The Psychology of Freedom. - book reviews

Mind, Jan, 1998 by Alfred Mele

By Thomas Pink. Cambridge: Cambridge University Press, 1996. Pp. x 284. 35.00 [pounds sterling].

Pink's announced task is "to determine what states of mind we need if we are to be free to act otherwise than as we actually do" (p. 1). He defends a "Psychologising conception of freedom" according to which "our control over our actions depends on our having a capacity to control how we act through free decision-making" (p. 80). At the heart of the book is his detailed view of decision-making.

Significant theses defended in the book divide into two related groups: theses about decision that do not explicitly mention freedom; theses about freedom. The former group includes the following:

D1. Decisions are "second-order" actions. They are actions that generate other actions "by way of constituting the formation of an intervening action explanatory psychological state" (p. 3).

D2. "Decisions to act are means-end justifiable--justifiable in terms of desirable ends which they themselves are likely to further." This is what "makes decisions ... actions" (p. 8). Decisions, like associated first-order actions, are "governed by reason in its distinctively practical ... form" (p. 271).

D3. Deciding to A "must be motivated by" a desire to A (p. 252).

D4. Decisions are "motivation-perpetuating". "Decisions to do A simply ensure" that the ends that motivated the decisions--typically, "ends which from the outset [the agents] expect doing A to further"--"thereafter motivate the decision maker to A. Decisions cause their execution by perpetuating their own motivation" (p. 274).

D5. Decisions are "rationality-preserving". "If a decision to act is taken rationally, then, in the absence of new information which warrants revising the assumptions on which the decision was based, the action which executes the decision will be rational too" (p. 93). By making decisions "we exercise future action control in a way that ensures our continuing rationality as agents" (pp. 6-7).

D6. Reason-Apply: "Any end E that justifies deciding to A must, supposing that decision is taken, also provide at least as much justification for doing A" (p. 153).

D7. Action: "Any justification at t for then deciding or intending to do A consists in the likelihood at t that so deciding or intending would further a desirable end E' (p. 145). Given that decisions are actions, Action is implied by

Justify: "Any justification for doing A consists in the likelihood that doing A would further a desirable end E" (p. 139).

D8. "Deciding now, rather than later or not at all, to do A typically increases what an agent is likely to gain were he to do A--and to increase what the agent is likely to lose were he not to do A", thereby increasing the "agent's justification for doing A later" (p. 240).

D9. The function of decision-making "is to ensure that the actions which we perform at any given time arc those which are justified given the actions which we perform at other times" (p. 209).

In defending D2, Pink is careful to observe that it does not entail that the reasons for which agents decide to A are reasons specifically for so deciding. Indeed, the reasons for which we decide to A are often just reasons we have for A-ing (pp. 145-6). What justifies us in deciding to A can go significantly beyond the reasons for which we decide to A. I might decide today to visit London next summer, and the reasons for which I so decide might be limited to my London-recommending reasons. What justifies me in so deciding, Pink contends, must conform to Action (D7). My deciding well in advance to visit London next summer is likely to "increase the benefit I derive" from visiting London (by prompting me to make plans that will enhance the value of my visit, cf. D8 and pp. 130-1). This fact is a justification for deciding today to visit London, even though it plainly is not a reason for visiting London.

Theses D3 through D6 are threatened by cases in which there is a reward for deciding to A that is not contingent upon one's A-ing. For example, in a version of Gregory Kavka's "toxin puzzle", an agent is promised a million dollars for deciding tonight to drink a certain non-lethal toxin tomorrow: the agent understands that he does not need to drink the toxin to gain the money (G. Kavka, "The Toxin Puzzle", Analysis, 43, 1983, pp. 33-6). In this case, it seems, if the agent does manage to decide to drink the toxin, that decision was not motivated by a desire to drink the toxin, might not be motivation-perpetuating, and would not be rationality-preserving. Furthermore, an agent who decides to drink the toxin apparently has justification for deciding to drink it without having (as much) justification for drinking it.

Pink devotes considerable ink to the puzzle (indeed, much more than his index indicates: see pp. 147-59, 177-83, 192-200; cf. pp. 201-4). The bottom line regarding Reason-Apply (D6) is this: we have no justification for deciding to drink the toxin; rather, we have justification for taking steps to bring it about that we decide to drink it and for deciding to take such steps (pp. 1989). If that is right, the threat to Reason-Apply is dissolved. But one may still wonder about D3, for example. Elsewhere (A. Mele, "Intentions, Reasons, and Beliefs: Morals of the Toxin Puzzle", Philosophical Studies, 68, 1992, pp. 171-94), I argued that a rather special agent can win the prize for deciding to drink the toxin under conditions that are even more stringent than those laid down by Kavka's billionaire; and my agent had no desire to drink the toxin prior to deciding to drink it. However, I should think that relatively modest modifications in Pink's account of deciding can accommodate the possibility of such an agent (assuming that it is a possibility): room can be made for special cases.