Source
Parfit (1971), “Personal Identity”, The Philosophical Review 80(1): 3–27, LIT-447, read in full in NOTE-378. The dilemma fission poses for psychological views, as the current literature states it: Olson (2002; substantive revision 2023), the SEP entry “Personal Identity”, LIT-454, read in NOTE-374.
What was actually shown
The premises. “Most of us would agree, after thought, that the resulting person is me. I shall here assume such agreement” (p. 5). Footnote 5 says those who disagree need a different, multiple case, which the paper does not give. People “have in fact survived with half their brains destroyed”, and each half is taken to be “exactly similar” (p. 5). Attitudes can be criticised for resting on a false belief or for being inconsistent (p. 9).
Division (§I, pp. 4–10). Each half of my brain is housed in a new body. Of three descriptions, “I do not survive” and “I survive as one” are “highly implausible” (p. 5). “I survive as both” is not absurd for a brief case: with control over the bridge between hemispheres one could divide, work two calculations and reunite, so a mental history “could be like a river, with islands” (p. 7). With two bodies for good, “it would become intolerable to deny that they were different people” (p. 7). The relation to each resulting person “contains all that interests us—all that matters—in any ordinary case of survival” (p. 10). The relation I bear to each is the one I would bear to the only survivor had the other half been destroyed, and that would be survival. So division lacks only uniqueness: “To regard these as the same is to confuse nought with two” (p. 10).
The inference (§§II, IV). “Identity is a one-one relation. Wiggins’ case serves to show that what matters in survival need not be one-one” (p. 10). “Identity is all-or-nothing. Most of the relations which matter in survival are, in fact, relations of degree” (p. 11). Fusion, in which compatible traits combine and incompatible ones cancel, is “not clearly survival” and “not clearly failure to survive” (p. 19).
The relations, described without identity (§III). Q-memory is a belief that seems like a memory belief, about an experience someone had, on which it depends “in the same way (whatever that is)” as a memory depends on its experience (p. 15). Connectedness, the holding of direct relations, is not transitive and comes in degrees; continuity, overlapping chains of them, is transitive (p. 20). A criterion survives branching: X and Y are the same person if psychologically continuous and no contemporary person is continuous with either (p. 13), offered as sufficient, not necessary (fn. 17).
The dilemma, as the survey states it. Fission forces psychological views either into multiple occupancy, on which two people were there all along (Lewis, Perry), or into a non-branching clause, on which fission is death. Non-branching makes survival depend on whether the other hemisphere is destroyed, which the entry calls “mysterious”, and it is why Parfit and others conclude that identity is not what matters (LIT-454 §5; NOTE-374’s C5, strong as a dilemma).
What could have come out otherwise. The relation to each product could have differed from the relation in a single transplant in something besides number. By the construction of the case it does not, and that is what the argument uses.
What this does not say
- That it binds those who deny transplant survival. The paper assumes agreement and concedes in fn. 5 that dissenters need a case it does not supply.
- That every question about personal identity lacks an answer. That is one of the paper’s targets, and it is made “implausible”, not refuted: the author “cannot see how to disprove” the belief (p. 4).
- That connectedness matters more than continuity. Asserted: it “would need more arguments than I have space to give” (p. 21).
- That self-interest has no rational force. §VI argues this in two pages with an analogy to nationalism. The reading rates it weak.
- That a copy, a recording or a transcript is a survivor. Q-memory needs the later belief to depend on the experience “in the same way (whatever that is)” as memory does. The paper leaves that open, and it is where cases of re-instantiation are decided. Birch (LIT-111, NOTE-183) argues that a transcript handed to a successor fails it. The paper says nothing about machines.
- That “I” should be given up. §V’s successive selves, drawn by degree of connectedness and “left to the choice of the speaker” (p. 25), are a proposal about how we could talk, offered as such.
Connections
- Rivals. Lewis (THEORY-476) grants both premises of the inference, that identity is one-one and all-or-nothing and that what matters may branch and come in degrees, and denies the conclusion by comparing the relation among stages of one person, not identity, with the psychological relation. Williams (THEORY-531), which Parfit’s fn. 11 answers, argues that from the first person the question “shall I survive?” must have an answer, carries a risk, and leans to the body.
- The criterion. The non-branching clause is the repair, set out in THEORY-546, that lets Locke’s psychological criterion survive division. Parfit’s fn. 2 says division is “implicit” in Locke’s II.xxvii.18.
- Presupposed by Dufner’s map of identity and ethics (THEORY-464), which takes the separation of identity from what matters as its premise.
- Dennett’s “Where Am I?” (LIT-405, NOTE-338). The brain and its program in step are Parfit’s divided mind before reunion; their drift is division without reunion.
- The record’s own claim. CLAIM-287 holds that organizational continuity “can branch and merge, so it is not an identity relation”. That is this account’s inference made for selves that persist through inactivity, and this reading corroborates it.
- AI consciousness. THEORY-023 draws on Birch’s Challenge One, which argues from Parfit’s later Relation R that a chatbot character has no persisting bearer. This paper is where the relations and their branching are first set out.
Note of 2026-10-29: what holds it back
Kept Proposed by the tenth status review. NOTE-378 read the paper in full. Its first premise, survival of a brain transplant, is assumed, and fn. 5 concedes that dissenters need a multiple case the paper does not give. Lewis’s reply (THEORY-476), which grants both premises of the inference and denies the conclusion, is unanswered in the record.