Blog Coordinator

« Two FW & MR-related Volumes Reviewed in NDPR | Main | Passing the Featured Author Torch! »



Feed You can follow this conversation by subscribing to the comment feed for this post.

Is there a relevant difference between this question:

A. "there is a core concept of free will that is aligned with X"?

and this question:

B. "the definition of free will requires X"?

And, even if they are different, how can we answer question A without answering question B?

It seems to me that most, if not all, of the free will debate requires an investigation into the precise definition of the term "free will" (putting aside the question of libertarianism, which involves neuroscience). And I am not sure how we can decide the definition of "free will" without asking lots of people, in a manner parallel to how Knobe and other ExPhi people ask questions about compatibilism and thought experiments.

Yet nobody seems to take on this kind of project of "what exactly is the definition of free will." Moreover, if they did, I am fairly (moderately?) confident that we would find a lot of variation and vagueness. And this variation and vagueness would probably be so strong that it prevents success compatibilism (or any other view) from carrying the day.

What I've described above, more or less, is Double's view on the non-reality of free will. Am I wrong? How?


Well, that was my cat's--William Jefferson Kitten's (Willie's)--contribution. Probably will make as much sense as anything I say.

BTW I like your draft paper on this Joe, and helped me see your viewpoint better.

I do have some sympathy for Kip's remarks. Frankly I think people are collectively/transtemporally inconsistent in their take on how FW relates to matters of MR. In your paper you seem to say that this indicates overall that the folk are compatibilist. And I have logical sympathy for that too. Incompatibilism after all is univocally crystal clear on one thing: freedom (or MR in that version) is a function of the truth of a metaphysical binary, determinism or indeterminism. So if the folk are incompatibilists they may not vary from that baseline. But since compatibilism forms an open-ended class of views that collectively are a denial of that (along with a denial of Double-like semantic skepticism about the concepts/terms FW/MR), then reliable reported X-PHI deviations from FW/MR incompatibilism prove that the folk are not, as a collective synchronously and/or chronically, adherents of that view. That makes clean logical sense. But it doesn't--and can't by itself--show a certain compatibilist view is true (or that Double-like skepticism is true or false), or that some subset of folk aren't consistently incompatiblist by itself--maybe further data could.

I'm still working through Manuel's book but I take it that all this motivates his revisionism, given problems with libertarianism and rejection of skepticism both Double-like and hard incompatibilist. Maybe what separates you from Manuel is methodological more than anything, especially given your shared rejection of libertarianism and Double/Derk skepticism. Manuel seems (so far as I've read) to come from a holistic/consequentialist standpoint that does not stress a monolithic view of the compatibilist agent. Your commitment to up-to-usness seems more focused as a metaphysically groundable property, but I wonder if Manuel would finally say that your view is consistent with his.

Time to feed Willie!

Kip: It can't JUST be tracking the usage of the term "free will" that matters when we try to understand the concept. This view strikes me as a kind of logical postitivism about meaning. It might be better to consider the meaning of "free will" relative to other theoretical constraints than just empirical adequacy: consistency, simplicity, fecundity, etc.

In any event, the view I'm trying to promote is supposed to be lean wrt meaning, leaving the philosophical issues as a matter of debate not definition. I don't see a problem here. Why not define "free will" as sourcehood or the ability to do otherwise or something fundamental. To the extent that you get empirical data suggesting otherwise -- folks who lean toward incompatibilism as being an essential feature of free will -- you could just say that there was some error of reasoning, or something like a category mistake. In other words, folks often suppose an essential connection when in fact there is not one.

What's wrong with explaining things this way? And this explanation is only needed if it turns out that Nichols et. al. are right and Nahmias et. al. are wrong, which I take it is still a matter of debate.

Thanks, Al!

I'm not sure I think the folk are compatibilists. Maybe they are in a sense, depending on how the Nahmias/Nichols debate works out. I want to say that compatibilism comes for free. Consistency is a theoretical constraint. You throw in your various commitments and if determinism and the free will thesis are among them and you accept those commitments, you've got compatibilism. It takes an argument -- or at least an intuition -- to say otherwise. So compatibilism is the de facto view.

Unfortunately, incompatibilist intuitions arise upon the slightest speculation. One needs only to mention the possibility of determinism -- of the future being contained in some sense in the past -- before worries arise. So incompatibilist "intuitions" are easy enough to produce. In this way, the compatibility problem is analogous to the problem of epistemological skepticism.

Two final remarks: First, in the case of epistemological skepticism, our intuitions about having knowledge return quickly once the skeptical worries leave. Maybe something similar happens wrt the compatibility problem.

Second, it might turn out that our attitudes about free will are a lot like Bacon's claim about our attitudes about theism: a little philosophy turns one away from God but a little more turns him back (to paraphrase).

We start with a core concept of free will as sourcehood. A little speculation gets us to draw other conclusions: free will requires the ability to do otherwise, free will requires ultimacy, free will is essential to moral responsibility, or free will is incompatible with determinism. In some cases, the speculations prove to be true. In other cases, they don't. Either way we should be cautious about folding the speculations into the concept.

Hi Joe-

Happy New Year!

FWIW, I'm currently inclined to think what you are calling Rev. 1.0 is too underdescribed to characterize as revisionist or not. I think it is a reasonable project to be involved in, but it seems to me an open question whether the term of art maps on to any folk usage, and if so, the extent to which it does (and does not).

On the difference between 1.1 and 1.2, I still don't know what "tweaking" means in this context. I take it that there is a big difference between prescribing a view that conflicts with received views (say, in the early 20th century, holding that race is not biological but social) and a view that makes more precise the coarse-grained commitments of some widely received view (for example, that agaves are a particular sub-species of plant—this supplements or refines a coarse view that agaves are plants, but it doesn't conflict with any standard views about agaves in the way that holding that agaves are minerals would).

I'm inclined to think the difference between revisionist and conventional forms of compatibilism is that the revisionist regards his prescriptive account as in conflict with folk commitments whereas the conventional compatibilist does not view her proposal as in significant conflict with any folk commitments. (This is, I think, the view that Dana and Eddy have, and I would have taken you to have.)

On the Alan/Joe discussion, I do think one thing the X-Phi literature has made clear is that there is not nearly so much univocality in commitments among the folk as philosophers used to think. However, the old conception of the problem at least implicitly (and sometimes explicitly) seemed to presume that the folk were all and always *really* compatibilist or incompatibilist. On my reading of the history, the vast majority of folks in 20th century analytic tradition seem to talk that way, from Schlick to Ayer to Strawson and so on. And, I think this way of framing things was difficult to avoid, given the broader philosophical posits of English-language philosophy for most of the 20th century.

So, I suspect our philosophical positions and terms of art grew up around the univocality presumption. However, in an X-Phi world, it seems to me less plausible (but not implausible) to hold the older archetype versions of views under which we could easily assert that the folk are *really* this or that. If I were to psychologize, my suspicion is that people who 25 years ago have been compatibilists would have also insisted that of course the folk are virtually all really compatibilists. This is much harder to say now, so beyond simply ignoring the issue, compatibilist are left needing to either show (as Eddy endeavors to) that the folk really are compatibilist, or they to find some way to sever their projects from questions about what the folk think—but then methodological and phil language-y questions arise very quickly, and not many folks have been happy to walk down those paths yet.

(Something similar is true of incompatibilists too, but I think there are some extra wrinkles on that side.)

In optimistic moments, I like to think the apparent fact of non-univocality of folk commitments makes revisionism and the theoretical challenges that plague it far more unavoidable than it would have seemed twenty years ago.


There is a lot I could say here, and you know more philosophy than me (and more philosophy of language). I'll just respond to a few points.

"Kip: It can't JUST be tracking the usage of the term "free will" that matters when we try to understand the concept"

I'm not saying that usage is EVERYTHING we need to study. But I think it is a major component, and likely an essential component, of what we study. By far the majority of how we understand what terms mean, is by tracking how the terms are used. There might be outlier case where usage is unimportant for definitions, but I have never seen anything beginning to show (persuasively to me) that this is the case for free will.

"It might be better to consider the meaning of "free will" relative to other theoretical constraints than just empirical adequacy: consistency, simplicity, fecundity, etc."

In my view, all of these concerns are misguided. Suppose that everyone uses free will to mean X. Suppose, however, that X is inconsistent, complex, and not fecund (i.e., does not produce a lot of interesting research). Does this mean that free will doesn't mean X? Not in my view. You can't go out in the public and say "free will means Y," where Y is consistent, simple, and fecund. That would be deceptive. All you can do is create a toy definition, free will*, and do work on free will*, which has no direct relevance to free will as people actually use the term.

"In any event, the view I'm trying to promote is supposed to be lean wrt meaning, leaving the philosophical issues as a matter of debate not definition. I don't see a problem here."

I admit that there are situations where you can leave terms lean wrt meaning, and still solve problems about them. For example, you might be able to solve a geometry problem by knowing that a square has more than 2 corners, without knowing that it has four corners (as one random example). But it is not clear to me that the free will debate, including the sub-debate here, admits of those kinds of solutions. It seems to me that the free will debate, including the questions you ask in the title thread, hinge upon exactly the kinds of attributes that you want to leave undefined.

In fact, there is a paradox here. Your thread begins by asking (more or less) how we should define free will. Each of revisionism 1.0, 1.1, and 1.2 asks something like "what relevant group of users define the concept of free will" and/or "which concept/definition of free will should be use." In response, I suggest that we need to embark on a project of figuring out what the definition and usage of free will is. Yet, when I do that, you resist. It's like you want to figure out the definition of free will, but not too much, because of reasons X, Y, and Z (e.g., there might not be much philosophy left to do afterward).

Kip: Thanks!

I think you're misunderstanding a point that I made. We can ask, What does X mean? I want to say it's a lot like asking, Which scientific theory about X is true? Of course the data matters. But there are other things that matter besides the data.

Manuel: In a previous draft I called the first view "revisionism" 0.0, so I admit that it is hardly a kind of revisionism. Rightly or wrongly it seems to be widely held. Do you agree?

The rest of your comments are very helpful. Thanks!

I agree that the view you describe as 0.0 is a widely held view.

However, I suspect it is held mainly in reaction rather than in consideration (i.e., it is explicitly espoused only in frustration with x-phi and metaphilosophical pressures) rather than adopted at the outset of theory construction.

In general, when it comes to methodological issues I think most of us don't think too hard about it until someone complains. Then, we tend to retreat to whatever view most easily lets us keep saying the bulk of what we had been saying all along, quite apart from whether the commitments thus described were the project we had in mind at the outset.

But I'm just speculating.

Manuel: I'm not sure it is a reaction but there is a failure of consideration. Thus, I agree with the general claim. I admit that I'm as much to blame as anyone.

Certainly we don't think much about methodology until criticisms arise but aren't methodological assumptions pretheoretical? Don't they shape theories even when they are not explicitly recognized?

Of course, I'm just speculating, too.


You write: "What does X mean? I want to say it's a lot like asking, Which scientific theory about X is true? Of course the data matters. But there are other things that matter besides the data."

As I said above, I agree that data (i.e., common usage) is not everything.

However, I think you will agree that there are two kinds of situations: A. situations where philosophers can do meaningful work despite terms having some ambiguity or vagueness ("lean wrt meaning"), and B. situations where attempts by philosophers to reach results are thwarted because the terms are *too* lean with respect to meaning - the terms are essentially under-defined, resulting in vagueness and ambiguity, and this vagueness and ambiguity occurs precisely where philosophers need clarity to answer their questions.

All I am suggesting is that the free will debate, including the questions you ask in this thread, are like situation B, not situation A. In support of that, I can cite two facts: (1) smart guys like Richard Double agree with me (and meta-skeptics like Tamler agree with my conclusion for somewhat different reasons), and (2) philosophy has not achieved consensus on solving the free will problem despite thousands of years of work.

I don't pretend to be able to prove to you that the free will debate conclusively is like situation B. But I think it is something worth considering. The hypothesis, if correct, would explain a lot.

Hi Joe-

I don't think methodological presumptions always have to be pre-theoretical, but I agree that they frequently are. And yeah, I absolutely agree with you that methodological presumptions can shape theories even when not explicitly recognized.

If we disagree on things here, it may only be about the frequency with which 0.0 is an unarticulated but operative assumption in theorizing, vs. the frequency with which it is an after-the-fact methodological position adopted in light of various pressures about ordinary meaning and the like.

My problem is that my intuitions are all over the map, and even meta-maps. Joe and Manuel and Kip all make good points here.

I'm not as skeptical as Kip is. Take a phi/sci analogue. Objects move. Why? Aristotle says constant pushing (vertically; yearning otherwise), Philoponos says transferred impetus, Galileo says it's just natural (and relative!) though circular once started (inertia-1), Newton says it's natural but linear force-free (inertia-2), Einstein says it's never a matter of forces but spacetime geodesics sometimes interpreted as forces. Let's see. That's over two thousand years of explaining (and mansplainin' as is apparent) one obvious thing. Objects move.

People choose, decide, deliberate and the like when options occur to them as we now call at least logical possibilities. What really--metaphysically--is going on? For over two thousand years we've had a lot of different answers. But people definitely do something behaviorally that we call choosing and deciding.

Double moves onto the scene somewhat like Kuhn does, and I think not accidentally about the same timeframe. Kuhn looks at all this change over history and declares an answer to "Will there ever be an end?" NO! Stand back from the process and see what the process itself shows us. Change. No inter-paradigmatic values to assess change. The beginnings of scientific antirealism, not just about motion, but the very discipline that assesses motion.

That's the service Double has done us, helped along a lot by people like Frankfurt, giving us a new paradigm of FW, along with PvI's reaction, and Fischer, and Watson to name big names recently, and he draws a reasonable conclusion at the same level as Kuhn. And justifiable.

But I see work like Manuel's--to continue the parallel--like van Fraassen's. As pragmatic considerations of what science is doing is important for BvF, what X-Phi reveals to us about choosing and deciding are on a similar methodological level for MV. You have to take in the data and make it relate to key questions of FW and responsibility, and identify values that are justifiable in delivering the answers that fit the data. Now I'm not suggesting Manuel as a revisionist is somehow a FW antirealist--I doubt he's taking revisionism that far. But methodologically I do see alliances with the BvF pragmatic strategy for dealing with how to answer the question: what's up with the motion of objects anyway?

Joe and I are probably more like scientific realists here (not to speak for you Joe, just grokking), urging some real and identifiable ground for claims about FW and MR that can satisfy empirical demands. But my recent thinking has really moved--whatever that means!--toward seeing that pragmatic considerations are ineliminable for meeting those demands.

"Now I'm not suggesting Manuel as a revisionist is somehow a FW antirealist"

I'm suggesting it!

But I admit, grudgingly, that MV remains the best judge of what MV believes.

I'm explicitly a quietist about theses issues, but I have offered characterizations of my view that go through under realist and antirealist views of the subject matter. In general, though, I'm very sympathetic to Al's characterization. What I'm about to say is surely too quick, but I think things stack up this way:

If anti-realism is true, then I think a pragmatic/normative reconstruction is all we've got available to us, and all we ever really had. I'm obviously happy to advance a theory under those constraints.

If realism is true, either normative considerations are part of the reference-anchoring features of any realist notion (at which point, my form of connotational revisionism is a plausible view, sez me) or normative considerations are not part of the reference-anchoring features of the realist notion. If the latter is true, then at that point there is little reason to object about to proposals that re-anchor reference in some practically significant notion. If so, then my proposal is a form of denotational revisionism, and thus, a concession that the old (irrelevant-for-practice) notion of responsibility in the strict sense doesn't obtain, coupled with the view that we should switch topics to something that does matter for our practices.

Thanks for the kind remarks Manuel.

Correction to my phi/sci/fw/mr account above: Aristotle's pushing is horizontal, of course--the yearning of final causation accounts for the ups and downs; typing too fast or thinking too slow.

A bit of a follow-up question though. I can't escape the sense that some form of pragmatism in service of large-scale norms does very heavy background work for revisionism. For example, in the BBB 121-26 section you cite, you say in a footnote that one way to become a denotational revisionist is "think there is some thing that better does the work of the concept than the kind of thing we have been referring to" (fn. 121). The word "better" in that passage in modifying "the work of the concept" seems to me to be referring to ultimately pragmatic justifications that aim toward prescriptive norms. Now this is just one passage, but that same word "better" is 1/3rd of the title too. Yet, the index doesn't have a listing for pragmatism or a cognate.

At least for the methodology of revisionism, is some sort of pragmatism in service of prescription a necessary condition for embracing that view? Can any form of revisionism do without some pragmatic superstructure?

Thanks in advance for helping me think through all this. I love the book and its fresh approach.

To paraphrase, Moliere, these ten years now I’ve been speaking in pragmatism without knowing it!

I suppose I'm somewhat skeptical that pragmatism is *necessary* for all forms of revisionism, only because it seems to me that one could have a view on which the proper reference-fixing features of 'freedom' and 'responsibility' involve particular various normative properties which might possess various non-pragmatist credentials. If we can identify those properties, we might still have reason for thinking connotational content drives theories in ways disconnected from the reference-fixing features. Nothing pragmatist there, so far as I can see.

In contrast, denotational revisionism is more plausibly a sort of view that will typically be motivated by broadly pragmatist convictions. But I'd have to think carefully about whether there are non-pragmatist considerations that could support denotational revisionism.

My exposition about this is probably more problematic than it should be, given that I'm frequently trying to walk two different lines simultaneously, i.e., (a) presenting a set of broadly revisionist possibilities and (b) identifying the one that seems most promising to me.

On, (a) it seems to me that there is probably good reason to reject the view that pragmatism is a requirement on revisionism in general. On (b), though, it is probably fair to say that I am attracted to broadly pragmatist approaches in this domain.

Perhaps it is the fact that I'm aware of having gotten all my pragmatism filtered through Rorty seminars that has made me uncharacteristically cautious about claiming a pragmatist mantle. That said, an alarming number of self-described pragmatists tell me that I'm really a pragmatist.

That was smooth and satisfying Manuel. You are the Bunnahabhain 18 Year Old single malt of bloggers.

Thanks Al, and thanks to Joe for sparking the conversation.

Kip: I'm not sure of the relevance to the fact that there is no consensus about (some) philosophical debates. David Chalmers has an article about this that he's working on but I haven't looked at it yet. This is true about ALL philosophical problems. There is nothing special about free will on this view. But I think that you hold that there is something special about free will. It is incoherent unlike other philosophical concepts.

Also, I have a lot of sympathy for situation B. It isn't so much the case that I believe that situation A is the case and that situation B is not the case. It is just that situation B is, for me, the challenge. Doing philosophy, for me, is confronting the challenge. But I’m not done yet. I might go with situation B in the end. For now, I’m interested in trying to meet that challenge.

Al: Just a quick point since I have other posts to respond to (some of yours!): I really like the “in the Fog” post, making analogies between various positions on free will and in philosophy of science. You’ve got my view right! Something to think about.

Wow! I wish my life was less complicated, and I could contribute more to the conversation that Al, Kip, and Manuel are having. Alas, there is time for just some brief (likely inadequate) comments.

I agree with Manuel that there is no 1-1 mapping of revisionism onto antirealism or pragmatism (broadly construed). And I think Al accepts this too, given the qualifications I express below.

Manuel is right when he says “denotational revisionism is more plausibly a sort of view that will typically be motivated by broadly pragmatist convictions.”

Just for some background, Manuel distinguishes between connotational revision and denotational revision, where the latter gives up reference fixing content and the former does not (2013, 88-9). Vargas advocates the former, so his concept of free will has some flexibility. Perhaps the folk believe that free will is essentially libertarian. Nonetheless, we can wean them away from this view without changing concepts.

Things are different for the denotational revisionist, who is committed to free will skepticism should it turn out that libertarianism is incoherent, for free will is an essentially libertarian concept. To the extent that a denotational revisionist believes that there is no libertarian free will, to that extent she is an antirealist about free will.

But I think that Al was speaking to a broader point. After all, it might turn out that free will is essentially libertarian – and this is all the more likely if we decide to base our decision on the empirical data. To the extent that any form of revisionism is bucking against a libertarian view, it just might be antirealist (given arguments against the coherence of libertarian free will).

The comments to this entry are closed.

Books about Agency

3QD Prize 2014: Marcus Arvan