As I’ve suggested in previous comments, I believe that the counterfactuals involved in understanding human choice should not be treated as special—that is, they are not best understood or analyzed in a way that is different in kind from the counterfactuals involved in understanding other contingent events. Here, I want to present this issue as a question for incompatibilists and then I’ll push it by considering fixity of the past principles. This will be a lot to do in a blog post (and the points are not original*), but if nothing else I’m trying to engage more directly with traditional incompatibilist arguments (perhaps appeasing those who think I’ve been “bypassing” the real issues ;-)
Terminology: Call an analysis of ‘can’, ‘could have done otherwise’, and/or ‘opportunities’, etc. as relevant to free will or choice reducible if it works using whatever turns out to be the best analysis of counterfactuals for contingent events in general (where that analysis does not itself include special features to deal with free-will related issues such as choices). Otherwise, the analysis is irreducible.
Question: What is the motivation or argument for irreducibility? That is, when we think or say that some contingent event (such as a die landing six, a leaf landing here rather than there, a dog just missing a Frisbee, a rat going left rather than right in a maze, perhaps even a soccer player missing an easy shot, etc.) could have happened otherwise (die landed five, leaf landed there, dog caught it, rat goes right, player makes shot), do we or should we understand such counterfactual claims as different in kind from cases where we think or say that some agent could have (freely) chosen otherwise? If the answer is yes, why? If the answer is no, then does that entail that indeterminism must be true if these claims about ‘non-agential’ contingent events are true?
(To be clear, I think humans clearly have special, or uniquely developed, abilities or powers—e.g., to envision various future outcomes as they depend on what we decide, to consider our reasons for each, to try to control the influence of some of our desires, etc. I just don’t think possessing those powers or exercising them requires an irreducible analysis; rather, it will involve an account of a bunch of cognitive capacities we possess, at least to some degree, and then a reducible analysis to consider when we could or could not exercise them).
Fixity of the Past: As one of the premises of the Consequence argument, van Inwagen stipulates NP, which means “P (some statement about the distant past before agents existed) is true and no one has a choice about whether P is true.” John Fischer (as usual) helps clarify the issues by offering a more precise statement of the Fixity of the Past (FP) premise:
FP: For any action Y, event e, agent S, and time t, if it is true that if S were to do Y at t, some event e which actually occurred in the past relative to t would not have occurred, then S cannot do Y at t.
(I think John argues that the CA can use this principle without using a transfer principle—is transfer built into it?—and I think he remains agnostic about whether FP is true, though he says it is highly intuitive; in any case, he clearly thinks agents can be MR even if FP is true and determinism is true.)
Now, consider a parallel principle to FP:
FP*: For any event Y, event e, and time t, if it is true that if Y were to happen at t, some event e which actually occurred in the past relative to t would not have occurred, then Y cannot happen at t.
Determinism entails (assuming we hold fixed the actual Laws of nature): For all events Y, events e, and times t, if Y occurs at t, then if Y were not to occur at t, some event e that occurred prior to t would not have occurred.
So, if determinism is true, accepting FP* would mean that for any event that does not actually occur, that event could not occur. And that means that every event that could occur does occur (does this entail that for every event that does occur, it had to occur?). According to FP*, when people think (and say) that some event that occurred in a specific way at a specific time—the die landing six, the leaf landing here, the dog barely missing the Frisbee toss, the shot going wide, etc.—but that it could have happened differently, they are either committed to believing that determinism is false or they are making a mistake of some sort. I see no reason to think we do (or should) assume determinism is false when we think and say that such events could have happened (or could happen) otherwise. And it seems strange to think we would have always been making such a mistake if the MIT physicists (correctly) announce that determinism is true. Isn’t it more likely to think that our understanding and use of such counterfactuals makes no commitments about the truth or falsity of determinism?
But if one thinks there is reason to reject FP*, why should we accept FP? Wouldn’t it make more sense to recognize that determinism simply means that if a contingent event (or action) had occurred differently than it actually does, some earlier events would have had to be different? Isn’t that what we normally assume anyway—that something (we often know not what) would have had to explain why things happened differently? (The next post will suggest that causal modeling and interventionist theories of causation can help flesh out these ideas.)
* In lieu of footnotes: I want to make clear that the general approach suggested here is obviously not original—it’s just one way of presenting an approach going back at least to Lewis, Horgan, and others responding to CA and picked up more recently by Perry, dispositionalists, and Vihvelin (whose new book I have not yet read).
I also want to recommend a paper by Christian List that offers a nice analysis of ability to do otherwise, one feature of which fills out an idea I’ve been playing with for a while—i.e., that multiple realizability can explain how, while holding fixed deterministic laws and the ‘high-level’ state of the universe, events (including choices) could have happened otherwise (because of micro-level differences).
I also want to suggest people check out Charles Hermes forthcoming paper in Phil Studies “Truthmakers and the Direct Argument,” where he raises some related issues in the context of developing really interesting points about the need for an “ontological turn” in the free will debates.