I still haven't had a chance to look at Bruce Waller's book, Against Moral Responsibility (2011), but I've been reading about it and related topics in my spare time a bit over the past several days. One reader, David Duffy, was kind enough to bring one of Waller's papers to my attention. It's called "Empirical Free Will and the Ethics of Moral Responsibility"(2003). In it, Waller claims that moral responsibility and free will are either conceptually wedded by definition (in which case, he says, we only get confusion) or there is some synthetic (empirical) connection between them. He then argues that there is no such empirical connection.
I question the claim that there is any confusion resulting from regarding a logical (analytic) entailment between moral responsibility and free will. Unfortunately, Waller does not support his assertion here, though perhaps he addresses the issue in his more recent book. I think the only conceptual confusion comes from trying to treat free will and responsibility as objects of strictly empirical concern. I will discuss this below and its relation to the Kantian notion of transcendental freedom, but first I want to point out another problem I have with Waller's paper. It concerns his treatment of Dennett.
Waller accuses Dennett of including "mysterious elements" in his attempt to "accomodate" moral responsibility. He says Dennett "ultimately appeals to choices that are difficult to square with a naturalistic framework," and he quotes Dennett as follows:
I am faced with an important decision to make, and after a certain amount of deliberation, I say to myself: "That's enough. I've considered this matter enough and now I'm going to act," in the full knowledge that I could have considered further, in the full knowledge that the eventualities may prove that I decided in error, but with the acceptance of responsibility in any case.
Waller comments: "If we really had full knowledge that in such circumstances, with the precise natural histories that shaped us, we actually could have considered further, then that would be knowledge of a choice inexplicable in the naturalistic framework."
I think Waller's point is this: If we really could have considered further, then our natural histories did not determine the length of our consideration. In that case, naturalism must be false. It looks like Waller might be conflating determinism and naturalism here, a charge which has been made against him in a related context. Still, there seems to be a problem with Waller's reading of Dennett.
I think Waller and Dennett would agree on this: What an account of free will does not need, and should not even want, is knowledge that we could have acted differently without any changes in our natural histories. According to Dennett, our actions cannot have been different without the physical causes of those actions also being different. Dennett's point is that any other variety of free will wouldn't be worth wanting. We should not want free will, if it means that we can have the exact same causal factors--the same beliefs, desires and experiences--and still choose differently. We should want a free will that allows us to choose based on who we are, and that includes all of the forces which have made us who we are. If those forces are irrelevant, then so are our beliefs and desires. In that case, free will would amount to a random generation of outcomes, none more likely to benefit us than any other.
Something odd is going on here. Did Dennett really say that people can have "full knowledge" that they could have acted differently? How could Dennett be so inconsistent on such a crucial point?
I don't think Dennett means what Waller takes him to mean. Perhaps Dennett was just being sloppy, though it is hard to say without seeing the context from which this passage has been taken. (I don't have a copy of Dennett's Brainstorms handy, but the passage is on page 297, if you want to check.) Contrary to Waller, I do not think Dennett "entangles his notion of free will with such remarkable choices." I think Dennett is usually quite clear on this: The idea that we could have acted otherwise does not entail that we could have done so against the causal nexus of our past. According to Dennett, the idea that we could have acted differently means that our decision was in some robust sense up to us. We are not passive in the decision-making process. We have agency.
This does raise another important question against Dennett, though: How can his compatibilistic conception of free will allow for real moral responsibility, and not merely the illusion of it? After all, isn't it an illusion that I could have acted otherwise? Isn't it an illusion that I am acting of my own, internal rational agency, without compulsion from external forces? Doesn't this illusion sustain our conception of free will?
Waller is in a bind here, too, I think: He wants to maintain the conception of free will and some minimal notions of responsibility and agency, but he wants to abandon the idea of moral responsibility that comes with robust rational agency. It's not clear to me how he can justify any notion of responsibility and agency without moral responsibility.
Let's go back to Dennett for a moment. He wants to keep the robust sense of agency without accepting some kind of transcendental freedom--that is, freedom to act independently of physical causes. I ended my last post on this topic by suggesting that Dennett is (or at least should be) closer to Kant than he might realize. Kant famously argued for a compatibilist position, allowing for both determinism in the realm of empirical causes and free will (and moral responsibility) in the realm of human action. I'm not a Kant scholar, but I've been consulting a few secondary sources of late. (I did read the Critique of Pure Reason when I was in grad school many moons ago, but I don't remember the relevant arguments so clearly.) I am still getting a better picture of Kant's thinking, but I don't want to get into any of it yet. However, I will point out one interesting paper I found which views Kant as a precursor to Davidson. It's by Todd D. Janke, and it's called "A Freewheeling Defense of Kant's Resolution of the Third Antinomy" (2008).
Janke's basic idea is that Kant can be read in a very appealing, if unconventional, way. I'm not sure Janke would agree with my interpretation, but here's how I currently see the situation: We have a certain way of making sense of human actions, including the notions of intention, desire, belief and will. This vocabulary does not reduce or unproblematically translate to a vocabulary of empirical observations. Though Janke doesn't mention it, Ryle made a similar observation in The Concept of Mind (1949): When we say that a person has succeeded in an action, we are not postulating an additional event on top of the events we could find in a purely mechanistic description. As Ryle (1949, p. 152) puts it, in a passage connecting the supposed mysteries of intentionality with those of perception itself:
Epistemologists have sometimes confessed to finding the supposed cognitive activities of seeing, hearing and inferring oddly elusive. If I descry a hawk, I find the hawk but I do not find my seeing of the hawk. My seeing of the hawk seems to be a queerly transparent sort of process, transparent in that while a hawk is detected, nothing else is detected answering to the verb in 'see a hawk.' But the mystery dissolves when we realise that 'see', 'descry' and 'find' are not process words, experience words or activity words. They do not stand for perplexingly undetectable actions or reactions, any more than 'win' stands for a perplexingly undetectable bit of running, or 'unlock' for an unreported bit of key-turning. The reason why I cannot catch myself seeing or deducing is that these verbs are of the wrong type to complete the phrase 'catch myself. . . .' The questions 'What are you doing ?' and 'What was he undergoing?' cannot be answered by 'seeing', 'concluding', or 'checkmating'.Ryle says a conceptual problem arises from confusing action verbs with success verbs. A success verb, like "see" or "conclude," describes an intentional action but it does not denote a particular event in a sequence of physical causes and effects. It therefore does not cash out in empirical terms. Yet it does not thereby require a break in the laws of nature. It does, however, seem to suggest a way of talking which does not neatly cash out in terms of causes and effects. Ryle's argument is that the concept of mind itself, including the language of intentionality and intelligence, is not a language of causes and effects. Yet, when people act, they are not thereby acting against the laws of cause and effect. All human movements can be explained in physical, cause/effect terms, but we are not speaking directly in those terms when we talk about human action.
This seems to be how Janke wants us to read Kant. The intelligibility of human action, including rational agency, does not make sense in empirical terms. Kant wants us to have it both ways: We can have talk of agency in non-empirical terms, which Kant calls "transcendental freedom", and we can have talk of determinism in the empirical world. Whichever we prefer, Kant says, is a practical matter. It cannot be decided by pure reason.
Dennett's view is similar. We can regard some events as intentional, as the products of beliefs, desires and rationality, or we can regard them in purely physical terms. This is a matter of practical, not theoretical, necessity.
So we can have practical reasons for believing in transcendental freedom. These are the very same reasons, the very same arguments, that justify our belief in intentional behavior, rationality and human agency in general. They do not require abandoning determinism. Confusion arises when we try to understand one vocabulary in terms of another. As Ryle put it, mental events do not happen alongside or in sequence with physical events. They are not events at all, in the sense in which physical occurrences are events.