Surrounding Free Will: A Response to Baumeister, Crescioni, and Alquist
This contribution to a symposium on an article by Roy Baumeister, A. William Crescioni, and Jessica Alquist focuses on a tension between compatibilist and incompatibilist elements in that article. In their discussion of people’s beliefs about free will, Baumeister et al. sometimes sound like incompatibilists; but in their presentation of their work on psychological processes of free will, they sound more like compatibilists than like incompatibilists. It is suggested that Baumeister and coauthors are attempting to study free will in a metaphysically neutral way and that, because this is so, the incompatibilist elements of the article are out of place.
Roy Baumeister, A. William Crescioni, and Jessica Alquist report that their interest in free will grew out of their research on self-control . My own interest in free will had a similar source. I had developed a view about the nature and functions of self-control in Mele , and the guiding question of my first in-depth treatment of free will  was this: What might be added to an ideally self-controlled agent to yield an autonomous or free agent? Naturally, I did not regard ideal self-control as a necessary condition for free agency, but I did regard it as an ideal place to start thinking about how to understand free will (or autonomy).1
Baumeister and his coauthors study free will from several angles. They study people’s understanding of the idea of free will, people’s beliefs about their own free will, the effects these beliefs have on their behavior, and “psychological processes” traditionally associated with free will, including rational decision-making, planning, and exercising self-control. In my opinion, this is an excellent approach. One expects well-crafted studies of people’s understanding of free will to provide useful data for on-going debates about folk conceptions of free will. Regarding studies of the effects of people’s beliefs about free will on their behavior, the finding that increasing people’s subjective probability that scientists have shown that free will is an illusion increases morally shoddy conduct has provided skeptics like me about scientific argumentation for the thesis that free will is illusory with additional motivation to explain why the data fall well short of justifying this thesis: on this, see Mele . Finally, given traditional associations between free will, on the one hand, and rational decision-making, planning, and self-control, on the other, careful studies of these psychological processes may be expected to shed light on free will.
My primary focus in this article is a certain tension between an element of Baumeister et al.’s discussion of people’s beliefs about free will and their position on psychological processes of free will. Some philosophical background on compatibilism and incompatibilism helps bring the tension to the surface. Sketching that background is the first item of business below.
Compatibilism and Incompatibilism
One major division in the philosophical literature on free will is that between compatibilists and incompatibilists. Compatibilism and incompatibilism are theses about the conceptual relationship between free will and determinism. Determinism, according to the standard philosophical definition of the term, is the thesis that a complete statement of the laws of nature together with a complete description of the condition of the entire universe at any point in time logically entails a complete description of the condition of the entire universe at any other point in time. (You and I are parts of the universe, and a description of what we are doing at this moment is part of a complete description of the universe at this moment.) Compatibilism is the thesis that free will is compatible with the truth of determinism. For reasons having to do with leading contemporary theories in physics, the great majority of contemporary compatibilists do not believe that determinism (as defined here) is true; but they do believe that even if it were true, that would leave it open that we have free will. Incompatibilism is the thesis that free will is incompatible with the truth of determinism. In the incompatibilist group, most answers to the question what free will means come from libertarians. Libertarianism is the conjunction of incompatibilism and the thesis that free will exists. Some incompatibilists argue that free will does not exist. They argue that even the falsity of determinism creates no place for free will.
The compatibilist thesis sounds self-contradictory to some nonspecialists. When people first encounter the pair of expressions free will and determinism, they may get the impression that the two ideas are defined in opposition to each other—that they are mutually exclusive by definition. But consider the following conversation between two police officers who have a notoriously stingy friend named Carl. Ann: “Carl gave $20 to a homeless man today.” Bill: “Why? Did he hold a gun to Carl’s head?” Ann: “No, Carl freely gave him the money.” As far as I can see, Ann and Bill do not need to have an opinion about whether determinism (as defined above) is true to have this conversation. If what Ann says is true—that is, if Carl freely gave away $20—and if free will is simply the power to act freely, then Carl has free will (or had it at that time). Even if free will is typically opposed to determinism in ordinary speech, he freely did it seems not to be. And even if he freely did it were typically opposed to determinism in ordinary speech, that would settle nothing. After all, in ordinary speech, deductive reasoning seems to be defined as reasoning from the general to the particular, and that would only jokingly be said to constitute an objection to a logician’s definition of deduction (according to which “Ann is a police officer; Bill is a police officer; therefore Ann and Bill are police officers” is a valid deductive argument). Just as deduction is a technical term in philosophy, so is determinism.
In the preceding paragraph, I mentioned the idea that free will is simply the power to act freely. I hope this sounds like a truism. Even if it is a truism, it is a rhetorically useful one. The expression free will has more of a tendency to conjure up the supernatural in readers’ minds than the expression free action does. Since neither Baumeister and his coauthors nor I wish to be doing any such conjuring, it will be useful to treat free action as the notion in terms of which free will is defined and to focus the discussion on acting freely.
Compatibilist theories of free action emphasize a distinction between deterministic causation and compulsion. If determinism is true, then my eating a ham sandwich for lunch today and my working on this article today were deterministically caused; and so were a compulsive hand-washer’s washing his hands dozens of times today, a delusional person’s spending the day trying to send a text message to the devil, an addict’s using his favorite drug while in the grip of an irresistible urge to do so, and a robbery victim’s handing over money to gunmen who convincingly threatened to kill him if he refused. But there is a difference. I am sane and free from addiction, and I received no death threats today. The basic compatibilist idea is (roughly) that when mentally healthy people act intentionally and rationally in the absence of compulsion and coercion they act freely, and an action’s being deterministically caused does not amount to its being compelled or coerced.2
Many compatibilists have been concerned to accommodate the idea that, for example, if I freely spent the day working, I could have done something else instead. They grant that, if determinism is true, then there is a sense in which people could never have acted otherwise than they did: they could not have acted otherwise in the sense that their acting otherwise is incompatible with the combination of the past and the laws of nature. But, these compatibilists say, the fact that a person never could have acted otherwise in that sense is irrelevant to free will. What is relevant is that people who act freely are exercising a rational capacity of such a kind that if their situation had been different in any one of a variety of important ways, they would have responded to the difference with a different suitable action . For example, although I spent the day working, I would have spent the day relaxing if someone had bet me $500 that I would not relax all day. This truth is compatible with determinism. (Notice that if someone had made this bet with me, the past would have been different from what it actually was.) And it reinforces the distinction between deterministic causation and compulsion. Offer a compulsive hand-washer $500 not to wash his hands all day and see what happens.
Like compatibilists, libertarians tend to maintain that when mentally healthy people act intentionally in the absence of compulsion and coercion they act freely, but they insist that the deterministic causation of an action is incompatible with the action’s being freely performed. (Recall that libertarians believe that determinism is false.) Some libertarian theories of free action assert that people never act freely unless some of their actions are indeterministically caused by immediate antecedents . Whereas the laws that apply to deterministic causation are exceptionless, those that apply most directly to indeterministic causation are instead probabilistic.3 Typically, events like deciding to run five kilometers in a fund raising event—as distinct from the physical actions involved in actually running—are counted as mental actions.4 Suppose that Ann’s decision to run in such an event is indeterministically caused by, among other things, her thinking that she should help raise money in this way. Because the causation is indeterministic, she might not have decided to run given exactly the same internal and external conditions. Some libertarians appeal to indeterministic causation to secure the possibility of acting otherwise that they require for free action.
What libertarians want that determinism precludes is not merely that agents have open to them more than one future that is compatible with the combination of the past and the laws of nature, but also that, on some occasions, which possible future becomes actual is in some sense and to some degree up to the agents. Causal libertarians want something that requires that people themselves be indeterministic in some suitable way—that some relevant things that happen under the skin are indeterministically caused by other such things. The focus is on mental events (or their neural correlates), as opposed, for example, to indeterministically caused tics—and, more specifically, on mental events that have a significant bearing on action (or the neural correlates of these events).
Quantum mechanics, according to leading interpretations, is indeterministic. But indeterminism at that level does not ensure that any human brains themselves sometimes operate indeterministically in ways that are relevant to the production of actions. Empirical discoveries that any indeterminism there may be in the brain is irrelevant to action production would show that we do not have free will on some familiar libertarian conceptions of it.
Why do philosophers continue to argue about whether determinism is compatible or incompatible with free action even though hardly anyone currently involved in the debate believes that determinism is true? Partly because of tradition. In any case, two different kinds of conception of free action have taken shape in this debate, and one kind—the incompatibilist kind—is considerably more demanding than the other.
In their discussion of people’s beliefs about free will, Baumeister et al. sometimes sound like incompatibilists.
In their discussion of psychological processes of free will, Baumeister et al. offer what may be regarded as evidence for the truth of the proposition that people sometimes act freely, but the evidence they offer includes no evidence that determinism (as I defined it) is false. Accordingly, among philosophers who are favorably impressed by this evidence, compatibilists will be considerably more impressed than libertarians. In short, in their presentation of their work on psychological processes of free will, Baumeister et al. sound more like compatibilists than like incompatibilists.
Why do I say that Baumeister et al. sometimes sound like incompatibilists? Consider the following passage from their target article: “The lack of counterfactual thinking in the no-free-will condition can be considered as a straightforward response that is consistent with determinism. After all, if nothing could have happened other than what actually happened, then there are no counterfactuals. Yet counterfactual thinking is an important form of thought for learning and for adaptation to social life.” Recall that compatibilists grant that if determinism is true, then there is a sense in which we could never have done otherwise than we did: our doing otherwise is incompatible with the combination of what happened before we acted and the laws of nature. (Similarly, compatibilists grant that there is a sense in which “nothing could have happened other than what actually happened”: something’s happening at a time t that did not actually happen at t is incompatible with the combination of what happened before t and the laws of nature.) Recall also that compatibilists contend that the fact that a person never could have done otherwise in that sense is irrelevant to free will and that some compatibilists claim that what is relevant is that people who act freely are exercising a rational capacity of such a kind that if their situation had been importantly different, they would have responded to the difference with a different suitable action. In the illustration of this idea that I offered earlier, although I spent the day working, I would have spent the day relaxing if someone had bet me $500 that I would not relax all day. I pointed out that this truth is compatible with determinism (and notice that the proposition in question is a counterfactual).
Imagine that Diane, who is angered by a remark Ed makes, injures him by punching him and that determinism (as I defined it) is true. In this deterministic scenario, such counterfactuals as the following are true: if Diane had kept her hands in her pockets, she would not have injured Ed; and if Diane had simply walked away as soon as she began to feel angry, she would not have injured Ed. Of course, because the scenario is deterministic, there is a sense in which Diane could not have kept her hands in her pockets—and could not have walked away—at that time: her doing so is incompatible with the combination of the past and the laws of nature. But according to compatibilists, this particular metaphysical sense of could not have is irrelevant to free will.
Whether compatibilism or incompatibilism is true is a conceptual question—not an empirical one.5 (Compatibilism, as I defined it, does not include the assertion that, in fact, people sometimes act freely.) But the following questions are empirical: (1) If it is assumed that incompatibilism is true, is it likely that people sometimes act freely? (2) If it is assumed that compatibilism is true, is it likely that people sometimes act freely? Scientists who take a stand on whether compatibilism is true or false do so as theorists, not as scientists. But scientists certainly can take a stand as scientists on questions 1 and 2.
One way is to study phenomena that are associated with free will and can occur both in deterministic and in indeterministic universes: for example, self-regulation and decision making. Presumably, if physicists were to discover that determinism is true, we would not conclude that no one has ever successfully resisted temptation or that no one has ever made a decision . . . . Incompatibilists would conclude that no one has ever freely done these things, but that is another matter. (, p. 332)
As I see it, Baumeister et al. are attempting to study free will in this metaphysically neutral way. But if that is right, the whiff of incompatibilism in their discussion of people’s beliefs about free will is out of place. I have no quarrel with the idea that people sometimes behave as though they were incompatibilists. My discussion here has been about hints of incompatibilist commitments in the thinking of Baumeister and his coauthors and about a tension between such commitments and their discussion of psychological processes of free will.
I am officially agnostic about whether compatibilism is true or false [4, 6]. (In Mele , I develop two overlapping views of free will, one for incompatibilists and another for compatibilists.) I should add that Baumeister and his coauthors sometimes sound like they might be fellow agnostics. They write: “We propose that several specific kinds of behaviors constitute the real phenomena relevant to the idea of free will. Whether they fully deserve to be called free will is beyond our expertise and may turn out to depend on semantic arguments and metaphysical assumptions. Again, our point is that, if free will exists, this is how it works, and if it does not exist, these are the real phenomena that are mistaken for it.” (The phenomena at issue are “rational choice, self-control, planning, and initiative.”) Perhaps they are sitting on the fence between compatibilism and incompatibilism here. According to incompatibilists, even a person who frequently makes rational choices, exercises self-control, makes plans, and exhibits initiative may lack free will; for such behavior is compatible with the truth of determinism. Were it known that some people engage in such behavior in a suitably indeterministic way, many incompatibilists may be happy to say that some people have free will. But, again, the evidence that Baumeister et al. describe about psychological processes of free will does not include evidence of indeterminism. Possibly, their recognition of this point about their evidence together with their uncertainty or agnosticism about a conceptual matter—whether free will entails the falsity of determinism (as defined here)—accounts for the passage I quoted about whether the phenomena they focus on “fully deserve to be called free will.”
Baumeister et al.  is an instructive, rewarding read. I applaud their efforts to surround free will, and I look forward to the publication of the unpublished work to which they refer.
The condition just offered is an alleged sufficient condition for free action.
So if the occurrence of x (at time t1) indeterministically causes the occurrence of y (at t2), then a complete description of the condition of the universe at t1 together with a complete statement of the laws of nature does not entail that y occurs at t2. There was at most a high probability that the occurrence of x at t1 would cause the occurrence of y at t2.
As I understand deciding to do something, it is an action of forming an intention to do it; and, as I see it, many intentions are acquired without being actively formed (see Mele , ch. 9).
Whether lay folk tend to conceive of free will in a compatibilist or an incompatibilist way is an empirical question. But it is a different question.