Contribution
The paper turns the division case from a puzzle about which answer is right into an argument that the question does not matter. Before it, fission was an objection to the psychological criterion (Williams’s reduplication argument). After it, fission is the case that separates identity from what matters, and the psychological relations become the object of concern in their own right. It also introduces q-memory into the identity debate, the connectedness/continuity distinction, and a vocabulary of successive selves.
Key insight
Ask what would make division worse than ordinary survival. The relation I bear to each resulting person is the one I would bear to the only resulting person if the other half-brain had been destroyed, and that would be survival. So the only thing division lacks is uniqueness, and uniqueness cannot be what I care about: “To regard these as the same is to confuse nought with two” (p. 10). Identity is what we use to imply the relations that matter, because in fact they come one-to-one.
Assumptions
- Brain-transplant survival. “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.
- Half-brain survival is possible. “People have in fact survived with half their brains destroyed” (p. 5), and each half is “exactly similar” (p. 5).
- Attitudes are criticisable. An emotion “can be criticized for resting on a false belief, or for being inconsistent” (p. 9). The division argument works by showing regarding division as death to be one or the other.
- Psychological continuity includes causal continuity (p. 11). Footnote 19 relies on it to answer Williams’s objection about exact similarity.
- The cases are imagined. The paper says the problem cases “could never occur” may be false, “some, for instance, might become scientifically possible”, but its claims do not need them to occur (p. 3).
Key results
- §I. Division (pp. 4–10).
- Three descriptions of Wiggins’s case: I do not survive; I survive as one; I survive as both. The first and second are “highly implausible” (p. 5).
- “I survive with two bodies and a divided mind” is not absurd. The exam case: with voluntary control over the bridge between the hemispheres, one could divide one’s mind, work two calculations, and reunite. “A person’s mental history need not be like a canal, with only one channel. It could be like a river, with islands” (p. 7).
- But with two bodies, permanently, “it would become intolerable to deny that they were different people” (p. 7). Keeping identity by letting two people compose a third (“the Pope’s three crowns”) changes the concept of a person; the duel in footnote 8 shows the cost.
- The paper’s first conclusion: the relation to each resulting person “contains all that interests us—all that matters—in any ordinary case of survival” (p. 10).
- §II. A criterion that survives branching (pp. 10–14).
- “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).
- Williams objected that a criterion must appeal to a logically one-one relation. Parfit answers that psychological continuity is “a ground for speaking of identity when it is one-one” (p. 12). He turns Williams’s own principle, that an important judgment be asserted and denied only on importantly different grounds, against the Guy Fawkes argument: if continuity is the important ground, the branching case should attach the importance to each branch (p. 13).
- The criterion: “X and Y are the same person if they are psychologically continuous and there is no person who is contemporary with either and psychologically continuous with the other” (p. 13). Footnote 17: a sufficient, not a necessary, condition.
- §III. Relations without identity (pp. 14–18).
- Q-memory: “I am q-remembering an experience if (1) I have a belief about a past experience which seems in itself like a memory belief, (2) someone did have such an experience, and my belief is dependent upon this experience in the same way (whatever that is) in which a memory of an experience is dependent upon it” (p. 15). Following Shoemaker’s quasi-memory.
- That the remembered experience was mine “is something that I automatically assume”, justified only because we do not in fact q-remember others’ experiences (p. 15). “Memories are, simply, q-memories of one’s own experiences” (p. 16); this “vindicates the ‘memory criterion’ of personal identity against the charge of circularity”.
- Q-intention: the original person in division could “q-intend, as one of them, to continue his present career, and, as the other, to try something new” (p. 17), with limits set out in footnote 28.
- A “bolder claim”, that experiences might be thought of “in a wholly ‘impersonal’ way”, is named and not developed (p. 17).
- §IV. Degree (pp. 18–21).
- Fusion: two people’s bodies grow into one; compatible tastes and intentions combine, incompatible ones cancel or weaken (Palladio and Giotto; red hair and votes). Fusion is “not clearly survival”, “not clearly failure to survive”, “and hence … what matters in survival can have degrees” (p. 19).
- Beings who reproduce by natural division form a “tree”. Connectedness “requires the holding of these direct psychological relations” and “is not transitive”; continuity “only requires overlapping chains” and “is transitive” (p. 20).
- “I suggest that connectedness is a more important element in survival. As a claim about our own survival, this would need more arguments than I have space to give” (p. 21).
- §V. Successive selves (pp. 21–25).
- “Descendant self” and “ancestral self” imply continuity; “past self” and “future self” imply connectedness and come in degrees, from “my most recent self” to “not in any way one of my past selves—just an ancestral self” (p. 22).
- For beings that fuse every autumn and divide every spring, “I” would cover too little; for everlasting beings with fading connectedness, it would carry none of its usual implications (pp. 23–24).
- The revision: draw the line between successive selves by degree of connectedness, “left to the choice of the speaker” (p. 25). Unlike Proust, “there is no underlying person who we both are”. Whether to say “I” or “one of my future selves” “is entirely a matter of choice”.
- §VI. Consequences (pp. 26–27).
- “The principle of self-interest has no force.” Its competitors are biased rationality (“do what will best achieve what you actually want”) and impartiality; it borrows its force from one or the other.
- Egoism, the fear of distant death, and regret that so much of “one’s only life” has gone by “are all strengthened by the beliefs about personal identity which I have been attacking” (p. 27).
- The closing question: can these be weakened without weakening “loyalty to, or love of, other particular selves”? It ends on Hume, hoping “more of what is bad depends upon false belief”.
Claims
| id | claim | strength | support |
|---|---|---|---|
| C1 | In division, each resulting person stands in all that matters in survival | strong given the assumptions (transplant survival, half-brain sufficiency, exact similarity) | §I; the argument from “double success” and the drug analogy (p. 10) |
| C2 | So what matters in survival is not identity | strong as a conditional on C1 | identity is one-one; C1’s relation is not (pp. 10–11) |
| C3 | Not every question about personal identity has an answer | moderate | division makes the belief “implausible”; the author says he “cannot see how to disprove” it (p. 4) |
| C4 | Non-branching psychological continuity is a sufficient criterion of identity | moderate | §II; continuity left unspecified (“we should need to explain what we mean”, p. 13) |
| C5 | Memory and intention can be redescribed without presupposing identity | strong for coherence of the concepts | q-memory and q-intention definitions (pp. 15, 17); the dependence clause is left open “(whatever that is)“ |
| C6 | What matters in survival comes in degrees | moderate | fusion (pp. 18–19) and connectedness over long spans (p. 21) |
| C7 | Connectedness matters more than continuity | asserted | ”would need more arguments than I have space to give” (p. 21) |
| C8 | Whether to call a later self “I” is a matter of choice once the facts of continuity and connectedness are fixed | asserted as a proposal | §V |
| C9 | The principle of self-interest has no independent rational force | weak | two pages; the nation analogy carries it (pp. 26–27) |
Concepts
- survival — at first equivalent to “will there be some person alive who is the same person as me?”; then used in a sense that “does not imply identity”, so that one can survive as two (pp. 9, 14).
- q-memory / q-intention — memory and intention minus the requirement that the experience or action be one’s own (pp. 15, 17).
- psychological connectedness — the holding of direct psychological relations (q-memory, q-intention, a lasting q-characteristic); not transitive, a matter of degree (p. 20).
- psychological continuity — overlapping chains of direct relations; transitive; runs in one direction of time (p. 20 and fn. 29).
- ancestral / descendant self — linked by continuity (p. 21).
- past / future self — linked by connectedness, by degree (p. 22).
Connections
- Williams, “The Self and the Future” (filed beside this reading). Footnote 11 credits it with showing how natural it is to believe that the question “Shall I survive?” must have an answer, all-or-nothing, with a “risk” of getting it wrong. Parfit says those beliefs must be undermined by their causes, and locates them in misunderstanding remembering and anticipating. §II answers Williams’s earlier reduplication argument (1956–57, 1960–61), not the 1970 paper.
- Locke (filed beside this reading). Footnote 2 says division is “implicit” in the Essay II.xxvii.18. §III’s q-memory answers the charge of circularity pressed against memory criteria “from Butler onward”.
- Lewis, “Survival and Identity” (filed beside this reading) is a reply to this paper, accepting C1 and C5–C6 and denying C2.
- Survey (LIT-130, NOTE-166). The survey’s C1 is this paper’s C1–C2. The survey cites Parfit 1984 for it.
- Birch (LIT-111, NOTE-183). Birch’s GP-transcript and no-scan teletransporter cases fail clause (2) of q-memory: the later beliefs do not depend on the earlier experiences “in the same way … in which a memory … is dependent upon it”. The paper leaves that dependence unspecified, which is where Birch’s argument has to do its work.
- Dennett (LIT-405, NOTE-338). The divided-and-reunited mind of the exam case is Hubert and Yorick in step; the permanent division is their drift. Neither cites the other.
- Goldstein and Lederman (LIT-206, NOTE-138). Their instance agents are branches without connectedness to each other; see the LIT entry.
Bearing on the record
It bears on THEORY-023 only at one remove: Birch’s Challenge One, which THEORY-023’s source reading includes, uses Parfitian persistence to deny that a chatbot character has a persisting bearer. This paper is where the relevant relations and their branching are first set out. It does not address machines and carries no instruction for machine-learning practice.
Limitations
What is shown and what is asserted:
- Shown, given its assumptions: C1 and so C2. The assumptions are stated, and footnote 5 concedes the argument needs a different case for those who deny transplant survival, without supplying it.
- Shown as coherent, not as true of us: q-memory and the impersonal description (C5). The bolder “impersonal” claim is named and set aside.
- Asserted: C7, explicitly; and C8, as a proposal about how to speak.
- Sketched: §VI’s consequences for rationality and the emotions. The argument against the self-interest principle is two pages with an analogy to nationalism.
- Not specified: what psychological continuity is, how much the criterion requires (p. 13), and the “same way” of dependence in q-memory.
Open questions
- What is the dependence condition in q-memory, and does a non-standard cause (a transcript, a copy) satisfy it? The paper’s parenthesis “(whatever that is)” leaves the question that later decides teletransporter and AI cases.
- If connectedness matters more than continuity, how much does it matter, and at what rate does prudential concern for a future self fall? The paper gives degrees without a measure.
Corrections
None. The brief’s citation is correct in every detail.