Blog Coordinator

« Some Varia about Free Will | Main | The Featured Author of the Month: Randolph Clarke »



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


When you ask “how luck could manifest in a physical sense”, perhaps here’s a way: Luck pertains to how different living systems within me and around me affect my path forward. Those living systems aren’t predeterministic in nature, and I don’t control them. Therefore, there’s luck involved regarding how they affect me.

BTW, thanks for the links, Marcus!

I've found that philosophers tend to use 'luck' pretty much as an antonym of 'control' (in the sense relevant to moral responsibility). A pool shot is lucky to the extent that you aren't in control of where the ball goes, and so on. The luck problem, at least informally (but mostly, I think, we are swayed more by the informal expressions of arguments than the formal ones), tends to amount to the question 'how does this enhance our control?'.

I've heard at least one physicist suggest that we have free will because this theory is correct:
I'm not sure if it was originally intended to provide for free will, rather than as an explanation of our consciousness, but this person thought that determinism rules out free will but that the indeterminism in the Orch-OR theory would be enough to supply it (I should also note that, in the original, I think the emphasis is more on the power of quantum computation rather than indeterminism per se). It seems to me that there is a sound luck argument at least against *this* position. If the *only* thing that you think prevents our choices from being unfree is their randomness, how is that any different from just taking an agent subject to determinism and introducing a mechanism that tosses a coin when they make a choice? How does the randomness alone add anything?

Of course, most libertarian philosophers are going to be more sophisticated than this, but at the very least luck arguments have a role in giving those philosophers the burden of showing their theories add to our control (add more than just randomness).

Josh, I just noticed I pretty much replicated your point on luck and control. Sorry! Either I accidentally skipped your comment before or I read it far too quickly and without paying attention. Let it be noted that we agree, anyway.

Kip, I can't help but reply to this: "I see the real, genuine, interesting debate to be the debate between skeptics and compatibilists. And that debate is essentially a debate about the definition of free will."

The reason is that I can't understand it at all. Say I have two exceptionally knowledgeable zoologists in a bar. They both know of all the species that currently exist, and all of those that existed but are extinct, and they know which are which. They start to argue about whether dinosaurs are extinct, or whether they continue to exist in the form of birds. Surely they will have learned nothing of any importance even if they somehow settle the question (maybe they run a bunch of surveys or something)? How is their debate real, genuine or interesting? (Perhaps they touch on some interesting questions as to what the natural kinds studied by zoologists are, but only incidentally.)

Do you really think that philosophers are in a similar position to my zoologists? If so, how can you find that interesting? Don't you think that, for instance, the supposed connections of free will to responsibility, desert, authenticity, and so on, provide exactly the sort of genuine stakes to the debate that don't exist in the case of the zoologists? (Perhaps you will tell me that surveys might show that free will is, by definition, completely unrelated to all of those things. But if that happened then the philosophical debate would really have been about something other than free will all along, because those stakes are much more important than anything else that could be tied up with the actual meaning of 'free will' as revealed by the surveys, and they are also what is really motivating the debate. It would be as if it turned out, after surveying the population, that 'bird' meant something with teeth - in that case, zoologists wouldn't really have been studying birds; they would have been studying what was really important, i.e. a particular zoological class.)

Further, can you really have a debate about definitions? Maybe in a bar, but not in a serious context. If it comes down to competing definitions, you're not really disagreeing any more. Insofar as you can give competing definitions A and B, you can just choose to study what satisfies A or what satisfies B - if anything does - and happily coexist with people studying the other; there's only a debate if there's some thing X you're both interested in and you're trying to show whether it satisfies A or B, but that's not a matter of definition any more.

(I should add that talking about the definition of 'free will' always seemed somehow infelicitous to me. It's not the sort of thing that gets a definition. Just look at other subfields of philosophy, like philosophy of mind. No-one talks about defining consciousness because it's not the kind of thing you define. You can distinguish types of consciousness, and give definitions for those, but only given an existing understanding of consciousness more broadly.)

There is a well developed account of luck from epistemology, which for some reason has had little effect on the free will debate despite the efforts of EJ Coffman and I. Here it is, in outline (and restricted to events for simplicity): an event is lucky for a subject if that event is chancy, significant, and beyond the agent's direct control. Significance and lack of direct control are reasonably clear (the direct control condition is only needed for abstruse reasons we can set aside for now). So let's focus on chanciness. An event is chancy if it occurs in the actual world but fails to occur in a significant enough proportion of nearby possible worlds, where a world is nearby if it differs from the actual world prior to the time of the event's occurrence in at most trivial aspects. It's easy to see this account gives us the right results. Getting hit by lightning is significant and beyond agents' direct control. It is also highly chancy: lighting wouldn't have struck A if she had sheltered under a different tree, or taken her walk just a second earlier, or if the tree nearby had been a fraction taller, or....Winning the lottery is significant, beyond B's direct control and highly chancy (had the balls rotated a second longer, or shorted - indeed, milliseconds longer or shorter- or had their surface properties been minutely different, or..... different numbers would have been drawn.

This account is neutral on the nature of the laws of the universe. And that's obviously right. If someone tells you that they know indeterminism is true because their friend had a stroke of luck, I hope you would not be convinced. Luck is a feature of the actual world, whatever the nature of the laws. And luck is a problem for everyone. There is no Luck Problem, if by that is meant a special problem racing libertarians alone. Libertarians do not face the luck problem alone, nor is there any good reason to think that necessarily they face it in a form that is harder to deal with (that said, many libertarian theories actually have features that make luck especially intractable. However, this is not a necessary feature of a libertarian view).

EJ and I will keep trying. I don't know what explains the free will debate's resistance to even acknowledging the existence of this account. Everyone thinks luck is important, but refers to their intuitions when it comes to deciding when a case involves luck (and if so, how). It's a good decade now since the account was fleshed out clearly enough to epistemically require people who write about luck to adopt, or criticise and replace, this account. But it doesn't look like happening.

Ah, luck as randomness makes sense to me. Thanks Josh, Marcus, and CJ.

I would agree with those that argue that free will is not rescued by randomness.

Just to speak to the red herring comment by Kip... Wegner's main contribution on this issue, I think, was to show that the feeling of will is extremely powerful and compels us to believe in personal freedom. Indeed, I think it is this feeling that fuels so much emotion (and blog posts) about the issue. We can argue that subjective experience is not *actually* important for the debate but I think it shapes our views even when we're trying to be careful and set it aside and I would guess it is used as primary evidence in the lay view.

And thus, thanks to Josh, CJ, and Marcus (all of whom reflect the actual usage in the debate perfectly well) ideas which obfuscate more than they enlighten get propagated a bit further. I give up.

Hi Neil,

Thanks for weighing in! As I said, I'm no luck expert, so was trying to give a sense of how others seem to understand it. But this Coffman/Levy account is interesting, and it connects up with and perhaps diverges from some things I've said about control (mainly in a paper just out online in Phil Studies). Will have to think more about this.

Neil, I wish you the best of luck with trying to promote discussion about the nature of luck in philosophy of action. But, while I haven't formed an opinion one way or the other, I can see why the account of luck relevant to epistemology might not be so appealing to someone working on e.g. moral responsibility.

Take the chanciness condition, for instance. Someone might say, "Even if you have libertarian free will, you're lucky that you've been raised in an environment that encourages the development of moral virtues and so on, so your responsibility for your good deeds is somewhat diminished compared to a hypothetical being whose behaviour was not influenced by its upbringing." I think such a person would genuinely be talking about (or trying to talk about) luck, but not in a chancy sense; the worlds in which I am raised very differently are quite distant because they change so much about me (especially compared to the worlds relevant to assessing responsibility for someone with libertarian free will, which are the very close worlds where everything is the same up to some undetermined choice on my part).

Now, it might be that a revisionist account of luck that rules out this kind of usage is appropriate. It might be that our intuitions about luck rest on a comparison to genuinely chancy events. In my comment above I compared something to flipping a coin, because I was trying to give a sense of how the argument goes and that's a move people genuinely make. There, I was relying on the chanciness of the event to indict it as lucky. Perhaps the only basis anyone ever has for the intuition that an event is lucky is that they are, explicitly or implicitly, correctly or incorrectly, thinking of it as a chancy event. If that were the case, I think that chanciness would have been shown to be essential to luck even in the context of action/responsibility. But that would require a lot of showing.

Sorry --missed James' response. James, thanks, that sounds reasonable, but to dig deeper I'm not sure I understand what you mean by predeterministic (deterministic ok?), "me," "living systems within me," or "I" in your response, at least not from a biological perspective.

Neil --I thought I had a handle on luck but apparently do not. Why is the concept so opaque?

Yeah, CJ. Might turn out that the account doesn't work so well in the context of the free will debate. Someone shoukd write a book, seeing whether it actually illuminates these particular set of issues, and in particular the one you mention.

Oh wait. I seem to recall that someone already did.

Neil, I apologise if my response came across as ignorant or dismissive. I was trying to understand how and why philosophers of action think about luck differently than epistemologists, not to pose a challenge. (Though it did strike me in the course of doing so that to say that philosophers of action are thinking about luck defectively rather than about a different type of luck, and should revise their usage accordingly, might actually require an interesting investigation into where intuitions come from.)

Also, books are, sadly, expensive. I remember reading chapters of the one in question online, but can't find them now. Are they still available?

It's not terribly opaque, Thalia. The possible worlds apparatus is a bit forbidding, but its trying to capture something quite intuitive. An event is lucky if it matters (for someone) and it could easily have failed to happen, where it could easily have failed to happen if it would not have happened if things had been different in some tiny respect.


I think I can state my position on luck more clearly.

If I place one 5-ounce weight on a scale, and then I place a second 5-ounce weight on that same scale, we can probably both agree that if the scale reads 10 ounces, there is no luck involved. Here’s a more formal way of stating that principle: When forces A1 and A2 add together in force field A and form a net force A3, there is no luck involved. The interaction is predeterministic in nature, and force A3 is simply a direct sum of A1 and A2.

With that in mind, in order for something “lucky” to occur in material reality, the event must be controlled by forces in addition to the four fundamental forces of physics (4FFOP). In other words, new emergent forces (i.e., those associated with life) need to be part of the equation (i.e., add into the mix and affect the path forward).

If a person believes in predeterminism, and they also believe that an event is attributable to luck, then they’re simply believing in luck in the “weak sense”, not luck in the “strong sense”, just as there is a weak sense (i.e., the ability to make a choice) and a strong sense (i.e., the ability to make a choice that isn’t controlled solely by the 4FFOP) to the term “free will”.

The reason I consistently make a distinction between predeterminism and determinism, is because determinism allows for new emergent forces to be part of what determines the path forward, whereas predeterminism advocates that only the 4FFOP exist in reality and those four forces are solely responsible for controlling the path forward.

Email me, CJ: I will see what I can so about getting you the book.


This is a great question. I think Gregg's distinction between external and internal threats to compatibilism is exactly right, as is his point that if cognitive science poses a threat, it must be an internal one.

Unlike (maybe?) Gregg, I am not really worried about unconsciousness or automaticity: I have no problem with blaming someone for an implicit expression of some bad attitude (even if they were in no position to know the attitude was bad). And I think skilled activity constitutes responsible agency par excellence, unreflective though it often is. Nor is this any big deal: Frankfurt is clear that our higher-order volitions don't need to be transparent to ground responsible agency, and lots of people have discussed how unreflective activity can be reason-responsive. (Railton's recent work is worth noting in the context because of how empirically informed it is.)

Still, I'm not sure whether philosophers not directly engaged with empirical moral psychology have considered the threat from situationism enough. Holding someone responsible for a response often presupposes the response to be _attributable_ to them, as embodying in some important sense the agent's identity or character. But I think it's an open question--which I'd think depends on what, exactly, is supposed to be significant about attributability--what sense that is. Depending on the answer, regarding a response as attributable to an agent may or may not involve unrealistic psychological assumptions.

For instance, Hume (or at least the version of Hume I seem to remember) thought that the significance of attributability was mainly evidential: if someone did something blameworthy, this was evidence that they were more disposed than (most?) other people to do blameworthy things in general. If this is right, then I take it that it is pretty clear from experimental psychology that we are not responsible in the sense that our blaming practices presuppose. (Not that this is anything that tons of astute pre-20th-century observers of human nature didn't already know!)

On the other hand, if the significance of attributability is not mainly evidential, it is a good question why we seem to care about it so much. This strikes me as an issue that might still be under-discussed.


I wonder if you haven't answered your own question about the zoologists, only to immediately dismiss the answer. "Perhaps they touch on some interesting questions as to what the natural kinds studied by zoologists are, but only incidentally." Maybe it's more than incidental. Suppose that the right answer is yes, birds are dinosaurs, and the reason why is that cladistic classification is clearly superior to morphological approaches, in satisfying the goals of zoology. And (suppose) in the case of birds, cladistics groups them with dinosaurs. That's pretty interesting and important.

The more generic, quasi-Quinean point I'd urge here is caution about dismissing "merely verbal" disputes.

The connections of free will to responsibility, authenticity and so on could play a role analogous to the goals of zoological classification. Given that responsibility and authenticity concerns influence our usage of "free will", I don't think it's implausible that surveys could tell a lot (though probably not all!) about how this constellation of concerns fits together and relates to causality, reason-giving, etc.

The comments to this entry are closed.

Books about Agency

3QD Prize 2014: Marcus Arvan