Tuesday, September 27, 2011

List on Free Will and Determinism (Part One)



I recently had the pleasure of reading Christian List’s new paper on free will and determinism (available here). List seems like an interesting guy. I haven’t read much of his previous work on group agency and epistemic democracy, but what little I have read I’ve always enjoyed. He adopts a formal and technical style which I’m finding more and more congenial and to which I now aspire in my own work.

Still, whatever the merits of his previous work, I wasn’t sure that he’d have anything of interest to contribute on the issue of free will. I say this not because I doubt his capacities to contribute to this field, but because I doubt everyone’s capacity to contribute to it. Having read a fair amount of the literature, it often seems like this issue has been so well-worn, and the analytical paths so well-trodden, that there’s no territory left to explore.

And yet, List’s paper does manage to contribute something. With his flair for the formal and technical style of analysis, List manages to render more perspicuous and cogen, an idea that has lurked in the background of the philosophical debate for some time now. What is this idea? Well, in essence, it is a new theory of what it means to say that an agent “could have done otherwise”. I want to share it with you over the next few posts and critically engage with some of its elements as I go along.

In this post, I begin by discussing the traditional conceptions of “could have done otherwise”, and continue by looking at List’s objections to these conceptions. If you’re a complete newcomer to the free will and determinism debate, I suggest reading this post by way of introduction; if you’re not a complete newcomer, then read on.


1. Compatibilism and Alternative Possibilities
The ability to do otherwise seems central to our conception of ourselves as free and responsible agents. Take a simple, everyday decision as an example: When I choose to make myself a cup of tea — as I often do — it certainly feels (from the inside of my decision-making) that I could have chosen to make an alternative beverage. That is to say, it feels, from the inside, like alternative possibilities are open to me.

The problem is that this aspect of my internal self-conception seems incompatible with another popular metaphysical thesis: determinism. Determinism is the view that physical events are causally connected in such a way that, at any particular moment in time, there is one and only one possible future. Since the ability to do otherwise seems to require more than one possible future, the incompatibility should be obvious.

List, like many before him, wants to challenge this apparent incompatibility. He wants to argue, contrary to the above, that it is possible for both determinism to be true and for us (agents) to have the ability to do otherwise. And he wants to argue his case by sticking to the most challenging (for a compatibilist) analysis of the concept of “could have done otherwise”, the modal analysis.

This implies that there are alternative analyses of “could have done otherwise” that are more favourable to the compatibilist. What are they and why does List reject them? We turn to that question now.


2. Conditional and Dispositional Analyses of Alternative Possibilities
There are two interpretations of “could have done otherwise” that have been championed by compatibilists over the years. List calls them the “traditional conditional” analysis and the “new dispositionalist” analysis. The descriptors “traditional” and “new” are used to indicate their relative ages, with conditionalism having been championed since the early part of the twentieth century and with dispositionalism having arisen more recently. I’m going to ignore these descriptors and simply refer to them as the conditional and dispositionalist analyses.

According to the conditional analysis, to say that an agent has the ability to do otherwise, is to say something like the following:

Conditionalism : If the agent had chosen differently, or had tried to do otherwise, then he or she would have succeeded.

In other words, according to conditionalism, an agent’s ability to do otherwise depends on the antecedents of their actions (captured by the “if” portion of the conditional) being different. So to go back to the earlier example concerning my choice of a cup of tea over an alternative beverage, the conditionalist would (hah!) argue that I could have chosen differently had my preferences been different.

This might seem trivial at a first pass (more on this later), but it also seems highly compatible with determinism. After all, although the determinist will argue that the antecedents of my action could never have been different, this doesn’t prevent the counterfactual statement being true. Thus, although it may be that my preference for tea could not have been different at that particular instance, it may still be true to say that if it had been different, I could have have chosen differently.

So much for conditionalism, what about dispositionalism? According to the dispositionalist, the ability to do otherwise can be parsed in the following manner:

Dispositionalism : The agent has the disposition to do otherwise when, in appropriate circumstances (to be spelt out further), he or she tries to do otherwise.

In other words, according to dispositionalism, an agent’s ability to do otherwise is dependent upon the set of dispositions they happen to have. If this set of dispositions is such that, when circumstances shift or change, they perform different kinds of actions, then it is meaningful to say that they have the ability to do otherwise. Thus, to use the tea example yet again, if my set of character traits and dispositions are such that, when circumstances shift or change, I choose a beverage other than tea, I can be meaningfully said to have been able to do otherwise.

Again, this seems to be compatible with determinism. Although the determinist will argue that circumstances were such that my other character dispositions could not have been exercised at the relevant time, it would nonetheless be possible to say that I have unexercised character dispositions that could be exercised in different situations.


3. Objections to Conditionalism and Dispositionalism
List identifies two sorts of objections to the conditional and dispositional analyses of the capacity to do otherwise. The first sort of objection focuses on intuitively compelling counterexamples. The second sort of objection focuses on a substitution principle. I’ll discuss both, although I’m much less comfortable with the second than I am with the first.

Looking at the counterexamples first, consider the following. Most of us will agree that if a person suffers from a psychological or neurological impairment like tourette’s syndrome, volitional insanity or severe addiction, they will not always have the ability to do otherwise. On some occasions, their condition will compel them to do something. Yet oddly, under a conditional analysis, such individuals — even on the occasions when their impairment is most active — can still be said to have the ability to do otherwise. After all, if their compulsions were different, they would act differently. This seems counterintuitive, and so the conditional analysis seems flawed.

Here’s another counterexample, this time targetting the dispositionalist. Imagine someone trying to act under the most severe possible external constraints. For instance, someone being asked to unlock a bank safe while a group of bank robbers have guns pointing at his head, or someone driving a car and suddenly and unexpectedly experiencing temporary paralysis. In neither case (although the bank robber case seems debatable) would it seem right to say that the person had the ability to do otherwise. And yet, under the dispositional analysis, unexercised dispositions to do otherwise could still have existed. Again, this seems counterintuitive and so the dispositional analysis seems flawed.

Turning to the substitution argument, List first proposes the following test of the adequacy of an interpretation of something:

The Substitution Test : A good test of the adequacy of an interpretation of some concept is to substitute the interpretation for its target and see whether ordinary meaning is preserved.

I won’t comment on whether this is actually a good test of the adequacy of an interpretation since I don’t know. I’ll take it for granted that it is. My beef really lies with List’s application of this test, which begins with the following argument:


  • (1) If the agent does not try to do X, then he or she cannot do X.
  • (2) The agent does not try to do X.
  • (3) Therefore, the agent cannot do X.


Which is the direct negation of…


  • (4) The agent can do X.


To put it more pithily, if (1) and (2) are true, (4) cannot be true. And, according to List, this is where the problem arises. Since (4) is just stating what the agent can do, it would seem fair to replace it with the compatibilist analyses of the ability to otherwise since they are also statements about what an agent could do. One would think that such a substitution would preserve the overall logic of the argument. But as it turns out this is incorrect because the conditional or dispositional versions of (4) would go like this:


  • (4*) If the agent were to try to do X, he or she would would succeed in doing X.
  • (4**) The agent has the disposition to do X when, in appropriate circumstances, he or she tries to do X.


And neither of these statements are inconsistent with (1) and (2). Hence they fail the substitution test.

I have problems with the line of reasoning. Mainly because I’m not sure I follow it entirely and I think the source of my discomfort comes from premise (1). As I read it, this premise is a principle stating a necessary constraint on the abilities on an agent. But for the life of me, I can’t see why I should accept that principle as fairly capturing what it is that constrains the abilities of an agent. Surely, the truth of a principle like (1) is something that is at issue in the free will debate? That is to say, surely (1) begs the question against conditionalist and the dispositionalist?


4. The Modal Analysis
I may well be wrong about this, and to a large extent it doesn’t matter. What’s most important about List’s article is not his dismissal of these old compatibilist theories, but rather his attempt to develop a novel one. And this novel theory is one that accepts the modal anaylsis of “could have done otherwise”. As follows:

Modalism : It is possible for the agent to do otherwise.

The modal analysis depends upon the agent having the outright possibility of doing otherwise when they act. That is to say, at the time of their acting, it is genuinely open to them — irrespective of counterfactual conditionals and hidden dispositions — to perform multiple kinds of act.

This modal analysis is the most challenging for the compatibilist because it seems to require more than one possible future. This, of course, is in direct opposition to the definition of determinism offered earlier in this post.

So List is setting the bar pretty damn high for his compatibilist theory. Can he rise to the challenge? We will find out in subsequent posts.

11 comments:

  1. I don't think I understand the first counterexample.

    Let's say my tic is scratching my ear. On my understanding, I can't avoid this behavior by choosing not to scratch. I can't avoid scratching even if I try as hard as a can. So tic behaviors seem to be precisely the kind of thing a conditional analysis counts as "could NOT have done otherwise."

    If I had a different compulsion, that wouldn't be a result of my choice or my trying. So it seems irrelevant to the conditional analysis.

    What am I missing?

    ReplyDelete
  2. I'm always impressed by your posts. I have been scraping the barrel for a new take on FW issues as well and came upon this piece by Timothy O'Connor: CAUSALITY, MIND, AND FREE WILL. http://php.indiana.edu/~toconnor/Causality_MFW.pdf

    I'd love to see a Danaher style treatment of it.

    ReplyDelete
  3. Gato,

    I'll certainly have a look. I'm always interested in papers in this area, but I read far more than I ever end up writing about so I can't promise anything. I feel like I already know what O'Connor is going to say, but I can always give him another shot.

    Robert,

    I guess it depends on the phenomenological nature of the tic. If it involves an overwhelming desire to do something despite conflicting desires, then I think the objection to the conditional analysis goes through. But if it involves behaviour that seems unconnected to what you actually desire, then the counterexample doesn't work.

    ReplyDelete
  4. Yeah, he's a dyed in the wool theist, so you probably have a good guess where he's aiming. He was highly recommended by Christian philosopher of religion and Plantinga protégé James Sennett (on Luke's podcast). I am terrible at the formal logic, but if his formal propositions on (presumably contra-causal libertarian) free will are as "cutting edge" as Sennett claims, I'd love to see what the all hubbub is about.

    ReplyDelete
  5. O'Connor's definitely a top-class philosopher and, going by his interview on Luke's podcast, he seems like a reasonable guy. My only point is that he has been pushing his agent causationist theory of free will for some time now so I'm not entirely sure if he has anything new to say. Maybe I'm wrong. I'll check out that paper.

    If you're really interested in free will, I'd recommend reading Robert Kane's A Contemporary Introduction to Free Will. It's a few years old now but it's a great little read and doesn't use much in the way of formalisations. Kane's a libertarian, but he does a great job summarising the arguments on both sides. There's also a discussion of O'Connor's agent causationist theory in there.

    ReplyDelete
  6. Thanks so much- I will definitely check it out!

    I'm trying to get through Thomas Pink's defense of libertarian FW right now (the OUP's Very Short Introduction series) and even after rereading the chapters toward the end where he finally gets into the meat of his argument, it seems incoherent. Anything you can say about his work? Does Kane address him? By the untenable propositional jumps Pink made, I was literally checking to see if some pages were missing. I want to give him a fair shot and I am not a professional philosopher, but I've been reading philosophy for 20 years and I'm not completely daft either. He is *too* informal to be understood. That's why I like your blog so much; it's a perfect middleground.

    ReplyDelete
  7. Yeah, Doyle's page on him doesn't say anything about that paper yet either: http://www.informationphilosopher.com/solutions/philosophers/oconnor (I presume you've mused through this great website. Doyle holds the Cogito two stage model). Perhaps it does try to reaffirm ye olde metaphysical quasi-materialistic substance. That seems like more of a task for scientists than philosophers though (if this quasi-mental substance really is reified).

    ReplyDelete
  8. Re: Thomas Pink,

    There's nothing about him in Kane's introduction and I'm not that familiar with his work (I've heard of him but I've never read anything he's written). So alas I can't comment about it.

    ReplyDelete
  9. Thanks for looking. After just finding some appropriate criticism of Pink's book, it seems I have been wasting my time trying, out of fairness, to discover something coherent beyond the increasingly typical pragmatic arguments against determinism. I'll move on. I am going to delve in deeper to O'Conner and Kane when I can though. Peace.

    PS- I noticed you also requested the promised links mentioned in the last Reasonable Doubts discussion of recent FW studies (Vohs, Baumeister, etc), Episode 89. I recently posted them all there in the blog if you still want them.

    ReplyDelete
  10. Ah yes. I'd kinda forgotten about that. Thanks for posting the links.

    ReplyDelete
  11. BTW, what seems intuitively wrong with "(1) If the agent does not try to do X, then he or she cannot do X" might be that it may be true that the agent cannot do X... but it doesn't *follow* that the agent can't do X *because they did not TRY to do X*.

    ReplyDelete