Roskies on Neuroscientific Challenges to Free Will

The unstoppable philosopher/neuroscientist Adina Roskies has a new paper on 'Neuroscientific Challenges to Free Will and Moral Responsibility.' 

In essence, she argues that it is a mistake to think that work in neuroscience could in any way threaten our sense that human choice is indeterministic.  If we learn that the universe is deterministic, it will have to be from discoveries in physics.  Perhaps physics will show that the universe is deterministic, perhaps it won't... but neuroscience won't end up settling the question either way.  (After all, theoretical models in neuroscience are always probabilistic.) 

I'd be interested to hear what the experts here at the Garden think of this argument.  Do you think that neuroscience really can provide a challenge to our conception of ourselves as free and morally responsible, or should we conclude that the whole issue is -- as Roskies suggests -- a red herring? 

Comments

[Kevin Timpe]

Posting on this paper has been on my To-Do list for a few days, but Joshua beat me to it (or, provided me with the motivation to actually put fingers to key-board, depending on how you look at it).

I found the article interesting, but a bit 'loose'. For example, Roskies seems to rule out classical compatibilism from the get-go: "If this is the case [i.e., if the universe is deterministic], then we cannot do other than we do, and so are not free" (419). I also have an issue with the diagram on page 420, which appears to commit compatibilists to thinking both that the universe is deterministic and that we do in fact have free will, when compatibilism per se is committed to neither.

Perhaps I'm being unfair here, but these are important distinctions, I think--particularly in an article that is likely to be read by people not as familiar with the contours of such debates.

All that said, I haven't yet addressed Joshua's particular question. I'm inclined to agree with Roskies regarding the following: "Understanding the brain mechanisms underlying choice and action merely incerases awareness of antecedently existing problems" (420).

[Tamler Sommers]

There also seems to be no place for the FW skeptic who is also an indeterminist (or is agnostic about the truth of determinism). My problem with the diagram is that hard determinism, as she describes it, is not a "main philosophical position" right now. Who, besides maybe Honderich, can be counted as one today?

In response to Josh: I think the primary threat neuroscience poses to our sense of ourselves as free and responsible is one that Adina only lightly addresses: (Greene and Cohen allude to this threat in their 2004 paper.)If Spinoza is right that we (wrongly) think ourselves as free because we're unaware of the causes of our actions, then neuroscience could pose a threat by making us aware of those causes. A convincing account of how people make choices would undermine our sense of ourselves as agent-causes, which might in turn undermine our belief in desert-entailing moral responsibility.

Of course, this story depends on a lot of controversial assumptions, namely (1) Spinoza was right, and (2) that one day the right experiment will show that most people see libertarian free will as necessary for desert-entailing moral responsibility. But neither of these assumptions is utterly implausible, so I think the view deserves attention.

Oddly enough, Adina briefly discusses and almost endorses a similar picture in box 2...

[Kip Werking]

1. First, I think this is an excellent introduction to the free will problem.

2. I think Roskies' argument, as related by Knobe, is undermined by her admission that:

"What neuroscience can indicate is that, regardless of whether or not the universe is deterministic, the brain effectively is [16]. That is, at some higher level than the motions and interactions of atoms and molecules, low-level indeterminacies wash out and the high-level operation of the system can be characterized by laws, so that its future activity can be reliably predicted on the basis of its past activity."

If this is true, then neuroscience seems to become relevant again. Or does she have additional arguments against its relevance?

3. Kevin: I think she is stating that argument as one threat, so characterized, and not necessarily one she agrees with. At least, that might be a charitable reading.

4. Tamler: Like Mele, I don't think Spinoza's explanation is right. Or if it is right, it is far too crude and simplistic. Something more like your evolutionary error theory is needed.

Personally, I think neuroscience (or at least, cognitive science) has a HUGE amount to say about the free will debate. I outline my own view, in detail, in my recent article The View from Nowhere through a Distorted Lens (sadly, but perhaps deservedly, unpublished).

To give a very brief summary:

1. many cognitive biases infect human thinking
2. a substantial fraction of these biases would be relevant to the free will debate, some more so than others (and some not at all)
3. surprisingly, virtually 100% of the ones that are relevant favor belief, and not disbelief, in free will
4. most importantly, these biases would infect not just our present thinking about free will, given reasonable concepts, but would also infect the concepts themselves, such that the concepts themselves are too demanding or overly strong
5. these biases can be grouped together and understood as part of faster, cheaper, but less accurate neural processes in the brain (read: heuristics), which exist in uneasy tension with slower, more expensive, but more general and more intelligent processes in other parts of the brain (e.g. the cortex). This would be a dual-process theory of moral agency, very similar to the dual-process theory or error theory that Josh Greene makes for deontology.
6. how these biases evolved can be largely explained by Error Management Theory because they are adaptive in the Environment of Evolutionary Adaptedness

These biases would include: the fundamental attribution error, the positive illusions (including the illusion of control), the just world phenomenon, trait ascription bias, asymmetrical attributions of blame or demonization processes, etc.
7. It seems to me that Roskies advocates, in her conclusion, a sophisticated sort of revionism, without mentioning Vargas.

[Kevin Winters]

I think Susan Hurley's analysis in Consciousness in Action is helpful here: neuroscience can pose such a problem only if we accept three (possibly fallacious) assumptions. 1) We conflate vehicles with content, 2) we accept the Cartesian Theater, and 3) we assume temporal atomism in relation to brain state/experience correlation. Even if the brain is deterministic that is no reason to assume that our actions or experiences are equally determined; such would be the case only if we assume materialistic reductionism and psychical internalism.

[Tamler Sommers]

Kip,

Right--the Spinoza explanation is certainly too crude. There has to be more to the story. My first assumption should probably be something like 'Spinoza was on the right track.' Incidentally, I do think his account is compatible with other more detailed error theories, including both of ours.

Kevin. Why (2)?

[Kevin Winters]

Tamler,

How (2) effects this debate is still underdeveloped for me (I haven't spent enough time on it). A few years ago I the idea that the lack of a neural central locus of consciousness means that neurocentrism is at least compromised. If Hurley's vehicular externalism is correct--and I think there are many things in its favor--then neural firings alone are insufficient for full-blown consciousness, or it is a highly diminished form of consciousness.

I myself come from a Heideggerian perspective where the inner/outer distinction is seriously questioned and put on new footing, so I'm sure that it influences the above train of thought. If there is no "inner" (and, as it relates to the duality, "outer"), if my intentions within the world and thoughts in general are infused with my interactions with the beings around me and cannot be understood without reference to them, then simple neural firings are not sufficient for giving me the kind of life that I do, in fact, have.

[Clark Goble]

Kevin, one should note that Heidegger probably only disputes the inner/outer distinction in certain ways. But clearly he found the notion of consciousness which tended to dominate philosophy of mind problematic: thus his move to Dasein.

Trying to reconcile Heidegger to more analytic thought is interesting. I think most agree that he's a content externalist ala Putnam. Whether he's a vehicle externalist is a tad more controversial I think.

As to how that ties to the free will issue is more problematic since Heidegger's use of freedom clearly isn't the analytic or common-sense use. It's probably more akin to the primordial requirements for freedom.

The big problem that I see with neurology and free will isn't whether it eliminates free will. I don't think it can for various reasons. One can always at minimum turn down to the chemical or even quantum level to show why it probably can't.

The issue though isn't whether it eliminates free will of the Libertarian sort. Rather the issue is more a matter of degree. Given how much we thought was open to our control is actually causally determined, to what degree do our common sense linguistic uses of freedom still apply. And I don't think the Libertarian responses are persuasive. That is if I find out that some biological structure biases my reasoning such that I reason in a particular fashion 80% of the time, am I still free? Maybe in absolute terms, but certainly I think from a common sense stance or a folk psychology stance one quickly finds one becoming uncomfortable.

Put in more narrow terms, consider a variation of the traditional thought experiments. Say there is some device in a brain that only works 10% of the time but that 10% of the time it controls your decisions. There is no way you can tell when it is working nor can anyone else. Now, are you free? In the absolute senses that are typically debated, clearly we are. But in an other sense I think most would start having pretty big qualms about the term.

[Adina Roskies]

Here are a few thoughts:

Kevin and Tammler: Apologies for the looseness. There is only so much I could do in 3000 words; of course there are more distinctions to be made. Also, I do think that hard determinism is a live position -- Derk Pereboom and Saul Smilansky come to mind.

Kip: There is some tension in my position. I think that we can come to think that we are not free on the basis of what neuroscience shows, but that ultimately we cannot be certain. For even if behavior seems to be largely law-governed, if freedom does rest on some low-level indeterminacies, for instance, there is always the chance that sometimes it won't follow the laws. Some, like Kane, only think that occasional instances of "self-forming actions" are sufficient to make us free and responsible. We can also come to think we are free when we are not, because we fail to understand the brain at a lower level. We never know when we have all the relevant information, which is one of the reasons that I think neuroscience won't resolve the issue. But that doesn't mean that I don't think it will affect our thinking!

I don't quite follow your own view -- is it that we have biases that influence our freedom, or our belief in freedom? And does the fact that they are biases entail that they lead to false conclusions?

Clark: I think I largely agree, but partially because I think the notion we have of freedom is incoherent. Where I do think neuroscience can make headway (not in telling us whether we are or are not free in the traditional sense, but to help us move beyond that sense) is to help articulate notions of control -- what does it mean for *me* to *control* my behavior? What is the 'I', and why do we think that my control is for some reason opposed to normal physical causation? I think that a revisionist notion of freedom is called for (or jettisoning the notion, and making moral responsibility ride on something else). But all this is pretty off the cuff -- I haven't really developed a view in any detail yet.

Anyway, glad you found it worth discussing!

[Tamler Sommers]

Hi Adina,

I'm pretty sure that neither Pereboom nor Smilansky is committed to the truth of determinism. (Pereboom calls himself a 'hard incompatibilist' now for this very reason, right?) My point was that today's free will denier tends to believe that it doesn't matter whether determinism is true or not. Double coined the term 'no-free-will-either-way' theory to refer to this position.

But that's a small point that only relates to your diagram. I'm more interested in why you don't think that the issues raised Greene and Cohen can't be broadened to affect more than just the criminal justice system.

[Clark Goble]

Adina, I think one ought look to revisionist accounts myself. However I think that a lot, if not most, analytic philosophy is still caught up in the linguistic turn. That is what we're primarily debating is language and not something "out there." A move to be more open to revisionism of linguistic terms would be very helpful in my mind.

[Kip Werking]

Adina wrote:

"I don't quite follow your own view -- is it that we have biases that influence our freedom, or our belief in freedom? And does the fact that they are biases entail that they lead to false conclusions?"

My view places emphasis upon accidents of birth and, regardless of these biases, we lack control over such accidents of birth. So my view is much more the latter: that we have biases that influence our *belief* in freedom, and not necessarily our freedom itself. What these biases do, I think, is give us a sense of false self-ownership, such that even though people had no control over the accidents of birth that helped make them who they are, they *feel* as if they did. One way of explaining this is the following: these biases bring us closer to thinking that we have "novelist* control", the sort of control that a novelist has over the lives of hir characters, but that we actually do not have this sort of control over our lives.

My discussion of cognitive biases (as in the article I wrote) involves my view on the relevance of neuroscience and cognitive science to the free will problem. My more general view on the free will problem is that libertarianism is a total non-starter and that compatibilism is too shallow (I am much more pessimistic about revisionism than you and Vargas seem to be). I also think that this eliminativist or non-realist view has certain ethical advantages.

When you ask "does the fact that they are biases entail that they lead to false conclusions?" That fact does not entail that they would lead to a false conclusion in any given situation, or that they always lead to false conclusions. But the fact that they are biases does entail that they lead to *some* false conclusions---or else they wouldn't be biases. But, if you are asking, "does the existence of these biases *prove* that free will or moral responsibility or autonomy doesn't exist?" then I think the answer is clearly no. My detailed theory about cognitive biases and Error Management Theory is one in need of empirical support (I outline a variety of predictions to test at the end of my article). I would defend my more general view on the free will problem with reference to thought experiments and perhaps surveys of folk attitudes and intuitions (like TNR).

[Joshua Knobe]

Tamler,

If I understand correctly, your suggestion is that any theory that helps us to explain human behavior presents a threat to our conception of ourselves as free and morally responsible. Hence, neuroscience could present a threat, but so could social psychology, cognitive science, ... (Indeed, my sense is that theories emerging from these other fields have done more to explain specific human behaviors and would therefore present more of a threat.)

I think this is a very plausible and interesting suggestion, but I'm not sure if it really conflicts in any way with the view Roskies offers in her article. In essence, it seems like the two of you could agree that neuroscience in particular does not present any special threat to free will -- any threat beyond what would come out of any other attempt to predict and explain human behavior. Does that seem right to you?

Kip,

The argument you offer here does seem like a powerful one. If our compatibilist intuitions were the product of unreliable mechanisms, then it seems that those intuitions do not themselves give us any reason to believe compatibilism. But I have the same question for you that I had for Tamler. It seems like the theories you are invoking all come from cognitive science, social psychology, evolutionary psychology... in short, from fields outside of neuroscience. So I wonder if you actually would be inclined to accept Roskies' view that neuroscience itself does not present a threat to our conception of ourselves as free and responible?

Eddy,

I felt like you had some really interesting suggestions on these topics in your paper on "Folk Fears about Freedom and Responsibility" (http://www.unc.edu/~knobe/Nahmias.pdf). So I'd love to hear any thoughts you might have on this question.

[Kip Werking]

Josh,

I'm no expert on neuroscience. I've suspected, like you do, that neuroscience may be so narrow as to exclude the sort of research and methods I've discussed. That is why I have phrased my posts in terms of "neuroscience *or* cognitive science", because perhaps I have only been talking about cognitive science.

But I also suspect that this understanding of neuroscience may be too narrow. Wikipedia describes neuroscience thus:

"Neuroscience is a scientific discipline that studies the structure, function, development, genetics, biochemistry, physiology, pharmacology, and pathology of the nervous system. Traditionally it is seen as a branch of biological sciences. However, recently there has been of convergence of interest from many allied disciplines, including psychology, computer science, statistics, physics, and medicine. The scope of neuroscience has now broadened to include any systematic scientific experimental and theoretical investigation of the central and peripheral nervous system of biological organisms. The methodologies employed by neuroscientists have been enormously expanded, from biochemical and genetic analysis of dynamics of individual nerve cells and their molecular constituents to imaging representations of perceptual and motor tasks in the brain."

This seems to include the sort of MRI studies that Josh Greene has done studying moral dilemmas, and which, by analogy, I hypothesize may be relevant to the free will problem.

But regardless of whatever words we decide to use, I think the content of my view (what I suggest is relevant to the free will problem) is clear.

[Tom Clark]

Hi Adina,

Glad that Joshua pointed us to your article. I think you’re on the right revisionist track to tie our concept of moral responsibility not to the ability to do otherwise but to some notion of being in control of our actions that’s consistent with “normal physical causation,” as you put it. But if we do this, our responsibility practices might change, as you suggest near the end of your paper.

Even if you think neuroscience doesn’t give us any additional reasons to doubt moral responsibility, it’s also true that understanding the causal story of behavior (even with some macro-level indeterminism included, although that’s unlikely) is no less undermining of punitive attitudes than it’s always been. As Spinoza put it “The mind is determined to this or that choice by a cause which is also determined by another cause, and this again by another, and so on ad infinitum. This doctrine teaches us to hate no one, to despise no one, to mock no one, to be angry with no one, and to envy no one.” So, theoretical considerations can (and should, I believe) influence our assessment of our own reactive attitudes, prompting us to second guess the high-affect emotional responses that (according to Nichols and Knobe’s research) drive the compatibilist judgment that people deeply deserve punishment, even if they don’t have libertarian freedom. And this, along with the reductive view of mind that Eddy Nahmias worries about, puts revisionist pressure on the responsibility practices premised on that judgment, such as non-consequentialist retribution (now being given pride of place in the Model Penal Code). You point to this possibility in Box 2 of your article, and as Tamler suggests the ramifications might extend beyond improving the criminal justice system.

As you say in your article, “our views about personal moral responsibility are robust.” We’ll necessarily continue to hold each other responsible, but as our views of agency “mesh with scientific understanding,” the ways we hold each other responsible, and their rationales, might evolve in a more enlightened direction. So a “theory of moral responsibility that is not predicated on paradoxical views of absence of causation or freedom from causal laws” might have considerable ethical impact.

["Q" the Enchanter]

Just a clarifying question (to Tammler or anybody who happens to know): Is Spinoza saying we are unfree because the (mental) reasons we cite for our actions are themselves caused by other (nonmental) causes beyond our control, or is he saying we are unfree because the reasons we cite for our actions are in fact confabulated, otiose rationalizations or folk theories?

[Tamler Sommers]

First of all, there is just one 'm' in my first name. It was one thing when only John Fischer would add the extra m. But when Adina (who I know) and "Q the Enchanter" start doing it, it's time to put my foot down. Is 'Tammler' that much more common a name than 'Tamler'?

'Q': I think it's the former. Spinoza was a true hard determinist. The 'otiose rationalizations'claim is just part of the explanation for why we THINK we have free will, even though we do not.

Josh: Your powers of reconciliation are strong--you'd be the worst moderator of one of those political pundit shows ever--but I still must cling to one point of disagreement. Adina is not only claiming that neuroscience doesn't pose a UNIQUE threat to our sense of ourselves as free and responsible. (If she were, you're right--I would agree. That dime in the phonebooth study is more rhetorically powerful than anything coming out of neuroscience.) But Adina is saying, if I understand her correctly, that neuroscience doesn't pose much of a threat at all. There I would disagree, since neuroscience does, as you say, hold the promise of giving us mechanistic explanations for human behavior. And if our belief in FW and MR hinge on our conception of ourselves as agent-causes, then neuroscience might, along with the other fields devoted to explaning behavior, contribute to the the dissolution of the belief.

Incidentally, looking back at the original post I see that you might be more of a muckraker than your reputation suggests. You start out the post by saying

"In essence, [Roskies] argues that it is a mistake to think that work in neuroscience could in any way threaten our sense that human choice is indeterministic."

Then you conclude the post with:

"I'd be interested to hear what the experts here at the Garden think of [Roskies'] argument. Do you think that neuroscience really can provide a challenge to our conception of ourselves as free and morally responsible, or should we conclude that the whole issue is -- as Roskies suggests -- a red herring?"

Note the implication. A belief in indeterminism is equated with a belief in free will and moral responsibility. That's the move that motivated skeptics to stop calling themselves hard determinists. Every skeptic I know of believes that the kind of indeterminism that physics may prove to be true or false has no bearing whatsoever on the free will problem. The debate in physics is only relevant to the very specific question of whether human choices are determined. Not whether human choices are free or responsible. (Of course Kane and the event-causal libertarians would disagree. But I'm talking about the skeptics.)

[Clark Goble]

Just a quick question. At this stage isn't there a pretty large overlap between neuro-science and cognitive science? I recognize there are some differences, especially if one looks at cognitive science focusing in on symbolist or their arch rival connectivist approaches to cognition. But most of what I've read seems to have a pretty blurry boundary. Sort of like the difference between fundamental chemistry and physics. Or material science and physics. It's so blurry a distinction as to be non-existent.

Sorry for the tangent.

["Q" the Enchanter]

Thanks for the note Tamler. Sorry for the misspelling of your name (and there you'd even observed the quotation marks in my pseudonym); pretty sure I imported the extra 'm' from your last name.

Anyway, if Spinoza's view is as the first gloss I gave it, I'm not clear why you think it poses a challenge to the contemporary compatibilist. Compatibilists have been willing to accept that we might well become aware of the ultimate *nonmental* causes of our actions since at least Frankfurt. (Or am I reading contemporary compatibilists wrong?)

On the other hand, I think on the second gloss (assuming it's true) you get a pretty thorough refutation of modern compatibilism, which rests on some kind of meaningful relationship between the reasons we think we have (or had) for acting and the actual causes that lead (or led) to our having acted. In short, if our Davidsonian rationalizations are merely confabulations or plausible theories generated after the fact to help us make sense of the careers of our actions, it would seem the relation that is supposed by compatibilists to bind causes to reasons to action is incorrigibly fractured. (Hard thing to prove, of course.)

This speaks to the issue raised in the original post. "Choice" supervenes on the same mental and neurophysiological features of human action as "freewill." As such, while neuroscience might not be in the business of assessing the status of determinism, it could very well turn out to be in the business of proving that our perceptions of agency are fundamentally unreliable.