Bratman on identity over time and identification at a time

Abstract

According to reductionists about agency, an agent’s bringing something about is reducible to states and events (such as desires and beliefs) involving the agent bringing something about. Many have worried that reductionism cannot accommodate robust forms of agency, such as self-determination. One common reductionist answer to this worry (which I call “identification reductionism”) contends that self-determining agents are identified with certain states and events, and so these states and events causing a decision counts as the agent’s self-determining the decision. In this paper, I discuss Michael Bratman’s well-known identification reductionist theory and his general strategy of grounding an agent’s identification at a time in the agent’s identity over time. I develop two constraints that an adequate identification reductionist theory must satisfy, argue that Bratman’s theory cannot satisfy both, and show that his general strategy for grounding an agent’s identification at a time in the agent’s identity over time is without merit.

Keywords:

• identification
• self-determination
• reductionism
• causal theory of action
• Bratman

Acknowledgements

Earlier drafts of this paper were presented at Biola University, University of Edinburgh, the 2014 Central Division Meeting of the American Philosophical Association, and the 2014 Eastern Division Meeting of the Society of Christian Philosophers. I am grateful to these audiences, as well as Justin Coates, Ben Mitchell-Yellin, Neal Tognazzini, and two anonymous referees of this journal for their helpful feedback.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes on contributor

Christopher Evan Franklin received his PhD from the University of California, Riverside, and works at the intersection of ethics and metaphysics, on problems of agency, free will, and moral responsibility. His publications have appeared in Australasian Journal of Philosophy, Mind, Pacific Philosophical Quarterly, and Philosophical Studies, among other venues. He is assistant professor of philosophy at Grove City College.

Notes

1. I think it is clear that reductionist accounts have an easier time accounting for these minimal forms of agency, although it is not clear that they can in fact adequately analyze them (cf. Steward 2012). My aim in this essay, however, is to argue that an important variety of reductionism cannot account for self-determination. We can assume, then, for the sake of argument, that reductionists can unproblematically account for minimal forms of agency.

2. The notion of self-determination I am after differs from the one recently discussed by Buss (2012). She gives a functional definition of self-determination or autonomy:

An agent stands in the self-governing self-relation that interests me here if and only if the roles she plays in forming her intention is such that if she has the general capacity to appreciate the force of moral requirements and if she has adequate opportunity to discern the moral significance of what she is doing, she is blameworthy if her action is morally wrong and praiseworthy if it is morally admirable. (2012, 649)

As I understand the literature on identification, the notion of self-determination at play there is not as intimately connected to moral accountability as Buss’s notion. For example, Velleman is explicit that one can be morally accountable for an action even if one does not self-determine it. For Velleman, moral accountability requires the capacity for self-determination, not its exercise (1992, 127 n. 13). For this reason some of Buss’s objections to these accounts of self-determination miss their target (cf. Mitchell-Yellin 2014)

3. Bishop (1989) is an example of a reductionist model that restricts the reductive base of states and events to desires, beliefs, and intentions, even when seeking to account for self-determination.

4. See Lippert-Rasmussen (2003) for an insightful critical discussion of all species of identification reductionism except care-based accounts. See Franklin (2015, n.d.) for objections to reason-based and care-based accounts. It seems to me that Watson (1987) himself has raised the most pressing problem for his earlier good-based account (Watson 1975), but see Mitchell-Yellin (2015) for a recent defense of good-based identification reductionism.

5. Jaworska (2007) does just this. She agrees with Bratman’s strategy for defending claims of identification and yet contends that it is an agent’s cares that he is identified with. Part of Jaworska’s aim is to defend an account of identification that allows for the possibility of marginal agents being identified with their attitudes, agents such as small children and Alzheimer’s patients, who may not yet have or have lost various higher-order cognitive capacities. Note also that Bratman (2000, 44 n. 60) contends that his account of identification entails that an agent is identified with his cares as defined by Frankfurt (1993, 1994). I raise a number of objections to Frankfurt and Jaworska’s (as well as Shoemaker’s) care-based identification reductionist theory in Franklin (n.d.).

6. Other offered glosses for mental states and events with which an agent is identified include: they “constitute one’s standpoint” (Watson 1975, 26), they have “authority to speak for the agent” (Bratman 2007a, 4), and they are “fully the agent’s own” (Jaworska 2007, 537).

7. The skeptic might be inclined to complain that he has no clear sense of what all this identification talk is about. While I find the notion of identification intelligible and important in its own right, any skepticism about its intelligibility simply casts further doubt on the tenability of identification reductionism.

8. It is important to note the order of the quantifiers in the scope constraint, the existential quantifier being within the scope of the universal. This principle does not entail that for every self-determined action there is the same state that the agent is identified with, but only that for every self-determined action, there is some state or other that the agent is identified with.

9. While Bratman usually tends to prefer the language of “agential authority” to the language of “identification” (cf. 2000, 24; 2005, 202; 2007a, 4), I will interpret him as offering an account of identification. Bratman claims that solving the problem of agential authority requires the specification of a relationship between the agent and attitude that confers authority on the attitude to speak for the agent. This suggests that in trying to solve the problem of agential authority Bratman can plausibly be interpreted as defending an account of identification. Moreover, see Bratman (2002, 82–85), where he does frame his theory as account of identification. Thanks to an anonymous referee for bringing this reference to my attention.

10. Bratman’s account, therefore, requires the truth of a psychological continuity account of personal identity. This may strike one as a cost since this account of personal identity (while it has its share of able defenders) faces serious objections (see, e.g. Olson [2007]). Basing a controversial account of identification on a controversial account of personal identity seems a precarious route to defending identification reductionism. However, my objections to Bratman’s theory will, for sake of argument, assume the truth of the psychological continuity account of personal identity.

11. Bratman’s conception of psychological continuities differs from Parfit’s (Parfit 1984, 206).

12. Olson (2007, Chapter 6) has argued that adopting a psychological continuity account of personal identity commits one to reducing the person to a set of states and events. If Olson is right, then Bratman is committed to reducing the person. Whether or not this is so, Bratman denies trying to do any such thing.

13. Not all desires will ground psychological connections. For example, the desire that it rain tomorrow is about an extra-mental event, and thus there is no psychological connection between these items. Only when the content of the desire is about other mental features of the agent can it ground psychological connections.

14. This will be true by definition if one endorses a functionalist theory of the mental. But even if one is inclined to reject functionalism, it is hard to deny that part of what makes a mental state a desire is its characteristic dispositions to bring about action.

15. Given the holistic structure of the mental, the desire to inhale cocaine will be connected to other desires, beliefs, and intentions. Although the atomistic desire to inhale cocaine may not speak for the agent, one might argue that the holistic structure of the addict’s mental life does speak authoritatively for the agent. In response, it is not clear what it means to say “the holistic structure speaks authoritatively for the agent”? On Bratman’s account, it is attitudes that speak for the agent, but a holistic structure is not an attitude, so how can it speak for the agent? But even supposing we can make sense of the idea of the holistic structure speaking for the agent, this point is of no help for Bratman. Bratman contends that an attitude grounding Lockean ties is sufficient for the agent being identified with the attitude. The above example of the unwilling addict shows that this is a mistake: the unwilling addict is alienated from his addictive desire even though this desire grounds Lockean ties. I am grateful to an anonymous referee for raising this worry.

16. Jaworska seems to be led astray on this point. She writes: “For example, a prior policy and the subsequent execution of the policy are so connected”, where by “connected” she means “a reference-based link between mental states occurring at different times in a temporally extended life” (2007, 550). Bratman never claims this, but only that the agent understanding the execution as an execution of the policy refers back to the policy. Jaworska further weakens the idea of referential connection between attitudes by introducing the idea of an “indirect” referential connection: “two mental states are referentially connected because they both refer (in a consistent way) to the same object” (2007, 553). I do not see how there is any kind of referential connection between two attitudes simply in virtue of each attitude referring to, or being about, the same object (even in a consistent way). I believe that the Lakers will be the next NBA champions and I desire that the Lakers will be the next NBA champions. Each attitude refers to the same object, but neither does the belief refer (not even indirectly) to the desire nor the desire to the belief. Jaworska’s notion of “indirect reference” is, I submit, not a notion of reference at all.

17. I am grateful to an anonymous referee for raising this worry.

18. While Bratman suggests that conditions (i) and (ii) are jointly sufficient for one policy’s undermining another policy, he does not say whether or not they are individually necessary or individually sufficient.

19. In fact, Bratman only offers this as a “preliminary proposal” of self-determination (2000, 35). He later qualifies it to include quasi-policies. On this extended account, an agent’s decision is self-determined just in case it is caused by a desire D and the agent has either a self-governing policy or quasi-policy with which he is satisfied that treats D as reason-giving in motivationally efficacious practical reasoning. Quasi-policies bear a “special relation to our temporally extended agency” and for this reason can also have authority to speak for the agent (Bratman 2000, 42). An example of a quasi-policy would be an agent’s ideal of citizenship. Such an ideal would resemble self-governing policies in that it would accord weight to desires that promote the ideal and discount desires that tend to violate it (Bratman 2000, 42). Quasi-policies, Bratman imagines, are unlike self-governing policies in that they do not involve intentions to do certain things or treat desires in certain ways, and so acting against a quasi-policy will not, like acting against a self-governing policy, involve a strong kind of inconsistency (Bratman 2000, 43). I will ignore this complication since it has no bearing on my criticism of Bratman’s theory.

20. The same problem besets Frankfurt’s own analysis of satisfaction. According to Frankfurt, when an agent is satisfied there is “an absence of restlessness or resistance” (1992, 103). While such an analysis of satisfaction may allow his account to satisfy the alienation constraint, it does so at the expense of violating the scope constraint: we can perform self-determined actions in the presence of restlessness or resistance. We can act self-determinedly even when we are dissatisfied.

21. The desire here is not a desire to spend time with his family or a desire for his family’s well-being, but rather a desire concerning a specific way to care for his family: namely by devoting his summer to them. I am not imagining that Kevin is in conflict about whether to care for philosophy or his family, but in conflict about how to spend his summer.

22. Two worries might spring to mind here. First, one might worry that Kevin’s decision is not self-determined because this is a case of weakness of will. To avoid this worry, we can stipulate that Kevin’s moment of awakening first causes him to judge that it is, all-things-considered, best to devote his summer to his family, and that he then makes the decision to devote his summer to his family on the basis of this judgment. Second, one might worry that in making the decision to devote his summer to his family, Kevin is “decisively identifying” with his desire to devote his summer to his family (cf. Frankfurt 1987, 168 [Reprinted in Frankfurt 1988, 159–176]). The problem with appealing to decisions as decisive identifications is that they do not help further the reductionist project. The identification reductionist is committed to there being, for all self-determined decisions, a prior motivation that causally contributed to the decision and with which the agent is identified. Even if Kevin’s decision thereby makes him identified with his desire to devote his summer to his family, Bratman cannot exploit this since he needs to show that Kevin was already (i.e. before he made the decision) identified with his desire to devote his summer to his family. I am grateful to an anonymous referee for raising these worries.

23. Another counterexample would be a case in which Kevin performs a self-determined action and yet this action is not motivated by any self-governing policy, let alone a self-governing policy with which he is satisfied. This strikes me as possible. However, some have told me in conversation that it does not seem possible to them. I thus offer the more complex counterexample above for those who think that all self-determined actions are motivated in part by self-governing policies.