By Bryan Caplan
I quite enjoyed Alfred Mele’s Free: Why Science Hasn’t Disproven Free Will (available for pre-order now). It’s a great exercise in the debunking of debunking.
My favorite case: Many psychologists (and laymen) argue that consciousness is epiphenomenal. In layman’s terms, unconscious urges, not conscious intentions, are the real cause of your behavior. The evidence: Some experiments where researchers manage to predict some trivial behaviors with greater than chance probability slightly before you experience a conscious intention to act.
Mele’s strongest counter-evidence:
There’s an important body of research on implementation intentions – intentions to do a thing at a certain place and time or in a certain situation. I’ll give you some examples. In one experiment, the participants were women who wanted to do a breast self-examination during the next month. The women were divided into two groups. There was only one difference in what they were instructed to do. One group was asked to decide during the experiment on a place and time to do the examination the next month, and the other group wasn’t. The first group wrote down what they decided before the experiment ended and turned the note in. Obviously, they were conscious of what they were writing down. They had conscious implementation intentions.
The results were impressive. All of the women given the implementation intention instruction did complete a breast exam the next month, and all but one of them did it at basically the time and place they decided on in advance. But only 53 percent of the women from the other group performed a breast exam the following month.
In another experiment, participants were informed of the benefits of vigorous exercise. Again, there were two groups. One group was asked to decide during the experiment on a place and time for twenty minutes of exercise the next week, and the other group wasn’t given this instruction. The vast majority – 91 percent – of those in the implementation intention group exercised the following week, compared to only 39 percent of the other group.
In a third experiment, the participants were recovering drug addicts who would be looking for jobs soon. All of them were supposed to write resumes by the end of the day. One group was asked in the morning to decide on a place and time later that day to carry out that task. The other group was asked to decide on a place and time to eat lunch. None of the people in the second group wrote a resume by the end of the day, but 80 percent of the first group did.
Neat stuff. My main complaint: Mele never clearly says what I want him to say, namely:
Though their methods vary, all of these free will researchers are playing a corrupt game of “Heads I win, tails I break even.” They never tell us in advance what would count as evidence in favor of free will. The reason, to be blunt, is that no matter how their experiments turn out, they will never announce, “The results confirm free will.” Yet Bayes’ Law then automatically implies that no matter how their experiments turn out, the scientists should also never say, “The results disconfirm free will.” After all, if P(hypothesis | A)>P(hypothesis), then P(hypothesis | not-A)<P(hypothesis).
Upshot: The trendy “science” of free will is question-begging, probabilistically illiterate philosophy.
But I suppose that’s asking too much. Despite his failure to express my own position, Mele’s written a fine book on a subject I considered all played out. Well done.
P.S. Just noticed Mele has a whole book on effective intentions.