My friend Russell Blackford continues the back-and-forth we’ve had on the issue of free will in a new piece at Talking Philosophy: “Jerry Coyne writes back—about free will.” Because it’s Monday and I’m lazy and haven’t yet had coffee, let me just post his opening paragraph that gives all the relevant links:
Over at Why Evolution Is True, Jerry Coyne recently wrote a post responding to my earlier post on his piece in The Chronicle of Higher Education. This debate can go back and forth a lot, but let me clarify a few things at least.
In private emails to me, Russell argues that I am “confused” on the issue, and perhaps I am. But I have to say that I don’t find Russell’s piece bringing a lot of clarity to the issue, either.
For one thing, he seems to contradict himself. On the issue of determinism, for instance, he says this:
. . . even if determinism is not true, there are well-known arguments as to why a mere mix of occasional indeterminism with determinism is unlikely to give us free will if we otherwise lack it (Jerry alludes to this in the Chronicle, and I agree with his brief comment on it). Moreover, even if determinism is not strictly true, it is difficult to see how I could be responsible for my own character, desires, etc., all the way down. Coming up with a picture of how this could work that is both coherent and plausible seems very difficult.
But he later argues that one can rightly blame someone for failing to save a drowning child. Note the word “rightly,” which assumes not just responsibility (which is okay with me, as blame changes future behavior, both of the “blamee” and onlookers), but moral responsibility. Russell certainly favors the idea of moral responsibility. But if he sees difficulty in understanding how one can be responsible for one’s own character (and he’s right: how could we be?), then whence the concept of moral responsibility?
This highlights to me a problem with compatibilism: not a philosophical failure, but a tactical one. If most “compatibilists”—that is, those who, like Russell, find free will compatible with determinism—really are determinists, why do they soft-pedal this important result? After all, shouldn’t they be telling people that their actions are determined by the laws of physics, not their own “spooky will”? Philosophers tend to de-emphaszie this important result—yes, a finding of science—in favor of showing us how we can haz free will after all. I’m not sure why the soft-pedalling of determinism, but perhaps some philosophers resent the incursion of science onto their turf. (I know that is not true of Russell, though.)
But never mind. Another problem is that Russell sort of agrees that he is a determinist, but doesn’t exactly trumpet that to the world:
Another point that should be made to try to get all this a bit clearer is that I am not especially reluctant to concede that causal determinism is true to whatever extent is required for arguments based on it to go through (assuming the arguments have no other problems). So Jerry misreads me when he thinks that I accept determinism “only grudgingly”. On the contrary, it would make the whole debate simpler for me if we knew that determinism is true. I’m not temperamentally opposed to determinism. Furthermore, I think that it’s probably true enough for our purposes.
That’s a lot of words to use for saying, “Yes, I accept determinism”!
But Russell, if you are a determinist, then why isn’t the whole debate simpler for you? What else could there be but determinism of character and actions—physical determinism by genes and environment, with perhaps a soupçon of quantum indeterminacy, which Russell agrees does not count toward “free will”? The fact that the debate has not become simpler for Russell is seen in his repeated statement that the “free will” case is complicated and fuzzy.
And if Russell is a determinist, why does he begin his last paragraph with this:
Perhaps we don’t have free will. Although there are no spooky forces controlling us, someone might argue that, for example, we all have deeply disordant sub-conscious urges which play much the same role.
No, those discordant subconscious urges cannot play the same role as a “soul,” because, unlike a soul, those urges have a material origin, and are hence part of the field of determinism that regulates our actions. They might be “spooky” because we don’t grasp them consciously, but subsconscious urges are not part of dualism.
But these are ancillary issues. Russell’s real beef with my argument against free will turns on the use of the words “can” or “could.” My definition of free will was this: if one replays the tape of one’s life back to the moment of a decision, with all the molecules and forces that acted previously still in play in an identical way, then “free will” is the notion that, given this situation, one could have made a different decision. Now most compatibilists, at least on this site, agree that it would indeed be impossible to have made a different decision. Russell is explicit on this point: if everything is the same on the replay, then only one decision was possible. But he takes issue with the word “could.” My definition of it in this context was “it was actually possible, not just logically possible” to make a different decision. Russell’s definition of “could have decided otherwise” is that “to an outside observer, you apparently had all the mental and physical equipment necessary for making an alternative decision.”
Re the drowning child, for instance, Russell says this:
I did, however, use the scenario of the drowning child to demonstrate how we ordinarily use such words as “can” (“can’t”, “could”, etc.). Let’s return to that.
Perhaps Jerry wants to use the word “can” in a special sense, but if so the word becomes equivocal in its meaning. Normally, when we say, “I can save the child” or “I could have saved the child” we mean something slightly (but not very) vague to the effect that I have whatever cognitive and physical capacities are needed, have whatever equipment is required, am on the spot, and so on. Perhaps it includes not being in the grip of a disabling phobia and not being coerced by someone with a gun. “Can” refers to a commonsensical notion – slightly vague, but no more so than most ordinary language – of having the ability to do something.
If all this applies, but I fail to save the child (perhaps because I dislike children or because I don’t want to get wet, or because I am just too lazy), it makes still makes sense to say that (speaking tenselessly) I can save the child but I don’t do so because I don’t want to. Here, the ordinary meaning of “can” is being applied correctly to the situation. If Jerry’s argument demands throwing out this ordinary usage, it’s in all sorts of trouble. If he wants to use “can” and “could” in some other sense, apart from the ordinary one, in the context of free will talk, I see no reason to believe that his conception of free will is much like what the folk have in mind when they say, for example, “Russell acted of his own free will.” The empirical research done to date, e.g. by Eddy Nahmias and his colleagues, does not suggest that the folk, or the majority of them, have some special meaning of “can”, “could”, and “ability to act” in their minds.
In other words, you have free will because you “could have saved the child,” although in reality there was no possibility that you really would save the child. You could have because you apparently possessed the cognitive and physical capacities to do so, even if it was impossible for you to exercise them.
I don’t see how this solves the problem. For one thing, where is the “freedom” here? There isn’t any: you still have only one course of action. The “freedom” is purely theoretical! More important, what does Russell really mean by “cognitive and physical capacities to save a child”? He rules out some aspects of cognition as part of the “could” here:
Perhaps it includes not being in the grip of a disabling phobia and not being coerced by someone with a gun.
Well, perhaps Charles Whitman didn’t have free will when he shot all those students in Texas because it was found, post mortem, that he had a brain tumor that could have produced a murderous aggression. So Whitman “couldn’t have” refrained from killing people.
But what is the difference between such a “disabling phobia” and the so-called “normal” neurological conditions that produce our behavior? Does someone who was repeatedly beaten and mistreated as a child have “free will” when he beats up someone else when he’s older? There are all sorts of environmental influences that can act deterministically (now or later in life) to produce a behavior—a behavior that was inevitable. What is the difference between these and a “disabling phobia”? Which phobias are “disabling” and which not—in other words, which neurological conditions show that one didn’t have free will and which truly instantiate free will? I see no relevant difference between Whitman’s brain tumor and a childhood trauma. They are both forms of “coercion” that are equally potent (but not nearly as obvious) as a gun pointed to one’s head. At what point does a “could not” become a “did not want to”? Does it lie between a tumor and a personality disorder?
(Let me hasten to add that I feel it is proper to remonstrate someone for failing to save a drowning child, for such remonstrations act to change behavior, and can lead to a better society. I just don’t think that such a person was “morally responsible” for that failure to act.)
Russell thinks that his conception of free will is better because it involves a use of the word “could” that is in more in line with our everyday parlance. When we say “I could have had a V8,” for instance, Russell thinks we mean, “There was V8 juice available and I didn’t have any mental disability that prevented me from ordering one, nor an inborn aversion to tomato juice.” Well, if all he means by “free will” is that to an outsider one seems equipped to make a choice (but which outsider? can the outsider see our neurons and how they operate?), perhaps I’ll go along with that. But not necessarily.
For there are other uses of the word “can and “could” that imply not just physical equipment, but the concept of want. Here are two. What do you you think when you hear these phrases?
“Hey, buddy, my car’s broken down—could you give me a lift?”
“I’m hungry. Can you give me a dollar for some food?”
According to Russell, in both cases you can tell the supplicant that yes, you could give him a lift or a dollar, but then walk away. After all, you were equipped to give a ride or a dollar (you “could have”), but according to Russell you just didn’t want to. But try telling that to those people. “Yes, buddy, I can give you a dollar. But I won’t.” I doubt that either supplicant would comprehend these meanings. In these cases the common understanding of “could” is precisely what I mean in my definition of free will.
The point is that there is often no clear distinction between “could have done” and “want to do”; indeed, under determinism I see no sensible difference between these concepts. But even in common parlance these two concepts elide. I think the onus is on Russell here to show how one can be “physically and neurologically equipped” to perform an act in all ways, but then not perform that act. If “free will” simply means that there is no reason obvious to an observer that one couldn’t perform an act, even though the character and constitution of the actor make it impossible for him to perform that action, then I see us mired in some sort of Free Will Wonderland. After all, an outside observer (unless she’s a neuropsychologist two centuries from now) will never have the ability to examine the molecular structure of our bodies and the neuronal connections that will enable her to see whether we’re mentally equipped to do something.