Understood as a composite of bodily and psychological or mental states, a person is one kind of particular, with a unique and uniquely identifying spatio-temporal path in the world. In this respect, persons are not unlike other spatio-temporal particulars, such as chairs. The reidentification of chairs is not always easy in practice, but admits of a simple enough rule: a chair is judged the same between time 1 (t₁) and time 2 (t₂) when the spatio-temporal continuity can be established linking its earlier and later instantiations. This matter is complicated, even with particulars like chairs, inasmuch as we utilize the same language to indicate qualitative as well as numerical identity. ‘The same chair’ may mean a qualitatively similar chair (same color, style) or the very chair (numerical identity). Context usually prevents us from becoming confused by these potential ambiguities when we are talking about such things as chairs; speakers of English are adroit at distinguishing ‘the same x‘ as bespeaking the spatio-temporal continuity of particulars, and the sameness which rests on similarity of qualities such as color and shape.

Some philosophers have insisted that the spatiotemporal continuity of the person’s bodily attributes are similarly the source of personal identity judgments. The body is thus said to be criterial for personal identity (Williams, 1973). On this theory, the continuity of persons is just as strict, no more and no less, and for the same reasons, as the identity of the other material objects we judge to remain the same through time. This may not be ‘identity’ in the strict numerical sense. But the workaday spatio-temporal continuity is all we ever need to reidentify particulars, be they chairs or persons, on this view, whether or not we describe the sameness we recognize to be ‘identity’ properly so called.

The normal person in everyday contexts has a high degree of continuity of several kinds. As well as the continuation of the body through its spatiotemporal path, continuity is also provided by enduring psychological states and dispositions, by memory, by the stream of consciousness, and by persisting capabilities and skills. Rejecting the view that bodily continuity accounts for personal identity, other theorists have identified some of these enduring or relatively enduring attributes as criteria for personal identity. Locke argued that the continuity of memory between earlier and later person or self stages explained personal identity. Several twentieth-century thinkers have pointed to psychological states and dispositions more generally understood to explain the person’s sameness or survival through time (Grice, 1941; Parfit, 1984; Shoemaker, 1979). As long as these attributes are taken to be psychological occurrences (particular, datable, introspectible events in consciousness), there will be interruptions in this continuity, such as those resultant from sleep. But by treating these traits as dispositions or capabilities with causal powers (we remain disposed to do, feel, remember, believe, or will, even while asleep) it is not difficult to countenance their providing uninterrupted continuity. On such an analysis, the continuity of psychological states explains the customary presumption that the self or person remains constant through stretches of time, forming a causally related and overlapping series in most people’s lives sufficient for us to speak of the same person surviving, if not as maintaining ‘identity’, strictly understood, through stretches of time.

The question of whether person individuation must be dealt with separately from person reidentification requires us to consider the person’s oneness or unity in two ways. The (synchronic) unity of the person’s experiences at a given time and the (diachronic) continuity of the person through time seem to give rise to questions of differentiation, or self-ascription, such as ‘On what basis do I know that all these experiences are my experiences?’ as well as to the questions of reidentification introduced earlier (‘Is this the same person between t₁ and t₂?’). But until Hume’s famous, fruitless pursuit of the self as an object of experience, these two kinds of question were apparently conflated. The Cartesian subject or self, for which Hume searched in vain, was understood to encompass a unifying principle, linking items in the imagination at a given time, and as well, because it remained unaltered through time, providing the continuity of perfect identity persisting throughout the individual’s life.

It was this Cartesian concept of subject or self which Hume believed he exploded. Rather than a simple, united whole which ensured differentiation and synchronic unity as well as reidentification and diachronic unity, Hume discovered an untidy bundle of heterogeneous and ever-changing impressions. Instead of the strict and perfect identity intrinsic to the self associated with its continuation through time, by which it was thought changeless, Hume pictured an imperfect and ‘imaginary’ or ‘fictional’ identity, imposed from without for the convenience of the observer.

Hume did not question the self’s unity. But he denied the basis for making claims about unity and continuity – the Cartesian notion that underlying these two kinds of identity there stood a metaphysical subject of experiences. If, as Hume concluded, there was no subject of experiences, then the supposed unifying functions of that subject would be lost. Nonetheless, synchronic unity and the consequent differentiation it provided remained – ensured, Hume believed, by the imagination. A loose and relative continuity, misnamed ‘identity’, based on the relations of contiguity (proximity or closeness), resemblance, and causation linked earlier and later experience stages in a way sufficient for us to speak of the person as remaining the same through time. Hume’s is a reductionist theory of the self, and if we speak of selves as unified both at and through time, it must be due to the rough empirical cohesion among and continuities between parts of the self, a cohesion ensured by the imagination.

Some philosophers (Kant was one) have rejected Hume’s empirical approach, and appeal is still made to a transcendental subject in explaining the self’s synchronic and diachronic unity (Korsgaard, 1989). Moreover, the imagination was hardly sufficient for the task Hume imposed upon it, as others came to recognize. Without the transcendental principle to at once unify the contents of consciousness at any given time and thus provide the means for differentiating one person’s experiences from those of another, and to unify the scattered parts of a person’s experience through time, permitting person reidentification, these two kinds of unity required for personal identity were in need of, and received, separate theoretical treatment.

Questions about synchronic unity such as the one noted above (‘How do I know these experiences are my experiences?’), and the means of differentiating one person’s experiences from another’s, concern what is known as self-attribution, or ownership. Different philosophical accounts share the conclusion that self-attribution is grammatically guaranteed: the concept of experience prevents our speaking of unowned experiences, thus ‘this experience’ entails ‘my experience’. Whether this is a trivial or an interesting feature of human experience, however, is disputed.

Some theorists have emphasized the psychological necessity of ownership. As a contingent fact, they suggest, humans have evolved to own all and only our experiences (Dennett, 1991). Others insist that self-ascription is logically guaranteed in any subjective or introspective report. Any first-person account is by its nature an account in which ‘this experience’ entails ‘my experience’. Examining the evidence from psychopathological symptoms of mental alienation, in which a patient appears to deny ownership of his or her experience, others have analyzed normal self-ascription into two parts, one concerning ownership (the assertion that experience X occurred within me), and the other agency (the assertion that experience X is my mental action, or a thought I think) (Graham and Stephens, 1994; Radden, 1996). This distinction between thoughts which merely occur in us, and thoughts we think, renders self-ascription and the grammar of ownership ambiguous.

Cut loose from the unifying function provided in the transcendental theories of personal identity which are largely rejected today in favor of empirical, psychological continuity analyses, the need for an account of self-attribution becomes more pressing. These efforts to understand self-attribution are important contributions.

If the self is experienced as relatively continuous through time, this may be because of our dispositions to remember, Locke believed. However, we do not remember perfectly. Recognizing this, Grice sketched a memory theory of personal identity which treats memory as an overlapping series (Grice, 1941). A ‘total temporary state’ (tts) is a set of simultaneous experiences of a single person, and Grice argued that a sequence of tts’s comprise parts of the same person (and thus constitute an identity) when certain relations obtain between them: each tts is disposed to remember at least one experience contained in the last, or contains an experience of which the next contains a memory. A single person’s tts’s thus form an interlocking series, defined by Grice as one in which no subset of members is independent of all the rest. Memory is understood dispositionally here. A tts is ‘memorable’ when it contains as an element some experience a memory of which would, given certain conditions, occur as an element in some subsequent tts. A tts is ‘memorative’ when it would, given certain conditions, contain as an element a memory of some experience contained in a previous tts.

Grice’s analysis calls for a very modest continuity of memories to ensure the continuation or survival of the same person through time. Though little else were to remain unchanged, one shared memory between two self stages may be sufficient for us to speak of those stages as earlier and later parts of the same person. Grice’s account is thus relatively conservative; that is, a (loose) identity or ‘singularity’ conserving theory. In all but the most extreme cases of discontinuity it allows us to maintain the traditional presumption that just one person inhabits one body for its lifetime (a presumption dubbed the ‘one to a customer’ rule: Dennett, 1991).

More encompassing than Grice’s theory, Parfit’s influential empirical continuity argument accommodates not only memory but also other forms of psychological continuity (Parfit, 1984). On this analysis, the ‘psychological states’ whose continuity explains language suggestive of personal identity include behavioral traits and tendencies as well as occurrent mental items datable to a particular moment – not only memories, dispositions, and personality traits, but also other, more general capabilities, such as how to speak or swim. Like Grice, Parfit uses the notion of an overlapping or interlocking series. Earlier and later self stages housed in a single body at different times have continuity when and to the extent that they overlap. Should psychological states A and B overlap in this way between t₁ and t₃, then there is ‘connection’ between A and B at t₂. Rather than ‘identity’ strictly understood, Parfit employs the notion of ‘survival’. If person P remains the same between t₁ and t₃ due to such overlapping connection, then P survives between t₁ and t₃. The more traits are continuous, the more survival there will be. So connections admit of, and can hold to, any degree. If between t₁ and t₃ person P’s connected traits numbered 100, and person Q’s numbered 50, we could conclude that between t₁ and t₃ person P survived more fully than person Q.

Parfit’s psychological continuity analysis permits a broad set of traits to determine survival. Thus, a person in whom brain injury has brought about total amnesia concerning her past life, together with radical personality change, may have retained no trait of memory and personality entitling us to judge her the same person after, as before, her injury. But while there remain some overlapping traits – she can still knit, let us say, or swim – then on Parfit’s account we may speak of survival.

The notion of a disposition is a causal one. This means that the continuity provided by interconnected dispositions is ‘causally grounded continuity’ (Shoemaker, 1984, p. 84). This causal basis of psychological dispositions has been particularly emphasized by Shoemaker, who regards the continuity of persons through time as an instance of the more general continuity of ‘continuants’, things which we treat as the same through apparent change. Direct psychological connection between earlier and later stages of a person is guaranteed by the later stage’s standing in the appropriate relation of causal dependence to a state contained in the earlier stage. (Theorists debate whether, for us to speak of personal ‘identity’ or survival, these chains of psychological continuity and connectedness must result from normal causal processes. The acquisition of artificially created ‘quasi-memories’ otherwise bearing the appropriate relation to a person’s current states puts pressure on this requirement.)

Because survival admits of degree, the determination that one self has succeeded another is an arbitrary matter calling for decision. How much variance of psychological traits through time ranks as sufficient for us to judge one self to have replaced (or joined) another in a single body will presumably depend on context and on the purposes and interests we bring to the decision.

Some argue that questions about personal identity are facts of the matter – to be discovered, not arbitrated, and this discussion reveals the contrasting approaches and methods found in discussions of personal identity.

In the tradition of Locke’s famous request that we imagine the minds of a prince and a cobbler, respectively, transferred to one another’s bodies, and allow our intuitions to determine whether the cobbler’s or the prince’s body housed the person of each, philosophical discussions of personal identity usually employ a thought experiment to establish their claims. As in Locke’s model, kinds of continuity are imaginarily eliminated and we are asked to consult intuitive convictions about the survival of the resultant entity and about our moral and other attitudes towards it. Elaborate thought experimental examples of fusion, branching, and brain transplant are constructed in support of these conclusions. For example, if A’s brain is transplanted into B’s body, we can ask whether reproaching A/B for A’s earlier crimes would be right.

Implicit in this methodology is the notion that a metaphysical fact of the matter exists, and can be discovered through patient theorizing. Such a view is fundamentally at odds with the approach which allows that survival is a matter for arbitration.

Some theories of personal identity have been subject to a different kind of challenge. Serious normative costs appear to be associated with the merely empirical self which explains identity language and presuppositions without resort to a numerically identical subject, and admits that several selves might coexist or succeed one another within a bodily lifetime. The effect of Parfit’s work, particularly, has been a marshaling of the reasons to value the unified and numerically identical self associated with individualism, and a rehearsal of the normative costs of rejecting it – costs which allegedly entail the loss of central moral and legal conceptions of autonomy, responsibility, agency, blame, praise, and so on. This reasoning takes the form not of discovering a metaphysical truth, but of emphasizing the normative tug of individualism, the several compelling reasons why, even if we could choose to adopt a survival threshold which violated traditional personal identity presuppositions, the costs of doing so in the loss of our most valued concepts, categories, and practices would be unacceptable. Initially raised by those who maintain allegiance to a transcendental subject which serves to ensure strict personal identity, such considerations also affect the position of those accepting an empirical theory of personal survival (Radden, 1996). Conceding the theoretical possibility that several selves or persons might inhabit one body through its lifetime, one might nonetheless argue for a more singularity-conserving survival threshold on the grounds that such normative costs are too high to be tolerated.

For all that they focus on the same issue, these contrasting approaches differ fundamentally. Whether something requires or, alternatively, invites us to uphold belief in personal identity and/or survival are very different matters, and if we are required to uphold that belief, then we do not need to trouble ourselves over the blandishments of normative individualism.

These four sources of fundamental disagreement provide a rough map of the lines of inquiry found in philosophical discussions of personal identity. As this map indicates, the significance for personal identity of theorizing about brain science will hardly be an uncontested or straightforward matter, but a number of the issues raised above intersect with, and may be affected by, parallel explorations in cognitive science.