A long-running feature of Jerry Coyne’s popular website has been his discussion of “free will”. Jerry sensibly rejects all notions of mind/body dualism or any notions of a supernatural “soul” which can over-ride the laws of physics.
However, Jerry is at odds with many of his readers in rejecting any notion of “free will” that is compatible with a deterministic universe. Such “compatibilist” stances have been advocated, for example, by Dan Dennett in his 2003 book “Freedom Evolves”, and recently by Sean Carroll.
In essence that dispute is simply about semantics, with both sides agreeing on the physical reality. To illustrate this, consider a laptop computer which looks at the type of a computer file and “chooses” the most appropriate program to open it with.
One could justly declare that, in such uses, the word “chooses” is purely metaphorical, since the computer’s actions are entirely determined by its programming. However, for everyone except those arguing for a supernatural soul which over-rides physics, all other “choices”, by humans or other intelligent animals, must be equally metaphorical, since they are also determined by the prior state of the system.
Thus we have two possibilities, either drop the words “choice” and “decision” from the English language, along with a whole slew of similar words and phrases (“control”, “attempt”, “option”, “plan”, “threaten”, “test”, “compel”, “consider”, “coerce” for starters), or accept that “metaphorical” choices are all there is in this universe, and thus that the word “choice” is quite properly used about deterministic machines when they make a selection from a range of options.
Dennett’s argument is that humans have evolved a sufficiently complex range of responses to any situation that it is sensible to regard them as autonomous agents making “choices” of their own “will”. The word “will” here indicates that it is their own internal properties that are making the “choice”, even though that choice is entirely determined by those internal properties combined with the laws of physics. The word “freedom” in Dennett’s title “Freedom Evolves” indicates that mammals have evolved the flexibility to respond in ways that suit themselves and their own purpose, going beyond the very narrow responses of a simpler or more primitive entity.
Coyne’s objection is that this is a “re-definition” of the traditional and still-popular conception of “free will” as residing in a dualistic soul which over-rides physics. And this is a fair complaint, except that Jerry would then have to argue for the eradication of the words “choice”, “decision”, “will”, “freedom”, et cetera, from the language. Unfortunately, so steeped is the English language in traditional dualism, that we don’t have available words that would clearly make the distinction.
Thus many “compatibilists” prefer to retain the words, and use them as they would about a computer program “choosing” actions as it has been programmed to do. This can be regarded as a “weak” meaning of “free-will” (contrasting with the “strong” meaning of the classical dualistic sense), but it is one that is easy to get used to. It seems to me as natural to regard a chess-playing computer as “choosing” a move, as it is to regard a young duckling as “choosing” whether to stray from cover in water reeds to acquire a piece of bread — even though one knows that both behaviours are determined (I am leaving out of this any possible role for randomness such as quantum indeterminacy; it doesn’t change the free-will debate since randomness can hardly be regarded as “will”).
Jerry Coyne, in railing against such a “redefining free will so that we can still have it”, also objects that:
“There is a continuum in animals from simple ones who make binary decisions based on only one input (i.e. swim toward the light and away from the dark) to ones that make decisions based on more inputs (“Did I hurt my knee the last time I ran?”). At what point does the complexity of input constitute a form of “free will”?”
But then many biological properties are also continua, including intelligence, species transitions over time, the juvenile to adult transition, and most likely consciousness. Being a continuum would indeed be a problem for the strong, over-riding-physics meaning of “free will”, since that either does or does not happen, but is not a problem for the weak meaning, indeed it is a positive feature, aligning it with other biological properties that emerge over evolutionary time.
While the above argument is largely semantics — the dispute is simply whether “free will” in the weak sense can be properly called “free will”, or whether it should be called something else (how about simply “will”?) — the related issue of moral responsibility goes beyond semantics.
“If our choices are completely determined by our genes and environments, according to the laws of physics, are we morally responsible for our actions? […] It’s my contention that, in light of the physical determinism of behavior, there’s no substantive difference between someone who kills because they have a brain tumor that makes them aggressive (e.g., Charles Whitman), and someone who kills because a rival is invading their drug business. We need to reconceive our judicial system in light of what science tells us about how the mind works.”
The short answer is that the difference between those two is that one is susceptible to social opprobrium and the deterrence of threat of punishment while the other is not. Evolution is pragmatic. And our social senses have evolved as a means of enabling and facilitating social cooperation in a species which exploits one of the most cooperative ways of life of any animal.
We have evolved anger and tears because of their effect on other people (and also, perhaps, on ourselves); we have evolved notions of justice, of ostracisation, of punishment, of atonement, because of how they affect relations between humans. And that depends only on the (entirely determined) reactions humans have to how other humans interact with them.
Since “evolution” will not have known or cared about whether we had “free will” when programming us with our social instincts, those instincts are not dependent on what we think about “free will”. Psychology says that much of what we tell ourselves about our moral judgements is a rationalisation, a surface commentary, one that is often at odds with the deeper, biological substance. Thus, we might rationalise our judicial system in terms of classical “strong” free-will, but that’s just a commentary, one that can readily replaced. Weak free-will, in the above sense, is entirely sufficient as a replacement commentary.
We get angry with a child who misbehaves because our anger will modify their behaviour. We don’t get angry with misbehaviour that results from a brain tumour because it won’t make any difference and so is pointless. Evolution is pragmatic; our social instincts are pragmatic; and our judicial system is, de facto, pragmatic.