THEORY-076: Diachronic self-governance is constituted by cross-temporal links of intention that are intrapersonal analogues of shared intention, not by narrative unity and not by intertemporal bargaining

Which unity this is about

Of the four unities the record keeps apart (ADR-024), this account is about self-governance: what makes an agent’s conduct over time governed by her, and not merely managed. It says something about the other two only by denying that they constitute it. Intertemporal conflict and its resolution belong to behavioral-integration, and the bargaining account is an account of that. A life’s unity as a story belongs to narrative-unity. The claim is that neither is what makes agency over time self-governed. It is not a claim that either does not exist.

Source

• Bratman (2018), LIT-567, read in NOTE-464 in the author’s draft of 10 July 2017. Page references are to that draft.

What was actually shown

Self-management is not self-governance. Diachronic self-governance needs synchronic self-governance at the relevant times, guidance by a standpoint “sufficiently coherent to constitute a clear place where the agent stands” (p. 2), but it is “not merely a concatenation” of it (p. 3). Ulysses at the mast manages himself over time “thereby blocking his self-governance when the sirens call” (p. 2). Sartre’s young man, shuffling between options he judges non-comparable, is self-governing at each time and still less self-governing over time (§§2, 7). So something links the times.

The link is the structure of acting together. Bratman’s conditions for shared intention are (A) interlocking intentions, (B) intended mutual responsiveness and meshing sub-plans, (C) correct belief in persistence interdependence, (D) expectation of success and (E) openness (pp. 8–9). In planned activity, my intention at t1 to A at t2 is meant to work through my later intention, my sub-plans then are meant to mesh with those I formed earlier, and so on. These are “intra-personal analogues”, so that the agent at each time “is acting ‘together’ with … herself — at other times” (p. 9). He grants that the analogy “will at most only be partial” (p. 10): there is no two-way causation across time, and condition (C) holds only in some cases (NOTE-464, C3).

Willpower without a bargain. In temptation, a prior resolution and a present evaluation cannot by themselves say why the resolution speaks for the agent. Bratman adds a third element, a “conditional, reflexive end” of one’s own diachronic self-governance in the standpoint, which is not a taste for “mere cross-temporal psychic tidiness” (§§6–7). The plan’s time-frame runs to “plan’s end” and includes later reflection. So sticking with a resolution is supported because one anticipates no regret at the end, and drinking Kavka’s toxin is excluded because one anticipates regret (§8; NOTE-464, C4–C5).

Against the two alternatives.

• Bargaining. McClennen’s “fair bargain between one’s different, time-defined selves”, which the essay groups with Ainslie’s picoeconomics, “seems to involve a kind of reciprocal inter-relation that is not possible here”, since “the person-stage at t1 will not be around” to respond when the later stage acts (p. 8). That is the whole argument (NOTE-464, C6).
• Narrative. A “story” that ties a life together “may apply to many cases. But according to the shared agency model, the more basic cross-temporal unity at work in diachronic self-governance is the unity involved when the agent is acting ‘together’ with herself” (p. 9). That is the whole treatment.

So the positive model is argued at length and by cases, and the two denials in the title are asserted. The account is the record’s statement of the positive model, with the denials kept at the strength the essay gives them.

What this does not say

• It does not say the analogy is exact. Bratman says it is partial. The account claims only that the analogues the essay names are what constitute the cross-temporal link, not that a person over time is a group.
• It does not say narrative unity is unreal or unimportant. Only that it is not what makes agency over time self-governed. McAdams’s narrative identity (THEORY-060) is a claim about a different unity, and the two can both be true.
• It does not refute the bargaining account as an account of self-control. Bratman’s objection is that a bargain needs reciprocity. Ainslie’s interests are successive and do not respond to each other either; his “bargain” is a repeated game played through self-prediction. Whether reciprocity is needed for that is not argued. The two accounts are rivals over what makes resolve self-governed, and over what carries it in practice they make different predictions (see promote_when).
• It does not show that agents have the reflexive end. It is posited by inference to the best explanation, and other ends can “trump” it when it is not “sufficiently important” to the agent (p. 22).
• It does not settle manipulation. The model is wholly structural. It has no condition on how the plans were acquired (THEORY-040).

Connections

• Ainslie’s picoeconomics, THEORY-067, declared this theory’s rival: a theory that self-control over time needs no faculty of will, because interests dominating at different delays bargain and resolve is the recursive staking of future choices on the present one. It rests on the précis LIT-586, which itself rivals this essay in the LIT scheme. The two are rival accounts of the same phenomenon, resolve against a temptation one foresaw. On Ainslie’s account resolve is credibility in an intertemporal game. On this one it is acting together with oneself, backed by an end of self-governance. The précis returns the challenge: it suggests Bratman’s “resolute choice” “may turn out to be another name for intertemporal bargaining” (NOTE-443). Neither tests the other. The rivals relation should be declared once that theory is complete.
• THEORY-029 (the regress). Bratman’s stopping point is structural: a coherent standpoint plus a reflexive end. Because the end can be outweighed, it is one weight among others and not a guarantor of ownership, which fits THEORY-029’s claim. Whether the end can be specified without presupposing the self-governance it is an end of is the question THEORY-029 asks, and this essay does not ask it.
• Korsgaard’s normative constitution (THEORY-065). Korsgaard reads Bratman’s Lockean claims as saying that “by conforming to the norms consequent upon intention, we constitute our identities” (LIT-587, ms. p. 23). On that reading the interlocking intentions here constitute self-governance because of the norms of consistency and means–end coherence they commit the agent to, and the two accounts agree. They part where Korsgaard reports that Bratman counts a post-hypnotically installed intention as the agent’s own (ms. p. 21), and where she holds that being active needs conformity to the categorical imperative, which nothing here requires. That is a disagreement about agency at a time, not about the link across times, so no rivals is declared.
• The empirical psychology of self-control (the behavioral-integration holdings) can tell this account from the bargaining one only through a test like the one in promote_when. The record holds none.