Translating the argument into symbolic terms quite literally, we get this: 'If F, then R. Not F. Therefore, not R.' That form of argument does indeed commit the fallacy of denying the antecedent: the premises don't logically imply the conclusion; the truth of the premises doesn't logically ensure the truth of the conclusion. The first premise says that F is sufficient for R; it doesn't say that F is necessary for R. In that case, R can obtain even if F fails to obtain. My hunch is that you're interpreting 'If F, then R' as 'R if and only if F': you're interpreting the conditional as a biconditional , i.e., as the claim that F is both necessary and sufficient for R. 'R if and only if F' and 'Not F' together imply 'Not R'. Your interpretation is understandable, because conversationally we often do intend to assert a biconditional when we use conditional language. A parent's 'If you clean your room, you can watch TV' usually means 'You can watch TV if and only if you clean your room...

You asked: "Could the issue be a conflict of opposing moral principles that may just be forever unsolvable by rational argument? So maybe you just can't produce arguments that can 'bridge the gap' between the two sides, i.e., the arguments just don't exist that would have the rational force and traction against the other side." I don't see it as a conflict of opposing moral principles. I think each side sees itself as trying to work out the implications of our shared concept of moral responsibility . One side thinks that our shared concept requires indeterminism; the other side thinks it doesn't. Or maybe our shared concept is inconsistent in both requiring and not requiring indeterminism, or we have two distinct concepts of moral responsibility, but even that I wouldn't classify as a conflict of moral principles. In any case, I'm not pessimistic about the possibility of making progress in this debate. Indeed, I think we've made progress in the last several years and will continue to. The new field of...

Your actuarial analogy seems similar to a point that quantum physicists often make. As I understand it, orthodox quantum physics says that the precise behavior of an individual particle is impossible to predict even in principle. Yet once you get zillions of such particles together, their collective behavior -- the behavior of the entire system -- becomes highly predictable. We can't predict the behavior of a particular atom belonging to the baseball, but we can predict with near-perfect certainty that the whole baseball won't suddenly tunnel through the earth and emerge on the other side. So individual unpredictability is compatible with collective predictability, both in the classical case you described and in the quantum case where the individual unpredictability isn't just a function of our ignorance. The flip side of this idea would be deterministic chaos, where individually predictable parts combine to make an unpredictable system -- unpredictable not because the system isn't still deterministic...

As a panelist, I'm glad you like the website. Spread the word! I don't think we need to invoke feelings in order to assign the burden of proof to those who challenge the existence of free will. In the literal sense, feelings have nothing to do with it: I don't think we have reason to believe, for instance, that causally undetermined actions have a characteristic "feel" to them that causally predetermined actions lack. But neither do we have reason to believe that they don't have a characteristic feel to them. So you're right that it's only fair to leave the phenomenology -- the literal feel of our actions -- out of it. There's also the metaphorical use of "feel," as in "I just feel that human beings sometimes act freely." Feelings in that sense too are irrelevant, I think. But the reason that those who challenge the existence of free will bear the burden of proof is that pre-philosophically -- i.e., before examining the issue philosophically -- we start with a widespread set of...

My diagnosis is different: I think the arguments claiming that causality (specifically, deterministic causation) rules out free will, or claiming that neuroscientific results cast doubt on the existence of free will, are bad arguments. This issue came up recently here: see the discussion at Question 4792 , particularly the reply by Professor Nahmias and the book review he links to.

You're right to be disturbed if the author never defines 'free will' in a book criticizing the concept of free will! One frequent obstacle to progress in discussing free will is that different parties to the discussion rely on different (and often unstated) conceptions of free will. Some think of free will as whatever is necessary and sufficient on the part of an agent to make the agent morally responsible for his/her action. Some think of it as requiring the ability to have acted otherwise than one in fact acted. Some think of it as requiring the ability to have acted otherwise than one in fact acted even under the same circumstances that preceded one's action. Those conceptions of free will aren't the same -- or aren't obviously the same. It's a good question which of those conceptions, or which other conception, best fits the way we use the concept of free will (assuming we use the concept in a stable way). It's another good question how well those conceptions stand up to scrutiny. In answering...

Eddy: Nice article! I'm glad they tapped you to review Harris's book. (I too suspect that it's the book the questioner is referring to.) --Steve

Could you or the Demon even understand what he tells you? The Demon tells you (a) that he has predicted what you'll eat for lunch. Of course, now that he's told you that, what he's really told you is (b) that he has predicted what you'll eat for lunch given that he's told you (a). Notice that (b) isn't the same item of information as (a). But wait. If he's told you (b), then really what he's told you is (c) that he has predicted what you'll eat for lunch given that he's told you (b). Notice that (c) isn't the same item of information as (b) or (a). And so the regress continues, forever. Therefore, I wonder if anyone can understand the story you've sketched well enough to see what the story implies or what's consistent with it. In this article , my co-author and I raised a similar worry in regard to Newcomb's problem, a famous problem in rational decision theory that also involves a predictor. But I'm not sure anyone else was convinced by our argument! In any case,...

I agree, and you'll find an ally in Hume: see this SEP entry . On the other side, philosopher Robert Kane invokes quantum indeterminism in his defense of libertarian free will; see his contributions to this volume .

You wrote, "But it is unclear to me how second order desires or volitions are genuinely up to me if they are causally necessitated by the relevant laws of nature and background conditions." Recall that, for compatibilists, how I act can in the relevant sense be up to me even if how I act is necessitated by the laws of nature and the prior conditions. If you grant compatibilists that much, then they're likely to say, "Why can't my second-order desires or volitions also be up to me, in the relevant sense, even if they're necessitated by the laws of nature and the prior conditions? Causal necessitation doesn't prevent those from being up to me any more than it prevents my actions from being up to me." Which invites the question you closed with: "Can any form of compatibilism, however sophisticated, survive the scrutiny of hard determinism?" The jury's of course still out on that one. But the more I think about the relation between freedom and determinism, the more it seems to me that...

I'd flag the word "ultimately" in your sentence "But this does seem to give weight to the notion that even conscious deliberation is not ultimately free." The search for "ultimate freedom," like the search for " ultimate purpose ," is doomed to fail, but only because the search itself is incoherent and hence ill-conceived. As you point out, ultimate freedom would require completing an infinite regress there's no reason to think we could complete. In that case, there's good reason to doubt that "ultimacy" is essential to the concept of freedom we ordinarily use and view as important especially in moral contexts. You summed it up nicely: "So you're free, but you don't have the impossibly infinite consciousnesses necessary to be ultimately free, right?" Right. Or at least the impossibility of ultimate freedom doesn't cast doubt on the possibility of freedom, just as the impossibility of an ultimate prime number doesn't cast doubt on the possibility of prime numbers. Let me recommend (again) ...