On the heels of Steve's post, which came on the heels of mine, which came on the heels of another one of mine, which came on the heels of Steve's post, which came off the heels of another one of Steve's posts, and which came off the heels of Peter's post, and which was caused - through prior factors that Peter had no control over - by my original post, I now bring van Inwagen's argument in its more developed stage: The Luck Objection.
Kane's requirement that the causation of a choice that is an SFW [pm: self-forming wills] be nondeterministic has drawn the objection that indeterminism located here would diminish the agent's control over the making of the choice. The objection is often couched in terms of luck. (It is so developed by Almeida and Bernstein , Ekstrom [2000: 105], Haji [1999a, 1999b, 2000a, 2000b, 2000c, and 2001], Mele [1998, 1999a, 1999b, and forthcoming], and Strawson .) If the agent's effort of will nondeterministically causes her choice, then, whichever choice the agent makes, there was, until the occurrence of that choice, a chance that it would not occur. If the agent's effort to chose in accord with her moral judgment happens to succeed, the objection goes, then her choice is at least partly due to good luck. In another possible world with exactly the same laws of nature and exactly the same history up until the occurrence of the choice, the agent's (or her counterpart's) effort fails; there, but for good luck, goes she. And analogously, if, in the actual world, the agent's effort fails, then her choice is at least partly due to bad luck. Either way, the choice is to some degree due to luck. And to that degree, the objection concludes, the control that the agent exercises in making the choice is diminished.
Kane's claim that indeterminacy precludes exact sameness has been contested (see Clarke [1999 and 2003b: 86-87] and O'Connor ). And Haji (1999a) and Mele (1999a and 1999b) contend that the argument from luck is just as effective if we consider an agent and her counterpart who are as similar as can be, given the indeterminacy of their efforts. Indeed, the argument might be advanced without any appeal to other worlds or counterparts: given that there is a chance that the effort will fail, the agent is lucky, it may be said, if it succeeds.
A further reply from Kane to the argument from luck appeals to the active nature of efforts of will. When an agent makes an effort to choose to do what she believes she ought to do, she actively tries to bring about a certain choice. When the agent makes that choice, she succeeds, despite the indeterminism, at doing what she was (actively) trying to do. And Kane points out that typically, when this is so, the indeterminism does not undermine responsibility (and hence it does not so diminish active control that there is not enough for responsibility). He describes a case (1999b: 227) in which a man hits a glass tabletop attempting to shatter it. Even though it is undetermined whether his effort will succeed, Kane notes, if the man does succeed, he may well be responsible for breaking the tabletop.
If left here, the reply would fail to address the problem of luck in a case where the agent chooses to do what she is tempted to do rather than what she believes she ought to do. In response to this shortcoming, Kane (1999a, 1999b, 2000b, 2000c, and 2002) has recently proposed a "doubling" of effort in cases of moral conflict. In such a case, he now holds, the agent makes two, simultaneous efforts of will, both indeterminate in strength. The agent tries to make the moral choice, and at the same time she tries to make the self-interested choice. Whichever choice she makes, then, she succeeds, despite the indeterminism, at doing something that she was actively trying to do.
This doubling of efforts of will introduces a troubling incoherence into cases of moral conflict. If an agent is actively trying, at one time, to make each of two obviously incompatible choices, that fact raises a serious question about the agent's rationality.
A final difficulty for agent-causal views accepts that all they require might be possible. The objection may still be raised that actions produced as required by such an account would be too subject to luck to be free actions. Van Inwagen has raised a similar objection to agent-causal accounts—though without referring to luck—on several occasions (see his 1983: 145 and 2000). Haji (2004) and Mele (forthcoming) present the objection in terms of luck as follows. Recall Leo's decision (section 3.1) to tell the truth. Until he makes the decision, there remains a chance that he will not decide to tell the truth, but will instead decide to lie. Likewise, until he makes the decision, there remains a chance that he will not cause a decision to tell the truth, but will instead cause a decision to lie. Then, in some possible world W with the same laws as those in the actual world, and with the same history up to the time of the decision, Leo decides at that time to lie, and he causes that decision to lie. The actual world, where Leo decides to tell the truth (and causes that decision), and world W, where he decides to lie (and causes that decision), do not differ in any respect until the time at which Leo makes the decision (which is also the time at which Leo causes the decision). There is, then, no difference between these two worlds to account for the difference in the decision, and likewise no difference to account for the difference in Leo's agent causings. Hence the difference between these two worlds is just a matter of luck. But if the difference between these two worlds is just a matter of luck, then Leo does not freely make his decision in the actual world.
Basically, control of action is required for moral responsibility and free will. If the action is due to luck, i.e., nothing determines wether an agent will A or B, and if he A'ed then he is worthy of prase but if he had B'ed he'd be worthy of blame, then it seems the control required for freedom and moral responsibility is only to be found in determinism and a compatibilism.
His philosophical task, then, is to show that a choice may be free, in the sense that it is something for which its agent is morally responsible, even though it was not guaranteed by antecedent mental processes. He must explain how the taking of a decision can be free while not the necessary outcome of the reasoning from which it issues. One might think, however, that I am responsible for choosing, e.g., to support a certain cause only if that mental act is secured by the exercise of my reasoning ability. Moreover, having weighed the cause’s pros and cons as I did, unless I was rationally bound to decide in its favor the outcome of this reasoning process seems inexplicable and, thus, not something for which I should be held accountable: praised or blamed. My decision must be rational if praiseworthy or blameworthy, but (unless I am in a situation like that of Buridan’s ass) how could it be rational if the reasons motivating it are consistent with the opposite choice? A mechanism that could take a set of reasons as the basis for more than one course of action appears erratic. Its exercise, thus, would fail to insure a rational result redounding to my credit or discredit.3 Kane, therefore, faces a dilemma: either some actions are undetermined, in which cases the control and rationality requirements of free agency are not satisfied, or a free agent’s conduct is always determined and explicable in terms of reasons that render irrational all but one course of action, in which case the alternative rational possibilities requirement of libertarian free agency can not be met.4 Alternatively, Kane must respond to the following chain argument:
1. If an act is free, then its agent has control over its performance
2. If an act’s performance is controlled by its agent, then it is the product of a reliable mechanism.
3. If an act is free, then it is rational.
4. If an act is rational, then it is the product of a reliable mechanism.
5. If an act is the product of a reliable mechanism, then it is the necessary outcome of the mechanism’s processing of its antecedents (specifically, the reasons arising in its favor).
6. If an act is the necessary outcome of a mechanism’s processing of its antecedents, then it is produced deterministically.
7. Thus, if an act is free, it is produced deterministically.
- Robert Allen