We Don’t Have Identities

1 Introduction

I will argue that We Don’t Have Identities. I will do this by suggesting that the type of identity philosophers are investigating when they consider `personal identity’ does not apply to persons. They seek instead an identity which I will term “set identity.” Set identity applies only to abstract items such as sets, numbers and the like.

Types Of Identity

That type of identity bears the usual defining characteristics of reflexivity, transitivity and symmetry. What they find, however, is something I will term `legal identity’. A legal identity is exemplified by the appropriate beneficiary of a bequest. It can be determined by legal procedures. Since the methodologies applied are legalistic, they can only find legal identity; moreover, they suffer from Sorites and other problems. In Sorites problems, a major binary change — the identity of persons — changes based on apparently insignificant analog changes in underlying quantities. That type of difficulty will drive one sort of problem for people denying that We Don’t Have Identities.

It has been argued before1 that there are different types of identity for items such as numbers. Number have strict identity, and persons could have identity only in a loose sense.

The general argument is that set identity for numbers or sets or any abstract entity is clear and trivial. However, identity of persons is unclear and important, suggesting that there is a distinction between the two types of identity.

That type of distinction would be consistent with the line I am proposing in this essay. But a better terminology is to allow identity to abstract items only, because what concrete objects have is something much more approximate. Their identity fails to have the key criteria for identity of symmetry, transitivity and reflexivity.


The mistake all the authors I consider here make is either to conflate the two types of identity, or as I would put it, to seek set identity where there is none. They find only what we may term “legal identity.” They finding the just recipient of an inheritance. That is just not the same matter as determining who I am.

Types Of Personal Identity

There have been three types of account of personal identity, as listed below.

  1. Psychological continuity accounts, relying on the importance of memory.

2. Bodily continuity accounts — the most well known of which is termed `animalism’ — relying on the importance of the body.

3. Combination accounts involving both of the above aspects.

Problems With The Accounts

Psychological continuity theories fail because they do not allow sufficient weight to the importance of bodily continuity. They also fail for Sorites reasons. So We Don’t Have Identities on that account.

Bodily continuity theories fail because they do not allow sufficient weight to the importance of psychological continuity. Again they also fail for Sorites reasons. So We Don’t Have Identities on that account.

Combination accounts fail because they attempt to weigh incommensurable quantities. Since this exhausts the list of possible accounts, no account can succeed. We must conclude that persons do not have identity. So We Don’t Have Identities on that account either.

We Don’t Have Identities: Brain Transplants

These problems arise from the brain transplant cases discussed by Williams.2 He describes a scenario in which we wish to agree that a brain transplant has taken place. A person has acquired a new body in what appears to be little more than a development of standard organ transplants.

This results from our tacit assent to the proposition that psychological continuity is a sufficient condition for personal identity. Williams is then able to reformulate the description of the scenario such that although it is the same scenario, we now agree that bodily continuity is the sufficient condition, and psychological continuity is not important.

He is able to obtain our agreement to inconsistent positions, and we are unable to decide which of the two conditions is the most important. So we should already be on notice that no resolution of the problem of personal identity will be available. I will discuss Williams’s account in more detail below as it relates to the different types of account.

Plan Of Work

The various problems with psychological continuity accounts are discussed in s2. I will look at bodily continuity accounts in s3. Combination accounts are covered in s4. I will then conclude in s5 that the reason that all of these accounts fail for one reason. It is that identity does not apply to any objects in the world subject to temporal change. Persons are just a special case of this.

2 We Don’t Have Identities: Psychological Continuity Accounts

Psychological continuity accounts allow body transfer. If the brain of a person were transferred to a different body, the brain now in that new body would have the memories of the person. They would think they were the person. However, supporters of a bodily continuity account must deny that the person in the new body is identical with the pre-transplant person.

Locke’s original psychological continuity account allowed that “person A is the same as person B if A can remember having an experience of B’s.”3 There is little space in psychological continuity accounts to allow any importance to the body. We then fall into the fatal problems that Williams describes in the first half of his paper. I will outline those next.

2.1 Transplant Problem

Photo by Pixabay on

In the scenario Williams describes, two persons A and B enter a machine. They subsequently leave it in states such that we would be inclined to say they had swapped memories. Williams adopts neutral terminology of the A-body person. That person now inhabits the A body whoever that is. This he does in order not to prejudge the issue. Likewise the B-body person is the person now inhabiting the B-body.

The A-body person has B’s memories and objectives, etc. The B-body person has A’s memories and objectives, etc. The question is now put. If it were known before the transplant that one of the bodies was to be tortured, to which body should one request that it be applied? This is assuming that one acts solely in self interest.

Under the description of the first half of the paper, and also under the influence of the psychological continuity account, one naturally agrees that A will say that the torture should be applied to the A body. He is about to `leave behind’ that body for his new life in the B-body. The problem now though is that Williams can reformulate the scenario in terms that make this decision appear irrational.

The Amnesiac Intuition

He does this by appealing to what is the central intuition behind bodily continuity accounts. It might be termed the “amnesiac intuition.” It is known that one’s memory is imperfect. Also there appear to be people who lose large parts of their memory altogether. If this process is too extensive, we may say that that person is no longer present. But generally we are happy to talk as though the same person is there even if they have forgotten some things. This intuition then leads us to think that the body is the key factor in deciding identity.

Williams does not describe the same scenario involving A and B as a memory swap. Neither does he describe it as an identity swap for adherents of a psychological continuity account. Instead, he uses a minor extension of the amnesiac intuition.

Wwe imagine that the torture is to be applied after deletion of all memories — i.e. externally imposed amnesia. This is to be followed by the insertion of someone else’s memories. But this does not improve matters. If anything, it makes them much worse. The situation is loss of all memories of personal attachments, projects etc followed by torture. Described thus, the pre-transplant A person would be irrational to select the torture for the A-body person because he would be that person.

2.2 Brave Officer Problem

There are other reasons why a psychological continuity criterion cannot work. The first point to note is that it must at least be improved from Locke’s version. That’s because of what is known as the brave officer paradox.

We can imagine that a General remembers leading a cavalry charge as a junior officer. The junior officer remembers school days even though the General can no longer remember school days. This scenario has the outcome that the General is identical to the officer. Also the officer is identical to the schoolboy. But the General is not identical to the schoolboy. That outcome is unacceptable since identity is transitive. Therefore, Locke’s simple criterion must be modified. If not, We Don’t Have Identities.

Perry4 gives an outline of Grice’s modified account, which he considers the most successful psychological continuity account.5

Total Temporary States

The concept of Total Temporary States — `TTS’ — is introduced to refer to the set of total experiences of a person. We are to find a sequence of TTS’s which are linked by common memory elements. A later TTS would, given certain conditions, contain a memory of an experience that was part of an earlier TTS. When we have found a complete set of TTS’s to which no further TTS’s can be added, we have a person.

We will immediately want to ask how a person can be identical to a set. Moreover — and we will see this problem again with Nozick’s account in 4.1 — this approach is an example of covertly introducing sets. Sets do have identity; but they are abstract. It is not possible to find an account of identity in the concrete world by considering identity criteria for abstract objects.

Perry6 gives an account of how we are to apply the Grice procedure. If Smith examines a green cube, and Jones is hypnotized into thinking he has, then both will have a later TTS which includes the belief that they examined a green cube. However, Grice will say that only Smith is identical to the earlier Smith and Jones is not. That’s because Smith is really remembering and Jones is not. We now have to bear in mind the `given certain conditions’ caveat mentioned above. These conditions will include necessary weakenings. Perhaps Smith has a weak memory. He can remember examining a cube, but not what color it is. Or weaker still, he can remember examining a solid object, but not its shape or its color.

A Sorites Problem

Here the Sorites problem emerges. We will examine various very similar counterfactuals in order to decide whether this is Smith. His identity will depend on how good his memory is of a particular event. This is a continuously variable parameter. This will inevitably lead to difficulties when used to determining binary questions. We will see this type of problem three times in this essay — with all accounts of personal identity. So again, So We Don’t Have Identities.

3 We Don’t Have Identities: Bodily Continuity Accounts

The animalist denies that a brain transplant could suffice for a transfer of identity. This of course is in direct opposition to the intuition behind psychological continuity theories. There can be no fully justified victory in a contest between intuitions. Only the current state of progress in medicine means that we are denied the sight of brain transplant cases wherein the new body walks, talks, remembers, feels very similarly to the way the person in the old body did. The vividness of that experience would go a long way towards dissolving the appeal of the idea that identity is a useful concept in relation to persons, because it would bring to life the difficulties I outline in this essay and the problems that Williams describes in the second half of his paper (see 2.1.)

Madden7 provides an animalist response to the brain transplant cases. He reformulates the question of identity8 in terms of reference. We are to establish to which person utterances of an `I’-token refer after the brain transplant. The idea is that there could be a gradual shift of reference of the `I’-token from the old animal to the new animal — this encapsulates Madden’s animalist approach whereby bodily continuity is decisive.

Reference Shift

The reference shift is of the type considered by Putnam.9 This allows little space for the importance of psychological continuity, and so all lines of this general type would fall victim to the challenges in the second half of the Williams paper. In addition, all accounts relying on analysis of physical parts will involve irresolvable Ship of Theseus-type problems.

The Ship Of Theseus

Ship of Theseus problems involve trying to establish the identity conditions for composite objects which may have some of all of their parts replaced over time. The problem becomes particularly acute if we imagine that all of the parts that have been removed from one ship as it is repaired over time are reassembled so that we have two ships. We have great difficulty in deciding which of the two ships is identical with the original ship.

The answer I would favor would be `neither’, because in fact even a single ship does not retain identity from one moment to the next. Persons are also composite objects. Crucially, it seems they can be split into brains and bodies in transplant cases. But also bodies have components. So the Ship of Theseus difficulties apply just as much to persons as to ships. So in fact we should not seek set identity for any real concrete composite objects. Once more, we see that So We Don’t Have Identities on that account.

The difficulties do not stop here however, as I shall now discuss.

3.1 Sorites Problem

A major difficulty with Madden’s account is that of Sorites problems. These arise when analogue changes drive binary distinctions. Imagine removing single pins from a heap of pins. We start with a heap. If we continue to remove pins, when we have none left, there are no pins so a fortiori, there is no longer a heap of pins. The difficulty lies in a line-drawing exercise. Did the removal of a single pin change the status of the group of pins from being a heap to no longer being a heap? That is a large change to hang on a single pin. Further, the difference between a heap of pins and a group of pins not quite large enough to constitute a heap is arbitrarily located.

Sorites problems are especially problematic in relation to personal identity. We find the pin examples confusing enough, but the consequences are more severe in questions of personal identity because those consequences are of such import. We cannot really allow changes in the answer to the question as to who we are to depend on unavoidably vague considerations of how many things we know.

A Principle Of Charity

Madden adapts a Davidsonian10 approach to assigning referents to terms. Davidson uses in radical interpretation a principle of charity whereby we compare assignments based on how many utterances they make true and prefer those which maximize the number of true utterances. Madden uses a variant principle of charity based on knowledge maximization rather than maximization of truth. In addition, we are to assign the referent of `I’-tokens based on finding which assignment maximizes the number of utterances that express knowledge.

There is a new objection to which the original principle of charity is not susceptible. In the original version, the test of whether an global assignment of referents is correct, or more correct than an alternative, is the extent to which it makes utterances true. This has a straightforward numerical measure: a simple count of true utterances constitutes the assignment’s figure of merit. On the new version, we need to count the number of utterances that express knowledge, and seek an assignment of personal identity that maximizes the number of utterances that express knowledge.


On Madden’s view, an utterance counts as an expression of knowledge if it is `safe’, viz. S has knowledge of p if S could not easily falsely believe p in the scenario where p is in fact false. So now we need to count not the number of true utterances, but the number of utterances where S believes p, p is true, but if p had been false in only a slightly altered scenario from the actual one, S would not have believed p.

Statement Counting

Counting statements that express knowledge is harder than counting true statements. We wish to determine a figure of merit for an assignment of personal identity by finding the ratio of utterances expressing knowledge to the total number of utterances. In assessing whether an utterance constitutes knowledge, we find whether it meets some specified level of safety. This means establishing the domain of possible worlds in which we require the utterance to remain true. So we need a specification of all the scenarios which qualify as only `slightly altered’ for each utterance under consideration. That is: which close possible worlds where the utterance must remain true to be knowledge.

We Don’t Have Identities: Knowledge Problems

Few deny that I have knowledge because there is a remote possible world in which I am a brain in a vat. However, many agree that I do not know that this is an elm tree even if it is because I cannot distinguish elm trees from oak trees; the world in which it was an oak and I said it was an elm is close.

Two different specifications might be very similar in that they differ only by excluding one additional scenario just for one utterance and are entirely identical in a myriad of other respects. They could differ only in relation to an utterance about trees on one particular day and remain unchanged for all other utterances and potential utterances on all other occasions. These two specifications differ in the most minor way if one measures difference by the number of differing scenarios divided by the total number of scenarios; yet this minor distinction would suffice to produce a quality distinction between two candidate assignments. This seems a fine level of subtlety; a great deal would turn on this very fine distinction.

Is That A Zebra?

Imagine you are a radical translator seeking to assign referents to the language I use, which happens to be English: you do not speak English. You see that there is a zebra in a field. I point at it and say `the zebra is in the field’. The original translator has few difficulties with this. An assignment of the term `zebra’ to the zebra produces an incremental mark for that assignment. My utterance has been made true by that appropriate assignment.

Do I have a safe belief that there is a zebra in the field? Answering this requires dealing with the usual questions such as whether it is a relevant possibility that someone placed carefully disguised donkeys in the field. Madden might respond that he is open only to the objections that he has imported along with the safety account of knowledge. It remains the case he has exacerbated that problem by multiplying it and he would have analogous problems with different accounts of knowledge.

The only real defence for Madden to adopt is the `dirty hands’ one, which is to observe that these problems apply to all other accounts also. That of course is the result for which I have argued here — since that defense is the only one available, no accounts of this type can succeed. So We Don’t Have Identities one more time.

4 We Don’t Have Identities: Combination Accounts

Nozick11 has observed the fatal flaws with both psychological continuity and bodily continuity approaches and this motivates him to try to combine the two. He also begins with Williams’s brain transplant cases, noting correctly that the problems therein derive from the fact that we do not know how to weigh up the competing significance of bodily and psychological continuity in determining questions of identity. He seeks a legal mechanism12 to do this. However, we cannot find such a mechanism, because a legal mechanism will find only legal identity. Moreover, the Sorites problem discussed in s3.113 still needs solving. All of this needs fixing or We Don’t Have Identities.

4.1 We Don’t Have Identities: Covert Introduction Of Set Identity

I will now argue that a major problem with Nozick’s account is that it covertly introduces notions of identity which are appropriate in the abstract world of numbers and sets but do not apply in the concrete world. I will once again conclude that We Don’t Have Identities.

The legal mechanism that Nozick chooses to determine identity — which he dubs the Closest Continuer theory { will basically involve looking at all the candidates for identity with something else and making a measured decision. This will mean that the identity criteria for an item will depend on other items that exist or potentially could, which is implausible. Why should it matter to who I am who you are, or what else exists? Would I not still be me in an otherwise empty universe? So much of Nozick’s effort will focus on attempting to make it plausible that external items can influence identity.

The Vienna Circle Story

Nozick introduces his proposed mechanism by way of a story that could have happened to the 20 members of the Vienna Circle, a group of philosophers who we are to imagine flee to separate locations where they are no longer in contact with each other or the outside world. Three members of this group escape to Istanbul where they believe wrongly that all the others are dead. As Nozick puts it:14 “[t]hey now are the Vienna Circle”. We may certainly agree that in these circumstances, were one to find a legal document naming the Vienna Circle, perhaps as the recipient of a bequest, these three would be entitled to claim it. That would be the case providing no better placed group could make out a stronger claim to the bequest, a caveat that will be important in Nozick’s development of the story, as we will see.

However, Nozick needs something more than this legal status, surely. We are discussing questions of identity. We should ask: which of the following claims is Nozick pressing?

The Potential Claims

1.The three members in Istanbul are identical to the same members in Vienna.

2.The three members in Istanbul are identical to the 20 members in Vienna.

3.The 20 members in Vienna are identical to the Vienna Circle.

4.The 3 members in Istanbul are identical to the Vienna Circle.

Of these claims, only 1). appears defensible. Claim 2). must be false because three persons cannot be identical with 20 persons. Claims 3). and 4). may appear prima facie implausible because it seems that a number of persons cannot be identical with a legal construct like that denoted by `the Vienna Circle’, which is presumably some kind of club or society with meeting rooms and a constitution. This question needs to be answered by anyone wanting to deny that We Don’t Have Identities.

We Don’t Have Identities: The Problem With The Story

However, to take this line would be unfair to Nozick; we must allow him to stipulate that `the Vienna Circle’ refers to nothing beyond the group of people who are its members. Yet this produces the real difficulty. We must now allow Nozick’s claim 3).; it is the case that the original group met, had meetings, comprised 20 members and those 20 members as a group were identical to the Vienna Circle because that term is just another way of referring to those 20 members.

We are only allowing that the term `Vienna Circle’ can be identical to its 20 members because we have allowed that that term can be another way of referring to those members. So in fact this is a case where the abstract realm in which identity does apply has been smuggled in — since we are talking about the identity which does indeed hold between a set and itself — under the appearance of making identity statements about items in the real world.

If the term `Vienna Circle’ changes over time — say one person dies and now the term refers to the 19 remaining members — the identity between the two uses of `Vienna Circle’ is not the set identity that obtains between the initial 20 members and the first use of `Vienna Circle’ because there is no identity between the 20 and the 19. There is only a legal identity between the two uses of `Vienna Circle’ because there is only a legal identity between the 20 and the 19.

We Don’t Have Identities: Argument Result

Nozick develops his story such that unbeknownst to the Istanbul three, there is also a California nine who have been conducting the same business as them. Both groups believed they spoke for the Vienna Circle. The outside world may decide that the nine are the Circle.15 However, this is merely to make a decision on legal status — it answers the question as to who can speak for the Vienna Circle. It achieves nothing connected to identity relations. Even if these legal questions could be definitively settled, none of these groups would be identical to the Vienna Circle in the required sense. So We Don’t Have Identities either.

5 We Don’t Have IdentitiesConclusion: Identity Is Abstract

A basic problem with all of the accounts considered here is that they seek the wrong type of identity. They have a legalistic approach which might be appropriate for determining who is entitled to what share of a legacy. They weigh up factors counting in various directions, and, in the case of Nozick’s account, allow also for factors not relating to the entity under discussion to affect that entity’s identity criteria. This cannot be the right way to decide who I am. A committee of lawyers cannot make the decision as to from which pair of eyes I regard the world. So We Don’t Have Identities on any account.

See Also:

Are We Allowed To Follow Our Personal Aims? Nagel says Maybe

Is Experience Time the same as Experienced Time?

The Importance Of Forgetfulness For Nietzsche

Merleau Ponty’s Phenomenology: What Is It And How Cogent Is It?


1Perry observes that JJC Smart takes this line at [2, p. 66].

2See [1].

3This is Perry’s formulation at [2, p. 84].

4See [2, Ch. 5].

5It might be objected that Grice is not making the error of seeking set identity when persons have only legal identity or no identity. This objection cannot succeed because Grice’s `logical construction’ approach: identities are sets (of experiences) for him. Thus if his account succeeds, it can only find set identity for sets rather than set identity for persons.

6See [2, p. 88].

7See [3, Ch. 9].

8It might be objected again that Madden is only seeking legal identity and not set identity. We can see this is false however. He describes the type of identity he is looking for as `numerical’ at [3, p. 287]. Further, he notes that in a transplant case “[t]he referent of these I tokens is a single thing” at [3, p. 293], confirming he is seeking set identity. No one thinks a legal heir is the same entity as a benefactor.

9See [4].

10Cf. [5, Ch. 9]. Davidson notes the difficulties inherent in attempting to make “finely discriminated distinctions amongst beliefs”.

Further Footnotes

11See [6, Ch. 1].

12It might be objected a third time that Nozick is only looking for legal and not set identity. However, we can see this is false by noting he observes that we are “not willing to think that whether something is us can be a matter of (somewhat arbitrary) decision or stipulation” at [6, p. 34]. He also admits that `nontransitivities’ are problematic at [6, p. 42] which is only the case because he seeks set identity.

13Nozick is aware of this; he writes: “But the interval can be varied gradually; it seems absurd that there should be some sharp temporal line which makes the difference to whether or not the person continues to live in the other body” at [6, p. 44]. He is discussing whether there could be a time period after which we allow that someone has moved to a new body but before which we do not. It will be apparent that this is the same objection as I raised in relation to Madden’s account in x3.1.

14See [6, p. 32].

15Even this is questionable though. We might imagine circumstances where the three are all of the major players making immense original contribution to philosophical debate while the nine are doing no work at all. This problem again highlights the con ation of the two issues of who legally may speak for an entity and what metaphysically is identical with it.


[1]B. Williams, “The self and the future,” The Philosophical Review, vol. 79, no. 2, pp. pp. 161{180, 1970.

[2]J. Perry, Identity, personal identity, and the self. Hackett Pub., 2002.

[3]K. Bennett and D. Zimmerman, Oxford Studies in Metaphysics. No. v. 6 in Oxford Studies in Metaphysics Series, Oxford University Press, 2011.

[4]A. Pessin, S. Goldberg, and H. Putnam, The Twin earth chronicles: twenty years of reflection on Hilary Putnam’s “The meaning of `meaning'”. M. E. Sharpe, 1996.

[5]D. Davidson, Inquiries into truth and interpretation. Philosophical essays, Clarendon Press, 2001.

[6]R. Nozick, Philosophical explanations. Harvard University Press, 1981.