David Lewis (1976), “Survival and Identity”, ch. I of Amélie Oksenberg Rorty (ed.), The Identities of Persons, Berkeley: University of California Press, pp. 17–40 — DOI-10.1525/9780520353060-002. Reprinted with two postscripts in Lewis, Philosophical Papers, vol. I (Oxford University Press, 1983), pp. 55–77.
On the citation. The brief’s details are right: author, title, Rorty’s volume, the University of California Press. Three things to add:
- Pages: 17–40, from Crossref and the volume’s contents page.
- Identifier. The DOI is the chapter in De Gruyter’s digitisation of
the UC Press book. Crossref dates it 31 December 1976, which is a
digitiser’s placeholder, not a publication day.
published:is the year only. - The reprint (Crossref DOI 10.1093/0195032047.003.0005, pp. 55–77) adds “Postscripts to ‘Survival and Identity’”. Postscript A answers Parfit’s reply in the same Rorty volume, “Lewis, Perry, and What Matters”. The postscripts are not this entry’s subject.
Key takeaways
- Two answers to “what matters in survival?”, and both are right. The philosophical answer is “mental continuity and connectedness”; the platitude is identity, “that relation that everything bears to itself and to no other thing” (pp. 17–20).
- They relate different things. Continuity and connectedness hold among “more or less momentary person-stages”; identity among “temporally extended continuant persons” (p. 20). The fair comparison is between the R-relation and the I-relation, the relation between stages of a single continuant person (p. 21).
- The I-relation is the R-relation. “Any stage is I-related and R-related to exactly the same stages”, in all possible cases, which gives a non-circular definition: “something is a continuant person if and only if it is a maximal R-interrelated aggregate of person-stages” (p. 22).
- Fission and fusion are partial overlap. In fission two persons share their pre-fission stages; the I-relation is then as intransitive as the R-relation (pp. 24–25).
- Count by tensed identity. To avoid “overpopulation”, count persons at a time by identity of their stages at that time. One person entered the duplication centre, counting by identity-yesterday; two all along, counting by identity (pp. 26–29). Roads that merge for a stretch are the model (p. 27).
- Degrees by supervaluation. If R comes in degrees, take every cut-off as a “delineation”; the degree to which two stages are I-related is the measure of the delineations on which they are. It then equals their degree of R-relatedness (pp. 35–36).
Standing in the record
Filed on 2026-10-02 at the owner’s request, to fill out the record’s coverage of personal identity. NOTE-373 is the reading, and it placed the work: Active. It is the standard reply to Parfit’s claim that identity is not what matters, and the place the stage view of persons was first put to work on survival. Read it for the move from comparing identity with R to comparing the I-relation with R, and for the price, which is stated: we must count persons by a relation that is not identity.
No anthology topic holds it, it carries no instruction for machine-learning practice, and the anthology does not hold it.
Where it meets the record:
- Parfit’s “Personal Identity”, filed beside it, is its target and its ally. Lewis accepts that paper’s premises, that identity is one-one and all-or-nothing and that what matters may branch and come in degrees, and denies only its conclusion.
- Goldstein and Lederman (LIT-206, NOTE-138) make the per-conversation instance the bearer of beliefs and desires and find “no more explanatory work” for a model agent. Lewis’s stage-sharing is a model on which both can exist without double counting: instances that share their earlier stages (the weights, the prompt so far) are distinct continuants counted as one until they diverge. The application is mine, not either paper’s.
- The survival-insurance case in Postscript A (1983), from Justin Leiber’s novel: the mind is recorded from time to time and played back into a fresh body after a fatal accident. Lewis says this satisfies “the weak desire” for survival, let at least one of us survive. Birch (LIT-111, NOTE-183) argues that a chatbot character recreated from a transcript has no persisting bearer. The two disagree about whether such playback is survival, though Lewis’s recording carries the causal link that Birch’s generative recreation lacks.
- Mereology. Lewis prefers “a mereological sum, so that the stages are literally parts of the continuant” (n. 4). The record’s mereology works, among them Meincke’s count of the pregnant organism as “between one and two” (LIT-142), meet the same problem of counting overlapping individuals. Meincke answers with processes, not stages, and neither engages the other.
- The survey (LIT-130, NOTE-166) gives the fission argument and the four views on what matters. Lewis’s is the view that identity and the psychological relation coincide, which the survey sets out through Parfit’s framing.
Williams’s “The Self and the Future”, filed beside it, is not discussed, and its non-branching body swap gives tensed identity nothing to do. Whether Reasons and Persons, filed beside it unread, answers Lewis could not be checked from its publisher’s description.