free will and necessitation

Since I also blog (I'm participating here in the verbing of America) on Prosblogion, the philosophy of religion blog, and since the entry there that has drawn the largest response there is one on free will and necessitation, I thought I'd blog the topic here and see what the real experts think...

I'm not as up to speed on the Frankfurt literature as I wish to be, so maybe I can learn something by provoking responses. Here's what I see so far. There are Franfurt-style counterexamples to PAP, and the most common way of defending the Kantian claim that responsibility requires being able to do otherwise is to talk about internal actions such as trying or intending. My uninformed response to this maneuver is that it doesn't help. If the internal factors are themselves actions, and if they have temporal duration, then Frankfurt-style counterexamples are still easy to construct: just have a demon watch, and interfere when the first segments of the internal action appear to prevent the person from trying or intending.

So, the only hope, as I see it, for defenders of the Kantian view is to talk about instantaneous internal factors which have no causally sufficient priors that could be used by the demon to anticipate the control exerted by the agent over that factor. Here things get murky for me. Could there be such instantaneous factors? It is hard to see how they could be actions, though I can see how they could be the beginning of an action; but I see no argument why there couldn't be instantaneous actions. I think the more natural route, however, is to think in terms of events such as the beginnings of an internal action--"beginnings" for short. Could the Kantian view be saved by insisting that beginnings must be such that we could have initiated a different beginning instead?

'Beginnings' is the wrong concept, however, since there could be an action A and its absence, ~A, that share the same beginning. At some moment, however, they will have to diverge. So maybe the instantaneous factors in question should be called "divergences". Let us formulate the Kantian view, then, as the claim that, regarding divergences, we must be able to proceed via either of the two divergences, and these divergences will themselves not have causally sufficient priors.

This is where my reflection ends. It looks to me that the rescue of the Kantian view requires talk of instantaneous divergences which are under the control of the agent in question, and it seems to me that Frankfurt-style counterexamples cannot be constructed against a "control over divergences" position. I'm uncomfortable with the position, however, in part because of the required instantaneous character of divergences.

So, finally, the point of the post: can the Kantian intuition be save from Frankfurt counterexamples by appeal to divergences? Or, perhaps, is there some other option I'm missing?

Comments

[robert allen]

Kvanvig:

What explains the divergence? If it is just a random occurrence, then you run into the same question that Prof. Kane faces in regards to his "self-forming actions," viz., why should we think that the agent in question has enough control over himself to be responsible for his action?

[kvanvig]

The explanation of the divergence I leave off for now, though it is a crucial question to answer. Right now, I just want to know about the scope of Frankfurt examples, and whether there is a way out of them. Even if there is, further questions such as yours may jeopardize the view on other grounds.

[John Fischer]

Jonathan: thank you very much for your extremely thoughtful post here and also for your very helpful posts on Prosblogion. I encourage Garden bloogers ("Gardeners"??) to have a look at Prosblogion.

Anyway, there are lots of complex issues in the neighborhood. But I would say that, on the indeterministic side, one might be able to come up with versions of the cases where you've got rid of any "divergences". Alternatively, the only divergences left are "mere flickers of freedom" and not sufficiently robust to ground ascriptions of responsibility. So, for example, you might have a look at the papers of Mele and Robb, and also David Hunt. (Some good resources here: my "Some Recent Work on Moral Responsibility," Ethics Oct. 1999, and the Widerker/McKenna volume published recently with Ashgate Press. Another promising indeterministic strategy is the "buffered alternatives strategy" of Pereboom, McKenna, and Hunt. This seems to work even in an indeterministic world: the residual alternatives are not robust.

Alternatively, I believe that the Frankfurt-type examples can work in a helpful way even in a causally deterministic world. They need not beg the question, properly understood. See, for example, my "Frankfurt-style Compatbilism", recently reprinted in Watson's OUP Readings on Free Will, or on the online "Papers on Agency" blog, the strategy is set out in the "Free Will and Moral Responsibility" paper.

[robert allen]

Prof. Kvanvig:

I also address the issue of robustness in "Robust Alternatives and Responsibility," (Journal Of Moral Philosophy, vol. 1.1, April 2004). I discuss Frankfurt cases in "Re-examining Frankfurt Cases," (The Southern Journal of Philosophy, vol. XXXVII, Fall 1999). Both papers are available at my website: http://home.twmi.rr.com/robertallen/.

[jon kvanvig]

Excellent comments and great leads! Now all I need is more time... As I take the comments, the fundamental issue will be whether ascriptions of responsibility can be based on the way of escape in question, not so much whether the Kantian position can be sustained as merely a claim that if you are responsible, you have to have been able to do otherwise. Still a deep problem, though.