Free will! Can I have one?

Forget free will and get on with empirical psychology.

Posted Feb 16, 2009

Despite this flight from free will, psychology has been plagued by a bad collective conscience. If free will is banished, something uniquely human seems to be lost. William James, Carl Rogers, and some of the some of the early social psychologists were among those who did not want to give up on freedom and dignity. The current champion of free will is Roy Baumeister of Florida State University. My interpretation of Baumeister's argument (e.g., Perspectives on Psychological Science, 2008) is that the automaticity school is mistaken in its attempt to overcome the two-system theory of mind. This theory assumes that behavior is generated by two different systems: one that is fast, reflexive, and automatic, and another that is slow, reflective, and controlled. If every imaginable behavior can be shown to arise from the automatic system, what is left for the controlled system to do?

I think that Baumeister has a good case for the claim that the automaticity school has overreached. Many psychological processes are slow, effortful, and infused with a subjective sense that we are doing something in our minds. But then his thesis gets murky. A defense of the two-system model of mind does not amount to a justification of free will. Let us consider three of his arguments.

First, Baumeister suggests that folk beliefs are evidence for the existence of free will. "If freedom and choice are completely illusions-if the outcome of every choice were inevitable all along-why must people agonize so over decisions?" (p. 14). Why should this argument be convincing? Often people do not agonize over their choices and yet perceive them to be free. Suppose you have a choice between fresh and spoiled food. You choose the fresh and fancy yourself to be free, although your choice was determined by the quality of the food and your preference for freshness that your ancestors have bequeathed on you. Conversely, experiencing agony over a decision does not imply free will. The agony is greatest in an avoidance-avoidance conflict. If you have to sacrifice one of your children, which one is it going to be? The choice given to Sophie in William Styron's novel is sadistic because she cannot set aside her belief in free will. She suffers because she cannot bring herself to think that whichever choice she makes is determined by forces predating her conscious experience of choosing.

It is generally a bad idea to justify scientific concepts on folk psychological grounds. Folk psychology can yield hypotheses, but it does not count as evidence. Anyone wanting to argue that collectively held beliefs are likely to be true should also respect the belief that the world was created by intelligent design, that souls are immortal, and that Sasquatch roams the woods of the Pacific Northwest. A more modest version of respect for folk beliefs is the idea that such beliefs can be considered true if their consequences are desirable. Kathleen Vohs and Jonathan Schooler (Psychological Science, 2008) showed that research participants who were induced (deterministically, by the way) to believe in free will were, when given the chance, less likely to cheat on an exam than participants induced to believe in determinism. Vohs and Schooler themselves, however, were careful to note that their finding says nothing about the existence of free will. The idea that the truth of a belief can be judged by its consequences is the hallmark of pragmatism. But even William James, the consummate pragmatist, justified free will only tongue in cheek. "My first act of free will," he declared, " is to believe in it."

Second, Baumeister notes that "the deterministic hypothesis-that every event is fully and inevitably caused by prior events and nothing else than what happened was ever possible-is itself unproven and even unprovable" (p. 15). This is true, but then, no scientific theory-let alone-meta-theory, is provable. Should we abandon hope and take Feyerabend's anarchistic "anything goes" attitude? Of course not. There are important asymmetries between the doctrines of determinism and free will that favor the former.

One asymmetry lies in the role of time. Experiments designed in a deterministic frame set up conditions (causes) to explore their outcomes (effects). Experimentation only moves forward in time, but determinism is bidirectional. While experiments tell us how one thing leads to another, they also suggest ways in which the first things arose to begin with. In contrast, the notion of free will is unidirectional. A free mind makes itself up and moves toward the future. The mind is free because there are no prior conditions that constrain how the mind makes itself up.

Another asymmetry is that determinism excludes the possibility of free will, whereas free will does not fully negate determinism. There is supposed to be a privileged domain in which the will is free. But how did this precious free zone open up? How did it emerge from an otherwise deterministic universe? And what are its boundaries? The self-congratulatory answer is that free will is uniquely human. Upon reflection, however, we are "determined" to realize that the boundaries are fuzzy. The behaviors of infants, senile or autistic humans show clear evidence of will, but that will does not appear to be free in the folk psychological sense. There is little reasoning, deliberation, or rationality.

I think that Baumeister's approach to the boundary problem lies in the role he accords perceptions of responsibility. This is his argument number three. The proposition is that if people have free will, then they are personally responsible for their actions. I do not argue with this proposition, but with its inverse. The fact that people hold humans (mostly others) to be responsible does not mean that there is free will; it does not even mean that they think there is free will. In fact, people hold others responsible even if they agree that the behaviors in question (e.g., heinous crimes) are determined by causes outside the person (Nichols & Knobe, 2007). If the allocation of rewards and punishments is an indication of perceived responsibility, people treat many animals as if they think these animals have free will. A similar argument can be made for power. Finding that many people pursue "the right to make decisions that may affect others" (Baumeister, p. 16) says nothing about the presumed freedom of those decisions. Many non-human animals are concerned about power, rank, and status, and they struggle to get it. Yet, they are widely regarded as automatons.

Baumeister does not seem to worry about such problems. His evidence for free will is that many healthy adults manage to self-regulate. The will is thought to be free if a person manages to overcome a short-term temptation for the sake of a greater, but later, value. Self-regulation raises a final asymmetry. Suppose you have a choice between slapping a misbehaving child and patiently discussing her behavior. Will you get free-will credit only for patient self-regulation? The answer appears to be no. If the will were free you could have chosen to yield to the impulse. If you yielded to the impulse, you could have achieved self-control had you tried harder. Hence, you are held responsible either way. But if you are rewarded for being patient and reprimanded for being impulsive, the mere availability of a self-regulatory option is no evidence for freedom of the will. It is only evidence for your a priori belief in it.

Baumeister's experiments on self-regulation and Bargh's experiments on automatic activation of behavior are fascinating. However, the former make it no more likely that free will exists, and the latter make it no more likely that it does not exist. Both research paradigms fit very well within a common deterministic framework. If pragmatism endorses the belief in free will because it generally increases happiness and kindness to others, realism suggests that this effect deterministically depends on specific conditions. For someone in Sophie's shoes, pragmatism is catastrophic. Empirically working scientists may want to leave pragmatism to laypeople and to those philosophers whose temperament inclines them to it.

Baumeister, R. F. (2008). Free will in scientific psychology. Perspectives on Psychological Science, 3, 14-19.

Nichols, S., & Knobe, J. (2007). Moral responsibility and determinism: The cognitive science of folk intuitions. Nous, 41, 663-665.

Vohs, K. D., & Schooler, J. W. (2008). The value of believing in free will. Psychological Science, 19, 49-54.