar choices must be made in the case of convergingimmanent-causal paths.)Others who think immanent-causal continuity is necessary for the per-

sistence of fundamental things may nevertheless suppose that, at least in


cases of branching and convergence (and probably elsewhere, too) there arefacts about the persistence of fundamental things that are not fully deter-mined by the facts about immanent-causal dependencies. (Chisholm him-self would likely have been in this camp.) In a case of branching paths, thereare (at least) three genuine possibilities: the original thing ceased to be, or itwent to the right while a new thing appeared on the left, or it went to the leftwhile a new thing appeared on the right. (Some will want to allow for afurther possibility: the thing became multiply located.) These philosophersposit a further fact about identity, at least for fundamental things – one thatis not fully determined by immanent-causal dependencies.

Now, where, within Parfit’s taxonomy, should one place an immanent-causal reductionist who believes, with Chisholm, that persons are a specialkind of fundamental entity – whether physical or immaterial (howeverexactly that distinction is to be made out)? Parfit clearly allows that onemay make use of psychological terms, when describing the facts in whichpersonal identity is supposed to consist, without violating clause (2) ofreductionism. So the immanent-causal reductionist could say: (a) a personis a thing of any fundamental kind that, at some time or another, haspsychological properties, or perhaps simply could have psychological prop-erties; and (b) a particular person’s persistence conditions just consist in theholding of immanent-causal relations between stages of the kind of funda-mental entity he or she (or it) happens to be. This philosopher seems prettyclearly to accept (1) and (2), and therefore to be a reductionist. But she alsoseems to satisfy Parfit’s example of a paradigmatic non-reductionist: she is abeliever in a Cartesian Ego or special physical substance, a “separatelyexisting entity.”

What should one say about a person who believes that we are funda-mentally immaterial or physical substances while denying immanent-causalreductionism, maintaining instead that there is a “further fact” about theidentity over time of such things? This person seems closer to rejectingParfit’s (1) (and thereby denying (2)). Still, suppose this philosopher believesin further facts about identity for all fundamental things, whether or notthey are capable of having psychological properties. She might very well bewilling to state very general necessary conditions for the persistence offundamental things, physical or not – namely, continuity of immanent-causal paths – but then she will say that necessary and sufficient conditionswill not be forthcoming, because there remains the further questionwhether the very same fundamental thing is present all along the path.She regards identity over time, for fundamental things, as a kind of “brutefact” above and beyond immanent-causal dependencies. But suppose she is

210 dean zimmerman

even-handed about the need to posit brute facts, not treating persons as inany way special within the category of fundamental things. She gives verygeneral, though not very informative, persistence conditions for fundamen-tal things of all kinds, ones that do not make use of terms that “presuppose”personal identity or “explicitly” invoke sameness of persons over time; andthe persistence conditions for persons are subsumed as a special case. Shouldshe really be said to deny (1) and (2)?Parfit writes as though no version of reductionism, in his sense, could be

consistent with our being “separately existing entities, apart from our brainsand bodies, and various interrelated physical and mental events” (Parfit1984, p. 216). It is not clear to me, however, that Parfit’s arguments againstnon-reductionism must count against views like Chisholm’s, if these theo-ries about personal identity satisfy (1) and (2) – as they would whenconjoined with immanent-causal reductionism, for example. The persis-tence conditions assigned to a Cartesian Ego or a “separately existing” newphysical particle by an immanent-causal reductionist may well consist incausal continuities that could, in principle, divide into equally viablestreams. If these circumstances require a “no-branching” clause amongthe facts in which their persistence consists, the resulting views wouldconfront the strongest of Parfit’s arguments for the conclusion that personalidentity is “not what matters” (Parfit 1984, pp. 261–6). They may well be noless (and, for that matter, no more) vulnerable to Parfit’s line of criticismthan the more usual forms of reductionism that he considers.Be that as it may, Parfit’s more informal characterizations of reduc-

tionism go beyond the idea that personal identity over time does notdepend upon further “impersonal” facts; he insists that the facts uponwhich identity over time depends do not include any “spooky” extra factsabout souls or special kinds of matter, but only facts about the mundanestuff our bodies have in common with all sorts of physical objects. Thefirst part of Parfit’s definition of reductionism seems effectively to bereplaced by:

(1*) that the fact of a person’s identity over time just consists in facts aboutordinary material stuff – the kinds of matter our bodies share withnon-living, non-sentient things.

I shall understand reductionism, or the complex view, as the combination of(1*) with Parfit’s (2) as a further condition upon the nature of the facts inwhich personal identity consists – it is needed to rule out a kind of furtherfact view to be discussed below.


I now articulate a reasonably precise thesis that should follow fromreductionism: namely, the supervenience of personal identity upon micro-physical facts, or upon microphysical facts supplemented only by “imper-sonal” psychological facts.

two doctr i n e s o f the su p e r v en i enc e o f p e r sona li d ent i t y u pon the imp e r sona l

Philosophers are more reluctant than they used to be to propose thatsomething is “analytic.” Even those who accept an analytic–synthetic dis-tinction will often admit that examples of truly compelling conceptualanalyses of interesting notions are thin on the ground. But most of us goon giving philosophical theories (of knowledge, right and wrong, objectivesimilarity, causation, etc.), and Parfit is no exception. Although we mightnot attribute analyticity to the core tenets of our theories, we do generallytake ourselves to be searching for necessary truths that, in some sense,explain the phenomena with which we are concerned. The ethicists(many of them, at any rate) want to know what makes something right orwrong, good or bad – what do these normative properties consist in; invirtue of what is an act right or wrong, or a state of affairs good or bad? Thesearch for the grounds for theoretically interesting truths continues in allfields of philosophy, though less frequently under the banner of “conceptualanalysis.” Parfit prefers to talk of one fact “just consisting in” other facts.This terminology is, I take it, in the same line of work as talk of the“grounds” of a certain phenomenon; or of what it is “in virtue of which”something is the case.3

When a fact is grounded in facts too complicated for us to fully grasp, wemay not be able to offer anything like necessary and sufficient conditions forthe obtaining of that sort of fact in terms of (what we take to be) its ultimategrounds. We may nevertheless manage to indicate the kind of facts uponwhich we think it must depend. And where there are several rival philo-sophical theories about some phenomenon, they may nonetheless agreeabout the kind of facts that, in one way or another, ought to providethe grounds for the phenomenon. Parfit’s reductionism (or, equivalently,the complex view) is a claim of this sort: whatever the truth is about theconditions that are necessary and sufficient for a person to exist at more thanone time, they must ultimately be grounded in facts about ordinary material

3 For discussion of various ways to understand a metaphysical “grounding” or “in virtue of” relation, seeFine (1994, 2001) and Schaffer (2009).

212 dean zimmerman

stuff, and facts that can be described in “impersonal” ways – ways that donot immediately involve the concept of sameness of person.When a philosopher claims that facts about the holding of a certain

relation just consist in facts about the holding of some other relations, he orshe is committed to a “supervenience thesis”: the pattern in which theformer relation is exemplified cannot “float free from” the patterns ofexemplification of the latter relations; there could not be two worlds justalike with respect to the latter but differing with respect to the former. “Justconsisting in” does not, itself, just consist in supervenience. But it entails it.Reductionism, with (1*) in place of (1), entails the supervenience of personalidentity over time upon the exemplification of certain patterns of propertiesand relations among bits of ordinary matter. I shall take a moment to spellout two supervenience theses in the neighborhood, one of which will appealto the physicalistically inclined reductionist; the other is the kind of super-venience that falls most directly out of reductionism.I begin with some definitions and simplifying assumptions.

Reductionism includes the claim that no “separately existing entity,” inparticular no kind of matter “not yet recognized in the theories of contem-porary physics,” need be introduced in order to have adequate groundingfor facts about persons. New kinds of particles are predicted and discoveredall the time, so there is no point trying to make this clause more precise bymeans of a list. But the atoms in our bodies are made out of particles thatcan be rearranged to make non-sentient non-persons; and they are subjectto a limited number of forces – gravity, electromagnetism, strong and weaknuclear forces – and these forces obey laws that, so far as we know, take nointerest in whether the universe contains sentient beings. One kind ofreductionism would be the claim that personal identity over time is fullydetermined by or grounded in the spatio-temporal relations among theparticles of these types (and particles of other kinds, so long as they do notoccur only in persons), plus the interactions among these particles that aregoverned by whatever impersonal forces show up in “final physics.” Thiswould be a kind of physicalist reductionism, a view that takes a physicalistgamble: that the introduction of consciousness will not require new laws ornew forces, let alone new kinds of matter or immaterial souls.To state a reductionistic claim about the identity over time of particular

persons, rather than simply affirming a global kind of physicalism, aphysicalist reductionist of Parfit’s stripe must say: the fact that some personpersisted from one time to another just consists in facts about sets of thesekinds of particles existing throughout that period, and their being arrangedin a certain way and interacting by means of these kinds of fundamentally


impersonal forces. It is a thesis about the conditions under which certainbatches of particles located at various times do constitute a single personexisting at all of those times, and the conditions under which they do not.The thesis entails a kind of determination of identity by the microphysical:when you have various collections of particles constituting the body of asingle person throughout a period of time, there must be impersonalphysical phenomena going on involving those microphysical particles andtheir surroundings which are sufficient to insure that a single person existsduring the period.

A canny physicalist reductionist will not want to say that whether someparticles go together to make up a certain person at a time is fully deter-mined by their intrinsic interrelations alone. Compare a pair of twins,Toothy and Gappy, who are absolutely identical except for the fact thatone of them has had the visible parts of all his front teeth removed. (I leavethe roots of the teeth intact, so Gappy’s gums are not behaving differentlyfrom those of Toothy.) The particles making up all of Toothy except for thevisible parts of his front teeth may be exactly the same, in their intrinsicarrangement, as the particles making up all of Gappy’s body. But, inToothy’s case, they do not constitute a human body, only a part of one –whether they make up a whole human body depends upon what relationshold between them and other particles, like the ones nearby in Toothy’sfront teeth. So, in looking for the underlying physical facts upon whichpersonal identity might depend, one must pay attention not only torelations among the members of a group of particles when asking whetherthey constitute the same person as someone existing at some other time; onemust also pay attention to whether they are embedded in a larger group ofparticles that could constitute that person instead.

I shall help myself to the notion of a certain kind of property ofcollections of particles: a “complete physical specification” of the “imper-sonal particles” in our universe – that is, kinds of particles that can existwhether or not there are any sentient beings. Two collections of particlesshare a complete physical specification if and only if there is an isomorphismbetween them that preserves all the instantaneous facts about: (i) spatialrelations among the members, (ii) all the interactions among them involv-ing basic (impersonal) forces, and also (iii) the kind of physical environmentin which they find themselves – for example, proximity to this or that kindof (impersonal) particle, or the presence of basic forces impinging uponthem from things not in the collection. The particles making up Gappy’sbody do not share a complete physical specification with the set of particlesmaking up Toothy-minus-front-teeth. Although there is an isomorphism

214 dean zimmerman

that preserves (i) and (ii), it does not preserve (iii).4 With this notion inhand, it is fairly easy to spell out a supervenience thesis that ought to be partof a physicalist reductionism about personal identity over time for creatureswith bodies made of our kinds of matter.

(D1) M is a micro-state =dfM is a pair consisting of a set of particles and acomplete physical specification.

(D2) Themicro-stateM occurs =df the set of particles inM exemplifies thephysical specification in M.

(D3) H is a micro-history =df H is a series of micro-states, each one ofwhich is paired with a different positive real number or zero.

(D4) The micro-historyH occurs =df for every pair of micro-statesM andM* in H, if H assigns n to M and n + m to M*, then M occurs mseconds before M* occurs.

(D5) A micro-history H is a continuous part of the history of a singleperson =df (a) H occurs, (b) H’s micro-states form a continuousseries ordered by their numbers, and (c) there is a person who, forevery time t at which a micro-state inH occurs, is wholly constitutedat t by the set of particles in that micro-state.

(D6) micro-historiesH andH* are indiscernible =df (a)H assigns a micro-state to some n if and only if H* does, (b) the micro-states which Hand H* associate with any n have the same complete physicalspecification, and (c) if there is a particle x that plays a certain role inthe micro-states that H associates with n and n*, then there is aparticle y that plays that same role in the micro-states H* associateswith n and n*.

Here is a first pass at a statement of the supervenience of personal identityover time upon the microphysical:

(PR) If a micro-history is a continuous part of the history of a singleperson in one possible world, then any indiscernible micro-historyoccurring in any possible world is a continuous part of the history ofa single person.

Not every materialist will accept (PR); I know that, were I a materialist,I should ask for at least two modifications.

4 See Zimmerman (1998) for a more detailed presentation of the kind of supervenience thesis sketchedhere.


If mental states do not supervene upon the physical, then one might wellthink that (PR) is false – that personal identity over time does not superveneupon the physical either. But one might still accept Parfit’s more generalreductionist thesis: personal identity over time supervenes upon the phys-ical plus the “impersonal” (i.e. personal-identity-neutral) mental states(quasi-memory, etc.) that could be associated with the history of a singleperson at a particular time. To add these to the supervenience base, eachstage of a history must include facts about whether, in the vicinity of theparticles, there is a thinker with impersonal mental states of certain kinds.The specifications of mental states should, like the physical states of theparticles, be complete specifications – ways that a mind could be in toto, atleast with respect to the most fundamental, non-supervenient, impersonalmental properties.

For this non-physicalist reductionist, a modified supervenience thesis canbe constructed:

(D7) M is a micro-psycho-state =df M is a triple consisting of a set ofparticles, a complete physical specification, and a completeimpersonal psychological specification.

(D8) The micro-psycho-state M occurs =df the set of particles in Mexemplifies the complete physical specification in M, and alsoconstitutes the body of a person who exemplifies the completepsychological specification.

Modifying the rest of the definitions is trivial, and allows for the formula-tion of a more general supervenience thesis (hence the “G” in “(GR)”) – onethat even non-physicalist reductionists could accept:

(GR) If a micro-psycho-history is a continuous part of the history of asingle person in one possible world, then, in any world in which anindiscernible micro-psycho-history occurs, it, too, is a continuouspart of the history of a single person.5

5 André Gallois (1998, pp. 248–54) articulates a related doctrine he calls “weak anti-haecceitism.” It ismore general than (PR) or (GR) (applying to anything made of parts, not just persons) and stronger inanother way too: it implies that if a micro-history is the history of a single person in one possible world,an indiscernible micro-history that occurs in any world is the history of that same person, so long as thesame particles play the same roles in each micro-state. Weak anti-haecceitism is, he thinks, “veryplausible.” One might worry that qualitatively similar but numerically distinct objects and eventsleading up to the beginnings of a micro-history might make a difference to the identity of the wholemade out of the parts caught up in the micro-history, even if the parts are the same. (PR) and (GR)remain silent on this matter.

216 dean zimmerman

A second reason to doubt (PR) (and (GR) as well) is the fact that nothing inthe definition of the occurrence of a micro(-psycho)-history requires thatany particular causal relationships hold between earlier and later states in thehistory. Suppose the history of the particles in my body could occur, butwith a causal “break” in it – a time at which the occurrence of subsequentmicro-states does not depend, causally, upon the occurrence of earlier onesin the way they actually do in my history. Arguably, that occurrence of mymicro-history would not be the history of a single person or even of a singleorganism.6

So long as the particles caught up in a micro-history do not all change atonce, the reasonable demand for immanent-causal connections whereverthere is identity over time will not allow my history to occur with a completebreak in causal connections. Still, a further fix-up may well be needed alongthe following lines:

(PR*) If a micro-history is a continuous part of the history of a singleperson in a possible world w, then any indiscernible micro-historyoccurring in any possible world w* is part of the history of a singleperson – so long as, for any n and n*, the micro-state correspondingto n is causally dependent (in one way or another) upon the micro-state corresponding to n* in w* if and only if they are dependent(in that same way) in w.

(GR*) will be understood to be the result of adding the same clause to (GR).Consideration of abnormal causal mechanisms illustrates the need for the

parenthetical “in that same way” to qualify the causal dependence amongmicro-states. Suppose time-travel is possible, and God exists. If all theatoms in my body at noon today had been caused to jump away from oneanother exactly then, this organism (and I, if I am this organism) wouldhave ceased to exist. This would be so even if, later on, God were to causethese atoms to travel back in time so as to seamlessly continue on thetrajectories they were tracing prior to their jump. A kind of causal depend-ence could still hold between the pre-jumping micro-state of my body andthe micro-states of the same particles, immediately post-time-travel. Godnotes the state of each particle in my body, just prior to its jump, andthen uses that information to send my scattered parts back in time to forma duplicate. God’s activity insures that there is causal dependence betweenthe trajectories of the particles prior to their jump and the trajectories of

6 Compare Armstrong’s discussion of “immaculate replacements” (Armstrong 1980).


the time-travelling particles. But the micro-states in the history of thispseudo-body would not display all the same kinds of causal dependence aswould the corresponding micro-states in the otherwise indiscernible micro-history of my body in the actual world.7

r e a son s to den y ( g r* )

Richard Swinburne accepts “the simple view” as a good name for his ownposition – a view he also attributes to Butler, Reid and Chisholm (Swinburne1984, pp. 17, 19 and 26). Here is the beginning of Swinburne’s statement ofthe simple view (“P1” and “P2” should be taken to stand for terms that refer tosome arbitrarily chosen person or persons – “Obama” and “Nixon,” say, or“Cicero” and “Tully”):

[A]lthough apparent memory and brain continuity are, as they obviously are,evidence of personal identity, they are fallible evidence and personal identity issomething distinct from them. Just as the presence of blood stains and fingerprintsmatching those of a given man are evidence of his earlier presence at the scene of thecrime, and the discovery of Roman-looking coins and buildings is evidence that theRomans lived in some region, so the similarity of P2’s apparent memory to that of P1and his having much the same brain matter, is evidence that P2 is the same person asP1. Yet blood stain and fingerprints are one thing and a man’s earlier presence at thescene of the crime another. His presence at the scene of the crime is not analyzable interms of the later presence of blood stains and fingerprints. The latter is evidence ofthe former, because you seldom get blood stains and fingerprints at a place matchingthose of a given man, unless he has been there leaving them around. But it mighthappen. So the suggestion is, personal identity is distinct from, although evidencedby, similarity of memory and continuity of brain. (Swinburne 1984, p. 19)

Swinburne is surely right in attributing a similar view to Chisholm.Chisholm uses the term “criterion of personal identity” to mean “a state-ment telling what constitutes evidence of personal identity,” and draws asharp distinction between the evidence one might have for identifying aperson who exists at one time with a person who exists at some other time,and the “truth conditions” for such claims (Chisholm 1976, pp. 108–13).Both Swinburne and Chisholm include all kinds of physical and psycho-logical continuities among the sorts of evidence one might have for same-ness of persons over time, and they insist that personal identity itself is“logically independent” of the holding of such continuities (Chisholm 1976,p. 112). Swinburne clearly agrees (Swinburne 1984, pp. 18–20). Their view is

7 This would be a sort of “immaculate replacement” without replacement of the fundamental materialparts.

218 dean zimmerman

not the triviality that the phrase “is the same person as” means one thing,and “satisfies such-and-such elaborate conditions for being the same per-son”means something else. Their thesis is muchmore radical: they reject allphilosophical theories that try to state persistence conditions for persons interms of physical and psychological continuities – at least physical continu-ities based upon the behavior of the ordinary matter found in human bodiesand brains, and psychological continuities that can be described in animpersonal way.Swinburne and Chisholm offer thought experiments to support this

conclusion – thought experiments that would undermine (GR*). Theydescribe one or another scenario in terms of the physical and psychologicalcontinuities that could hold between persons existing at different times, andthen they invite us to see that there are two or more genuine possibilitiesconsistent with the events as so far described – possibilities involvingdifferences in which persons exist at which times. There are some possibleworlds in which these physical and psychological facts hold, and a singleperson exists at the beginning and the end of the events described; and thereare others, exactly the same with respect to physical and psychologicalcontinuities, in which the same person does not make it through thewhole episode.Accepting the failure of (GR*) on these grounds has drastic consequences:

if any statement of necessary and sufficient conditions for the persistence ofpersons over time could be given in terms of physical and psychologicalcontinuities, (GR*)-style supervenience would have to hold. So all theoriesof personal identity that try to formulate such conditions are doomed.Some philosophers report that, when they conduct these thought experi-

ments, they do not even generate, for them, the appearance of more thanone possibility. And those who feel the pull of the stories will frequentlydeny that this appearance is veridical or trustworthy. For the record, I agreewith Chisholm’s relatively modest claim about the deliverances of thesethought experiments: “They seem to me to be worthy of being taken, atleast provisionally, as data in our philosophical enquiries about the person;in other words, we should affirm them until we have very good reason forrejecting them” (Chisholm 1970b, p. 188). In this chapter, however, I amnot primarily concerned with the status of the thought experiments.8 I shall

8 In a recent essay, Swinburne defends the deliverances of these sorts of thought experiments. Heprovides an analysis of the notion of a metaphysically possible world in terms of logically consistentmaximal descriptions couched in terms of what he calls “informative designators”; and he develops asubtle and interesting argument for the conclusion that our failure to see any inconsistency in the


instead explore the structure of a certain argument based upon the con-clusion that there are two possibilities in these cases. I ask whether theconclusion – the failure of supervenience – can be used to support a robustform of substance dualism. Examination of the argument turns up arelatively clear doctrine that could plausibly be called “emergent material-ism”; and I point out that emergent materialists can make their viewconsistent with (GR*) by positing immanent-causal differences due to“emergent properties.”

two thought e x p e r iment s

Since the stories told by Chisholm and Swinburne have more bite whenimagined from one’s own perspective, I will use the first person to statethem. One thought experiment involves my fission – the division of mybody and brain in a way that results in two equally qualified candidates forbeing me, either of which would understandably be taken for me had theother half of me been lost through an injury that simultaneously destroyedall the organs on the other side of my body. Here is a summary of the – bynow familiar – story:

Fission

I could have been more symmetrical than I am, and could have undergone anoperation that brought about my perfect fission. Were that to happen, itwould be possible for me to be the person who has the left hemisphere ofmy brain and other organs from that side; and not to be the person who wouldbe thinking with what had been my right hemisphere. But the reverse is alsopossible.9

There is another line of thought that, although not fully explicit inChisholm and Swinburne, can be used to support the failure of personalidentity to supervene upon physical and psychological continuities. Bothphilosophers insist that personal identity is a matter of all or nothing,immune to the effects of conventional decisions about where to draw thetemporal borders of a thing. But the kinds of physical and psychologicalcontinuity that can hold between a person at one time and a person at

thought experiments proves that there are metaphysically possible worlds, in his sense, that arephysically and psychologically indistinguishable while differing with respect to facts about personalidentity. See Swinburne (2007).

9 Versions may be found in Chisholm (1969, p. 106; 1970a; 1976, pp. 111–12) and Swinburne (1973–4;1984, pp. 13–20).

220 dean zimmerman

another time “may be satisfied to varying degrees” (Swinburne 1984,p. 16). When considering just how much of one’s body or brain couldbe lost at once, or just how much sudden psychological change could beendured by a person, the proponents of persistence conditions for personsstated in terms of biological or psychological continuity inevitably end upfacing borderline cases of the kind of continuity they have selected. Thepresence of these borderline cases will lead to a failure of supervenience inthe following way.

Fuzziness

Take any kind of mental or physical continuity that could be used in a theoryattempting to state necessary and sufficient conditions for the persistence ofpersons. It will always be possible to describe a series of similar events involvingthe loss of the relevant kind of continuity, a series that satisfies these con-straints: (i) at one end of the series, the event in question is a loss of continuitythat seems irrelevant to my surviving; (ii) at the other end of the series, theevent is a loss of continuity that seems very likely to be incompatible with mysurviving; and (iii) the difference in degree of continuity between each pair ofneighboring cases in the series is always too small for it to be necessary that inthe one I definitely survive, while in the other I definitely do not. Still, theremust always be a definite fact of the matter as to whether I survive anyparticular episode. So, for at least one of the events in such a series, there arepossible worlds in which I survive that degree of loss of continuity, and thereare also possible worlds in which I do not survive that very same degree of lossof continuity.

One way to resist fission would be to say either: in fission, I would bepartially the one, partially the other; or, in fission, I would necessarilydie. One way to resist fuzziness would be to say: although the differ-ences between some neighboring cases in the loss-of-continuity spec-trum may be too small to necessitate that, in one case, I definitelysurvive and, in the next, I definitely do not, nevertheless, there will bea transition from cases of continuity sufficient to necessitate definitesurvival to cases that merely necessitate its being less than definitely truethat I survive. Chisholm and Swinburne emphasize the difficulties ofthese lines of resistance (see e.g. Chisholm 1976, pp. 110–13; Swinburne1984, pp. 13–19).What the two thought experiments most immediately support is a non-

supervenience claim: where persons go is not settled, given all the “imper-sonal facts” – personal identity could “float free from” such facts, and so(GR*) is false.


f i s s i on and fuz z in e s s v s . s u p e r v en i enc eu pon the imp e r sona l

Fission stories are supposed to undermine both (PR*) and (GR*). In thecase of perfect symmetry, (PR*) and (GR*) require that either the persondies or goes both ways; but he or she cannot go one way in one possibleworld and the other way in the other.

In fuzzy cases, the facts about ordinary matter and impersonal psycho-logical continuities included in micro-histories are imagined to varysmoothly along all the dimensions that could be relevant to sameness ofperson; but whether something is the body of the same person is (we aresupposed to feel) sharp. So there must be worlds in which the same physicaland psychological continuities hold, but in one of them the micro-psycho-history is the history of a single person, and in the other it is not.

Clearly, accepting the conclusion that (PR*) and (GR*) fail due to thesepossibilities is to treat personal identity over time as involving some kind of“further fact.” But does countenancing these possibilities support the ideathat, in addition to the particles that make up the ordinary matter in ourbodies, there is also a further substance – be it immaterial or not – thateither is the person or is crucial to the ongoing existence of that person?

It has always seemed to me that recognizing these seeming possibilities asgenuine does support a robust dualism according to which a person is notmade entirely of the kinds of matter that appear in the micro(-psycho)-histories of our bodies. I have accepted what I shall call the “conditionalnecessity of (GR*)”: necessarily, if persons were entirely material, (GR*)would be true. And I am not alone in being attracted to arguments basedupon this idea.10 Here are two, based upon fission and fuzziness:

Fission argument for dualism

1. A perfectly symmetrical division of my brain and body could result in twoindiscernible micro-psycho-histories, which would overlap until the divi-sion, and then follow the two halves of my body into different rooms; thehistories would be mirror images of one another with respect to causaldependencies among stages; and the procedure would have (at least) the

10 Martine Nida-Rümelin makes use of it, appealing just to the case of fission (Nida-Rümelin 2010, esp.pp. 208–9). She does not regard her support for (her version of) premise 2 as consisting of a“conceivability argument,” but rather as an argument that denial of two genuine possibilities in thecase of fission would imply that we (along with any other beings capable of attributing consciousthought to themselves or others) would be radically mistaken about all sorts of important matters.

222 dean zimmerman

following two genuinely possible (incompatible) outcomes: (a) the left halfof my body ceases to be a part of me, and I survive with a body composedentirely of the matter that formerly constituted my right half; and (b) theright half of my body ceases to be a part of me, and I survive with a bodyentirely constituted by (what was) my left half (based on fission).2. If I had no parts other than the particles making up the ordinary

matter in this body, then the perfectly symmetrical division resulting inindiscernible micro-psycho-histories (with similar causal dependencies)would not have the following two genuinely possible (incompatible) out-comes: (a) I survive the loss of my left half, becoming constituted by myright half; and (b) I survive the loss of my right half, becoming constitutedby my left half (from the conditional necessity of (GR*)).Therefore, I have a part (perhaps an improper part) other than the

particles making up the ordinary matter in this body.

Fuzziness argument for dualism

1*. There is a possible continuation of my actual micro-psycho-history inwhich all the same kinds of causal dependencies hold among stages, butthere are two genuinely possible conclusions: (a) I definitely still exist at theend of it; and (b) I definitely do not still exist at the end of it (based onfuzziness).2*. If I had no parts other than the particles making up the ordinary

matter in this body, then there could be no continuation of my actualmicro-psycho-history with the same causal dependencies and two genu-inely possible conclusions: (a) I definitely still exist at the end of it, and(b) I definitely do not still exist at the end of it (from the conditionalnecessity of (GR*)).Therefore, I have a part (perhaps an improper part) other than the

particles making up the ordinary matter in this body.Most materialists accept a kind of global supervenience of all facts about

persons upon the fundamental physical facts about thematter thatmakes upourbodies; and that sort of supervenience will lead to acceptance of the conditionalnecessity (and actual truth) of both (PR*) and (GR*). But most materialists arewilling to resist these arguments by simply denying their initial premises: fissionand fuzziness have their seductive appeal, but they only seem to be possible.A few materialists take a different approach. They regard physical fission

as a genuine possibility for creatures relevantly similar to us; and they believethat, even if both resulting persons were equally similar to the original, therewould be a “further fact” about which (if either) was the same as the person


who underwent the division. (I find less explicit discussion of fuzziness bythe materialists I have in mind; but their responses to fission suggestanalogous things that could be said about fuzziness.) So these materialistsaccept the initial premise of the fission argument. But they deny theconclusion. I will appropriate the over-used, multivalent label “emergent”as a qualifier for their brand of materialism about persons. The question Iwant to ask is: Where do the emergent materialists think the arguments gowrong?

p e r son s a s en t i r e l y ma t e r i a l bu t “ emerg ent ”

One natural place for the emergent materialist to resist is by giving up (GR*)(and the conditional necessity of (GR*)), enabling her to reject premises 2and 2* of these arguments. (In the next section, I offer the emergentmaterialist a different strategy: accept (GR*), and its conditional necessity,but deny the first premises of the arguments from fission and fuzziness.) Letme put some words into the mouth of this sort of emergent materialist:

I have none but physical parts, consisting of ordinary types of matter. Nevertheless,at least in cases of fission and fuzziness, one and the same micro-history or micro-psycho-history can be the history of a single person in some possible worlds, butnot in others. There is a further fact –whether I go to the left or the right in a fissioncase, or when exactly I cease to exist in a fuzzy case – that is not settled by all theimpersonal facts. Since my comings and goings are not completely determined byfacts about my physical parts, I can rightly be said to be something “over andabove” my parts; I show it by my independent spirit. But do not take themetaphorical phrase seriously. Strictly speaking, there are no such things as spirits(or at least no human spirits, if we want to leave God out of this).

Imagine a perfectly symmetrical but otherwise human-like being whoundergoes an operation that produces two mutilated but conscious personswaking up in rooms 1 and 2. There is one micro-psycho-history consistingof the micro-psycho-states of the particles in the original patient and thoseof the patient in room 1; and there is a mirror-image micro-psycho-historythat begins with the same micro-psycho-states but is continued by particlesin room 2. The emergent materialist supposes that, in one possible world,the first micro-psycho-history is that of a single person but the second is not;while in another possible world, the reverse is the case. In an operation likethis, she says, there is simply a “further fact” about whether the patientloses the right half of his body and moves into room number 1, or losesthe left half of his body and moves into room number 2; and there is a

224 dean zimmerman

corresponding “further fact” about whether a new person is waking up inroom number 2 or room number 1.Suppose that, as a matter of brute “further fact,” the person recovering in

room 1 is the actual survivor of the operation; and that the patient in room 2came into existence upon the operating table. If the emergent materialistwere to explain the situation to the newman in room 2, the latter might wellprotest. He can make an equally strong case for the claim that he was thepre-operative person. But the emergent materialist will say:

In many respects, you do seem to be an equally good candidate for being that man.But the fact remains that there was one person who went into the operating room,and he is now waking up in room number 1; and there is no person who went intothe operating room and is now in your room.

These remarks will be more convincing if the emergent materialist canfollow them up with the claim that there is nothing that went into theoperating room looking like a surprisingly symmetrical human being andnow lies recuperating in room number 2. But an emergent materialist whoaccepts the doctrine of temporal parts cannot say this – at least, not if hermetaphysics recognizes arbitrary fusions of the temporal parts of persons.When the symmetrically fissioning person shrinks to half his original sizeand leaves the operating room for room number 1, and the mere fission-product wakes up in room number 2, this sort of emergent materialist mustadmit that there is something that went into the operating room in the formof a whole person, and that is now waking up in room number 2. Her claimmust be that the only person who entered the operating room is now wakingup in room number 1. The character in room number 2 will want to know:“Why does my word ‘I’ not refer to a person who existed before theoperation? After all, there is a thing that existed then, and that was fearfullyanticipating the operation, and that is now uttering these words; how couldit fail to be a person – in fact, the person I am?”A temporal parts theorist who denies the existence of arbitrary fusions

could give a perfectly good answer to the patient waking up in room 2 whenhe asks for a reason to think that he was not the patient who entered theoperating room: “Your word ‘I’ doesn’t refer to something with those pre-operative stages as parts because there is nothing that has both your currentstage and those stages as parts. Those temporal parts and your temporalparts do not form a larger whole.” By contrast, an emergent materialist whoaccepts arbitrary fusions of temporal parts will simply have to regard being aperson as a property that does not supervene upon the micro-history ormicro-psycho-history of an object.


How bad would this be? Worse, I think, than having to accept the failureof microphysical supervenience for the reasons given by Trenton Merricks.Merricks raises difficulties for supposing that the intrinsic properties ofcollections of particles, plus spatio-temporal and causal relations holdingjust among the members of the collection, provide a sufficient superven-ience basis for facts about whether some particles make up a conscioushuman being (Merricks 1998a). As Sider has pointed out in his criticisms ofMerricks’s argument, the materialist is likely to think that being a humanperson is a spatially maximal property, like being a rock (Sider 2003, pp. 142–3). A large part of a rock is not a rock – although it would be one, were it notembedded in more rock. Similarly, a large part of a human body, such as abody-minus-its-left-index-finger (to use Merricks’s example (1998a, p. 62)),might be the sort of thing that would be a person, were there not moreorganic matter attached to its hand. Both properties display “border-sensitivity”: whether something qualifies as a rock or (given Sider’s materi-alist assumptions) a human person “depends on what is going on around itsborders” (Sider 2003, p. 139). Suppose that a materialist wants to regardbeing conscious as intrinsic, and also to follow Sider in treating being a humanperson as a maximal property due to border-sensitivity as in the case of therock – a human person is the largest of ever so many would-be persons. Inthat case, she must see a host of conscious non-persons in the vicinity ofevery human being.11 Materialists can avoid this conclusion if they followSider in accepting the extrinsic nature of consciousness.12

One might have hoped that an analogous move could be made by theemergent materialist who believes in temporal parts and arbitrary fusions;that she could say the failure of supervenience was due to the fact thatpersonhood is an extrinsic matter, and a true supervenience claim can beformulated by appeal to a broader supervenience base – one that is sensitiveto the presence of nearby candidates for being a certain person. But theemergent materialist’s denial that being a (single) person supervenes uponmicro-histories or micro-psycho-histories cannot be rendered benign byclaiming that personhood is an extrinsic matter, and only fails to supervene

11 The moral can be drawn from Merricks (1998a) and Olson (1995).12 A materialist like Sider can sensibly conclude that consciousness must be extrinsic, in virtue of its

analytic ties with being a person – something a materialist will naturally take to be border-sensitive.The move is not without its costs, as Merricks points out (2003). I would argue that it is least costly forthe materialist who regards thinking and experiencing as no more natural or fundamental than suchobviously non-fundamental activities as executing a karate chop or shaking hands with someone.Arguably, body-minus-its-left-index-finger does not do either of the latter things, although it wouldbe doing so were there no left index finger attached to it. Materialists about persons who are propertydualists about the phenomenal may not have it so easy. But that is another story.

226 dean zimmerman

upon micro(-psycho)-histories due to their overly intrinsic nature. A micro-history includes all the facts about the disposition of matter surroundingthe collections of particles caught up in it. So one can tell, just by looking atthe micro-history of the particles that make up the person who survives theoperation, that his other half was removed intact and is now recoveringnicely in room 2.I conclude, then, that emergent materialism and the doctrine of temporal

parts make poor bedfellows, barring adoption of mereological principlesdisallowing arbitrary fusions. Those who reject temporal parts altogetherseem to me to be in a dialectically much stronger position. It is notsurprising to find that all the clearest examples of emergent materialistsare also opponents of temporal parts – for example, Trenton Merricks, TimO’Connor, Jonathan Jacobs and Lynne Rudder Baker.13

hol i s t i c c au s a t i on to the r e s cu e ?

In response to the arguments from fission and fuzziness, an emergentmaterialist must either reject (GR*) (and its conditional necessity) or findsome way to deny premises 1 and 1*. She accepts fission (the case upon whichI will focus): there is one micro-psycho-history that corresponds to thestages in the life of a person who loses half his body and then wakes up inroom 1, and another that corresponds to this person’s pre-operative stagesplus the stages of a different person who comes into existence on theoperating table and wakes up in room 2. To accept fission and reject premise1, she must find a causal difference between otherwise exactly similar micro-psycho-histories. Might the causal relations holding between the wholebody and just one of the immediately post-operative persons be of acategorically different nature than the ones holding between that bodyand the other post-operative person? If so, the two micro-psycho-historiesmight be indiscernible; but they would differ in the kinds of causal depend-ence holding between stages at the time of the surgery. (GR*) could then beaccepted, but premise 1 denied.Consider the fissioning symmetrical person, who goes to the left and ends

up in room 1. In the fission of this organism into two persons, it would benatural to expect that the properties of each resulting person’s body,

13 SeeMerricks (1997, 1998b); O’Connor and Jacobs (2010, pp. 80–2); and Baker (2000, pp. 132–41). Fortheir opposition to temporal parts, see Merricks (1999); O’Connor and Jacobs (2010, p. 74, fn. 5); andBaker (2000, p. 22). If I understand Stephen Davis’s appeal to the will of God in determining identity(Davis 2010, pp. 26–7), he, too, is committed to the falsehood of (PR*) and (GR*). Though I lackhard evidence, I would bet that he, too, rejects the doctrine of temporal parts.


immediately after the operation, would be adequately causally explained bydescribing two things: the features of just one half of the pre-operativeperson, and the procedure that separated the halves. On this naturalassumption, there is no additional property that belongs only to the wholepre-operative body immediately prior to surgery, and that causes the post-operative persons to have some intrinsic feature directly – i.e. not in virtue ofthe feature’s correlation with contemporaneous intrinsic properties belong-ing to just one or the other half of the person. I will call this supposition “noholistic causation.” On this hypothesis, any effect the whole person hadupon the intrinsic nature of the person who woke up in room 1 would bemost directly caused by just a part of the pre-operative person. And this doesnot seem to me to be consistent with supposing that the causal relationshipbetween the whole person and the new guy in room 2 is of a radicallydifferent nature.

Prior to the operation, the left half of the person is not itself a person; andit is causally responsible for the nature of the person in room 1. The righthalf is not a person either; and it is causally responsible for the nature of theperson in room 2. Considered intrinsically, the two halves generate the twoseparate persons by an exactly similar process. Assuming no holistic causa-tion, the whole person “works through” one of its halves to bring about theexistence of a person with certain intrinsic features in room 1; he causessomething by having a part that causes it. And the whole person also has anexactly similar part – his right half – which, when considered by itself, doesexactly the same thing. If something (in this case, a person) uses two exactlysimilar agents (its two halves) to bring about two exactly similar results bymeans of exactly similar causal processes, and the thing stands in exactlysimilar relations to the two agents, then the thing does not cause these tworesults in significantly different ways.

But relax the no holistic causation assumption, and things look altogetherdifferent. Suppose that the man is feeling queasy right before the operation;that this feeling of queasiness causes subsequent states of a similar sort; andthat it does not cause them in virtue of its correlation withmore direct causesof subsequent feelings, causes attributable to intrinsic properties of thebrain. When the man’s pre-operative queasiness causes his post-operativequeasiness, it does not do so purely in virtue of intrinsic properties of thehalf-brain that he kept (since, in general, queasiness is now supposed to haveeffects that are not entirely due to the physical properties of the man’sbrain). This opens up space for a difference in the kinds of causal relationsholding among micro-psycho-states in the two otherwise similar micro-psycho-histories. The pre-operative queasiness of the whole man could

228 dean zimmerman

directly cause the queasiness felt by the man who is wheeled off to room 1;while the similar queasiness presumably felt by the man wheeled off to room2 would have to be caused in some other way – perhaps by the physicalconditions of the right hemisphere that, upon removal, has become hisbrain. There would be a kind of causal – or at least counterfactual –dependence holding between the original patient’s pre-operative queasinessand that of the new man in room 2. If the original patient had not feltqueasy, he would not have had the kind of right hemisphere that wouldtend to produce queasiness in the new man in room 2. But, on the holisticcausation hypothesis, this would not be the same kind of direct causaldependence that holds between the post-operative queasiness of the manin room 1 and the pre-operative queasiness of the whole man.It may well be no coincidence that some of the main defenders of

emergent materialism (e.g. Merricks 2001, ch. 4; O’Connor and Jacobs2010) posit “emergent”14 properties exemplified by the person as a whole –properties that are causally efficacious, and not just in virtue of the causalpowers exercised by the smaller parts of the body. Indeed, the description ofa case of perfectly symmetrical fission given by O’Connor and Jacobsresembles – at least abstractly – the story I told about the whole man’squeasiness.15 Of course, if the holistic causation in question can holdbetween pre-operative patient and both fission-products, the emergentmaterialist is once again forced to deny the conditional necessity of (GR*).

f i s s i on ing s im p l e s and pr em i s e 2 o f the f i s s i onargument for dua l i sm

The emergent materialist asks me – as a defender of the argument fordualism based on fission – why I believe that, if something is a mere materialobject, then in a case of perfectly symmetrical fission, with indiscerniblemicro(-psycho)-histories, the object must go either both ways or neither.The question seems especially pressing when I reflect upon the very

14 I should note that Merricks does not himself use the term “emergent” to describe the properties invirtue of which an “object’s causal contribution would be . . . independent of what its atoms were likeand so, presumably, independent of what its atoms cause” (Merricks 2001, p. 90), but the label isoften applied to such properties by other advocates and opponents of the idea that wholes haveproperties satisfying this description.

15 See O’Connor and Jacobs (2010, pp. 80–1). Their “emergent level states” seem to consist mainly of the“particularities” of individuals, however (pp. 74–7), and not phenomenal properties. I am not a fan ofthe metaphysics of “thin particulars”which they use to explain the notion of a “particularity”; so, wereI an emergent materialist, I should look for other properties that might have causal powers in someway independent of the activities of a person’s proper parts.


different reaction I have to the case of a simple substance – be it a particle, anintelligent monad, or what-have-you – undergoing apparently symmetricalfission.

Simple Fission

Imagine a particle, x, that is buzzing along, minding its own business. Suddenly, itis zapped by some kind of energy, and two particles exactly like x appear, headed offto the left and right of x’s original trajectory in what looks like a case of symmetricalfission. Suppose the counterfactuals are the same on both sides: the particle on theleft (“Lefty” will be shorthand for this descriptive phrase) would have had spin uphad x had spin up immediately prior to the apparent fission, Lefty would have beenmoving more rapidly had x been moving more rapidly then, etc.; and similarly forthe one on the right (“Righty,” another abbreviated definite description).

What, if anything, is wrong with supposing that there are three genuinelypossible endings to simple fission that do not differ in the facts about theparticle stages, their environments, and the relations of causal dependenceholding between particle stages: (i) x ceased to be, replaced by two newparticles; (ii) x is Lefty and Righty is a new particle; and (iii) x is Righty andLefty is a new particle? Personally, I find myself inclined to say that these dorepresent three distinct possibilities, each consistent with the case asdescribed. But then why should I have such a different reaction to appa-rently fissioning composite things, like the emergent materialist’s physicalpersons? If I accept these three outcomes as genuine possibilities for simplefission, have I not thereby given up the idea that, when an object is entirelyphysical, its persistence through time should supervene upon facts that donot directly imply anything about the persistence of the object in questionand that concern only its smallest parts and the relations of causal depend-ence among its stages? If so, have I not lost the right to insist upon theconditional necessity of (PR*) and (GR*)?

These supervenience principles describe the way complex wholes dependupon the histories of their parts. When one reaches the level of a partlesssimple, it is not at all clear that its persistence through time should bethought to consist in anything other than the simple’s existing at more thanone time. What would a supervenience principle for the persistence ofsimples look like? It would have to find more fundamental facts that donot directly imply anything about the identities of the kinds of simple inquestion, and exhibit relations at this level of facts that could plausibly bethe grounds of simple-persistence. The friends of temporal parts will have

230 dean zimmerman

no difficulty finding subjects for such facts, namely, instantaneous temporalparts; but, as shall appear, they should probably resist the idea that simplefission has the three possible endings. Others, who do not regard temporalparts as inevitably accompanying all persistence through time, may wellaccept that there are three possible endings; but either insist that this is dueto the partlessness or fundamentality of the object in question (whichdistinguishes the case from that of a complex material person), or claimthat the story leaves out important details about the kind of causal depend-ence holding between x and its offspring, Lefty and Righty.Those who believe in the doctrine of temporal parts deny that a persisting

simple really is partless. They are likely to accept the supervenience of thepersistence of even the smallest particles upon facts about their instanta-neous temporal parts – i.e. to be immanent-causal reductionists aboutsimples. A temporal parts metaphysics may or may not leave room for adeep distinction between immanent causation and other sorts of causaldependence. On a very thin notion of immanent causation, the counter-factuals mentioned in simple fission would settle all the facts about whatcaused what, and by what means – leaving no room to suppose that the spinand speed of Lefty, say, but not Righty, depended in a special immanent-causal way upon x’s state at the point of fission. This metaphysician reallyshould accept immanent-causal reductionism, which requires symmetricaltreatment of perfect fission; she should deny that the three endings aregenuinely possible.A friend of temporal parts might have a thicker notion of immanent

causation; she might think that there is a special kind of causal relation thatalways holds between stages of the same particle, a relation that is notautomatically present whenever there is the kind of counterfactual depend-ence described in simple fission. She need not immediately draw the con-clusion that immanent causation must either fork or stop when x is zapped;for all she has been told, the counterfactuals on just one side might be due toimmanent causation, the others being the product of another kind of causalrelationship. If she thinks that immanent causation is, nevertheless, arelation that can hold between the earlier states of one particle and thelater states of another, then immanent-causal branching can simply beadded to the story. And, again, this temporal parts metaphysician shouldbe an immanent-causal reductionist, not admitting the three possibilitiesfor x once all the causal facts are stipulated to be symmetrical as well.Some of us do not believe that persisting through time involves, as a

matter of necessity, the possession of new, short-lived temporal parts. Wehave a right to the reaction I reported as my own: simple fission is


consistent with each of the three endings. Some of us may deny that anyspecial kind of causal dependence must hold between the way a simple is atone time and the way it was at earlier times – or that any causal dependencemust hold at all. These opponents of temporal parts will reject super-venience of the persistence of particles upon other facts. I do not see,however, that doing so should be thought to undermine their confidencein principles like (PR*) or (GR*), which pertain to complex wholes.

Other opponents of temporal parts have at their disposal a sort of back-handed way of saving a supervenience principle even for simples. Rejectionof temporal parts is naturally coupled with the idea that a robust kind ofimmanent causation is involved in persistence. When there is genuinepersistence of a perfectly natural kind of thing, there is a very distinctivesort of causal relation holding among stages of the thing – a way for stages inthe history of a single particle to depend upon one another that is inevitablyaccompanied by identity, but that stands at a slight conceptual remove fromidentity nevertheless.

The state or event of a certain fundamental object’s having a certainfundamental property at a time can have a tendency, in propitious circum-stances, to bring it about that there will subsequently be a fundamentalobject of the same kind with a fundamental property from the same“determinable” family. But the opponent of temporal parts believes it ispossible to have this sort of tendency for two very different reasons: a thingcan have it in virtue of a propensity to generate a new thing of the samegeneral type, or it can have it in virtue of a propensity to merely hangaround. Simple fission does not specify whether this second sort of propen-sity is being exercised at all at the point of fission; nor whether, if it isexercised, the effect occurs toward the right or the left. If this propensity wasnot triggered by the zapping, x ceased to be – however similar to it Lefty andRighty might be, and however robust the counterfactual connectionsbetween x’s states and theirs.

I find this second response quite plausible: there is a sort of necessarilyimmanent relation among stages in the history of a fundamental thing, itconstitutes a distinguished sort of causation, and it implies the identity ofthe thing in question. So when I hear the story about x’s apparent fission, I seegenuinely distinct possible ways in which x may have caused the stages ofRighty and Lefty; and these differences seem highly relevant to the questionwhether x survived, and, if so, whether it is Righty or Lefty. The super-venience of physical persistence upon the microphysical and the causal,when enriched with this kind of causal dependence, becomes unproblematic.If one supposes that x is Lefty, the micro-history of x-up-to-zap-then-Lefty

232 dean zimmerman

is not exactly like the micro-history of x-up-to-zap-then-Righty; one kind ofcausal dependence holds between early Lefty stages and x’s stage at the zap –immanent causation, involving the exercise of a propensity to hang around –and some other kind of causal dependence holds between the early Rightystages and x’s stage at the zap.So I have found a way to explain whymy reaction to the fissioning simple

does not betray a lack of commitment to the supervenience of the persis-tence of physical things uponmicrophysical histories plus (the right kind of)causal dependencies within those histories. Since the causal dependence isso closely tied to persistence through time, it would not satisfy reduction-istic attitudes toward the persistence of simples. But I do not feel much pulltoward the idea that identity over time, for simples, must consist in some-thing else.The differences in direction of immanent causation used in response to

simple fission suggest a way for an emergent materialist to resist the fissionand fuzziness arguments for dualism without quite giving up (GR*) (or itsconditional necessity): she can deny premise 1 (and 1*, though I will focuson fission) by positing an important immanent causal difference betweenotherwise indiscernible micro-psycho-states.In response to simple fission, I gave a sketchy account of a special kind of

causation that an opponent of temporal parts might believe in, consisting inthe exercise of a propensity to keep existing. Those of us who posit such acausal process may well deny that it occurs in anything other than afundamental entity. (I should want to leave it a largely empirical questionwhether subatomic particles persist through time, whether the actual fun-damental physical entities are small or large, and whether they include fieldsor a substantival space-time manifold or both. For all I know or care, allfundamental physical entities may consist of series of momentary things,just as the friends of temporal parts allege. I simply see nothing inevitableabout this.) In order to give the emergent materialist a chance, here I shallsuppose that complex organisms can exercise the deeply different kind ofhanging-around causation that I have posited at the level of fundamentalpersisting things. In particular, I shall allow that it may hold between thestages of an object while that object suddenly loses a large part.If some state of the pre-operative person immanently causes some state of

the person in room number 1, and no state of the man in room 2 is similarlyimmanently caused, premises 1 and 1* could obviously be denied by theemergent materialist. This strategy may make fission consistent with theletter of (GR*); but it is a nice question whether the supervenience principlestill captures something of Parfit’s reductionism, once immanent causation


is posited as a distinctive kind of causal dependence.16 On the face of it,appeal to causation that only works when an object persists would seem toseriously undermine reductionism about personal identity. However, ifimmanent causation can be characterized without any special mention ofpersons, or of psychological properties that imply that a person exists,perhaps the emergent materialist will not have violated the spirit ofParfit’s reductionism after all. What is needed at this point is a more preciseaccount than I have given so far of two things: the nature of the “hangingaround” kind of causation, and the motivations for advocatingreductionism.

Loose ends and confessions

Two of the central topics of this chapter are begging for a fuller treatmentthan I have been able to give them: (1) I am convinced that there is muchmore to be said in favor of the conditional necessity of (GR*), and againstthe emergent materialist who would resist arguments for dualism simply bydenying (GR*); (2) I am disappointed by the sketchiness of my attempt tocharacterize immanent causation, and to distinguish it sharply from otherspecies of causal dependence. I would hope to identify a kind of causalrelation definable in terms of the exercise of “strongly identity-entailingconditional powers” – a notion that proved useful when I was grapplingwith Shoemaker’s distinction between “thin” and “thick” properties(Zimmerman 2009, pp. 699–703).

I conclude with a couple of confessions. The conditional necessity of(GR*) seems to me to glow with the light of Truth, and fission and fuzzinessappeal strongly to my imagination. But if the arguments from fission andfuzziness provided my only reasons for believing in a dualism of persons andbodies, I would regard the case as inconclusive. Conceivability may be animportant source of evidence for possibility, but it can lead us astray –especially when used as evidence for the ascription of essential properties toparticular individuals, as in the present case.17

But other considerations push me toward dualism – or at least away fromall the more plausible kinds of materialism, including the versions ofemergent materialism canvassed here (Zimmerman 2003, 2004, 2010). In

16 I am grateful to Marco Dees for posing this nice question.17 For my criticisms of one argument for dualism that relies upon conceivability in an unsafe way, see

Zimmerman (1991).

234 dean zimmerman

the actual circumstances, then, and despite my qualms about conceivability,the arguments from fission and fuzziness are not irrelevant. Dualismreceives a small epistemic boost for me, in virtue of the fact that it respectsthe conditional necessity of (GR*) and allows me to trust my modalinstincts about fission and fuzziness.18

18 I learned much from the questions and suggestions of participants in the conference “DiachronicPersonal Identity: Complex or Simple?,” organized by the Department of Philosophy at theTheological Faculty of the University of Innsbruck, July 2010. I am especially grateful to RichardSwinburne and Hud Hudson. Much later, and in the nick of time, Ofra Magidor showed me thatI had somehow begun to use “(GR*)” to mean two quite different things; a reading group at Rutgersuncovered some technical flaws; and Marco Dees helped me through some confusions.Unfortunately, other comments from these sources, and from audiences at Oxford, Yale andBuffalo, raised important issues and objections too large to be addressed before this book went topress.


chapter 1 3

The morphing block and diachronic personal identityHud Hudson

the cr i t e r i on of d i a chron i c p e r sona l i d ent i t y

One perplexing philosophical puzzle concerns how best to complete thefollowing sentence:

(C1) Necessarily, if x is a human person at a time, t, and y is a human person at a distincttime, t*, then x = y if and only if and because ______________________________.

Defending any particular candidate for filling in the blank would amount toendorsing a constitutive (rather than an epistemic) criterion of diachronicpersonal identity and would yield a more informative claim than a merenecessary bi-conditional insofar as it would provide illuminating conditionsthat serve as the grounds for the persistence of a human person.

Plausibly, however, that way of putting the puzzle unfairly privileges onemetaphysics of persistence over another at the outset, for although theendurantist is now ready to go on a hunt for the blank-filler, the perdur-antist (more on this distinction on p. 237) first requires a reformulation:

(C2) Necessarily, if x is a momentary stage of a human person at a time, t, and y is amomentary stage of a human person at a distinct time, t*, then x is a momentarystage of one and the same human person as y if and only if and because________________________.

Several – but not all – prominent theories of personal identity can beintroduced and explored by way of this challenge to properly completethe sentence appropriate to one’s general metaphysics. My goal here, how-ever, is not to furnish an overview of current theories of personal identity,but rather to identify and pursue a problem for many of them that areroutinely classified among the so-called complex views of personal identity.

As a bit of disclosure, I share Eric Olson’s suspicion that there is no singleand straightforward thesis separating the simple and the complex views,despite the wide-ranging use of those terms, but whether or not the labels doin fact mark out a significant distinction between theory types, there is an

236

intriguing worry worth investigating that threatens several of the mostpopular current examples of the putative complex views.1

two comp l e x v i ew s o f d i a chron i cp e r sona l i d ent i t y

Introducing our targets:2 discussions of diachronic personal identity shouldbegin, I think, with the question “What is a human person?” Although notcontroversy-free, I favor a materialist metaphysics of the human person,according to which a human person is identical to a certain, highly organ-ized, material object.3 Some variety or other of materialism is the dominantview these days, but just which material object we are to identify with ahuman person is hotly contested. The animalists favor identification with anentire human organism, while theminimalists select the least inclusive thingthat houses the relevant psychology: say, a proper part of a human brain.4

Next (as was indicated by our two options for formulating a criterion onp. 236 above) let us ask the question “How does a material object persistthrough time?” Again, although hardly controversy-free, perdurantism andendurantism remain at the top of the candidate answer list. Perdurantiststake a persisting material object to be extended in four dimensions (threespatial and one temporal) and to be composed of temporal parts (as well asof spatial parts). Persistence across an interval of time is a matter of beingpartly present at the different times in that interval (i.e. of having differenttemporal parts located at the different moments in that interval).Endurantists take a persisting material object to be an entity extended inthree dimensions (all spatial). Persistence across an interval of time is amatter of being located at the different times in that interval (i.e. of beingmulti-located at a series of non-simultaneous three-dimensional regions).5

1 See Olson’s “In search of the simple view” (Chapter 2 in this volume). I borrow the “if and only if andbecause” connective from Olson’s discussion. Hereafter I drop the hedging “putative” when referringto the complex views.

2 The following three paragraphs describing two representative complex views contain material fromHudson (2012).

3 For an extended discussion of exactly this question and a defense of this reply, see Hudson (2001).4 There are many other flavors of materialism, as well, but these two will do for now. For a representativedefense of animalism, see Olson (1997b). For a critique, see Hudson (2007). For a representativedefense of minimalism see Hudson (2001).

5 For a representative defense of perdurantism, see Sider (2001). For a representative defense ofendurantism, see Van Inwagen (1990a). Since human persons are extended in three spatial dimensions,I am presenting our two alternatives against that assumption, but (strictly speaking) each theorist canaccommodate minima, too, and offer the very same explanations of the persistence of material objectsextended in only two or one or zero dimensions.

The morphing block and diachronic personal identity 237

Finally, we might ask the question “Does our preferred answer to theprevious two questions give us any guidance on filling in the blanks in (C1)or (C2)? – that is, on selecting a criterion of persistence for the special kindof material object that counts as a human person, a theory that will allow usto determine the beginning and ending of a human person’s lifespan?”Animalism together with endurantism, for example, may steer us in thedirection of a biological criterion of persistence for human persons, a theorythat will allow us to determine the contours of a human person by locatingthe precise beginning and ending, shape and size of a certain organism.Alternatively, minimalism together with perdurantism may lead us to apsychological criterion of persistence for human persons, a theory that will tellus that to focus on a human person is to trace out the history of a certainpsychological profile, whether it stays put in a single body, or transfers fromone human animal to another, or jumps across a temporal gap, or even shiftsfrom an organic to an inorganic host.6

Accordingly, we may hereafter focus our attention on two clear andrepresentative examples of proponents of complex views of diachronicpersonal identity: the animalist–endurantist–biological criterion theorist,who will fill in the blank in (C1) with his favored way of indicating samenessof human animal, and the minimalist–perdurantist–psychological criteriontheorist, who will fill in the blank in (C2) with his favored account of apsychological continuity and connectedness relation between human per-son-stages.7

But before generating problems for these theorists, we first have somemetaphysical stage-setting to do.

f i v e theor i e s o f t im e

A popular metaphysical picture can generate a worry for these complexviews of diachronic personal identity. Consider five theories in the philos-ophy of time: presentism is the view according to which only present thingsexist. The growing block theory is the view according to which only presentand past things exist. The shrinking block theory is the view according towhich only present and future things exist. The disappearing branch theory isthe view according to which past, present and future things exist (and in

6 For the arguments from animalism and endurantism to the biological criterion and from minimalismand perdurantism to the psychological criterion, see Hudson (2001, ch. 4).

7 Olson is an example of the first kind of theorist and I am an example of the second. For the record, (C1)need not be so narrowly construed, for one could certainly think (as Olson does) that “human person”is a phase sortal and thus that a thing can cease to be a human person without ceasing altogether.

238 hud hudson

which the future consists of a proliferation of equally real branches thatsuddenly disappear as soon as time flows along any path excluding them).Eternalism is the view according to which past, present and future thingsexist (with no additions, subtractions or disappearing branches).Other theories compete alongside this field of five, but our quintet is

representative enough. The literature on the advantages and disadvantagesof these competing theories is massive, and the references given throughoutthis chapter comprise only the tip of that philosophical iceberg which is thedebate on the nature of time.In this chapter, I am interested in exploring a feature shared by three of

these views in particular – the growing block, the shrinking block and thedisappearing branch theories – and I will develop my thought primarily incomparison with the growing block theory, since it is the best known of thethree.8

As it is standardly characterized, the growing block theory offers us apicture of the universe featuring a space-time volume that increases as timepasses. At any given moment exactly one time is special – the time asso-ciated with the hyperplane on the surface of the block in the direction of itsgrowth. The outermost surface, so to speak, is the new kid on the block; itdid not exist moments before, and although it will continue to exist, it willnot remain the outermost surface moments hence. During its brief instantin the spotlight – before becoming ever more imprisoned in the block’sinterior – its stock of facts and events are present. Soon they will becomeforever past and take their eternal places frozen in the block, but for oneshining moment they are privileged – balanced on the very edge of Being.(A description of the basic idea behind the shrinking block theory can behad straightaway just by thinking about its name and adjusting bits of thestory just told accordingly.)On some scorecards, the growing block theory is thought to combine the

best features of its primary rivals. Like presentism, it proclaims the unique-ness of the present, recognizes objective and irreducible temporal proper-ties, takes tense seriously, and countenances the genuine passage of time.Like eternalism, it furnishes truth-makers for past truths, provides relata forcross-temporal relations, and acknowledges the existence of many objectsthat are not present. Of course, this very combination of commitments isalso alleged to be the source of its decisive refutation, with opponentscomplaining that if it were true, we (absurdly) could not know it is now

8 For formulations and discussions of the growing block theory see Broad (1923) and Tooley (1997). Thenext two paragraphs describing the growing block theory contain material from Hudson (2012).


now or else would risk living amongst zombies in the block (i.e. amongstnon-present existants with no consciousness).9 Other critiques balk at aprivileged present and allege incompatibility with special and general rela-tivity or else target the endorsement of time’s passage, maintaining that theview must be supplemented by a hypertime in which the growing of theblock occurs and against which its rate of passage can be measured.10

Whatever its other defects, I judge Ned Markosian to have shown thisfinal criticism to be in error.11 And, although I am quite intrigued by thecompossibility of the growing block theory and the hypothesis of hypertime(for current purposes), I am happy to concede that if there is somethingmetaphysically amiss with the growing block, the shrinking block, or thedisappearing branch theories, it is not for want of being embedded in asecond temporal dimension.

So just what is so striking about these views that calls for specialattention? Well, the catch-phrase associated with them is fairly remarkable:namely, that the volume of space-time differs at different times. But that isnot really a careful enough slogan, since if (in the growing case) time has aninfinite past or (in the shrinking or disappearing cases) time has an infinitefuture, the total volume need not increase or decrease even with theaddition or subtraction of hyperplanes. Perhaps, then, it is better to putthe shared commitment like this: on growing, shrinking and branchingmodels, different regions of space-time exist when different times arepresent. That is, to the extent that these three theories of time are takento be live metaphysical possibilities, so, too, the theses that points of space-time can come into existence and go out of existence are also taken to be livemetaphysical possibilities. But just how is this proposal to be understood?

I think the most promising response to that question is to presuppose asubstantivalist space-time, identify times with hyperplanes, and incorporatea certain independence or recombination principle for space-time points.Substantivalism takes space-time to be a concrete particular with an onto-logical status not reducible to relations between material objects (or anyother entities). Such a view identifies space-time and its extended sub-regions with either pluralities or else fusions of uncountably many, simple,unextended, space-time points, and it countenances a perfectly natural

9 On the alleged obstacle to knowing whether it is now now, see Bourne (2002) and Braddon-Mitchell(2004). For the zombie problem, see Forrest (2004) and Heathwood (2005). For an attempt toundermine the motivation for the growing block theory, see Merricks (2006).

10 J. J. C. Smart challenges the notion of temporal becoming and raises pointed questions about theneed for a hypertime (Smart 1949, 1998).

11 Markosian (1993) answers Smart’s challenge without resort to hypertime.

240 hud hudson

relation of occupation that connects material objects (the guests) with suchpluralities or fusions (the hosts).12Quite unlike the proponent of Leibnizianmonads, then, it would seem that the substantivalist proponent of a space-time capable of augmentation or diminishment does not regard its pointlocations as modally bound to one another. Instead, different combinationsof space-time points coexist at different times.Well . . . if that is so . . . there has been something of a failure of

imagination when it comes to dynamic space-time.

the morph ing b lock theor y

Consider what appears to be a growing block that began with a slice that wasa duplicate of our own block’s slice from a billion years ago. When thecalendars on its outermost surface say October 14, 1066 the block has onevolume and when they say April 19, 1775 it has another, such that the firstplurality of hyperplanes have been joined by uncountably many others overthe 709 year interval that separates the two occasions; reality is growing.Then a surprise . . . new hyperplanes steadily appear at both ends of theblock continuing to duplicate the relevant ever-increasing portion of ourown block’s history; reality is growing at both ends. Then a reversal . . . thelatest hyperplane remains fixed as more and ever earlier ones are tacked on;reality is growing once again, but in the other direction. An alarmingdevelopment . . . hyperplanes at both ends begin to disappear; reality isshrinking, shrinking, shrinking until there is but one hyperplane in exis-tence. “Then” the world is non-spatio-temporal.If the creation-and-annihilation stories of the growing block, the shrink-

ing block and the disappearing branch theories are live metaphysicaloptions, then why not this block that grows at both ends or shrinks atboth ends or grows at one while shrinking at another and then reverses?Better yet – why not a block that acts just like a properly behaved growingblock but at some point in its development dutifully adds outermosthyperplanes while recklessly subtracting sub-regions from its inner hyper-planes, i.e. moving (roughly) from block to hourglass shape (or, rather,transitioning between their four-dimensional equivalents)?Why not a blockthat (perhaps tired of being the only instance of its kind) suddenly annihi-lates an entire inner hyperplane and thereby brings it about that actuality“now” consists of a pair of disconnected space-times, each of which grows inthe old familiar way?

12 On substantivalism (and its opponents) see Earman (1989), Nerlich (1994) and Van Cleve (1987).


Perhaps here is why not: note the scare-quoted “then” and “now” in thelast two paragraphs. Of course, there is no time such that the world is non-spatio-temporal then, and there is no time such that before that timeactuality has but one space-time and afterwards it has two! True enough,but nothing hangs on pretending otherwise. The scare-quoted temporalterms do help to get the relevant ideas across but are dispensable (as will beshown below).

Another try – perhaps here is why not: blocks do not cause themselvesto undergo such alterations, and nothing else could cause a block to suffersuch bizarre augmentation or diminishment either! Well, maybe Godcould . . . but true enough. All the business about the block’s being areckless agent or tired of its lonely status was just rhetorical fun, but thisobservation about causation has not disqualified the growing block or theshrinking block or the disappearing branch theories which also featureuncaused creation and annihilation of space-time regions. Where do allthose regions go that make for the shapely new hourglass figure of ourmorphing block? Presumably to the same graveyard that contains theshrinking block’s spent hyperplanes and the untrodden paths of thedisappearing branches. Whatever annihilationist story those theoriststold which managed to keep their views amongst the metaphysicallypossible theories of time simply gets told again.

Another try – perhaps here is why not: on the growing block and theshrinking block and the disappearing branch theories we have prettycompelling stories to tell about what it takes for a world to be temporal,what it takes for a time to be the present, what it takes for time to pass, whatthe direction of time turns out to be, which propositions express the laws ofnature, and how to understand the causal relations between the occupantsof the block; your unpredictable morpher cannot compete with that! Well,given the ongoing disagreements in the literature, maybe “compellingstories” is a bit overstated . . . but true enough.13 Depending on its symme-tries, a morphing block might well turn out to be a spatial-only (i.e. non-temporal) world. Alternatively, a morphing block might turn out to havemore than one distinguished hyperplane that deserves the title “present.”Then again, in a morphing block timemight reverse its flow or perhaps timemight happen to run in more than one direction.

My goal here, however, is not to provide and defend answers to suchquestions. I am content to introduce the morphing block theory and to

13 For a representative discussion of some of these far-ranging topics, see Skow (2007) and a number ofthe essays in Callender (2011).

242 hud hudson

acknowledge some of the troublesome questions that immediately arise. Asindicated above, motivation for taking the morphing block theory seriouslycan be had straightforwardly by reflecting on common presuppositions ofthe growing block, the shrinking block and the disappearing branch theo-ries and by maintaining that the primary controversial features of themorphing block theory are already to be found amongst these rivals.Accordingly, to ignore or exclude it while embracing its cousins is to playmetaphysical favorites with no adequate justification. In the remainder ofthis chapter I would like to explore some of the philosophical consequencesof taking the morphing block theory seriously. Anyone who wishes toundermine the metaphysical possibility of a morphing block – whileretaining that favored status for growing, shrinking or branching blocksand while pleading innocent to the charge of unjustified metaphysicalfavoritism – can simply consult the three previous paragraphs for an arrayof targets at which to take aim.

cur i ou s con s equenc e s o f mor ph ing

Recall our two representatives of the complex view from p. 238: the animalist–endurantist–biological criterion theorist, who fills in the blank in (C1)with his favored way of indicating sameness of human animal, and theminimalist–perdurantist–psychological–criterion theorist, who fills in theblank in (C2) with his favored account of a psychological continuity andconnectedness relation between human person-stages.Despite their considerable philosophical differences, these theorists

(along with many of their like-minded, complex-theory-of-diachronic-personal-identity brethren) tend to share two particular commitments.First common commitment: each theorist is responsible for explaining theappropriateness of the adjective “human” in the label “human person.”Usually, this is accomplished by acknowledging some significant relationbetween the person and a human organism (although, as we have seen, therelation need not be identity). Moreover, human organisms are materialobjects that have a certain sort of history: that is, nothing that came intoexistence suddenly and looking for all the world like Achilles at age twentywould be a genuine instance of the kind “human person.” Even though itmay exemplify all the cognitive and moral features we ordinarily associatewith human persons, to the extent that it had the wrong historical proper-ties, it would not be a human person. Second common commitment: when weask after the animalist’s account of sameness of human animal or theminimalist’s account of the psychological continuity and connectedness


relation, we will find that they often appeal to certain theses about causationand causal dependence. Without resort to the right sort of causal depend-ence of later features of an enduring person on earlier features of that person(alternatively – of later person-stages of a perduring person on earlierperson-stages of that person), there are no compelling ways to divide thecases of genuine persistence of a single person from cases of the seamlessspatio-temporal replacement of one person by a distinct but extremelysimilar one.

Consider a formulation of the causal requirement stated in languagefriendly to the minimalist’s perdurantist metaphysic:

(CR) Necessarily, if a human person, P, persists throughout an open temporalinterval, T, then for every instant, t, in T, there is an open interval of time, T *, witht as its point-limit such that the sum of P ’s temporal parts that exist during T * is apartial cause of P ’s temporal part at t.14

Finally, we have the ingredients for the problem facing our complex theoriesof diachronic personal identity.

Suppose ours is a morphing block space-time which (up until now) hasbehaved as would a growing block space-time – monotonously and uni-directionally adding new hyperplanes to reality as the moments tick by,while also faithfully preserving everything once added since the Big Bang.Our friend, Althea, a human person, was born in 1975 and is standing herealong with us (in virtue of her most current momentary person-stage) on thevery edge of Being: that is, on the hyperplane that is presently on theoutermost surface of the block. In five minutes, however, our morphingblock will show its true colors and will no longer masquerade as a meregrowing block, for in five minutes, a portion of the hyperplanes identifiedwith the 1980s will disappear. In particular, some portion of those hyper-planes that contain all of Althea’s 1980 temporal parts will simply beannihilated. Moreover, in ten minutes, a portion of the hyperplanes thatcontain all of Althea’s 1970 temporal parts will be annihilated as well.

I suppose a number of people would not regard the obliteration of certainportions of the 1980s or the 1970s as a tragedy, but what of our friendAlthea? What will become of her five and then ten minutes hence? Forstarters, we should note that at the five-minute mark our space-time will

14 This principle of immanent causation for human persons is a modified version of a more generalprinciple in Zimmerman (1999), in which “human person” is replaced by “material object.”WhereasI regard the general principle appropriate for most familiar sortal terms, “material object” can apply tofusions of objects with no causal connection whatsoever. On this last point, see Hudson (2006, ch. 5).

244 hud hudson

have a new shape – “a block with an internal cavity” is not quite right, sincecavities are themselves located at regions, but it gets the idea across. Withthat rough and ready observation on shape change, then, let us turn ourattentions to Althea.First let us note that the 1980s will still exist; the hyperplanes identified

with those times will have a hole in them (so to speak), but otherwise theywill be as they were when they occupied the outermost positions on theblock. If we could examine the contents of the block from 1975 to thepresent, we would find a temporally connected object spanning somemoment in 1975 to the first moment of 1980 that looks much like ahuman person from conception to the tender age of five. We would alsofind a temporally connected object spanning the first moment of 1990 to thepresent that looks much like a human person from age fifteen to thirty-seven. We would be inclined to declare that these are two temporallyextended person slices of one and the same human person (and indeed,were we to make that assertion now – premorph – we would be unpro-blematically right). But in just a few minutes we would face an obstacle –namely, that there will be no temporally connected object with the appear-ance of a five- to fifteen-year-old to be found anywhere in the 1980s that hasthe slightest chance of standing in the slice-of-the-same-person-as relationwith our other two objects. If Althea (who will still be standing with us fiveminutes hence) was born in 1975 (as she will continue to profess), then shewill be a temporally gappy person who will suffer from false memories of analleged ten-year period in her life that simply never took place.Some clarification: since (by hypothesis) the regions that contain Althea’s

1980 temporal parts will be obliterated, all the objects and events theycontain (e.g. her first kiss, her middle-school graduation, that scar fromher hiking accident) will also be removed from the world’s stage.Significantly, though, nothing at all happens to her temporal slices thatoccupy the 1990s onward. Accordingly, whereas she can now recall that firstkiss almost as vividly as she experienced it, in five minutes she will bemistaken to believe that she had that experience at all or that she attendedmiddle school or that the scar above her left eye is the result of anunfortunate tumble down a cliffside.Is she fated, then, to be temporally gappy and deluded by apparent

memories? Two questions there. Whether or not she is temporally gappywill be determined (in part) by whether or not she will still satisfy the causalrequirement introduced above. According to (CR) this would require that“the-conception-to-five” slice is the right kind of partial cause of “the-fifteen-to-thirty-seven” slice, despite not having temporally intermediate


objects to carry the requisite causal signals and despite the presence of manyfeatures in the latter object that seem to lack a cause altogether.

Whether she will be deluded by apparent memories is a tricky question.Note that on the model on offer what it takes (from the perspective of sometime, t) for an event to have happened is that when t is the present moment,the event is located in the block at times earlier than t. Accordingly, as thingsstand now – premorph – the memory of her first kiss is veridical; that eventreally happened. And accordingly, five minutes hence – post-cavity – thememory of her first kiss will be merely illusory; it will be true that that eventnever really happened. A sad state of affairs, but Althea is a strong humanperson – perhaps she can come to live with it. But even worse things are instore for her.

At the ten-minute mark, our space-time will have morphed again, and wecan ask after Althea a second time. Again, although a decade’s worth of morehyperplanes will now sport holes (of a sort), the 1970s and 1980s will stillexist, and otherwise those times will be as they were when they successivelyoccupied the outermost positions on the block. Undoubtedly, Althea willstill report 1975 as her birth year (but she will be wrong). Lots of things willhave happened in the 1970s but none of them will have happened to Althea,for she will not have any temporal parts located in any portion of that ten-year span. In fact, it will be true that Althea came into existence in 1990looking and thinking and feeling much like a fifteen-year-old girl (albeit onewho is wildly mistaken about her past). She will not be temporally gappy,but she will not be human either – and this, despite the fact that nothingwill have happened to her.

Why will she be non-human and how can we say nothing will havehappened to her if she becomes non-human? Two questions there: as wenoted above, our persistence theorists share a commitment to explaining theappropriateness of the adjective “human” in the label “human person,” andthis is accomplished by somehow relating a human person to a humanorganism. But human organisms are material objects that have a certain sortof history. Even though a thing may exemplify all the cognitive and moralfeatures we ordinarily associate with human persons, to the extent that ithad the wrong historical properties, it would not be a human person. Altheawill have the wrong historical properties – the wrong origins: that is why shewill be non-human.

Yet nothing will have happened to her. It is not as if she will undergosome change or transition from human to non-human (where “change” ishere read as qualitative difference between her temporal parts). For reasonsexactly similar to those which explain why five minutes hence she will be

246 hud hudson

deluded by apparent memories, ten minutes hence it will be true that shehas never had any pre-1990 temporal parts, for when the hyperplaneidentified with the time ten minutes hence is present, that is all of Altheathe block will contain. And the features had by the block precisely whenthat moment is present completely determine what is true of the past whenthat moment is present.These problems seem to be rooted in a simple observation. Just which

temporal parameters (and parts) a thing has are sensitive to the machina-tions of a morphing block, and the consequences of such sensitivity aresevere.

s e v e r i t y and s k e p t i c i sm

Let us return to our discussion of Althea’s fate five minutes hence, for it isthat scenario in which we most clearly face the difficulties for our complextheories of diachronic personal identity. To recap: if growing or shrinking orbranching blocks are taken with all metaphysical seriousness, then a morph-ing block is worth considering seriously as well. But such consideration hasconsequences for judgments of diachronic personal identity. In our morph-ing block scenario, whether it will be true that Althea was born in 1975depends on whether she will satisfy the causal requirement (CR) – or at leastthat is what our minimalist–perdurantist–psychological–criterion theorist(from p. 238) contends. Our animalist–endurantist–biological criteriontheorist (also introduced on p. 238) maintains that it depends instead onwhether there is the right sort of causal dependence of the (enduring andwholly present) woman who will be standing next to us five minutes henceupon a certain (enduring and wholly present) little girl in 1979. But whetherit ultimately relates to mere slices or whole persons, the shared commitmentto causal dependence remains the best strategy for these theorists when itcomes to telling apart cases of genuine persistence from cases of seamlessreplacement by a doppelgänger.Taking the morphing block theory seriously, however, may generate

rather unpalatable skeptical scenarios. Anyone who accepts as an epistemicpossibility that his is a morphing block world will shortly find himself in theuncomfortable position of being in serious doubt about whether he was infact born in the year he currently believes is his birth year. Moreover, this isno “maybe-the-world-was-created-five-minutes-ago” hypothesis, for – andthis is the truly surprising bit – such doubt can reasonably arise even if werealize that he currently knows perfectly well the year of his birth. That is, it


will put him in the position of endorsing an analogue of this speech offeredby one of Althea’s worldmates:

Even if I assume that I currently know that Althea exists here and now and thatI also know that she is numerically identical to that little girl who existed there andthen, five minutes hence I may not possess that knowledge (despite the fact thatnothing of significance will have happened to me during the interval and despitethe fact that “Althea” and “that little girl” will continue to successfully refer toindividuals in space-time).

A morphing block scenario would, of course, generate a large number ofsuch skeptical theses, but this seems a particularly impressive one. To theextent that we think such a skeptical doubt in Althea’s worldmates would beuntenable (or crazy) we face a choice: either reject the metaphysical possi-bility of the morphing block that gives rise to the causal worry or else rejectthe theories of personal identity that share the causal dependencerequirement.

Should the metaphysical possibility of the morphing block seem securedby the metaphysical possibility of the growing or the shrinking or thebranching blocks (and by their relevant similarities), we thereby acquiresome motivation for a simple view of diachronic personal identity. Aproponent of a simple view need not resort to the causal dependence claimsarticulated in (CR), since the little girl may simply be the grown woman,and that is all there is to say – there simply are no further, deeper facts thatare relevant to judgments of identity. Consequently, a proponent of asimple view can countenance a wider range of theories about the natureof time without facing bizarre threats to the identification of herself and herfriends with certain individuals in the past.

That is a modest but interesting advantage in the personal identity wars.15

15 Thanks for delightful comments, conversation and criticism to Ross Cameron, Peter Forrest, MeghanSullivan and Ryan Wasserman.

248 hud hudson

References

Adams, Robert 1979. “Primitive Thisness and Primitive Identity,” Journal ofPhilosophy 76: 5–26.

1987. “Divine Necessity,” in Robert Adams, The Virtues of Faith and Other Essaysin Philosophical Theology. New York: Oxford University Press, pp. 209–20.

Armstrong, David 1980. “Identity Through Time,” in Peter van Inwagen (ed.),Time and Cause: Essays Presented to Richard Taylor. Dordrecht: Reidel,pp. 67–78.

Baker, Lynne Rudder 2000. Persons and Bodies: A Constitution View. CambridgeUniversity Press.

2005. “Death and the Afterlife,” in William J. Wainwright (ed.), The OxfordHandbook of Philosophy of Religion. Oxford University Press, pp. 366–91.

2007a. “First-Person Externalism,” The Modern Schoolman 84: 155–70.2007b. “Naturalism and the First-Person Perspective,” in Georg Gasser (ed.),How Successful is Naturalism? Heusenstamm: Ontos Verlag, pp. 203–27.

2007c. “Social Externalism and the First-Person Perspective,” Erkenntnis 67:287–300.

2007d. The Metaphysics of Everyday Life. Cambridge University Press.2008. “AMetaphysics of Ordinary Things andWhyWeNeed It,” Philosophy 83:5–24.

2009. “Identity Across Time: A Defense of Three-Dimensionalism,” inLudger Honnefelder, Edmund Runggaldier, and Benedikt Schick (eds.),Unity and Time in Metaphysics. Berlin: Walter de Gruyter, pp. 1–14.

Barnett, David 2009. “Is Vagueness Sui Generis?,” Australasian Journal ofPhilosophy 87: 5–34.

2010. “You Are Simple,” in Robert C. Koons and George Bealer (eds.), TheWaning of Materialism. Oxford University Press, pp. 161–74.

Bealer, George 1996. “On the Possibility of Philosophical Knowledge,” inJames Tomberlin (ed.), Philosophical Perspectives, vol. x. Cambridge: Blackwell,pp. 1–33.

Benovski, Jiri 2010. “Relational and Substantival Ontologies, and the Nature andthe Role of Primitives in Ontological Theories,” Erkenntnis 73: 101–21.

Black, Max 1952. “The Identity of Indiscernibles,” Mind 61: 153–64.

249

Boethius, Anicius Manlius Severinus 1988. Contra Eutychen et Nestorium, inTheological Tracts, ed. M. Elsässer. Hamburg: Meiner.

Bolinger, D. 1972. Degree Words. The Hague: Mouton.Bourne, Craig 2002. “When Am I?,” Australasian Journal of Philosophy 80: 359–71.Braddon-Mitchell, David 2004. “How Do We Know It Is Now Now?,” Analysis

64: 199–203.Broad, C.D. 1923. Scientific Thought. London: Kegan Paul.Butler, Joseph 1975 [1736]. “Of Personal Identity,” in John Perry (ed.), Personal

Identity. Berkeley and Los Angeles: University of California Press, pp. 99–106.Callender, Craig (ed.) 2011. The Oxford Handbook of Philosophy of Time. Oxford

University Press.Chalmers, David 1996. The Conscious Mind. Oxford University Press.Chisholm, Roderick M. 1969. “The Loose and Popular and the Strict and

Philosophical Senses of Identity,” in Norman S. Care and Robert H. Grimm(eds.), Perception and Personal Identity. Cleveland, OH: Case Western ReserveUniversity, pp. 82–106.

Chisholm, RoderickM. 1970a. “Identity Through Time,” in Howard E. Kiefer andMilton K. Munitz (eds.), Language, Belief, and Metaphysics. Albany: StateUniversity of New York Press, pp. 163–82.

1970b. “Reply to Strawson’s Comments,” in Howard E. Kiefer and MiltonK. Munitz (eds.), Language, Belief, and Metaphysics. Albany: State Universityof New York Press, pp. 187–9.

1976. Person and Object: A Metaphysical Study. La Salle, IL: Open Court.1986. “Self-Profile,” in Radu J. Bogdan (ed.), Roderick M. Chisholm. Dordrecht:

Reidel, pp. 3–77.1989. On Metaphysics. Minneapolis: University of Minnesota Press.

Cross, Richard 2010 [2003]. “Medieval Theories of Haecceity,” in Edward N. Zalta(ed.), The Stanford Encyclopedia of Philosophy (Fall edition), available at:http://plato.stanford.edu/archives/fall2010/entries/medieval-haecceity/.

Davidson, Donald (ed.) 1980 [1969]. “The Individuation of Events,” inDonald Davidson, Essays on Actions and Events. Oxford: Clarendon Press,163−80.

Davis, Stephen T. 2010. “Resurrection, Personal Identity, and the Will of God,” inGeorg Gasser (ed.), Personal Identity and the Resurrection. Farnham: Ashgate,pp. 19–31.

Descartes, René 1984 [1641]. “Meditations on First Philosophy,” inJohn Cottingham, Robert Stoothoff and Dugald Murdoch (eds.), ThePhilosophical Writings of Descartes. Cambridge University Press.

Earman, John 1989. World Enough and Space-time. Cambridge, MA: MIT Press.Evans, Gareth 1978. “Can There Be Vague Objects?,” Analysis 38: 208.Field, Hartry 2003. “No Fact of the Matter,” Australasian Journal of Philosophy 81:

457–80.Fine, Kit 1994. “Ontological Dependence,” Proceedings of the Aristotelian Society 95:

269–90.2001. “The Question of Realism,” Philosopher’s Imprint 1: 1–30.

250 References

http://plato.stanford.edu/archives/fall2010/entries/medievalaecceity/

Forrest, Peter 2004. “The Real but Dead Past: A Reply to Braddon-Mitchell,”Analysis 64: 358–62.

Foster, John 1991. The Immaterial Self: A Defence of the Cartesian Conception of theMind. London: Routledge.

Frege, Gottlob 1953 [1884]. The Foundations of Arithmetic, trans. J. L. Austin.Oxford: Blackwell.

1960 [1892]. “On Sense and Reference,” in Peter T. Geach andMax Black (eds.),Translations from the Philosophical Writings of Gottlob Frege, 2nd edn. Oxford:Blackwell, pp. 56−78.

French, Steven 2006. “Identity and Individuality in Quantum Theory,” in EdwardN. Zalta (ed.), Stanford Encyclopedia of Philosophy (Spring edition), availableat: http://plato.stanford.edu/archives/spr2006/entries/qt-idind/.

Gallois, André 1998. Occasions of Identity: A Study in the Metaphysics of Persistence,Change, and Sameness. Oxford University Press.

Geach, Peter 1976. “Two Kinds of Intentionality?,” The Monist 59: 306–20.Gilbert of Poitiers 1966. The Commentaries on Boethius, ed. N.M. Häring.

Toronto: Pontifical Institute of Mediaeval Studies.Gracia, Jorge J. E. 1994. Individuation in Scholasticism: The Middle Ages and the

Counterreformation. New York: State University of New York Press.Hasker, William 2001. The Emergent Self. Ithaca, NY: Cornell University Press.Hauser, Marc 2009. “The Mind,” Scientific American (September edition): 44–9.Hawthorne, John 2006. “Epistemicism and Semantic Plasticity,” in JohnHawthorne

(ed.),Metaphysical Essays. Oxford University Press, pp. 185–211.Heathwood, Chris 2005. “The Real Price of the Dead Past: A Reply to Forrest and

Braddon-Mitchell,” Analysis 65: 249–51.Hick, John 1990. Philosophy of Religion, 4th edn. Englewood Cliffs, NJ: Prentice

Hall.Hirsch, Eli 1982. The Concept of Identity. New York: Oxford University Press.Hossack, Keith 2006. “Vagueness and Personal Identity,” in FraserMacBride (ed.),

Identity and Modality. Oxford University Press, pp. 221–41.Howard-Snyder, Frances 2005. “It’s the Thought That Counts,” Utilitas 17: 265–81.Hudson, Hud 2001. A Materialist Metaphysics of the Human Person. Ithaca, NY:

Cornell University Press.2006. The Metaphysics of Hyperspace. Oxford University Press.2007. “I Am Not an Animal!,” in Peter van Inwagen and Dean Zimmerman(eds.), Persons: Human and Divine. Oxford University Press, pp. 216–34.

2012. “Transreceptacle Identity,” in Shieva Kleinschmidt (ed.), Mereology andLocation. Oxford University Press, in press.

Kanzian, Christian 2009. Ding – Substanz – Person. Heusenstamm: Ontos.Kripke, Saul 1981 [1972]. Naming and Necessity. Oxford: Blackwell.Leonard, Henry S. and Goodman, Nelson, 1940. “The Calculus of Individuals and

its Uses,” Journal of Symbolic Logic 5: 45–55.Lewis, David 1976. “Survival and Identity,” in Amelie Rorty (ed.), The Identities of

Persons. Berkeley: University of California Press, pp. 17–40.1979. “Attitudes De Dicto and De Re,” Philosophical Review 88: 513–43.

References 251

http://plato.stanford.edu/archives/spr2006/entries/qtdind/

1980. “A Subjectivist’s Guide to Objective Chance,” in Richard Jeffrey (ed.),Studies in Inductive Logic and Probability, vol. ii. Berkeley: University ofCalifornia Press, pp. 263–93.

1981. Philosophical Papers, vol. i. New York: Oxford University Press.1983. “NewWork for a Theory of Universals,” Australasian Journal of Philosophy

61: 343–77.1984. “Putnam’s Paradox,” Australasian Journal of Philosophy 62: 221–36.1986. On the Plurality of Worlds. Oxford: Blackwell.1997. “NewWork for a Theory of Universals,” in D.H. Mellor and Alex Oliver

(eds.), Properties. Oxford University Press, pp. 173–87.Locke, John 1975 [1690]. An Essay Concerning Human Understanding, ed.

P.H. Nidditch, Oxford: Clarendon Press.Lowe, E. J. 1988. “Substance, Identity and Time,” Proceedings of the Aristotelian

Society 62: 61–100.1989a. “Impredicative Identity Criteria and Davidson’s Criterion of Event

Identity,” Analysis 49: 178–81.1989b. Kinds of Being. Oxford: Blackwell.1989c. “What is a Criterion of Identity?,” Philosophical Quarterly 39: 1–21.1994. “Primitive Substances,” Philosophy and Phenomenological Research 54:531–52.

1995. Locke on Human Understanding. London and New York: Routledge.1996. Subjects of Experience. Cambridge University Press.1997. “Objects and Criteria of Identity,” in Bob Hale and Crispin Wright (eds.),

A Companion to the Philosophy of Language. Oxford: Blackwell, pp. 613–33.2000. An Introduction to the Philosophy of Mind. Cambridge University Press.2001. “Identity, Composition and the Simplicity of the Self,” in Kevin Corcoran

(ed.), Soul, Body, and Survival: Essays on the Metaphysics of Human Persons.Ithaca, NY: Cornell University Press, pp. 139–58.

2003. “Substance Causation, Persons, and Free Will,” in Christian Kanzian,Josef Quitterer and Edmund Runggaldier (eds.), Persons. Vienna: OeBV-hpt,pp. 76–88.

2009. More Kinds of Being: A Further Study of Individuation, Identity and theLogic of Sortal Terms. Malden, MA: Blackwell.

Lund, David 2005. The Conscious Self. Amherst, MA: Humanity Books.Lutz-Bachmann, Matthias 1983. “Person und Natur in den Opuscula Sacra des

A. M. S. Boethius,” Theologie und Philosophie 58: 48–70.Madell, Geoffrey 1981. The Identity of the Self. Edinburgh University Press.Markosian, Ned 1993. “How Fast Does Time Pass?,” Philosophy and

Phenomenological Research 53: 829–44.Mavrodes, George 1977. “The Life Everlasting and the Bodily Criterion of

Identity,” Noûs 11: 27–39.Merricks, Trenton 1997. “Fission and Personal Identity over Time,” Philosophical

Studies 88: 163–86.1998a. “Against the Doctrine of Microphysical Supervenience,” Mind 107:

59–71.

252 References

1998b. “There Are No Criteria of Identity over Time,” Noûs 32: 106–24.1999. “Endurance, Psychological Continuity, and the Importance of PersonalIdentity,” Philosophy and Phenomenological Research 59: 983–97.

2001a. “How to Live Forever without Saving Your Soul,” in Kevin Corcoran(ed.), Soul, Body, and Survival: Essays on the Metaphysics of Human Persons.Ithaca, NY: Cornell University Press, pp. 183–200.

2001b. “Realism About Personal Identity over Time,” in “Metaphysics,” specialissue of Philosophical Perspectives 15: 173–87.

2003. “Maximality and Consciousness,” Philosophy and PhenomenologicalResearch 66: 150–8.

2006. “Goodbye Growing Block,” Oxford Studies in Metaphysics 2: 103–10.Mithen, Steve 2004. “Review Symposium of Andy Clark’s Natural-Born

Cyborgs,” Metascience 13: 163–9.Nagel, Thomas 1965. “Physicalism,” Philosophical Review 74: 339–56.

1986. The View from Nowhere. New York: Oxford University Press.Nerlich, Graham 1994. What Spacetime Explains. Cambridge University Press.Nida-Rümelin, Martine 2006. Der Blick von Innen. Frankfurt am Main:

Suhrkamp.2010. “An Argument from Transtemporal Identity for Subject−Body Dualism,”in Robert C. Koons and George Bealer (eds.), The Waning of Materialism.Oxford University Press, pp. 190–211.

2012. “The Conceptual Origin of Subject−Body Dualism,” in Annalisa Colliva(ed.), Self and Self-Knowledge. Oxford University Press, pp. 39−73.

Noonan, Harold W. 1989. Personal Identity. London: Routledge.2003. Personal Identity, 2nd edn. London: Routledge.2011. “The Complex and Simple Views of Personal Identity,” Analysis 71: 72–7.

Nozick, Robert 1981. Philosophical Explanations. Oxford University Press.O’Connor, Timothy and Jacobs, JonathanD. 2010. “Emergent Individuals and the

Resurrection,” European Journal for Philosophy of Religion 2: 69–88.Oderberg, David S. 1993. The Metaphysics of Identity over Time. New York: St.

Martin’s Press.Olson, Eric T. 1995. “Human People or Human Animals?,” Philosophical Studies 80

(2): 159−81.1997a. “Relativism and Persistence,” Philosophical Studies 88: 141–62.1997b. The Human Animal: Personal Identity Without Psychology. New York:Oxford University Press.

2002. “What Does Functionalism Tell Us about Personal Identity?,” Noûs 36:682–98.

2007. What Are We? A Study in Personal Ontology. Oxford University Press.2010. “Personal Identity,” in Edward N. Zalta (ed.), Stanford Encyclopedia ofPhilosophy, available at: http://plato.stanford.edu/entries/identity-personal/.

Parfit, Derek 1982. “Personal Identity and Rationality,” Synthese 53: 227–41.1984. Reasons and Persons. Oxford University Press.

Perry, John 1972. “Can the Self Divide?,” Journal of Philosophy 69: 463–88.1979. “The Problem of the Essential Indexical,” Noûs 3: 3–21.

References 253

http://plato.stanford.edu/entries/identity-rsonal/

2008 [1975]. “The Problem of Personal Identity,” in John Perry (ed.), PersonalIdentity. Berkeley: University of California Press, pp. 3–30.

Pinker, Steven 1977. How the Mind Works. New York: W. W. Norton.Putnam, Hilary 1975. “TheMeaning of ‘Meaning’,” in Hilary Putnam (ed.),Mind,

Language and Reality. Philosophical Papers, vol. ii. Cambridge UniversityPress, pp. 215–71.

Quinton, Anthony 1962. “Spaces and Times,” Philosophy 37: 130–47.Reid, Thomas 1975 [1785]. “Of Mr. Locke’s Account of Personal Identity,” in

John Perry (ed.), Personal Identity, 1st edn. Berkeley and Los Angeles:University of California Press, pp. 113–18.

2002 [1785]. Essays on the Intellectual Powers of Man, ed. D. Brookes. UniversityPark: Pennsylvania State University Press.

Rorty, Amelie (ed.) 1976. The Identities of Persons. Berkeley: University ofCalifornia Press.

Schaffer, Jonathan 2009. “On What Grounds What,” in David Chalmers,David Manley and Ryan Wasserman (eds.), Metametaphysics. OxfordUniversity Press, pp. 347–83.

Shoemaker, Sydney 1963. Self-Knowledge and Self-Identity. Ithaca, NY: CornellUniversity Press.

1970. “Persons and their Pasts,” American Philosophical Quarterly 7: 269–85.1979. “Identity, Properties, and Causality,” Midwest Studies in Philosophy 4:

321–42; reprinted in his Identity, Cause, and Mind, expanded edn. Oxford:Clarendon Press, 2003.

1984. “Personal Identity: A Materialist’s Account,” in Sydney Shoemaker, andRichard Swinburne, Personal Identity. Oxford: Blackwell, pp. 67–132.

1985. “Critical Notice of Reasons and Persons,” Mind 94: 443–53; reprinted as“Parfit on Identity,” in J. Dancy (ed.), Reading Parfit, Oxford: Blackwell, 1997.

1994. “Introspection and the Self,” in Quassim Cassam (ed.), Self-Knowledge.Oxford University Press, pp. 118–39.

1999. “Self, Body, and Coincidence,” Proceedings of the Aristotelian Society,supplementary vol. 73: 287–306.

2007. Physical Realization. Oxford University Press.Shoemaker, Sydney and Swinburne, Richard 1984. Personal Identity. Oxford:

Blackwell.Sider, Theodore 1996. “Naturalness and Arbitrariness,” Philosophical Studies 81:

283–301.2000. “Criteria of Personal Identity and the Limits of Conceptual Analysis,”

Philosophical Perspectives 15: 189–209.2001. Four-Dimensionalism: An Ontology of Persistence and Time. Oxford

University Press.2003. “Maximality and Microphysical Supervenience,” Philosophy and

Phenomenological Research 66: 139–49.2012. Writing the Book of the World. Oxford University Press.

Simons, Peter 2006. “Real Wholes, Real Parts: Mereology without Algebra,”Journal of Philosophy 103: 597–613.

254 References

Singer, Peter 1978. Practical Ethics. Cambridge University Press.Skow, Bradford 2007. “What Makes Time Different From Space?,” Noûs 41:

227–52.Slote, Michael 1985. Common-Sense Morality and Consequentialism. London:

Routledge.Smart, J. J. C. 1949. “The River of Time,” Mind 58: 483–94.

1998. “The Space-Time World: An Excerpt from Philosophy and ScientificRealism,” in Peter van Inwagen and Dean Zimmerman (eds.), Metaphysics:The Big Questions. Oxford: Blackwell, pp. 94–101.

Sorenson, Roy 2001. Vagueness and Contradiction. Oxford University Press.Strawson, Peter F. 1959. Individuals: An Essay in Descriptive Metaphysics. London:

Methuen.Suarez-Nani, Tiziana 2009. “Singularité et individualité selon Pierre Auriol,” in

Stephen F. Brown, Thomas Dewender and Theo Kobusch (eds.),Philosophical Debates at Paris in the Early Fourteenth Century. Leiden: Brill,pp. 339–57.

Swinburne, Richard 1973–4. “Personal Identity,” Proceedings of the AristotelianSociety 74: 231–47.

1984. “Personal Identity: The Dualist Theory,” in Sydney Shoemaker andRichard Swinburne, Personal Identity. Oxford: Blackwell, pp. 3–66.

1986. The Evolution of the Soul. Oxford: Clarendon Press.1995. “Thisness,” Australasian Journal of Philosophy 73: 389–400.1997. The Evolution of the Soul, rev. edn. Oxford University Press.2007. “From Mental/Physical Identity to Substance Dualism,” in Peter vanInwagen and Dean Zimmerman (eds.), Persons: Human and Divine. OxfordUniversity Press, pp. 142–65.

2010. “In Defence of Logical Nominalism: Reply to Leftow,” Religious Studies 46:311–30.

Thomson, Judith Jarvis 2008. “Persons and their Bodies,” in Theodore Sider,John Hawthorne and Dean Zimmerman (eds.), Contemporary Debates inMetaphysics. Oxford: Blackwell, pp. 155–76.

Tooley, Michael 1997. Time, Tense and Causation. Oxford University Press.Unger, P. 1990. Identity, Consciousness, and Value. Oxford University Press.Van Cleve, James 1987. “Right, Left and the Fourth Dimension,” Philosophical

Review 96: 33–68.van Inwagen, Peter 1990a. “Four-Dimensional Objects,” Noûs 24: 245–55.1990b. Material Beings. Ithaca, NY: Cornell University Press.2006. “Can Mereological Sums Change their Parts?,” Journal of Philosophy 103:614–30.

Wasserman, Ryan 2004a. “Framing the Debate over Persistence,” Metaphysica 5:67–80.

2004b. “Indeterminacy, Ignorance and the Possibility of Parity,” PhilosophicalPerspectives 14: 391–403.

Wiggins, David 1980. Sameness and Substance. Oxford: Blackwell.2001. Sameness and Substance Renewed. Cambridge University Press.

References 255

Wilkes, Kathleen 1988. Real People: Personal Identity without Thought Experiments.Oxford University Press.

Williams, Bernard 1970. “The Self and the Future,” Philosophical Review 79:161–80.

Williamson, Timothy 1994. Vagueness. London: Routledge.Wright, Crispin 1987. “Further Reflections on the Sorites Paradox,” Philosophical

Topics 15: 227–90.Zimmerman, Dean 1991. “Two Cartesian Arguments for the Simplicity of the

Soul,” American Philosophical Quarterly 28: 217–26.1997. “Immanent Causation,” in “Mind, Causation, and World,” special issue of

Philosophical Perspectives 11: 433–71.1998. “Criteria of Identity and the ‘Identity Mystics’,” Erkenntnis 48: 281–301.1999. “The Compatibility of Materialism and Survival: The ‘Falling Elevator’

Model,” Faith and Philosophy 16: 194–212.2003. “Material People,” in Michael Loux and Dean Zimmerman (eds.), The

Oxford Handbook of Metaphysics. Oxford University Press, pp. 491–526.2004. “Should a Christian Be a Mind-Body Dualist?,” in Michael L. Peterson

and Raymond J. VanArragon (eds.), Contemporary Debates in Philosophy ofReligion. Malden, MA: Blackwell, pp. 315–27.

2009. “Properties, Minds, and Bodies: An Examination of Sydney Shoemaker’sMetaphysics,” Philosophy and Phenomenological Research 78: 673–738.

2010. “From Property Dualism to Substance Dualism,” Proceedings of theAristotelian Society, supplementary vol. 84: 119–50.

256 References

Index

agency 174amnesia 6, 28animal

human 4–5, 88, 89, 90, 92, 94, 238, 243non-human 190

animalism 4–6, 7, 91, 131, 145–6artifacts 108

Baker, Lynne Rudder 13, 59, 179, 181, 182, 184, 191Barnett, David 17, 73Benovski, Jiri 193block theory of time

growing 238–40, 241, 242morphing 236, 241, 242–3, 244, 247–8shrinking 238–40, 242

Boethius 195boundary between simple and complex view 44,

50, 55brain

damage 5, 48identity 151, 152transplantation of 6, 7, 88, 95, 151

Brown and Brownson 6, 88, 90, 95, 100Butler, Joseph 60, 85, 123, 148, 158, 207, 218

Cartesian Ego 210, 211Cartesianism see dualismcausation

holistic 227–9immanent 209, 231–4

Chisholm, Roderick 15, 54, 60, 63, 64, 85, 158, 180,185, 187, 207–11, 218–21

circularity of personal identity 125, 134–6, 143, 148,149, 151–5

co-consciousness 86, 87coincident entities 98, 100, 101complex view 3–10, 15, 44, 46–7, 49, 50, 51, 52, 54,

56–8, 60, 63, 64–6, 71, 73, 82, 85–8, 95, 108,123, 124, 127, 132–3, 134, 135, 136, 153, 179, 180,192–3, 204, 206–7, 211, 212, 236–8, 243

composition 17, 35, 186, 188, 201, 203, 205

compound of body and soul 51, 52conceivability 164, 234, 235conditions of identity see identityconnectednessphysical 65psychological 63, 65, 238, 243

constitution relation 181, 187, 191constitution view 191constitutional basis 158–65, 173continuer 94psychological 91–3

continuitybiological 90, 135bodily/material 3, 8, 12, 13, 44, 45, 51, 52, 55, 56,

58, 65, 87, 90, 108, 114, 115, 120, 179, 206,207, 220, 221

of brain 106–8, 119, 120of form 203of life 5, 203of memory 106–8, 114, 115, 132of stuff 56, 87psychological 3, 6–9, 12, 13, 44, 51, 53, 55–6, 60,

63, 65, 86, 87, 89–92, 95, 108, 114, 132, 179,206, 207, 218, 220, 221, 238, 243

spatio-temporal 127, 128, 129, 130, 200counterparts 125perfect 159, 161–73, 174, 175

criteria of identity see identity

Davidson, Donald 143, 149, 150, 152Descartes, René 144dualismCartesian 49–50, 52, 55–6, 57, 58mind–body 123

duplication problem 184, 185

embryo 45emergentism 190, 204and materialism see materialism

endurantism 237, 238epistemicism 71–4

257

eternalism 238, 239Evans, Gareth 72, 89experiencing subjects 175

first-person memory 130, 131, 176first-person perspective 10, 15, 16, 144, 180–2, 184,

185, 188–91first-person reference 89, 90, 91, 92, 182fission and fuzziness 222, 223, 224, 227, 233,

234, 235fission argument 222, 224, 229fission case 11, 15, 16, 148, 224fission scenario see fission caseform

individual 192, 200, 202–3, 205Foster, John 58, 180, 183four-dimensionalism 11, 13, 53, 133Frege, Gottlob 140, 141, 142, 155French, Steven 119fusion 76, 148, 171, 185, 225–7, 240, 241fuzziness 221, 223–4

Gallois, André 121Geach, Peter 207Gilbert of Poitiers 194Goodman, Nelson 185graduality 8, 13, 15

Hasker, William 204–5Hauser, Marc 190Hick, John 59Hirsch, Eli 201Howard-Snyder, Frances 78human vegetable 7Hume, David 56, 57, 85, 193, 206

identityanti-criterionalism of 49–52biological criterion of 198, 238constitutive criterion of 47, 131, 133,153, 154

informative criterion of 3, 14, 17, 134, 137, 151,152, 153, 156

Locke’s criterion of 89–93, 135, 146–54, 155non-circular criterion of see identity,informative criterion of

one-level criterion of 141, 142, 147, 151,152, 155

psychological criterion of 16–17, 198, 238specific criterion of 61transitivity of 11, 13, 127, 140two-level criterion of 87, 141–2, 151–2,155, 156

immaterial substance 144, 145, 180, 183, 191, 208immaterialism 180

impersonal facts 63, 221, 224indeterminacy

epistemic 97ontic 97

individuation of conscious states 132, 153

Jacobs, Jonathan D. 227, 229

Kant, Immanuel 58Kripke, Saul 116, 118

laws of nature 31, 39, 242Leibniz’s law 138, 140Leonard, Henry S. 185Lewis, David 8, 54, 77, 82, 85, 87, 89, 92, 94,

180, 206living being 4–7, 45, 52, 88, 135, 181, 182, 189, 190,

194, 196, 198, 201, 202, 217, 233, 238location 241Locke, John 52, 132, 137, 139, 143–6, 149, 151, 152,

153, 154, 193, 198, 206Lowe, E. J. 3, 14, 17, 132–4, 141, 143, 144, 146, 150,

152, 153, 156, 158Lund, David 10Lutz-Bachmann, Matthias 196

Madell, Geoffrey 15materialism 237

and dualism 206emergent 220, 223–7, 229, 233, 234

mereology 185–7Merricks, Trenton 47, 50, 60, 226–7, 229Mithen, Steve 189

Nagel, Thomas 4neo-Lockean account of personal identity see

identityNida-Rümelin, Martine 16Non-reductionist account of personal identity

207–8, 210Noonan, Harold 8, 10, 13, 56–7, 179,

183, 184Nozick, Robert 10

occupation see locationO’Connor, Timothy 227, 229Oderberg, David S. 201Olson, Eric 1, 4, 89, 145, 236only x and y principle 10organism see living being

human 2, 4, 6, 145, 171, 181, 189, 196, 202, 227,237, 243, 246

Parfit, Derek 6, 9, 11, 63–4, 70, 73, 85, 89, 96, 148,179, 180, 206–9, 213, 216, 233, 234

258 Index

partsessential 120material 202mereological 185–8, 191temporal 2, 11–13, 14, 53, 54, 75–6, 133, 209,225–7, 230–4, 237, 244–7

part–whole relation 185Perry, John 12, 13, 93, 206personal identity see identitypersonality change 6, 32, 38, 39, 40, 42personhood 2, 4, 46, 54, 84–9, 97, 144, 146, 180,

188, 191, 194, 195, 196, 197, 201, 226person-stages 7, 8, 54, 76, 87, 133, 134Pinker, Steven 190possibility

logical 108, 112, 115, 120metaphysical 15, 109, 115, 164, 166, 243, 248

presentism 238, 239primary kind 181properties

emergent 189, 220, 229essential 122, 234intrinsic 226, 228mental 15, 16, 49, 53, 54, 94, 105, 107, 108, 120,121, 122, 131, 135, 159, 172, 208, 209, 210, 216,229

physical 16, 105, 106, 108, 120, 121, 122, 128, 131,135

sortal 127, 128thin and thick 234

Putnam, Hilary 117

quasi-memory 132, 148, 216

reductionist account of personal identity 63, 136,206–7, 211, 213–14, 216, 231, 233

Reid, Thomas 60, 63, 68, 85, 114, 123, 147, 158, 183,207, 218

relationscausal 50, 59, 119, 124–7, 129, 131, 210, 217, 226,227, 228, 231, 232, 234, 242

spatial 37, 127, 213, 214temporal 121, 127, 213, 239

replacementof brain 115of parts 115of person 244

replacement machine 65, 66, 70, 73, 74, 75rigid designators 116, 117Rorty, Amelie 2

samenessof brain 146of person 211of soul 50, 55

Shoemaker, Sydney 6, 7, 10, 53, 54, 60, 85, 86, 88,89, 91, 92, 93–5, 96, 100, 119, 132, 134, 153–5,180, 207, 209, 234

Sider, Theodore 11, 12, 77, 226similarity 55, 65, 130, 141, 212, 218Simons, Peter 187simple view 1, 3, 14–15, 17, 44–6, 47, 48, 49, 50–2,

56–8, 60, 62, 85–6, 88–9, 105, 107, 108, 120,123, 124–7, 128, 129, 130, 136, 158, 179–80,182–3, 184, 185, 188, 191, 193, 204, 205, 206,207, 208, 218, 248

simples 188, 229, 230, 231, 232, 233Singer, Peter 196, 197skepticism 14, 130, 153, 247sortal term 85, 192, 194–6, 198, 199–200Strawson, Peter F. 149, 153substance dualism 51, 120, 122, 196, 220supervenience 212–13, 214–16, 219, 220, 221, 223,

226, 230–4Swinburne, Richard 15, 49, 50, 51, 52, 54, 55, 56,

64, 85–7, 88, 96, 152, 158, 180, 207, 218–21symmetry 16, 100, 140, 222

teletransportation 27–8, 32, 38, 39, 40, 41–2temporal gap 238theories of time 238, 240, 242thinking animal problem 4–5, 89, 94Tibbles and Tib 99too many thinkers problem see thinking animal

problem

Unger, Peter 52, 54unity of consciousness 204–5unity relation 124, 126, 127, 129, 156

vague objects 96–101vagueness 54, 71, 72, 76, 95, 98, 100, 111van Inwagen, Peter 52, 54, 88, 187

Wiggins, David 46, 148Wilkes, Kathleen 11Williams, Bernard 10Williamson, Timothy 71–2, 74, 75Wright, Crispin 72

Zimmerman, Dean 50, 59, 209, 234

Index 259

Documents

Personal identity: complex or simple?