For many people, the idea of free will is bound up with the notion of “could have done otherwise”. By their lights, if only one future is possible for a person — that is, if the person cannot do otherwise — then free will is an illusion.
Philosopher Christian List — author of the recent book Why Free Will is Real — proposes an interesting species of free will based on the claim that while physics may be deterministic, behaviors at the agent level are not. Agents can do otherwise, according to List, and this is enough to ground free will even if physics is deterministic.
I think List is mistaken, but I’ll save my criticisms for the comment thread.
Readers can find List’s argument in this paper:
Free Will, Determinism, and the Possibility of Doing Otherwise
See you in the comment thread.
In case you or lurkers think my original answer is too coy:
Assessing a deduction requires agreement on the meaning of terms.
So we need to understand what metaphysical connotations List takes from supervening and the word ‘determines’ in the text you quote. For example, saying X MUST happen due to supervening relation to Y assumes a metaphysical connection of necessity that he might deny.
But I don’t know what he thinks and I don’t have any more guesses beyond my first one that he won’t accept a definition that provides an easy counter to his arguments eg because it immediately disqualifies ontological NRP.
IMHO, given the complexity of supervenience as per the SEP overview, there’d need to be an email exchange with him to resolve the metaphysical connotations of supervening to his argument. Further, he does have other related papers on this issue, eg on NRP itself, on the Consequence Argument, and on Kim’s causal exclusion argument against NRP for mental (ie psychological) properties. So he might choose to refer to those as well.
Do you think you just got asked if you accept the bride’s hand in marriage? It would not totally surprise me given the level of non sequiturs you emit.
petrushka,
Yes. Notions of moral responsibility are intimately connected to views on free will. Many people feel uncomfortable about punishing someone for something they couldn’t not do. If determinism holds at the agent level, this is exactly what happens.
Bruce:
List:
Bruce,
Given List’s statement above, I can’t see any plausible interpretation of “determined by” that invalidates my counterargument. Can you think of one?
Last night Bruce linked to a novel argument for free will from philosopher Michael Huemer.
The gist of the argument is conveyed by these two quotes:
And:
I think the argument fails, but I’ll refrain from giving my reasons until readers have had some time to ponder it.
This is discussed in SEP; as I said I have not studied that in detail and so will not try to rehearse here.
In any event, the point of my reference to Graziano in the original coy response was that whether or not I could think of one was irrelevant. What matters is what he thinks.
Personally, I find it hard to believe that over the years he has submitted refereed papers on this argument, and considered feedback to them, that there would be a easy counter argument by deduction that everyone missed. But I guess that is just an argument from my personal incredulity (but incredulity I think is quite justified!)
You may want to listen to the podcast where Shermer tries to steelman the Coyne and esp Harris arguments for List to respond to. But I have to say it sticks to a fairly basic level and I cannot remember it getting beyond the detail in the paper. Possibly even less detail.
I will now stop posting on this subtopic.
Re the Huemer argument, I’ll just drop a hint: My main objection is more direct and simpler than the four objections he addresses in the paper.
Bruce,
We’re not just trying to figure out what List thinks; we’re also trying to figure out whether his argument is sound. I’ve presented a counterargument derived from List’s own statements. The question is whether anyone at all — not just List — can find a flaw in my counterargument. I’ve tried, and I can’t find one. Can you?
You’ve suggested that List might mean something different by “determines” than I do. If you can think of any alternate meanings that would invalidate my counterargument, I’d be interested in hearing them.
Bruce,
Even smart people can overlook things at times.
Anyway, I’ll happily consider any proposed rebuttals of my counterargument, regardless of their source. Even if it’s Nonlin. 🙂
Okay, back to the Huemer argument.
A major flaw in the argument becomes apparent in step 7:
The problem is that “MFT is true” is equivalent to “determinism is false”.
Substituting, we get
…which is obviously nonsensical. Huemer is “proving” a contradiction.
So why is “MFT is true” equivalent to “determinism is false”? Consider the definition of MFT:
Under determinism, a person can only ever take one course of action. The only way the MFT can be true, then, is if determinism is false.
Huemer confirms as much when he writes:
The argument undermines itself by “proving” a contradiction.
Right, but isn’t that his point? From top of page 3:
“Thus, we see that determinism is self-refuting, in the sense that, modulo certain true premises, determinism implies its own contradictory (MFT). Any proposition that thus implies its own contradictory is false, so determinism is false, and MFT true.”
(Posting question: I dislike long stretches of italics so I avoid blockquote tag. Is there a way to indicate a quote without italics?)
ETA: Have a look at the people cited in footnote 4 to Heumer for a coincidence.
When considering rebuttals, do you mean only those posted to the blog or do you also include arguments in the philosophical literature that List directly or indirectly cites?
For me, it would be the latter. Based on our past exchanges, I think it is the former for you. That thought was another thing my coy reply was alluding to.
ETA: So for me, to formulate a full reply to your argument implies significant philosophical research. If that research does not interest me, then I do not have anything more to say at TSZ.
.
The argument reminded me of the ontological argument: there must be something wrong with it but it is not obvious where exactly.
It is strange that the argument turns on at least one person honestly believing MFT But that is not a fatal flaw; it just means you’d need a different argument for MFT if no one honestly believed it.
I also have not checked the validity, which he formalizes in footnote 3. It is suspicious to me that there is not a more explicit appeal to modal logic, since an argument involving ‘can’ and possibility would seem to need it. But I’ll take his word for the validity.
However, I do think there is merit in objection 3: that shoulds in epistemology are about justification, not truth. But that epistemology stuff is more of interest to Walt (and you?) so I have not pursued that source of suspicion.
Speaking of Nonlin, I noticed from other threads that you are dealing with interpretations of QM, determinism, and free will.
If you want more reading on that, Kane does try to make a naturalistic, libertarian argument for free will using a quantum event. Of course, he is aware of the usual objections to free will that involves indeterminism and so tries very hard to deal with them.
Dennett spends a whole chapter in Freedom Evolves critiquing the argument.
John H and I had an inconclusive exchange on Kane over at PS in the same thread that someone posted the Heumer stuff.
https://www.ucl.ac.uk/~uctytho/dfwVariousKane.html
Ted Chiang’s Exhalation collection contains the story “Anxiety is the Dizziness of Freedom” which turns on moral responsibility and free will in a many worlds universe. That collection also has “Omphalos” which relates to TSZ topics: how to get meaning from life in a world where science strongly supports recent supernatural creation after astronomy turns up something unexpected.
Of course. But consider the following hypothetical situation:
One very smart person SP1 who specializes in a field repeatedly submits his or her hypothesis for formal review to other very smart specialists in the field No knock down objections to SP1’s hypothesis emerge from the process (since he or she continues to publish papers on the hypothesis and be cited on it in peer-reviewed journals).
Another very smart person SP2 who is not a specialist in the field reads one paper on the hypothesis and a day or two later claims to have a knock down argument for its hypothesis.
What should a rational third party who is not an expert in the field believe about the status of SP1’s hypothesis? In particular, who is most likely to have done the overlooking?
BruceS,
Yes, It is really hard to avoid hubris when one thinks one has come up with a good argument for or against something. As you say, peer review is useful for this very reason. And there is a (hard to describe) difference between something that dawns on one when reading an internet piece and something one has worked out over a period of months while reading a ton of literature (pro and con) on a specific subject.
Of course, it remains true that anybody can be wrong at any time and anybody can come up with a brilliant new theory or argument or refutation. But, especially wrt issues (like libertarian free will) that have been hashed out by a lot of very smart people for for centuries, a lay person’s “proof” is not likely to be worth much, however they may feel about it. It’s not that somebody else must be “smarter” than them; it is, as you say, largely a matter of division of labor.
As far as I know, this works:
<blockquote><em>Text to be quoted</em></blockquote>
I’m testing that in this post, with:
If you are interested and have time, I’d welcome your thoughts on Huemer’s argument, since it uses ideas in one of your fields, epistemology, eg what should we believe?
http://www.owl232.net/papers/fwill.htm
OK, but that is even uglier to me!. I am hoping there is an HTML tag or combination that works in wordpress that just does block indenting without changing the font.
It sounds interesting, and I’ll download it along with the List paper, but I’m struggling to get CHOICE and the chapters that surround it into some kind of shape. I will mention again that keiths’ comments have been helpful and they–along with simply rereading my always sloppy, slipshod writing– are causing me to have to make a number of revisions–some difficult. So I don’t know when….
As far as I know, “blockquote” is supposed to do that. But there is such a thing as a theme, which changes how the final document looks. And then there is the matter of how the browser handles it.
Some artist somewhere has decided that the theme for this blog requires that quoted blocks appear in italics. It doesn’t make sense to me, but then I was never any good at art.
Ugly!? That’s nicer! I wish the whole site used that font!
Bruce:
I just checked, and WordPress (or the comment plugin) strips out any instances of the <font> tag, so that option isn’t open.
The only remaining option I’m aware of is to change the CSS for blockquotes via the font-family and font-style attributes. An admin would have to make that change, and it would apply to everyone’s blockquotes (including in OPs).
Bruce,
Any rebuttals from any source. Are you aware of anyone who has made the same deductive counterargument that I’ve presented here, to whom List has responded?
I’m not sure why, since my counterargument depends on nothing but List’s paper for its premises.
That’s fine. No one is obligated to respond to anything at TSZ. If you’re not aware of a flaw in my argument, and you’re not interested in looking for one, we can leave it at that.
Bruce,
It’s always possible that I’ve overlooked something. That’s why I’m soliciting rebuttals!
Bruce,
Yes, I ‘ve read Freedom Evolves and found myself agreeing with Dennett’s criticism.
Thanks for the Ted Chiang recommendation.
Let me explain why I disagree with that claim:
Supervenience is only a logical relation; it can be formalized completely as a mapping between microstates (base) and macrostates (states supervening on the base). State changes at the microstate level can then be formalized as mappings between the powerset of the microstates and the powerset of the macrostates. (See section 3.2 of the Butterfield paper List cites). All that List’s definition requires is that logical mapping.
So, as List’s diagram illustrates (and Butterfield formalizes), it is conceptually possible for microstate dynamics to be deterministic while macrostate dynamics is indeterministic. Let me emphasize that is only a logical claim about the mathematical mappings between the respective state spaces; that is what I mean by conceptual.
What List claims and what you argue against concerns metaphysical, not conceptual, possibility. The issue is the relation between metaphysical possibility at the microstate level and metaphysical possibility at the macrostate level. I read you to say that the logical relations defined by supervenience imply something about metaphysical possibility. Namely, no alternative metaphysical possibility at the microstate level entails no alternative metaphysical possibility at the macrostate level. List claims otherwise, based on ONA and NRP.
Because the SEP article on supervenience indicates that the metaphysical implications of supervenience are still an open issue, I claim your argument is not a knock down argument against his claim. More is needed on why supervenience must have metaphysical implications for List’s argument. I think that to be complete that argument against List should transparently consider the important arguments that claim to separate the conceptual and metaphysical possibilities of supervenience in the philosophical literature.
ETA: Link to Butterfield paper:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3262312/pdf/rsfs20110052.pdf
Bruce,
Right. Here’s how I would put it:
If every microstate corresponds to exactly one macrostate, as List confirms, then every fixed microstate sequence corresponds to a fixed macrostate sequence. There are no other possibilities. It doesn’t depend on the broader metaphysical implications of supervenience; it just depends on the fact that the mapping from microstates to macrostates is surjective (aka “onto”).
The indeterminism that List identifies at the macrostate level is just epistemic indeterminism. That is, if all you know is the current macrostate, then you don’t (necessarily) know the next macrostate. But that’s just due to your lack of knowledge about the microstate. In reality, the system is progressing through a fixed sequence of both microstates and macrostates.
keiths:
Bruce:
Ah! I missed that part.
So given that it’s a proof by contradiction, its soundness depends on the truth of the other premises and the validity of the logic.
Okay, then let me fall back to my next objection, which concerns premise 2:
I think that’s incorrect, because “should” in this case does not imply metaphysical possibility. “Alejandro should do X” is really just shorthand for “it would be better if Alejandro chose to do X”, and that is true even if it’s not metaphysically possible for him to do so.
Perhaps I should be saying “physical” instead of “metaphysical” in the comment above. In another possible world, it might indeed be possible for Alejandro to do X, and I suppose that counts as metaphysical possibility.
Another issue:
Notice how Huemer phrases Premise 1:
And notice how he phrases steps 5 and 6:
The qualifier “with respect to the free will issue” is not needed. Huemer might as well say
Any falsehoods at all. That’s a pretty extreme conclusion, and it’s easy to see that it’s false. There’s nothing about determinism that precludes false beliefs.
Nor anything that precludes true beliefs. Or anything else.Determinism has no entailments that are violated by observation.
I can’t see how the same isn’t true of free will.
petrushka,
Free will might have some observable entailments, depending on the exact definition.
I remember some interesting experiments along those lines. I’ll see if I can track them down and post links.
Many people intuitively believe that our experience of free will is an accurate reflection of what’s going on inside our heads. That is, they think that when we perceive ourselves as freely willing an action, we really are willing it, and when we don’t perceive ourselves as willing an action, we really aren’t willing it.
A couple of interesting experiments show a double dissociation: that we can perceive ourselves as willing an action we aren’t responsible for, and as not willing an action that we are responsible for.
In the first experiment (Wegner and Wheatley, 1999) two people — one a subject, the other a confederate — jointly moved a computer mouse in slow circles via a square board mounted on top of it. (Reminiscent of a Ouija board setup.)
The experimental design was complex, but the upshot was that people could be manipulated into thinking they had stopped the motion of the mouse when in fact it was the confederate who had done so. They had the experience of freely willing the mouse to stop, but they weren’t actually stopping it.
The second experiment (Schlegel et al, 2013) involved fist clenching. Via the clever use of hypnosis, subjects could be convinced that the experimenters had caused them to clench their fists (via a set of electrodes attached to their forearms) when the subjects had actually initiated the clenching themselves.
This gets to the heart of the matter for me.
My phrasing of your point: macrostates just reflect our human perceptual limitations. Once we use science to overcome these limitations, we understand that the microstates of our best fundamental physics are where the action is. They control what happens next in the world.
But then: Why should any level control what happens next? In fact, fundamental physics is time symmetric, including the Schrodinger equation of QM (ETA: Need unitary stuff too).
There is no notion of one instant controlling the next in the equations themselves. In the equations on their own, we only find larger t values and smaller t values. But the layout of the the time axis simply reflects our conventions.
The notion of next comes from our human perceptions of time. But that brings us back to our epistemic limitations as reflected in 2LT: the passage of time is merely an aspect of our explanations, presented at the macrostate level, reflecting our limited knowledge. We are back to the issue we started with.
In the equations, there is no force of nature saying this must turn into that. There is only pattern and regularity. Fundamental physics implies the block view of spacetime.
That gets us into the metaphysical status of laws of nature. Are they simply human descriptions of patterns we find in nature? Or is there some metaphysical, non-scientific force built into laws that compels the world somehow to act according to the equations we create. Do laws reflect some “mystical” necessity in reality which is not captured in our physics?
I think Ismael does a much better job of confronting this issue and how it relates to free will than does List. If you want to spend some time with her book or her Youtube videos, I am happy to discuss. Not easy going, at least not for me. For videos, I recommend this one and the Q&A.
There are also some interviews on eg Closer to the Truth which give popularizations.
I am going to take a break from posting to try to catch up on my sleep. I’ll respond later if you want to talk about Ismael.
All the points you make here and other post bother me too. There is some equivocation or definitional issue, or some hidden assumption, that the argument turns on. But, like the ontological argument, it’s hard to see exactly where it is, at least for me. Possibly it is some second-order interaction?
I never had the patience to do logic/math puzzles or to play chess well. (OTOH, I enjoyed programming because of the quick feedback on whether I was right or wrong).
My limited patience also applies to philosophical arguments like this one. There don’t seem to be any big ideas, just some tricky logical error. So I am not going to pursue it.
Heumler’s web page on his philosophical preferences indicates that he non-physicalist, rationalist, libertarian about free will who leans atheist. That aligns with his use of arguments like the above.
https://philpeople.org/profiles/michael-huemer/views
BTW, I think EricMH is a rationalist too. Specifically, he thinks armchair, a priori mathematical reasoning trumps empirical science. I posted that opinion in a Joe F thread which was about challenging Eric to apply his ideas using KC and Levins result so they applied to population genetics. Eric did not disagree with it when he pointed to my post as a reasonable summary of his ideas.
Eric did do a thread with an effort at presenting his ideas in a biological framework, but has not followed up to address issues brought up in various responses. But, as you can see, he loves discussing the math.
Here is something he might bring up if he has read some modern Aristotelian causation on free will:
Ismael spends the first part of her book describing what cognitive science tells us about the nature of the self. That description is an important part of her arguments. IIRC it is not a big part of the videos I linked, although I think there are other YT videos where she describes those ideas, because she had an earlier book devoted to the topic.
I read the Huemer (it’s shorter than the List), and I’m with those who think the first premise begs the question. His definitions/explanations of question-begging are bad. I don’t have the energy/interest to do a better job at it, and it’s probably not easy to do right. But those three attempts obviously suck. And the first premise seems to me the clear culprit.
I think it’s in there because he’s worried about situations in which there are moral considerations compelling false beliefs. Doesn’t do any harm, I don’t think. The problem is the first premise and his crap definitions of question-begging.
I’ve now read the first few pages of the List. On page 7, I don’t think 3* is either “equivalent to 3” or whatever he means by a “substitute for 3*”. I think it’s better than 3. That is, it captures what we mean by having free will better than 3 does. I stopped there, and I haven’t read any of the commentary here on his paper, so I don’t know if you’ve already discussed this, and I don’t know if it matters very much. I just want to say that my own sense of personal autonomy is that if doesn’t mean something very like 3*, it’s not what I’m talking about. So if his paper is about being able to do something whether or not (and because) one wants to, I don’t think I’d care too much about it.
Again, I’m not sure how this cuts. Maybe he’s trying to prove something even more difficult for the compatibalist. Something which, if he proves it, we get 3* anyhow. But it seems like a wrong turn at the starting gate in any case.
I read a few more pages. Doesn’t float my boat.
In the absence of any hope of resolving the issue scientifically or theologically or philosophically, that’s what most people mean by free will. Can I do what I want to do.
If I have appetites, or emotions that lead to long term harm, can I learn to manage my behavior to avoid self harm?
This is really the most important issue facing most people. A related issue is how to teach self management to children.
It isn’t a philosophical problem, and it has nothing to do with overpriced ideas about cause and effect. It’s just something like: I eat too much; can I learn self control.
petrushka,
There is hope for investigating free will scientifically, which is why I cited those two studies. They falsify the claim that the conscious experience of will is an accurate indicator of what’s going on “under the hood”.
Depends on a) what the word ‘can’ means here, which is where philosophy comes in, and on b) whether physics is in fact deterministic, which is where science comes in.
keiths:
walto:
It’s not false; it’s just unnecessarily restrictive, and it masks a problem with Huemer’s argument.
#6a raises eyebrows in a way that #6 doesn’t:
As with the previous issue, I think the problem traces back to premise #2:
The right way to understand “should” is something like the following:
“Ursula should refrain from believing falsehoods” just means “it would be better for Ursula if she chose to refrain from believing falsehoods”. It doesn’t mean that it is physically possible for her to do so.
Objection 2 deals with that. I think he agrees with you on shoulds.
I read the “can be done” part as saying we do not hold people responsible for not taking some option that they could not possibly have taken, eg using a rock instead of a hammer to nail when no hammer was available. He needs 2 for the should-can link in 5.
Last few paragraphs of objection 4 deal with issues of what it means to judge people in a deterministic world where what they were going to do is what they did, regardless of shoulds.
Try substituting “In the matter of the existence of unicorns” and “I honestly believe in unicorns for “in the matter of free will” and “I believe in MFT”. I think that should show why the constraint is needed. (I hope it does — I have not done that analysis because of above- stated personal limitations).
I don’t think he would accept your bare 6a as he would agree some people do believe in determinism even though he has proved it a falsehood.
ETA: Had to get above off my mind to help sleep tonight! I will try very hard not to even look at TSZ now for a few days.
He is trying to show that his modal approach — real possibility of alternative futures — is needed because those weak-compatibilist ideas do not capture what is meant by free will.
For example, some compatibilists accept that X had no possibility of doing something else because of determinism, but argue that instead it is enough to have the truth of the implication “if X could have done something else, X would have” (roughly).
For them, It does not matter that determinism means the antecedent must always be false in that actual world.
I agree that this sort of compatibilism pretty thin gruel.
So, as far I as I can tell, that section is just trying to motivate the need for his argument; it has nothing to do his argument’s content or soundness.
walto,
That concern could be accommodated with the following qualifier…
…which leads to…
…which is almost as sweeping as #6a was and should still raise eyebrows.
The experience of having choices is no more debatable than the experience of perceiving green. It’s an aspect of consciousness.
Science can scan the brain and declare that some pre-consciousness element of the brain does the deciding, but that is marginally relevant and useful.
From the standpoint of society, the question is what to do about bad behavior. The philosophical question isn’t relevant. The societal question is, can people learn from experience; can consequences change behavior. And if so, what kind of consequences are effective?
The failure to figure this out results in a lot of useless debate about free will.
Bruce,
Sometimes I dream about what we discuss at TSZ. Haven’t had any Kekulé eureka moments, though. 🙁