Nahmias and King in The Philosophers' Magazine

The Philosophers Magazine has two new book reviews:  The first is a review by our own Eddy Nahmias of Sam Harris's "pithy pamphlet" Free Will.  The second is a review by Matt King of our own Tamler Sommers's Relative Justice.

Comments

[Kip]

Here's a reply to Nahmias:

1. Nahmias cites research to show that Harris definition (whatever it is) of free will is wrong. Nahmias summarizes the conclusion: "Nearly 70% agreed with both statements (with about 20% disagreeing and 10% neutral)."

Two concerns here.

First, even if Harris' definition is wrong (suppose it is), that does not make compatibilism right by default.

Consider the data that Nahmias cites to undercut Harris' definition. This data, albeit impressive, is hardly decisive. Nahmias writes "nearly 70%", which I take to mean less than 70%? Regardless of whether 70% was reached, that still leaves roughly 30% of dissent and ambiguity.

My point is not that Nahmias hasn't undercut Harris's definition, or that Nahmias hasn't done something impressive - my point is that, at the end of Nahmias' critique, we're still left guessing. Less than 70% agreement isn't enough for me to agree on the definition of anything. It suggests to me that free will is simply poorly defined or incoherent - not that compatibilism will win the day.

Second, the more important concern is that Nahmias is essentially asking people to deny that they have something (seemingly) very valuable. People will have huge, overarching, emotional reasons for refusing to deny that - regardless of whether they actually do have it. Loss aversion, reactance, positive outcome bias, etc., would likely all come into play.

Analogy: a person believes he has a winning lottery ticket. You show the person evidence that the person lost the ticket. The evidence mounts over time, becoming more and more persuasive. Do you think that this person is really going to have an objective, unbiased view of whether they lost their lottery ticket? Or do you think the person will be like HIV patients, who dramatically overestimate their life expectancies (according to one famous study)? Don't you think the lottery-ticket-loser might be a little skewed toward "I'm still rich"?

2. Nahmias writes: "Organic chemistry did not make life disappear by explaining how living processes work. Copernicus did not explain away the earth by explaining how it moves. The only reason to think free will can’t be explained is to define it such that it must be inexplicable – to assume that people demand the impossible."

But we shouldn't cherry pick examples to suggest that science never eliminates superstitions. Copernicus did kill the geocentric model. Chemistry and biology did kill the phlogiston theory. Darwin did kill the view that God created the world in six days. Sometimes, we have precious beliefs about ourselves and the world, and science just kills them dead. This is the eternal danger for people who view their philosophical mission as preserving one or more precious beliefs about ourselves.

3. "People think free will includes conscious deliberation and self- control. But we don’t lack these capacities, and it won’t improve our lives or society if we start thinking we do."

How can people ever doubt that they have consciousness or self-control? They immediately experience them! You can tell me that I don't have free will all day long, I will never worry that I cannot experience the sensation of seeing red - I experience it all the time!

It simply amazes me that Nahmias worries that people might think that they lack consciousness because someone said that they lack free will.

More worrisome, however, is that Nahmias' error theory doesn't capture sophisticated skeptics. Did Einstein deny free will based on a simple dualism? Did Bertrand Russell? Does Derk Pereboom? Do sophisticated skeptics like these doubt that free will exists because they think that all consciousness bypasses actions, and is just epiphenomena? Nahmias' error theory does not predict why the smartest skeptics would make his alleged error - because there are deeper, more subtle, and more intriguing reasons for doubting that free will exists than the bypassing theory.

4. Nahmias concludes in part: "Science is likely to show we have less free will than we tend to think, and learning this may move us towards Harris’s practical goals."

This is a remarkable, and welcome, concession. For years, I have argued that errors in human thinking about free will and responsibility are widespread and pervasive. In fact, the idea that our thinking about responsibility attribution is flawed, in many ways, is (I believe) not that controversial among neuroscientists and legal theorists without a vested interest in compatibilism.

My biggest complaint about compatibilism has been that it has virtually nothing to say about all of these errors and flaws. At the end of the day, one typically gets the impression that, according to compatibilists, human thinking about free will and responsibility is perfect and flawless...

Thankfully, that is an impression that Nahmias avoids here.

That said, I still wonder... if Nahmias is willing to admit that so much of our thinking about free will is mistaken - that we have "less of it" than we think we do - why still use the label "free will"? When Nahmias says that people have free will, they are going to think that they have 100% free will, not 70% or something else. At best, it sounds odd to say that free will admits of degrees. So why should Nahmias, as a compatibilist, be able to say that? Does Nahmias intend to qualify his compatibilism every time by saying that "Free will exists, and is compatible with determinism, but only 65%?"

This returns us to the larger problem, undercutting Harris' muddled and extravagent definition of free will doesn't replace it with a better one. Nahmias does a remarkable and admirable job of conceding that much of our thinking about free will is wrong, and that we therefore have less of it than we thought.

The question is: now what? If we only have 65% of the free will we thought we had, how should we address that situation? My proposal is that we shouldn't address this situation by saying that free will exists.

[Neil Levy]

Eddy, your claim about the ignorance of neuroscientists wrt to
consciousness is true, at most, for phenomenal consciousness. And whatever the scientists think, their experiments are rarely about phenomenal consciousness. At most, they infer, on the basis of intuition, that where there is reportability there is phenomenal consciousness (there are exceptions, but they are few). So far as access consciousness is concerned, we have a pretty good theory about the neural basis and functional role. Since it is access consciousness that is relevant to Libet/Wegner and co, a much stronger response to Harris is available. Not "wait, we might find out that consciousness does stuff" but "you're plain wrong: we know consciousness does stuff".

[Tamler Sommers]

In reply to Matt's review, I agree that I don't give a sufficiently well-delevoped account for how moral responsibility assignments would work among people with different core intuitions. That's a fair point. But it's a little misleading to say that "Sommers seems to liken the conditions on moral responsibility to those that attend matters of taste." I never make that analogy, nor do I think it's appropriate. At one point I talk about "good Oliver Stone movies" as an empty concept (in my view) but only in reply to specific objection that metaskepticism must collapse into first order skepticism.

The reason I don't (as far as I know) compare responsibility judgments to taste judgments is that I think there is a lot more factual information to discover regarding the former than there is with the latter.

(Aside from that, I think Matt did a great job for such a short review.)

[Brent]

So besides your possibly valid concern that Harris (and neuroscientists in general) put too much weight on consciousness bypassing arguments, your main objection seems to be with his libertarian definition of free will: “Harris’s definition of free will is mistaken.”

Assuming Harris is defining the type of free will required for moral responsibility, I actually prefer his over yours, even if he is a mad scientist. Harris’s definition is basically the definition used by all free will / moral responsibility skeptics. Just a reminder, and Kip mentioned this above, there are respectable philosophers who hold essentially the same definition.

So with all due respect:

Nahmias’s definition of free will is mistaken. :)

[tnadelhoffer]

Brent,

You have misread the thrust of Eddy's objection. He's not claiming that Harris's definition of free will is mistaken because it's incompatibilist and incompatibilism is mistaken. Whether we ought to be compatibilists rather than incompatibilists is an orthogonal issue. Eddy's criticism of Harris is that he is wrong, descriptively speaking, concerning the ordinary or folk concept of free will. According to Harris, the ordinary concept is both incompatibilist and dualistic--which is a claim made by most neuroscientists who write about free will. Eddy provides evidence that this claim is mistaken (albeit the evidence is even more mixed that he makes it sound--but that is a story for another post!). So whether you or other philosophers find the compatibilist definition plausible is simply beside the point. This particular issue is an empirical one that must be settled with the tools of social psychology and not one that can be resolved from the armchair. And Eddy's surely not mistaken about that even if he is mistaken, normatively speaking, that we ought to be compatibilists! After all, I, too am a skeptic about free will :)

[Brent]

Thomas,

I agree that ONE of Eddy's main points is that Harris is wrong to assume that the folk are inherently incompatibilist/dualist. But I still think the very first point Eddy makes in the opening paragraph is actually an independent objection, in addition to, Harris’s perceived misjudgment of folk intuitions. Here’s the passage:

“Given his other books, one would expect science to drive Harris’s conclusions, but here his argument is conceptual. Step 1: Define free will in such a way that it is impossible. Step 2: Remind us that we cannot have what is impossible. My response is just as simple: Harris’s definition of free will is mistaken. To have free will, people don’t need the impossible; nor do most people think free will requires the impossible.”

There are two objections being made here. The first is that Harris’ definition itself is actually a trick—of defining something out of existence (this objection would seemingly be applicable to incompatibilists in general, and doesn’t necessarily have anything to do with assumptions about folk intuitions). The second objection, which turns out to be the main thrust of the review, is only mentioned in the last clause there: “nor do most people think free will requires the impossible.” Again, you’re right that the balance of the review is probably more devoted to this second point. But the first independent objection IS there. And again I don’t think it scores many points. As I said in my first post, Harris’s definition, regardless of whether he assumes the folk share that intuition, is far from controversial. It’s not only a cheap trick of defining the concept out of existence. Most philosophers who are skeptics DO assume that we would need some sort of magical unknown thingy for moral responsibility to make sense (though some obviously don’t even think that would save it). And if Harris’s conception of free will is plausible to most incompatibilists, then I don’t think Eddy can so lightly dismiss it with “Harris’s definition of free will is mistaken.”

[Brent]

In any case, I shouldn't have said that it was Eddy's main point. Obviously it was not, as Thomas rightly points out.

[Tom Clark]

Neil,

In an earlier thread on Eddy's review I made somewhat the same point as you, that we can already say consciousness does stuff:

"Conscious processes are just those associated with being phenomenally conscious (having experience), and we already know they are neural processes and that they play important roles in higher level behavior and cognition, http://www.naturalism.org/kto.htm#Neuroscience So it seems to me we already have a secure place for conscious processes (and therefore for compatibilist free will as Eddy defines it) in our explanations, it's just that we don't know the ins and outs of their neural mechanisms."

But you're saying it's access consciousness, which needn't involve or go along with phenomenal experience, that's doing the work. I'm wondering why, if experience is left out of the picture altogether (even as a possibly non-functional accompaniment), why it's called access *consciousness* and not simply global accessibility or some such thing. Put another way, what is it about access consciousness that makes it conscious processing?

[Gregg Caruso]

I agree that when one makes a claim about the folk or ordinary conception of free will, one needs to do more than consult their armchair. That said, I believe it is unfair and way too premature for Nahmias (and I believe others) to claim that the empirical issue has already been settled. It should at least be acknowledged that there *are* experimental findings out there that also indicate that ordinary folk are committed to an essentially libertarian conception of free will. There are, for example, the findings of Shaun Nichols that indicate that ordinary folk believe that their choices are not determined and that we do not live in a deterministic universe; there are the findings of Deery et al. that found (with regard to phenomenology) that ordinary folk experience an ability to do otherwise and report that their experience was incompatible with determinism; there are the findings of Bryony Pierce that indicate that (as she puts it) “the folk concept of free will is libertarian – perhaps so libertarian it looks compatibilist, as my hypothesis is that folk participants reject the idea of determinism, even in principle, and give their responses without taking the deterministic nature of the scenarios properly into account”; there are also the findings of Sarkissian et al. (2010) and others. Of course, the apparently compatibilist findings of Nahmias and others should also be addressed, but one should not assume (or claim) that the issue is anywhere near settled yet!

Personally, I tend to agree with Bryony Pierce when she writes in her recent 3:AM interview (posted here a week ago): “I am inclined to take the folk concept to be a libertarian concept of free will in which we are free to choose what to decide/how to act and the outcome of the decision-making process is not determined.” But even if I’m wrong about this, one should at least be willing to admit that the empirical findings thus far have been ambiguous!

I also think compatibilists should be willing to admit the non-naturalistic tendencies of ordinary folk. Given that 92% of Americans believe in God, 4 in 10 believe in strict creationism, 80% believe in miracles, 85% believe in heaven, and 84% believe in the survival of the soul after death, I think it would be a big mistake to assume that ordinary folk are completely comfortable with a naturalistic account of agency! I personally think that if we are going to probe into the folk-psychological and folk-metaphysical commitments of ordinary people, we need to broaden the investigation beyond the narrow compatibility question. We should also ask: “Are ordinary folk dualists?” (BTW, hasn’t Paul Bloom provided some evidence that they are?); “Do ordinary folk view a naturalistic account of the mind as incompatible with free will?”; “Do ordinary folk view the universe as deterministic?”; Etc. If such investigations reveal what I believe they will reveal, then beating up on a libertarian conception of free will may not be as useless a task as Nahmias makes it out to be.

That said, I think there are many shortcomings to Harris’ book. It clearly is directed toward the general public and as a result fails to address a number of substantive issues.

[tnadelhoffer]

Gregg,

I completely agree and I will be posting some recent data here on Flickers in the coming days that further muddy the waters. For now, let's just say the folk are far more dualist and libertarian that Eddy makes it seem. He should know, the data I am going to discuss is from our collaboration for Mele's BQFW! But more on that later.

[Gregg Caruso]

Thanks Thomas. I look forward to hearing more about your new findings!

[Mark Young]

//[T]here are the findings of Bryony Pierce that indicate that (as she puts it) “the folk concept of free will is libertarian – perhaps so libertarian it looks compatibilist, as my hypothesis is that folk participants reject the idea of determinism, even in principle, and give their responses without taking the deterministic nature of the scenarios properly into account”//.

While it may be that people generally have a libertarian /theory/ of free will, the results suggest that their /conception/ of free will is compatibilist. When asked to imagine a determinist world, they retain freedom rather than libertarianism. Their belief in freedom (of the appropriate sort) is /stronger/ than their belief in libertarianism: if determinism ever were proved true, these people would become compatibilists (or semi-compatibilists) rather than reject free will/moral responsibility.

Given that they would give up libertarianism before free will, it must be that their conception of free will allows it to exist in a deterministic world, even tho' they don't believe it would. Compare: I am a materialist, and I have a materialist theory of humanity, but I don't have a materialist /conception/ of humanity. If immaterial souls were proved, I'd reject my materialism, not my belief that humans exist.

[tnadelhoffer]

Mark,

I am not sure about your understanding of the flip floppers. For one thing, on your view, flip floppers are compatibilists and not incompatibilists. That's fine as far as it goes I suppose, but then we would need to call folks like Van Inwagen compatibilists--which is odd. What I think we should say is that if flip floppers were faced with a deterministic universe, they would give up their robust notion of free will and moral responsibility and settle for some watered down compatibilist substitute. What they wouldn't and couldn't do is continue to insist that humans have contra-causal freedom, the unconditional ability to do otherwise, etc. So, it's strange to call pre-flop flip floppers compatibilist.

That being said, I also want to make a quick point about the experimental research on folk intuitions about agency and responsibility. I can't speak for other folks who work in this area, but I can say that at least when it comes to my own work, I am constantly surprised with how many people miss the manipulation check--which usually means they don't believe that the determinism in the scenarios really operates in the way we've specified. In my paper with Feltz and Coakley, we call these people free will no matter what theorists because they're so convinced of free will that they can't imagine that the agent couldn't have done otherwise in the unconditional sense. So, they are simply blind to the determinism we build into the scenario. Now, of course, because these participants fail the manipulation check, we don't use their data--but it is often a sizeable number of participants (e.g., in one study it was nearly 40%, if my memory serves me correctly). But this just means that a sizeable number of folks I would call libertarians don't show up in the reported results. Indeed, if we count these folks as incompatibilists-as I think we should--and we add them to the number of incompatibilists who got the manipulation check correct, there are as many incompatibilists as compatibilists even in studies where the latter seemed to heavily outweigh the former. In my eyes, no one has yet adequately figured out a way of identifying flip floppers or free will no matter what theorists among the folk--which is an issue that probably merits its own post!

[Brent]

Mark,

What I think Bryony Pierce and Greg (and Thomas, who just responded with a much superior post while I was writing this) are saying is that the data gleaned from that experiment is unreliable in the first place because, as Pierce puts it, “folk participants reject the idea of determinism, even in principle, and give their responses without taking the deterministic nature of the scenarios properly into account.” In other words, the apparent revelations of the accompanying paper have about the same value as verdicts from a mistrial.

You say “the results suggest that their /conception/ of free will is compatibilist. When asked to imagine a determinist world, they retain freedom rather than libertarianism.”

But the folk are a) so committed to their (libertarian) freedom and b) so biased against even the possibility of determism that they can’t honestly go along with the parameters of the case. Just because they answer “yes we are free in a deterministic world” doesn’t mean they’ve actually changed the type of free will they are committed to (if they did, they would have indeed arrived at compatibilsm). But I think it’s more likely that they are just settling on being incoherent (maintaining libertarianism with determinism). If that doesn’t make sense, well, it’s because the participants by and large aren’t making sense.

By taking the results of that paper at face value you are assuming that the folk in these experiments are aspiring to coherence and consistency in a) their assessments of the cases and b) the intuitions/conclusions they eventually report.

[Eddy Nahmias]

I can't respond to everyone's interesting comments now. But I did want to point out that my arguments against Harris were not based solely on x-phi results--there are lots of reasons to reject the often ridiculous definitions of free will Harris employs (and note that he rejects compatibilist views based on his armchair suppositions about what people think about free will and without engaging with the philosophical arguments). And to the extent that I did use x-phi results to respond to him, I mentioned my own studies about compatibilism very briefly, but focused more on the recent work I've been doing on Frankfurt cases, but we'll be testing more directly the sort of case Harris uses to support his assertions about ordinary thinking about free will. Finally, the stuff about dualism is largely irrelevant to this discussion because Harris is not using the dualistic definition of free will to make his case and I'm not calling him out on that.

[Kip]

Gregg wrote:

"Given that 92% of Americans believe in God, 4 in 10 believe in strict creationism, 80% believe in miracles, 85% believe in heaven, and 84% believe in the survival of the soul after death, I think it would be a big mistake to assume that ordinary folk are completely comfortable with a naturalistic account of agency!"

Exactly! This is why I wrote:

"My biggest complaint about compatibilism has been that it has virtually nothing to say about all of these errors and flaws. At the end of the day, one typically gets the impression that, according to compatibilists, human thinking about free will and responsibility is perfect and flawless..."

However, it's also important to remember that Nahmias does acknowledge this. Nahmias explicitly states that science will probably show that we have less free will than we think we have! This is a remarkable concession. I don't remember hearing other compatibilists, like Dennett, making similar concessions.

[tnadelhoffer]

Brent,

You're overstating my point. There is no monolithic folk concept, libertarian, compatibilist, or otherwise. Instead, some people genuinely seem to believe in good ol' fashion libertarian contra-causal free will, some believe in compatibilist-style free will, and others are much harder to pin down, experimentally and conceptually--e.g., flip floppers and those who are in the free will no matter what camp.

I was simply pointing out that while far more people appear to have compatibilist-friendly intuitions (at least in some contexts) than I predicted when I started probing folk intuitions with Eddy and others back in 2003, our reported results don't tell the complete story. Of course, neither do the other studies that have been run. Free will and moral responsibility are complicated, so we ought not to be surprised that probing them is complicated as well!

So, when you speak of "the folk..." I think you're overstepping. Some folk are so firmly committed to their libertarianism that when they are presented with deterministic scenarios, they deny the agent has freedom and responsibility. In some studies, these folks are in the minority. In other studies, they are in the majority. So, we ought to avoid talking about the folk in the abstract--or, at least, that's the lesson I've learned from nearly ten years of probing their intuitions!

Finally, I just wanted to point out that in the kinds of cases you mention--namely, what I am calling the free will no matter what camp--it's hard to know whether we should place them in the compatibilist or the incompatibilist camp. After all, if you believe in a non-physical soul, then you could conceivably believe that we have contra-causal freedom even in a fully determined physical universe since the seat of your free will--namely, your immaterial soul--sits outside (or above) the physical universe even if it somehow interacts with it. So, while they think we can be contra-causally free even in a deterministic universe, this is a view that neither incompatibilists nor compatibilists defend. After all, both parties to the debate think that if the universe is determined we can't have the unconditional ability to do otherwise. Ultimately, the traditional philosophical debate boils down to whether the conditional ability to do otherwise is enough. But no one thinks you can have the unconditional ability to do otherwise in a determined universe. Given that I think some people do hold this view--quite a few more than Eddy lets on--this is why I think people like Sam Harris and others define free will in terms of impossible powers. In short, lots of people think we have impossible causa sui powers! The difficult is trying to figure out how to get at these intuitions! Yet more work for us to do!

[Brent]

I'm wondering whether folk conceptions of free will are stable enough or consistent enough to survive the process and scrutiny of being categorized so specifically. That’s a pretty sophisticated list of possible positions that you claim folk reliably express in these experiments: "good ol' fashion libertarian contra-causal free will", "compatibilist-style free will", "flip floppers", and the "free will no matter what camp," probably among others.

I agree that there is no uniform folk conception of free will, but I think any differences are simply different symptoms of the same overall commitment to libertarianism. It’s uncontroversial, I think, to say that all religions, no matter how different, are expressions of the same basic human attraction to the transcendent. Well, I propose that all different categories of responses to these surveys are actually expressions of the same libertarian identity crisis. In their rush to reconcile libertarianism with determinism (probably in some cases for the very first time), they produce various categories of unsuccessful defenses. I don’t think there’s a chance in heck that they would either understand most of the categories you listed, or self define themselves as those once you’ve explained them. But I guess this is all just armchair babble. :)

[Neil Levy]

Tom, your response is quite common among philosophers today: why should we *that* consciousness? Well I don't care what we call it. But I think it has as good a claim on the word as any other: its what Freud meant by 'conscious'. I think this is a good example of how quickly our intuitions can be pushed by theoretical debates" Chalmers and Dennett talk like the the problem of consciousness is the problem of qualia, and suddenly everyone's convinced that that's what the word always meant.

[Mark Young]

Thomas: //So, it's strange to call pre-flop flip floppers compatibilist.//

If I understand the paragraph that that sentence ended (which is somewhat doubtful), then I agree. The pre-flip-flop flip-floppers are libertarians, not compatibilists. (Likewise, I am a materialist, not a dualist.) The distinction I was aiming for was "theory" versus "conception". These people are libertarians because their theory is libertarian.

But if they truly would switch to compatibilists definitions of "free will" and "moral responsibility" when convinced that the world is deterministic, then libertarianism is not part of their *conception* of free will and moral responsibility. Only those who would reject free will and responsibility under those circs are libertarian right down to the conceptual level.

Brent: //Just because they answer “yes we are free in a deterministic world” doesn’t mean they’ve actually changed the type of free will they are committed to (if they did, they would have indeed arrived at compatibilsm).//

Quite right. Maybe what I wrote above will clarify my intent (even if you still disagree with it). But if they truly do say (and mean) that "we are free in a deterministic world" then they /theoretically allow the co-occurance/ of freedom and determinism, even if they do not believe that it can co-occur (just as I theoretically allow the existence of immaterial souls while holding that no such things exist). And if they do theoretically allow coexistence, then libertarianism is not *conceptually* linked to free will for them -- not like it is for those who completely reject free will for deterministic entities.

My point, I guess, is that the situation is not quite so binary as it's siometimes made out to be.