Akrasia and Compatibilism

Akrasia, or weakness of will, is a complicated phenomenon. It may be best to describe it with an example than to offer a philosophical definition of it, which is sure to be controversial.

(WW) Merrie is convinced that is it, all things considered, better for her to study than to go to the party. Merrie freely goes to the party.

Some people have argued that, in light of the actuality of akrasia, compatibilism is false. “Compatibilists can’t handle cases like (WW),” it is said, “therefore, compatibilism is false.” I’m not sure these typical arguments succeed. The reason is simply due to a technical matter. Let me make that point first, then I’ll table it and address a contemporary expression of the argument against compatibilism from akrasia.

Here’s the problem with the above form of reasoning:

Akrasia is a contingent fact, if a fact at all. Yet, the conclusion that compatibilism is false, is a necessary truth. To say that compatibilism is false is to say that no possible world, w, is such that, in w, someone—Hugh—is (non-derivitively) determined to do an action—his logic homework—and yet Hugh freely does his logic homework. But the argument from akrasia notes that it is a fact about our world—the one that is fact the actual world—that agents sometimes do weak-willed actions (here I skip over the distinction between weakness of will construed as a character flaw and weak-willed actions, most of the relevant literature is concerned with the latter). But granting this, it does not follow that the homework-world mentioned above, w, is such that w contains agents who do weak-willed actions. To get the strong modal conclusion regarding the falsity of compatibilism, we would instead need a premise that akratic action is possible in all worlds containing free agents. So far as I can tell, there’s nothing about the nature of free agency—libertarian or otherwise—that entails an akratic action must be possible.

It may be replied that the argument from akrasia gets the anti-compatibilist conclusion through another route, namely:

1. In the actual world, @, there exists a free akratic action A. (premise)
2. If an A were determined, A wouldn’t be free. (premise)
3. Therefore, the akratic action A in @ is undetermined. (1 and 2)
4. If a free act is undetermined, it is a libertarian free act. (premise)
5. Therefore, libertarian free will is true in @. (3, 4)
6. If libertarian free will is true, incompatibilism is true. (def. of LFW).
7. Incompatibilism is true. (5, 6)
8. Incompatibilism, if true, is necessarily true. (premise)
9. Incompatibilism is necessarily true. (7, 8)
10. If incompatibilism is necessarily true, compatibilism is necessarily false. (analytic)
11. Compatibilism is necessarily false. (9, 10)

However, the problem with this argument is that it’s unsound. Premise (5) is false. A free act’s being undetermined is only a necessary condition and not a sufficient condition for that act’s being libertarian free. So this argument won’t work. (5) would work if one had an argument that any free action is incompatible with that action being (non-derivatively) determined. But if one already had that argument, the argument from akrasia becomes otiose.

At this stage I will put this flaw in extant contemporary arguments from akrasia to incompatibilism to the side. I want to focus more narrowly on an instance of this argument as found in J.P. Moreland’s and William Lane Craig’s book (M&C), Philosophical Foundations for a Christian Worldview (IVP, 2003: 277–278). In the course of their scattergun approach to arguing against compatibilism, M&C write:

According to libertarians, weakness of will is a real phenomenon to be understood as an occasion where a person fails to act in keeping with his own personal preferences (his values, desires, beliefs, etc.) or acts against them.

For the compatibilist, people never act against their actual preferences, so if akrasia is defined in this way there is no akrasia. Actions are always caused by the strongest preferential state in the agent. People do have conflicting preferences…and sometimes one set of preferences is stronger and wins out over another set. If by akrasia we mean that one set of preferences we desire to act on to some degree is sometimes overridden by a stronger, sometimes immoral set of preferences, then there is such a thing as akrasia. In sum, libertarians and compatibilists differ over whether or not there is such a thing as akrasia and/or over how we are to define it.

In context, M&C intend this to be an argument against compatibilism. While it’s not clearly stated, we can construct their argument as so:

13. Akratic actions exist. (premise)
14. If compatibilism is true, then no agent A fails to act against A’s actual preferences. (def. of compatibilism)
15. If akratic actions exist, then some agent A fails to act according to A’s actual preferences. (def. of akratic action)
16. Some agent A fails to act according to A’s actual preferences. (13, 15).
17. Compatibilism is false. (14, 16)

This argument fails right off the bat for those reasons stated above, though we will not dwell on that. But we should not despair of finding problems, this argument has plenty. Let’s briefly rehearse some of them

i) According to M&C, weakness of will is to “be understood as an occasion where a person fails to act in keeping with his own personal preferences (his values, desires, beliefs, etc.) or acts against them.”  This definition is hopelessly flawed. First, to be clear they mean a weak-willed action (or omission) is to be understood thus and so, not merely “weakness of will,” which is ambiguous between the former and a character trait. Now, they define a weak-willed action either a failing to act on or acting against the person’s own preferences.  Presumably, “preferences” are “values, desires, beliefs, etc.,” Apparently, a “preference” is simply a summary term for either something a person values or believes or desires. Thus, on this definition, I act akratically if I fail to do something I desire, or act against a desire I have. So let’s give a case that meets their criteria:

BEACH: In BEACH, I desire to go to the beach and lay out in the sun while reading a good book. However, I just it best, all things considered, to go to work and complete my project.

In BEACH, I have a desire to go to the beach, and I act against it by going to work, thus I act akratically. But this is simply false. M&’C’s understanding of acting akratically, which is “to be understood as an occasion where a person fails to act in keeping with his own personal preferences (his values, desires, beliefs, etc.) or acts against them,” is to not understand akrasia. On their view, virtually every action would be an akratic one (!), because I often act against or fail to do something else I desire, value, etc. This is why akrasia has typically been defined as acting against an all-things-considered judgment that it is better to A than B (leaving out cases, for simplicity, where the former is true but I do B since I also judge it better to do B and C than A).

So premise (15) is false.

Now take premise (14). This isn’t true either, for similar reasons as those I cited above. But perhaps M&C mean that if compatibilism is true, I never act against my strongest actual preference at a time. Again, ‘preference’ is vague and obfuscates things. For M&C it’s just a summary term for values, desires, beliefs. But why think compatibilism as such entails this? Suppose God sets up the world such that, at time t, my strongest desire (preference) determines my action, but at time t*, my second strongest desire does. If I am free in both of these cases, then since they’re both cases of determinism, compatibilism would be true, yet it would be false that, at both times, my strongest preference determined what I’d do.

Now, there are other arguments M&C could give here. Suppose they define akrasia along more sophisticated lines. And suppose they say that, on compatibilism, a free agent always does what he judges best at a time. Why think this? Suppose that if I didn’t do what I judged best at a time, I would be irrational. Then I couldn’t be free, since I’d fail to meet a rationality constraint. But notice here that this is a problem for both libertarian and compatibilist views. On libertarianism, I wouldn’t always do what I judge best. But if that’s irrational, then even though I can do otherwise than what I judge best, I don’t do so freely, since it’s irrational. Of course, M&C could “Chisholm” this some more, eventually getting to a sophisticated challenge. It’s not my aim to address such challenges, but just note that they have been addressed in compatibilist friendly terms (see Alfred Mele Irrationality: An Essay on Akrasia, Self-Deception, and Self-Control (1992) and Backsliding: Understanding Weakness of Will (2013).