¹ Like the role in which James casts ‘stream of experience’, ‘the transition of one experience into another’, ‘coconsciousness’, ‘conjunctive relations’. See Essays in Radical Empiricism, (Longmans, 1912).

² See footnote 4 below.

³ Cf. Bennett, Kant's Analytic, (Cambridge, 1963) p. 117.Google Scholar

⁴ A defender of Locke might be momentarily misled into thinking that this definition came to the same thing as one which did mention that, if he failed to see that he must decline what Butler offers to him under the guise of a concession —that ‘consciousness of what is past does … ascertain our personal identity to ourselves’. The offer is unlikely to have been disingenuous, but it must certainly be refused. See Shoemaker, Sydney, Self Knowledge and Self Identity, (Cornell, 1963).Google Scholar

⁵ See Flew's, A. G. N. ‘Locke and the Problem of Personal Identity’, Philosophy, 24 (1951).Google Scholar

⁶ Op. cit, p. 55.

⁷ The argument needs to be reinforced by examples free of the complications of scope imported by definite descriptions. To go into these and the other important syntactical ambiguities latent in such examples would be diverting but irrelevant, as would the larger task of assessing the putative equivalence of A remembers X-ing, and A remembers his X-ing.

⁸ Cf. Williams, B. A. O. ‘Imagination and the Self’, Proceedings of the British Academy, (1966).Google Scholar

⁹ ‘P’ (or ‘Q’) stands in for a definite description of a continuing person, not of a ‘slice’. The subscripts tj, tk … index the time at which the definite description applies to the continuing person. The subscript will be omitted where this does not need to be indicated. The notation leaves it an open question whether ‘P’ and ‘Q’ stand in for the same or different descriptions.

¹⁰ Even though it is a possible elucidation of ‘extend’ in Locke's words at Essay, II, xxvii, 2 (p. 449 Fraser).

¹¹ The reader may have believed it to be an objection to Ip that Ip is already *C is already an asymmetrical relation. Here we must guard against a misconception of what was meant by P_tj and Q_tk. What such descriptions stand for are not time-slices of people or ‘person-moments’. They are people, persisting three-dimensional things which are born, live for some time, and then in one manner or another die. If P_tj and Q_tk are the same person then, regardless of the fact that these descriptions may pick out the person by reference to predicates which hold of the person at the different times tk and tj, their references are one and the same. Everything true of one is true of the other and, with certain merely grammatical adjustments, the designations are everywhere intersubstitutable salva veritate.

¹² Cf. Ishiguro, Hidé, ‘The Person's Future and the Mind-Body Problem’ in Phenomenology and Linguistic Analysis (ed. Mays, Wolfe and Brown, Stuart) (London 1971).Google Scholar

¹³ Cf. Flew, p. 67 and Williams, B. A. O., ‘Personal Identity and Individuation’, PAS, (1956–1957).Google Scholar

¹⁴ Nor would it be right to insist on redescribing the situation as one where a single dissociated or scattered person did X and also, with another part of him, did not X. Such insistence might be based on an analogy with the way in which, without prejudice to his unity, a normal person can fidget with his left foot and not fidget with his right. There may be ideas of what a person is which allow this sort of redescription. But it is in the spirit of the Lockean notion, and of our own ordinary notion of person—the notion of a three-dimensional thing whose only genuine parts are spatial, parts—to disallow it. When P_t1 splits into Q_t3 and R_t3, it splits into two whole, persons, and Q_t3 does not share R_t3's consciousness of doing X. Even if Q_t3 and R_t3 both do X, their consciousnesses of doing X may be as distinct as if they were two people with no common origins at all. And they will communicate inter-personally.

¹⁵ Cf. Leibniz, , Gerhardt, IV, p. 460.Google Scholar

¹⁶ This is to suggest that just as the functions of remembering and intending and much else play a crucial role in regulating the individuation of persons, so perhaps identity itself regulates the correct application of the predicate remembers X-ing. To the critic who sees in this relationship of reciprocal regulation the chance of reviving Butler's objection to Locke, I offer the following a priori refutation of the generally received account of how clockwork functions in a timepiece. It is said that the mainspring unwinds, and in unwinding affects the hairspring. But it is also said that the hairspring affects the speed and manner of unwinding of the mainspring. How can that possibly be? If the normal operation of the mainspring presupposes the normal operation of the hairspring how can the normal operation of the hairspring presuppose the normal operation of the mainspring? Well, it can and it does. Presupposition like mechanical regulation can be reciprocal.

¹⁷ ‘Remembering’, Philosophical Review (1966).

¹⁸ Cf. Grice, , ‘The Causal Theory of Perception’, PAS, Vol. 35 (1962)Google Scholar—an analogy to which I was already indebted in the analysis of memory at Identity and Spatio- Temporal Continuity pp. 45, 55 and footnote 55.

¹⁹ This is not a circular procedure: but even if it were circular that would not matter for my purposes, which relate to the necessary conditions of remembering.

²⁰ For a more careful statement of this see my ‘Towards a Reasonable Libertarianism’ in Essays on Freedom and Action, Ed. Honderich, T. (Routledge, London, 1973), p. 49.Google Scholar

²¹ It scarcely improves things to think of the memory trace as an immaterial imprint on immaterial stuff. This can only help to the extent that immaterial mind is made intelligible by being modelled on the material, and distinguished from it only by the apparently vacuous contention that it is immaterial.

²² See e.g. Analysis 21, 3 (1961).

²³ An impossibility if we subscribe to Leibniz's Law. See my Identity and Spatio- Temporal Continuity (Blackwell, Oxford 1967) p. 3. In rejecting this possibility I distance myself yet further from Locke's own development of his theme. Locke tries to overcome the standard difficulties of C (amnesia, sleep etc.) by distinguishing questions of identity of man from questions of identity of person. This is a thoroughly unsatisfactory part of his discussion. However well one makes the distinction between the concepts man and person, this can hardly show that nothing falls under both concepts (under which is John Locke?), or that identity can be so relativized as to make the two identity questions independent of one another.

²⁴ It may be said that we can and do distinguish between the corpse of a man freshly dead—he is still here but dead—and his earthly remains. When ashes (say) are all that is left then he is no longer there, it may be suggested. But the whole distinction which is relied upon by the objector is parasitic upon the point of distinguishing between life and death. Mere material continuity is not sufficient. And if life or its absence gives the point of these distinctions, then the principal distinction is between being live and being dead, and the best overall view will make existence or non-existence depend upon the principal distinction.

²⁵ Cf. Identity and Spatio- Temporal Continuity p. 45 para. 2.

²⁶ Nor, contrary to the purely physiological view of the question, is the seat of memory and consciousness just one part of the body: although, purely physiologically speaking, I suppose that the difficulties of replacing a brain are simply more severe than the difficulties of replacing a kidney or an eye. The physiological view would not be a doctor's view, but this is part of what people have in mind when they say doctors must be men of science and something else besides. Amongst the first to have thought into what thought-experiments involving change of bodily parts really involve is Stefan Themerson. See the last chapter of Bayamus: a semantic novel (Gaberbocchus, London 1949).Google Scholar

²⁷ Cf. Identity op. cit. (Appendix) 5·7 and 5·6 and (Part Two) 4·3 p. 54.

²⁸ For simplicity and because they raise special and even graver difficulties, I exclude from consideration here the fusion of lines of consciousness.

²⁹ The conception of person by which our own everyday life sets such store registers only in a rarely-used nonsense names. The names which matter are birth order names, kinship terms, teknonyms and status titles. Geertz, Clifford, Person, Time and Conduct in Bali S.E. Asia Studies, 14 (New Haven, Yale, 1966).Google Scholar

³⁰ L'idiot de la famille: Gustave Flaubert de 1821 à 1857 (Gallimard, Paris 1971).Google Scholar

³¹ See especially the vivid and extraordinary description of the Balinese practice of drama in Geertz.

³² Cf. my ‘Being in the Same Place at the Same Time’, Philosophical Review (January 1968).

³³ Metaphihsophy No. 3 (1970). See also Language, Belief and Metaphysics (Vol. 1 of the International Philosophy Year Conference at Brockport), (Albany, New York, 1970)Google Scholar. Similar conclusions have been arrived at independently by Rogers Albritton and Saul Kripke, and related ones by Mr Vernon Pratt of University College, Cardiff, in ‘Biological Classification’, British Journal for the Philosophy of Science (1972) pp. 305–327Google Scholar. See also Quine, W. V. ‘Natural Kinds’ in Essays in Honour of C. G. Hempel (Rescher ed.) (Reidel Dordrecht, 1969).Google Scholar

³⁴ Cf. p. 67, Ziff, Paul ‘The Feelings of Robots’, Analysis, Vol. 19, No. 3 (January, 1959)CrossRefGoogle Scholar, and compare Hobbes'; problem of Theseus' ship, cited at my Identity 2.1, p. 37. There are some telling considerations one can use to try to resolve disputes about that ship. (Pure conventionalism is not a very good story even for artifact identity.) But if doubt or dispute persists this is at least partly because one party is looking for an archaeological relic and the other for a functionally persistent continuant. Is one party wrong and the other right? There is scarcely anything to discover to vindicate either conception against the other. I do not mean the decision is arbitrary—only rarely is it entirely arbitrary—but nothing, as it would in the case of a natural kind, compels it. The antiquarian who favours the reconstructed ship has a different interest from the priest who favours the continuously repaired continuant. But both are stuck with the identification ship. Neither can base his view upon the natural development of a ship, or suggest a programme of research to resolve the question.

³⁵ ‘Quantities’, Philosophical Review (1969). I should make it clear that nothing in my discussion here of natural things versus artifacts is meant to rule out the possibility of borderline cases—think of a wasp's nest or (in another category and in relatively undeveloped parts of the world) cheese or bread: and that the discussion is intended to find a distinction with respect to essence between natural kinds and other kinds which will supersede the principle of distinction implicit in the etymology of ‘artifact’.

³⁶ See Williams, Bernard ‘Are Persons Bodies?’ in Spicker, S. (ed.) The Philosophy of the Body. (Quadrant Books, Chicago, 1970)Google Scholar; also Miri, N., ‘Personal Identity and Memory’, Mind (January, 1972).Google Scholar

³⁷ As Derek Parfit has in several ways suggested. Cf. his ‘Personal Identity’, Philosophical Review, Vol. LXXX, No. 1 (1971).Google Scholar

³⁸ These psychologically interesting differences between different members of a species will presumably give out well before we reach the oyster, to which Lucian thought it was so comical that Aristotle attributed a ‘psychology’.

³⁹ Cf. Shorter, J. M., ‘Personal Identity, Personal Relationships, and Criteria’, PAS (1970–1971).Google Scholar

⁴⁰ I assume that the functional specification is not trivialized by being continued to a point where it coincides with the full set of attributes which human beings (and …?) actually have in virtue of their biologically etc. given nature. Cf. the penultimate paragraph of VI above. This would immediately concede to me everything I am pleading for—and more.

⁴¹ I owe this view of needing to Miss Sira Dermen. It is prefigured in one form in Aristotle at Metaphysics Δ.5. More generally, see e.g. MacPherson, in Blackburn, (ed.) Social Science and Ideology (Fontana, London, 1972) pp. 25 ff.Google Scholar

⁴² Aeschylus, , Prometheus Vinctus, 440 ffGoogle Scholar. and 250 ff.

⁴³ ‘Freedom and Resentment’, Proceedings of the British Academy, XLVIII (1962).Google Scholar

⁴⁴ A responsible review of the situation made in the light of the theory of individuation outlined here could not, simply by explaining the source of the perplexities they create, evade the problems posed by surgical and mechanical interventions which fall short of imperilling the application of person and related concepts. It would have to distinguish between the (implausible) necessary conditions and the (perhaps less implausible) sufficient conditions of personal identity which *C could be used to frame. I only remark that it should take nothing for granted about how well we really understand brain transfers of the kind described by Shoemaker. How do we fit the brain to the physiognomy of the new body which is to receive it? (Cf. Williams, , ‘Identity and Individuation’, PAS 1958–1959Google Scholar). How is the existing character expressed in the new body? We are deceived by the quality of the actors and mimics we see on the stage if with the help of greasepaint and props they have made us think this is as (relatively) simple as the transposition of music from one instrument to another.