Blog Coordinator

« Fischer on Freedom and Death at 3:AM Magazine | Main | Free Will in the News »



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

Hi Matt & Chandra,

In my reply to Matt, I focused just on clarifying my concessions to Chandra. I should have emphasized that I don't think that Matt has muddied the waters, and that I fully agree that this new way that Matt & Chandra are formulating manipulation arguments can be helpful.

In my view, we've now moved over to what I've called "Meta-manipulation Arguments" (I did a poster on these last year at RoME; manuscript in draft). In this type of argument, you build your metaphilosophical principles right into the major premises of the argument. Given standard presuppositions about the evidential role of intuitions, presumption of their rationality in the absence of defeaters, etc., it seems to me that meta-manipulation arguments are the fallback for incompossibilsits when their pet *persuasive* manipulation argument fails. Do your meta-manipulation argument right, and (I think) you can knock bullet-biting/hard-line replies & stalemates off the table.

Matt: I see your recent paper on manipulation arguments as focused on persuasive manipulation arguments, but Tognazzini, in his reply to your paper, seems (at least in parts) to covertly switch the conversation to meta-manipulation arguments. The dialectics are different for each type of manipulation argument. Once we distinguish meta- from persuasive manipulation arguments, I think you could show that some of Tognazzini's replies to you are off the mark.

Chandra: Okay, I grant you a temporary reprieve on what your cases do/don't show...but, spoiler alert: the same issue will pop up again in other contexts this month!

Bayesian PvI:

Postulate no conceptual change, and let F mean at least some humans have free will and D mean determinism is true. Suppose PvI starts with priors such that P(F)=0.999999 (just 1 in a million shy of unity) and P(D) = 0.00001 (10 parts per million). Despite the seemingly convincing argument for incompossibilism - reflected in the low probability for D - given the overwhelming probability of F, PvI can still reasonably suppose that P(D & F) = 0.000009 (9 parts per million). So on learning D, P(F) becomes 0.9.

Look what you've done, you've gone and made me defend PvI.

Paul: Cool! I started out my reply to Kip thinking that I was going to defend PvI's reasoning, then talked myself out of it.

Let me see if I've got this now:
PvI would be well within his (Bayesian) rights to assign a super low prior to D--although he seems to have no reason to give D such a low prior *other than* that he believes both that free will and that (based on the Consequence Argument) determinism must be false for free will to exist. So, PvI's reasoning only *seems* fishy because the very argument that he used to set his very low prior for D is the one that he *rejects* as unsound when he finds out that D is true. By Bayesian standards, though, this is fair play. Yeah?

Perhaps, if Eddy gets a chance, he'll let us know: Have we totally gotten rid of that fishy smell?

Thanks Kristin -

You explained the point much better, and with more precision, than I did. I still wonder how PiV can be so confident that free will exists, especially if he uses "free will" as a placeholder for a number of different candidate concepts (in the manner you describe).

Here's another thought: if philosophers like PiV and Mele are willing to shift from compatibilism to incompatibilism (or vice versa) based on confidence than free will (or autonomy, or moral responsibility) exists - then aren't they revisionists of some sort? Wouldn't Vargas call them allies? Is anyone in this debate *not* a revisionist of some sort?

Hi Damir,

About your proposed case: I'm not sure I see how putting the indetermination into Ernie's reasoning would help. Among my worries with Chandra's case is that people (even ideal targets) might fail to get the "right" intuition that Ernie lacks freedom and responsibility in his "manipulation cases" because he's watered down the appearance of manipulation so much that there is virtually no appearance of manipulation at all. By the same reasoning, I would say the same about your case. That said, I feel that I might be missing your point somehow...if so, could you clarify?

Hi Kristin,

I expected that you will find manipulation in my case too weak. However, my idea was that we don’t need manipulation of the same strength when the target is the lack of self-creation as when it is determinism. In particular, I thought that manipulation (or design more precisely) in the former case does not have to extend all the way to action. It is enough that one is designed to be a certain sort of person at the moment of making a decision. I thought that that might be enough to make it intuitively clear that the person did not create herself before making the decision. But maybe nothing short of total manipulation is sufficient to make it intuitively clear that one is not self-created. Anyway, I hope it is a bit clearer now what I wanted to say.


I didn't interpret what Kristin said as taking PvI to be using FW as a "placeholder". Instead he is using it as [the thing we are arguing about]. Logically speaking, if compatibilists and incompatibilists were talking about different things, then there'd be no incompatibilism between compatibilism and incompatibilism. The very fact that we consider this a debate entails that we are all working with the same assumption -- that "free will" means the same thing for everyone.

Alas, I've been trying to formalize the fishiness in PvI's flipflopping for a few years and have failed (so far). If nothing else I want to publish a paper where I get to use "van Twinwagen" (PvI's doppelganger who is convinced determinism is true and hence convinced that Transfer is invalid). Paul's Bayesian defense might work (though one of my students, Ben Freed, tried to use Bayesianism to show PvI's view was indefensible). I'm inclined to think that the thesis of determinism is not "connected" to our beliefs about free will or transfer principles in a way that suggests these beliefs should influence our priors regarding the truth of determinism. I think the truth of determinism is dependent on what our best theories of physics discover. And given what I know about these theories (not enough!), I'd say we should think determinism is possible (with credence of about 30%).

A bit off topic, but if anyone saw the van Persie goal against Spain yesterday, it was a thing of beauty ... and seemingly a thing of contingency. There were so many ways it might not have worked out to glance off his head in just the right way to loft over the keeper. We often talk about how the truth of determinism might conflict with the ability to do otherwise. But of course, the same reasoning entails that, in some sense, what happens *has to happen*. So, in that sense, van Persie could not have missed. Wow! It's amazing to me that some thesis that may or may not be true for all we know is such that if it's true no one could miss the goals they make. Seems better to me to leave our analysis of abilities, could, counterfactuals, and such be less dependent on how the physics turns out (or what is different, to develop analyses that can 'go both ways').

Ugh, if your grammaticality intuitions tell you that some of my sentences in the previous comment are ugly, I say they are truth-tracking. On that note, I should add that the discussion of the zygote argument has been really helpful. For the most part, I think Chandra's right. And Oisin Deery and I are working on a paper to validate the (truth-tracking) intuitions that Ernie is unfree but Bernie is potentially free (we'll present a short version at SPP next week).

fwiw, I have a paper coming out in the volume edited by Timpe and Speak in which I seek to pinpoint the problems with PvI's Flipflopping. Glad to share it with anyone who is interested.

Also, fwiw, I think that there is a similar problem of manipulation (initial design) for the libertarian. That is, manipulation/ID is also a challenge for a libertarian, since the manipulation/ID can be set up in an indeterministic world. So I don't think that our intutitions are (or should be) driven by the fact that Ernie is causally determined.

Distinguishing regulative control from guidance control is crucial here (as elsewhere). Ernie may lack regulative control, but he doesn't lack guidance control, and that's all the freedom he needs for moral responsibility. I'm not so concerned what the folk say about ANY of these issues, because intuitions are so much affected by nuances in how questions are posed. Further, philosophy isn't a poll (as Tim Scanlon pointed out on this blog some time ago--and it bears repeating). But I stick my neck out more than is necessary here.

Matt wrote:

"I didn't interpret what Kristin said as taking PvI to be using FW as a 'placeholder.' Instead he is using it as [the thing we are arguing about]."

Matt, I see your point, and I don't mean to misrepresent either Kristin's or PvI's views, although I'm sure I have to some extent.

I think we need to distinguish between different sense of "placeholder." In your quote above, it is obvious to me that "[the thing we are arguing about]" is a kind of placeholder. Hence the brackets. Consider an analogy: detectives are looking for a criminal, but they don't know who it is. They call the person "the perp," but they don't know his/her identity. The "perp" is a placeholder, in one sense, but not a placeholder (in your sense), because it refers to "the person who committed the crime."

With that distinction in mind, I think we can make our different points above consistent with each other.

I don't want to beat a dead horse, but the reason why I highlight the "placeholder" aspect of the argument is that, traditionally, compatibilist free will and incompatibilist free will are *very* different things. The refusal to acknowledge this fact (per PiV) creates a lot of the cross-talk and churning in the debate.


I agree with you, that we still have a fundamental issue when talking about free will; there are two very different meanings to the term and we’re constantly mixing them up during our discussions. During Al Mele’s post last month on Big Questions Online, he referred to the two different kinds as “modest free will” and “ambitious free will”. If we start being specific here on FoF regarding which type of FW we’re talking about, I believe we’ll make more progress since the cross-talk and churning would be reduced, and we could clearly understand one another. Perhaps we could call them MFW and AFW for short.

I think it’s fair to say that almost all of the contributors and readers here on FoF agree that MFW exists, and that it’s really AFW that we’re debating the existence of.

If there ever comes a day when science proves that “ambitious free will” exists, then we could drop the “ambitious” and the “modest” and simply go back to using the term “free will”, since the modest sense would effectively be covered by the ambitious sense, and the distinction would no longer be required.

Matt, Kip, and James:

I tend to agree with Matt, that we're all *trying* to think/talk about the same thing *free will* in this debate. I also agree with Kip and James that there is something of a "methodological pluralism" at work, which can be good in some contexts, but in practice often leads to confusion.

This reminds me of a keynote address I heard at one of my first professional philosophy conferences. I listened in shock as van Inwagen passionately chastized an audience of his elite peers for driving the debate off course with sloppy use of technical terminology, e.g. failing to articulate what *they* meant by "free will" when they used it. At the time, I thought "What nerve!", but I've come to appreciate his frustration.

About that screed of PvI's--I was there too. It was interesting. And we should all be careful about the use of terminology. I have for a long time sought to distinguish different kinds of freedom often conflated or not sufficiently distinguished.

Something I have been thinking about recent with respect to PvI's terminology. He defends "restrictivism", which is the doctrine that we are rarely free to do otherwise (but that our moral responsibility traces to those rare occasions). I don't accept his argument, but, ok, cool. But then he also defends in his book the Free Will thesis, and he points out that what he means by "Free Will" entails that when a person has Free Will, he *often* has a specific dual power (involving freedom to do otherwise).

Go figure! How does this fit together? Practice what you preach?

James - I agree with everything you wrote. You captured what I was trying to say well.

Kristin - I don't think I was at that keynote, but I've heard PvI make similar outbursts at other conferences. The irony, I think, is that PvI insists that there is just one free will, which is what everyone is talking about, without room for variation or subtle differences. If only he were here to defend himself.

John - your comment made me laugh!

One point I've made in a NDPR review (ironically) and in previous posts is that we ought to attribute the origin of "restrictivism" to CA Campbell's work in the 50s. He was the first to my knowledge to downplay the relative frequency of the use of libertarian dual-power choice, allowing that one reason determinism bacame a popular view is that, in fact, causality governs the overwhelming number of our actions. My sense is that restrictivism is the default view of libertarians, whether of van Inwagen's sort or Kane's reliance on SFAs, but I do wish Campbell would get more credit for advancing if not founding the contemporary libertarian strategy.


Thanks--I sometimes think we in philosophy don't have much of a sense of humor!

btw, I think that PvI's view is that we shouldn't proliferate "senses" of "Free Will"--otherwise, the debate wouldn't be conceptualized in such a way as to allow genuine disagreements (we'd all be talking past each other). This however leaves room for lots of differences, including different *kinds* of freedom.

I think however that there is an interesting tension (as above) between PvI's restrictivism, his definition of "Free Will" (or, if not definition, a claim about implications of "Free Will"), and his main argument in An Essay on Free Will to the effect that we have Free Will.

The comments to this entry are closed.

Books about Agency

3QD Prize 2014: Marcus Arvan