Ryle on the will (part 2)

Last time, I examined Ryle’s attack on the concept of volition, which according to Ryle is proposed by the “official doctrine” in order to explain the distinction between voluntary and involuntary actions. In this post, I’ll look at Ryle’s positive account of the voluntary/involuntary distinction, and his comments on the problem of free will. I’ll argue that Ryle’s arguments on both of these points leave much to be desired.

The voluntary/involuntary distinction

In the technical language of philosophers, the terms “voluntary” and “involuntary” apply to every act. Thus it’s considered a perfectly sensible question whether I read a book voluntarily, or whether I went for a walk voluntarily, or whether I solved the mathematical puzzle correctly voluntarily. Ryle thinks that this is a mistake. In ordinary language, Ryle argues, we describe actions as voluntary or involuntary when there is a question of blame, i.e. when somebody has done something wrong. Frank and Vincent are both accused of screaming in a library; we say that Frank’s scream was voluntary, because he was trying to startle Vincent; whereas Vincent’s scream was involuntary, because he couldn’t help it, having been startled by Frank. We wouldn’t ask, however, whether Frank and Vincent voluntarily sat silently in the library. Sitting silently is what you’re supposed to do in libraries, so the question of blame, hence the question of whether the act was voluntary and involuntary, simply doesn’t arise.

(One caveat that Ryle rightly notes is that we also use “involuntary” to refer to actions done under compulsion. We might say that Frank and Vincent sat silently in the library involuntarily, if they were forced to sit silently by somebody holding a gun to their heads. This kind of use is of course not really relevant to the present topic.)

The voluntary/involuntary distinction is connected to competence or ability. If a boy ties a granny-knot instead of a reef-knot, we can show that he tied a granny-knot voluntarily, i.e. that it was his fault that he tied a granny-knot, by establishing (a) that he knew how to tie a reef-knot, (b) that his hand was not forced in any way, and (c) that he was in good conditions (rather than e.g. freezing with numb fingers). To establish that the act was performed voluntarily doesn’t require showing that it was preceded by some special inner mental event; rather we simply demonstrate that the boy had a certain ability and that he was in the right conditions to exhibit that ability.

This isn’t to say that an act’s being voluntary has nothing to do with the mind. Ryle’s point is that qualities of mind are exhibited directly in the action. To say “Frank screamed voluntarily” is not to report two things, the screaming and the volition to scream; it reports one thing performed in a certain way. Compare “the woman walked seductively”: nobody would suppose that walking seductively involves two occurrences, the walking, and also some special inner “seductive state” that somehow confers seductiveness onto the walking. Walking seductively is doing one thing, namely walking, but in a specific manner or style. The same is true, according to Ryle, for doing things voluntarily.

Some support for a Rylean analysis, at least his suggestion that the voluntary/involuntary distinction is properly applied only when there’s a question of blame, might be found in a famous result from experimental philosophy, the Knobe effect. For those of you unfamiliar with this, Knobe (2003) presented people in a Manhattan park with a simple story. The first group of subjects were told the following story:

The CEO of a company is sitting in his office when his Vice President of R&D comes in and says, “We are thinking of starting a new programme. It will help us increase profits, but it will also harm the environment.” The CEO responds that he doesn’t care about the environment and says to go ahead with the programme.

The subjects were then asked: did the CEO intentionally harm the environment? 82% said yes. The second group are given exactly the same story, except the Vice President says, “It will help us increase profits, but it will also help the environment.” In this case, when asked whether the CEO intentionally helped the environment, only 23% said yes.

A straightforward explanation of this asymmetry is that the moral character of an action influences whether or not it’s judged to be intentional. We’re more likely to judge an action to be intentional when we judge it to be morally wrong. This meshes rather nicely with Ryle’s claim that the question of whether an action is voluntary is properly introduced only when somebody has done something wrong. There’s nothing wrong with helping the environment, so it seems odd to ask “did the CEO intentionally [i.e. voluntarily] help the environment?”

Actually, as Holton (2010) points out, things aren’t quite do simple. In some scenarios, the Knobe effect occurs “the other way around”: the good action is judged to be intentional, and the bad action is not judged to be intentional. Holton cites the example: “a profit-driven executive whose actions have the side-effect of violating a pernicious Nazi law – surely a good outcome in most subjects’ eyes – is judged to have acted intentionally; whereas a similarly driven executive whose actions have the bad effect of fulfilling the requirements of that law is not judged to have acted intentionally.” On Holton’s analysis, we’re more likely to believe an action to be intentional when we judge that it violates a norm, and judging that a norm has been violated does not depend on judging that the norm is good; Nazi laws are bad norms. This is, I think, still entirely compatible with Ryle’s view; but the important point is that we have to be careful with appeals to the Knobe effect because it remains controversial exactly what the Knobe effect shows.

Let’s turn to some problems for Ryle. First, regardless of findings like the Knobe effect, I’m not convinced that ordinary usage restricts the voluntary/involuntary distinction to those actions that are considered blameworthy. It strikes me as being quite natural to say e.g. that I voluntarily ate the sandwich, that I voluntarily went for a walk, that I voluntarily typed up an essay. Or consider the following sentence: “If you get a friend to flick their finger near your eyes, you will blink involuntarily.” That sentence is, surely, perfectly reasonable. Yet there is nothing blameworthy about blinking. Perhaps Ryle would say that such sentences seem natural to us because ordinary language has become infected with technical theory; just as we’re happy to use technical terms like “H2O” and “sound waves”, similarly the philosopher’s extended use of “voluntarily” and “involuntary” has crept into everyday use. This of course is one of the perils of tying one’s analyses to ordinary language. Ordinary language changes.

A second difficulty with Ryle’s account is that it just doesn’t get the phenomenology right. In at least some cases of voluntary actions, it really seems as though there are two things going on, a mental occurrence followed by the behaviour. Frank is in a library and chooses to disturb people by screaming. Vincent is in a library, but is startled, screaming accidentally. On Ryle’s analysis, both Frank and Vincent do only one thing, namely scream, but they each do it in different ways. We don’t need to suppose that some special event occurs in Frank’s mind that doesn’t occur in Vincent’s mind. However, if we imagine ourselves in Frank’s place, it certainly seems that there’s something going on in his mind that causes the scream. Note that we can’t account for this by pointing out that Frank desires to scream in the library, whereas Vincent doesn’t. Frank’s scream is produced not simply by his desire to scream, but also, it seems, by his actually executing a decision to scream now. Perhaps a better to put the problem would be to ask: why does Frank scream when he does, and not five seconds earlier or five seconds later? The volition theorist has an easy answer: he screams when he does because that’s when he executes a volition to scream. What can Ryle say? (Bear in mind, of course, that the same kinds of difficulties don’t arise in Vincent’s case. Vincent screamed when he did because that’s when he was startled.)

One option for Ryle would be to grant that in this particular case, there is something going on in Frank’s mind that’s not going on in Vincent’s. It’s not a volition, though; rather, the difference between Frank and Vincent lies in the fact that Frank is deliberating. It’s not easy to scream in a library. It seriously violates social conventions concerning libraries, it’s embarrassing, it could get you into trouble, etc. Frank has to think through all of this stuff; he has to, in a sense, persuade himself to go ahead and scream. (Wouldn’t positing such internal mental activity violate Ryle’s general behaviourist commitments? No, I don’t think so. While Ryle was a behaviourist of sorts, he was happy to allow that private mental activity does occur. He talks about people engaging in “internal monologues” and “silent soliloquy”: see e.g. pages 16, 28,70, and 160.) However, such mental activity is not necessary for voluntariness: sometimes we perform voluntary actions without deliberation, as when I see some ice cream, I want it, and I go and get it. No detailed thoughts need precede this. Indeed, probably most of our everyday actions are performed without deliberation. So we can’t explain the voluntary/involuntary distinction by appealing to inner mental activity.

This response is fair enough, but one still feels that it doesn’t quite get to the heart of the objection. I think that the deeper problem with Ryle’s analysis is that it leaves us hanging re the “springs of action.” For many of our actions, even those for which we don’t engage in any deliberation, it certainly feels as though there’s a point where we say to ourselves “right – move now.” What can this be, if not a volition? Appealing to competence or ability is surely of little help here. Abilities need not be occurrent. The mere fact that a boy is competent to tie a reef-knot tells us nothing about when he’ll actually go ahead and tie a reef-knot.

Presumably Ryle would say that this is simply beside the point as far as he’s concerned. Remember, he denies that we need talk of voluntary/involuntary distinction except in cases where there is a question of blame. So his analysis simply isn’t concerned with the general “springs of action.” The trouble with this response is that the volition theory is concerned with this; volitions are proposed to account for e.g. the difference between not just Frank’s scream and Vincent’s scream (where they’ve both done something wrong: noise in a library), but also Frank’s raising his arm because he intended to and Vincent’s raising his arm because of a muscular twitch (here, there’s no question of blame). The concept of volition gives us a very intuitive, straightforward explanation of the difference. If Ryle wants to do away with volitions, he owes us an alternative explanation. And while I tend to share Ryle’s skepticism of volitions, it’s not clear to me how to develop Ryle’s own analysis to provide a satisfactory explanation on this point.

Free Will

In the final part of the chapter, Ryle deals with the problem of free will. The sciences have shown that everything in the external world is rigidly governed by physical laws. Everything follows a deterministic and purposeless path. But then what becomes of our appraisal concepts? How can we describe people’s behaviour as good or bad, how can we say that charitable man is worthy of praise while the greedy man deserving of blame, if everybody’s behavoiur is determined by rigid laws? The notion of an inner mental world, containing special acts of volitions, allowed a way out of this problem, since physical laws need not hold sway in the mental realm.

Ryle’s alternative solution is that while physical laws may govern everything that happens, they don’t ordain everything that happens. He draws an analogy to chess. Observers who are unacquainted with chess, but who are allowed to observe and study it, may arrive at the conclusion that the players are governed by rigid laws. Every time a players picks up a pawn, the observers say, we can predict what move he will make. Some moves might seem strange or surprising, others might seem to evince ingenuity, but all moves are governed by clear rules, and we can expect that further study will reveal still more rules that explain the unexpected moves. Hence there isn’t really any intelligence or purpose in the game; every move that is made is the product of prior states of the game plus the rigid laws.

What has gone wrong with this argument? What these observers have missed, says Ryle, is that all moves governed by the chess rules, but no move is ordained by them. The claim that every move is explicable wholly in terms of the chess rules is simply a mistake, or at least it conflates two different kinds of explanation. Explaining a move requires appealing not just to the chess rules, but also to tactical principles and to the intelligence of the player. The sense in which a rule explains a move is not the same as the sense a tactical principle explains a move. One single process, such as the movement of bishop, can be in accordance with two principles of different logical types: the rule and the tactical principle. Neither principle is reducible to the other. The chess rules and tactical principles are fundamentally different things. So the fact that chess follows rigid rules doesn’t in any respect diminish the ability to demonstrate intelligence and ingenuity. Indeed, good chess playing requires the chess rules; if there were no rules, it would be impossible to evince any skill. Similarly, intelligence, ingenuity, skill etc in other areas of life presupposes physical laws: we couldn’t display intelligence if everything behaved randomly.

This whole argument simply misses the point. The problem of free will arises from the fact that “ought” implies “can”: we can’t, we assume, hold someone responsible for something that they couldn’t have avoided. Now the worry that determinism poses is that, if my actions are determined, then there’s an important sense in which I couldn’t have avoided doing them. If decision to eat a sandwich is a result of the state of the universe at some earlier time plus the physical laws, then my decision to eat a sandwich isn’t really the product of my free choice.

Consider again Ryle’s chess analogy. The chess rules stipulate that I can only move the bishop diagonally. But it’s entirely open how far the bishop moves, or whether it moves forwards or backwards. Given only the chess rules, we can’t even in principle predict exactly how the pieces will move. The problem of free will, however, is that given certain initial conditions plus the laws of nature, we can predict, not in practice but in principle, exactly how a person will act. So a better analogy to illustrate the problem of free will is to compare humans not to the chess players but to the chess pieces. The mere fact that there might be different kinds of explanation doesn’t help here. Indeed, given determinism, it looks like the “tactical” explanation just leaves things out, or at least gets the reasons for the actions completely wrong.

For what it’s worth, I’m a compatibilist, so like Ryle, I don’t see the discovery of physical laws as posing any threat to free will or to our use of appraisal concepts. But Ryle’s argument for this is obviously inadequate.

Holton, R. (2010) “Norms and the Knobe Effect”, Analysis, vol. 70, no. 3, July, pp. 417-424.

Knobe, J. (2003) “Intentional Action and Side Effects in Ordinary Language”, Analysis, vol. 63, no. 3, July, pp. 190-194.

Ryle, G. (2000) The Concept of Mind, London: Penguin Classics.