It’s my honor to comment publicly on Shaun Nichols’ “How Can Psychology Contribute to the Free Will Debate?”, since I’ve been consulting his work with great profit for years. As is the philosophical convention, I offer repayment in the form of “higgling and haggling.”
My haggling commences early, with Nichols’ opening query. Questions like “Are people free and morally responsible?” (1), invite the unwary to suppose that two questions “Are people free? And “Are people morally responsible?” require one answer; the implication is that the fate of freedom (in some to be specified metaphysical sense) and the fate of moral responsibility are intermingled. But as Nichols is well aware, not everyone accepts this commingling of philosophical fates; like lots of folks, I’m inclined to answer “No and Yes”; people aren’t free, but they are (sometimes) morally responsible. My reasons for thinking this are numbingly familiar: I’m (something like) an incompatiblist about determinism and freedom and (something like) a compatiblist about determinism and responsibility (see Fischer 1999), and I share the suspicion, voiced by people like Pereboom (2001) and Sommers (2005), that it wouldn’t help much if determinism were false. I’m moved by considerations in the neighborhood of what Nichols (9) calls “Hobbes’ libertarian dilemma,” and I join the chorus of cranky metaphysical philistines in claiming to find agent-causal libertarian accounts of freedom verging on unintelligible. In short, I’ve P-Strawsonian sympathies of a decidedly unpanicky sort, and that makes me think that the tractable questions in the areas of action theory and moral psychology have to do with responsibility rather than freedom, and that these tractable questions are all the questions we need.
Enough about me. What do I think about Nichols (1) Three (interrelated) Projects? By way of kicking off our discussion, I’ll say something about the descriptive, prescriptive, and substantive projects, in that order.
The Descriptive Project
As Gardeners have already occasion to observe, I, together with my friends Joshua Knobe and Rob Woolfolk, think “the folk” are variantist about responsibility, meaning I endorse the view flagged by Nichols (3n4) in a footnote: it is unlikely that ordinary people’s responses to questions in this area are consistent by the light of systematic philosophical theories like libertarianism and compatiblism. I’m pretty sure it goes this way for responsibility, because I think we’ve learned a good bit recently about responsibility attribution, but I’m quite unsure of how non-philosophers think about notions like “freedom” and “free will,” in part because I’m quite unsure of these notions when I see them deployed by philosophers. In general, I think it’s not at all easy to glom how non-philosophers think about the slippery metaphysical notions that exercise philosophers. In doing empirical work, we need to take pains, as Nichols’ (6) notes, over what sort of freedom subjects have in mind; in particular, we need to take pains to see if they are operating with a “Kantian” notion of freewill (henceforth KFW), where “could have done otherwise” gets cashed in some suitably indeterminist way. Hard work, that. For now, I’m just not sure how important KFW is to the action theorist on the street, and I don’t think the current state of evidence justifies much confidence one way or another. Of course, similar complaints might be (and have been) raised about the experimental scrutiny of most any philosophically interesting term. But I’m inclined – perhaps because of my aforementioned metaphysical philistinism – to suspect that things are especially knotty in the area of KFW. It’s also true that complicating associations are likely to run hot and thick here; when George Bush suggests that terrorists “hate freedom,” I doubt he means that terrorists have a strong antipathy to KFW.
The Prescriptive Project
Whatever one thinks about the outcome of the descriptive project, it’s likely to impact one’s thinking on the prescriptive project. (Hark! Was that the sound of another plank being laid on The Bridge Over the Is/Ought Gap?) For what one thinks folks ought be doing is going to be informed by what one thinks they are doing; strange for me to think you ought stop acting like a jerk, if I don’t think you’re acting like a jerk.
In the present regard at least, I don’t think folks are acting like jerks; I’m (more or less) on record, again with Knobe and Woolfolk, as a conservative; in the area of responsibility, I think the presumption is against philosophically pressed revisions in ordinary thought and practice. On the particular question of “what would happen if people stopped believing in libertarian free will [here, KFW],” I find myself a bit at sea. As I’ve said, I don’t know that people much believe in KFW, but I do believe that people often enough, if not always, attribute responsibility like good compatiblists (see Woolfolk et al. 2006; Nahmias et al. 2006), and this makes me (for better or worse) inclined to doubt that KFW is deeply implicated in people’s thinking about their selves and social world. Like Nichols (13-14), I’m therefore not much inclined to worry that “anarchy and despair would ensue if people knew” – assuming they don’t already – “there is no libertarian free will.” As Nichols (14) notes, this is an empirical question -- an empirical question, it seems to me, that it would be rather difficult to sort out.
We might try cross-cultural studies, and attempt to see if cultures manifesting pervasive belief in KFW are detectably better off than cultures not manifesting such belief. We might look for results, say, indicating that the high murder rate or the racist death penalty in the US is associated with weaker belief in KFW than in countries less afflicted with such pathologies. Now we might find that belief in KFW is pancultural, in which case we couldn’t do such studies. But that would be a result worth having: it would certainly get my attention to be shown that belief in a Western philosophical theory -- a theory I think is deeply problematic – is pancultural.
I expect you can guess where I’m heading: the work in cultural psychology done by people like Dick Nisbett and his outfit makes me doubt that belief in KFW is pancultural. Now the needed empirical work on culture and KFW has not, so far as I know, been done. But suppose we found, as I’m guessing we might, that belief in KFW is weaker in some East Asian cultures, and further suppose we find that these cultures exhibit higher rates of certain pathologies than do cultures where belief in KFW is stronger. We might then, with some theoretical work, tell a story linking the belief states and the pathological states. I doubt it would come out that way, but it’s not completely fantastical to think it’s empirical work we might make a start on.
Alternatively, we might proceed on the intracultural level. We might try to identify the believers and non-believers about LFW, and see how their lives go. Is there reason to suspect that the non-believers would exhibit more in the way of pathologies like suicide and child abuse? I’ve no evidence on this score, but I’m strongly inclined to doubt we’d find differences. My skepticism is initially sourced in reflecting on the general form: “If people became convinced of the falsity of [fill in the philosophical position of your choice], social and psychological catastrophe would ensue.” Uh . . . ok. Perhaps free will is different, say, than statue counting, and this is one philosophical debate with sharp practical teeth, but I find myself doubtful.
Of course, if the folks who think – unlike me – that the folk are pretty uniformly believers in LFW are right, we couldn’t do the sort of field study I propose, as the subject pool would be too homogeneous for comparative work. But maybe we could do a sort of controlled experiment, and induce skepticism about LFW in a group of subjects, and compare their ensuing mental hygiene to non-skeptical controls. But wait! If there’s any credence to be given the conjectures of people like Smilansky (2002: 500, 505n7), we should hesitate to perform such a study, and we should expect (or hope!) that human subjects boards would withhold approval. For by inducing skepticism about LFW, we’d also be inducing psychopathology, and that wouldn’t be right.
Hold on. Don’t some of us – myself included – try to induce skepticism about LFW in the classroom? Sounds like we need to do some follow-up studies, and see how our students – victims? -- are faring. Of course, I may be flattering myself; probably I’ve never convinced my students of anything. But some students are convinced by skepticism about LFW; I was one of them myself. And while I’d be the first to admit that my life is something of a mess, I’m rather inclined to doubt that this sorry spectacle is due my adherence to a philosophical position. I’d also be the first to admit that I can’t be sure about this, at least barring a few more years of therapy. But given the stakes, it seems to me caution is in order: no more teaching skepticism about LFW. No there’s a prescription for you!
The Substantive Project
On the substantive question, I seem to be what Nichols calls an eliminativist about free will: I contend, “[Kantian] free will doesn’t exist” (1). But I’m not – aforementioned concerns about my mental health notwithstanding -- much inclined to worry about this, because, as I’ve said, I think questions of free will and responsibility are detachable, and we needn’t throw out the ethical baby with the metaphysical bathwater.
But it has long been my vague impression that many psychologists are guilty of such infanticide. One unfortunate result of this is that some pretty fabulous psychology can get transmogrified into some pretty unpersuasive philosophy. As a devotee of the human sciences, particularly experimental social psychology, I think this is unfortunate, so I’d like to linger a bit over Nichols’ (9-11) discussion of Bargh and colleagues, and the very interesting – and philosophically significant – work on automaticity and dual process theory. It’s a big topic, and I can at present only sound a few notes of caution -- amplifications, really, of some things Nichols says.
Although I won’t argue it here, I’m convinced that the automaticity literature makes for heavy philosophical weather, but the storm is not brewing in the vicinity of causal determinism. As Nichols notes, psychologists sometimes leave us with the contrary impression. For example, Bargh and Ferguson claim that “behaviorists and cognitive (and social-cognitive) scientists have accumulated evidence of determinism by their many demonstrations of mental and behavioral processes that can proceed without the intervention of conscious deliberation and choice” (2000: 925), and conclude that automaticity experiments “provide… rather obvious evidence that even controlled mental processes are themselves controlled and determined” (2000: 939). There are numerous problems with this line of thought.
The argument looks like an inductive one: psychologists can tell deterministic stories of the sort that are threatening to notions of freedom and agency for a wide variety of systematically observed behaviors. Furthermore, they can tell such stories even in cases where the actors themselves would reject the deterministic explanation and explain their behavior by reference to their own agency. (This is the problem of confabulation; see Hirstein 2005.) Therefore, we are supposed to conclude that a deterministic causal story may be told for all unobserved behavior, even when the actors insist on their agency.
Mercifully, the overwhelming majority of human behavior is not subject to observation and analysis by psychologists. Thus, the induction base for the inductive argument is miniscule; the percentage of all behaviors that are observed in the relevant sense is vanishingly small. Talk about meager input and torrential output! This might be put off to the magic of induction, but there’s further reason for hesitation.
In relatively few psychological experiments are the manipulations “100 percent effective”: not all experimental behavior, even in cases of smashingly successful manipulations, is determined in the way the investigators are trying to determine it. Grabbing at the first examples that pop to mind, Milgram’s (1974) classic obedience paradigm was “65% effective” while Darley and Batson’s (1973) high-hurry manipulation in their Good Samaritan study was “90% effective.” What are we to say of the participants who don’t – dammit – do what they’re supposed to?
KFWers will be quick to see the hope of free action here. After all, KFWers don’t need to deny that many, or even most, actions are determined in the problematic sense; they just need to deny that they all are. I don’t see how the psychological experiments, with their mixed effects, can rule this possibility out. It’s tempting to suppose – as many investigators likely do – that there must be a deterministic causal story for the awkward percentage of recalcitrant participants, even if we can’t yet sort out the details. But this is another instance of leaping inductive faith, and it is not obvious we are compelled to convert. Indeed, the determination (forgive me!) to do so is likely a function of a prior theoretical commitment to determinism, and that’s precisely what’s at issue.
These cautions, although worth noting, are not the main reason for my caution. Instead, it’s that I cling to a theoretical commitment of my own: as I’ve been saying, I’m (something of) a compatiblist about responsibility, meaning (near enough) that I think that the truth of determinism is compatible with the existence of the sort of human agency presupposed by ethical practices such as punishment and responsibility attribution (Doris 2002: Ch. 7; Doris and Stich 2005; Woolfolk et al. 2006). And if you’re my kind of (near enough) compatiblist, you think that the philosophically fertile trouble lies not in the fact that actions and decisions are caused (even deterministically caused), but in the facts of how they are caused. In short, some causal stories make trouble for responsibility, and some don’t.
Interestingly, while compatiblism is a conventional, if controversial, option in philosophy, it does not generally seem to be a theoretical option that much exercises psychologists, even the psychologists doing experimental work that is profoundly relevant to the philosophy of human action. This seems to me a pity, not only because of my theoretical predilections, but also because the compatiblist perspective can help us to focus on the psychologically important question of “How caused?” – a question that is important whatever one thinks of determinism.
That’s what I think the issues aren’t, on the empirical side of the substantive project; my view here is a particular instance of my conviction, stated at the outset, that the problems we moral psychologists and action theorists should take ourselves to have are not the problems of freewill and determinism. But that’s more than enough Doris for today; concerning what I think the issues are, I encourage you to eagerly await the completion of my “How to Build a Person.”