Is libertarian free will a necessary assumption for any decision theory because “one has to suppose that one has a choice to make”? It seems to me that decision theories needn't rely on the formulation that agent1 can x or ~x regardless of preceding states-of-affairs, but that it can equally well rely on agent1 x'ing because state-of-affairs1 determines that agent1 x's or agent1 ~x's because state-of-affairs2 determines that agent1 ~x's. The point is not whether the agent can make that decision with exactly the same preceding state-of-affairs, but whether the agent could make both decisions, however the decisions are brought about. My reason for this position is that if libertarian free will is a necessary assumption for any decision theory is correct, determinists should not make normative suggestions ever. One would suggest this perhaps, because they'd claim determinists cannot ever make normative suggestions coherently. I believe this is wrong because normative suggestions a determinist makes...

Like you, I don't see why rational decision theory must assume that agents have libertarian freedom. Indeed, if memory serves (it's been a while), Richmond Campbell argues that fundamental principles of rational decision theory rule out libertarian freedom on the part of the agent: 'Moral Justification and Freedom', Journal of Philosophy 85 (1988): 192-213. I think you're right to emphasize that if there's a deterministic causal chain leading to an agent's choice, that chain can, and typically will, include the agent's rational deliberation about which choice to make.

It has been said that if there is human freedom, then we are responsible for our actions. By this, it seems natural to suppose that "given that there is no human freedom (let's just suppose for the sake of argument) then it would follow that we are not responsible for our actions." But this seems an instance of what is called the "fallacy of denying the antecedent". Is this really an instance of the fallacy or is it an exemption to the case because personally I don't see any error in the form of the argument.

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...

I'm interested in the issue of whether people would have moral responsibility under determinism. So if a person in a deterministic universe would happen to commit murder, some people would say that they are morally responsible for the action, and others would disagree. When I speak of "moral responsibility" here I'm thinking along the lines of whether the person would deserve blame and retributive punishment. (If it actually happened that we lived in a deterministic universe, I assume that we would have to hold people morally responsible in some sense for practical reasons. We would have to punish to protect society and to deter future crime; but some might give up on the idea of retributive punishment and see criminals rather as unfortunate victims of the blind process of nature.) I'm not expecting a solution to the question, "Would people be morally responsible under determinism?". Rather I'm going to ask: could the issue be a conflict of opposing moral principles that may just be forever unsolvable...

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...

More of an observation than a question, about "compatibilism" in the free will "versus" determinism debate. In the short run, there is a strong correlation between life expectancy tables and the number of people who die in a calendar year. Somehow, even though on the level of the individual, many of these events may be due to "luck" (wow, that train just missed me; or wow, what a freak combination of factors to lead to such a bizarre accident); on the level of the population, the total number of deaths in a year can be "predicted" fairly well even if no individual death can be predicted. In the long run, life expectancy tables do change over time: collectively, each individual person uses what they learn about diet, exercise, cigarette smoking, etc. and makes adjustments in their day-to-day lives; and the aggregate results over time do reflect these changes. It seems to me that there is a good question buried here in this analogy but I can't quite figure out how to unearth it. Any thoughts from a...

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...

Hi, what an awesome website! I have another free will related question to add to the heap! I saw an interview with Steven Weinberg, the Nobel Prize winning physicist and, I don't remember the precise phrasing, but he said something like 'I know of no law saying that nature is here to make physicists happy.' He wasn't referring to free will, but it got me thinking about something... From what I've read and heard in papers and talks (which is certainly not nearly exhaustive), it seems that their is a tendency for those who chime in on the free will issue (even professional philosophers) to approach it from the perspective that the challenge is to show that free will does not or cannot exist. What I mean is that there seems to be a tacit presumption that since we "feel" free, the burden of proof is on those who contend we are in reality not free. I understand this perspective (and it is not unique to the free will debate), but it seems to presuppose some kind of rule that says that our feelings...

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...

It has long been recognized that free will appears to be incompatible with the causality observed in the rest of the universe. There is now evidence from neuroscience that free will does not really exist. Does the fact that many people - including many philosophers - find this conclusion absolutely unacceptable constitute a manifestation of the limits of the human mind's abilities?

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.

Recently I read a newly published very short book criticisng the concept of Free Will. I thought the book made some good points and some not-so- good points, but what really disturbed me is that the author didn't ever carefully define what he meant by Free Will. Is the definition of Free Will so obvious and clear that there is no need to define it in a book intended for lay readers?

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

I have a question about determinism, prediction and conscious choice. Suppose we live in a deterministic universe such that some epistemically-juiced Demon could predict future events with absolute certainty long in advance. When he sits observing, he's always right about what people are going to do. But, suppose, the Demon gets a little bored decides to try to impress some humans with his gift of prophecy. He tells me that he can predict any of my actions: for example, what I'm about to eat for lunch. He gives me an envelope and tells me to open it after I've made my lunch. I do and he's right about the sandwich I was just about to bite into. But at that point can't I just as well change my mind and eat something else? And isn't that true no matter what prediction is made, provided I'm aware of it sufficiently in advance of its "coming true"? Of course, the Demon could have made auxiliary predictions about how his telling me would affect my choice. And those could be true. But if I'm privy to...

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,...

Hierarchical compatiblism says that I have free will if I have the will I want to have. The theory claims to show that my desires can be up to me. I understand how the theory improves upon classic compatiblism by showing that the absence of external constraint is not sufficient for freedom. 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. Can any form of compatiblism, however sophisticated, survive the scrutiny of hard determinism?

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...

Pages