New Midwest Studies in Philosophy

Hi Gardeners,
Check out the new Midwest Studies in Philosophy volume on "Philosophy and the Empirical."  You may be able to access the articles here or here.  There are a lot of interesting articles, and several articles on free will and moral responsibility, including ones by Shaun Nichols, Dana Nelkin, Thomas Nadelhoffer & Adam Feltz, and yours truly (with co-authors Justin Coates and Trevor Kvaran).  Our article (which I link below in case you don't have access) includes the largest survey yet of intuitions about FW and MR (over 1000 subjects, 8 scenarios, 20 experimental questions), the results of which seem to support my claim that determinism per se is not intuitively threatening to FW and MR unless it is presented in a way that suggests reductive mechanism (but there are some other intruiguing results summarized in the Table and briefly canvassed in the appendix).  At some point we hope to set up a webpage with more info on the studies and a link to the surveys so we can collect data from philosophers. 
All the best, Eddy

Download nahmias_coates_kvaran_final.pdf

Comments

[Tom Clark]

Eddy,

Interesting paper, thanks, lots to chew on. Question: You say that “It is a challenge to describe the philosophical thesis of determinism in a way that makes it both understandable to non-philosophers and precise enough to properly test their intuitions about its compatibility with FW and MR while also being careful to avoid begging any questions about whether determinism entails, for instance, that decisions have to happen the way they do.”

But if determinism is the thesis that, as you put it, "holding fixed the laws of nature and the state of the system at one time, there is only one possible state of the system at any other time” doesn't that entail that things have to happen as they do, given the initial conditions and laws of nature? Doesn't it mean that things, including decisions, couldn't have happened otherwise?

If so, I would amend the scenarios as follows to make determinism more salient, for instance using the last paragraph of the Ertan/psychological scenario:

"So, once specific earlier events have occurred in an Ertan’s life, these events will definitely cause specific later events to occur *and no other later specific events could have occurred*. For instance, once specific thoughts, desires, and plans occur in the Ertan’s mind, they will definitely cause the Ertan to make the specific decision he or she makes *and the Ertan could have made no other decision*."

My PI interpretation of the data is that the mechanistic scenarios reduce (incompatibilist) attributions of FW and MR simply because mechanisms reveal cause-effect determinism more saliently. Put another way, person-level psychological explanations have built into them a folk-metaphysical assumption of indeterministic or libertarian slack that the person can exploit to have done otherwise, whatever their mental states. If you amended the scenarios in the way I suggested, then this assumption is blocked, in which case my guess is the differences between the neuro and psych cases in attributions of FW/MR would be less.

[Joe Campbell]

Tom,

You write: "But if determinism is the thesis that, as you put it, "holding fixed the laws of nature and the state of the system at one time, there is only one possible state of the system at any other time” doesn't that entail that things have to happen as they do, given the initial conditions and laws of nature? Doesn't it mean that things, including decisions, couldn't have happened otherwise?"

You have to admit that this is contentious, right? Not everyone accepts this. If the senerio were described as you describe it I might report that I'm an incompatibilist!

[Tom Clark]

Dunno Joe, seems to me like it's a straightforward implication that if determinism as Eddy defined it holds, such that given initial conditions and laws of nature there's only one possible state of the system (in this case an Ertan or human being) at any other time, then the decision couldn't have been other than it was. What am I missing?

[Eddy Nahmias]

Tom, the issues here are tricky. What everyone agrees on is that determinism does not entail that decisions have to happen the way they do (or that decisions could not be otherwise)--that is, if I do X, it is false that NECESSARILY I do X. (This would be a bad modal fallacy.) And everyone also agrees, that determinism DOES entail that if I do X, then this is true: NECESSARILY, given the actual past and laws, I do X.

Now, incompatibilists argue (usually through something like the consequence argument) that determinism entails that if I do X, then in the relevant sense of ability (or can or power), I am unable to do otherwise (in that sense, then, "the decision couldn't have been other than it was"). That is, I would not have the sort of control (or free will) to do other than X that is supposed to be required to be genuinely morally responsible for my actions.

But compatibilists (in general) disagree because they do NOT think determinism entails that if I do X, then in the RELEVANT sense of ability (or can or power), I am unable to do otherwise. (I would argue that even "semi-compatibilists" don't think this if we read "relevant" in the right way, but that's for another post). The reason we compatibilists disagree is because, first, we rightly believe that determinism allows that things could be otherwise in the sense that, if they were, the past and/or laws would be different than they actually were, and of course, the past or laws could have been different. And then we suggest that this fact is enough to secure the relevant freedom and responsibility so long as our actual actions are governed by the sorts of processes that are properly sensitive to the relevant differences (e.g., if I had had different reasons--were the past different--then I would have [or might have?] made a different decision).

All of this is to suggest that I think it is problematic to study ordinary intuitions about determinism by begging what I take to be a central question in the debate--whether determinism entails that we lack the *relevant* ability to do otherwise. (Regulars here will know that this is why I think that Nichols and Knobe's surveys of the folk don't show that incompatibilism is intuitive because, in my view, they present determinism in a way that suggests we could not do otherwise.)

In all of my studies I have tried to test how ordinary people understand the relationship between determinism and the ability to do (or choose) otherwise, and I have consistently found that the folk seem as conflicted as we philosophers seem to be about this.

[Cihan Baran]

Eddy,

(1) Assuming that the phrase "have/has to happen" is question begging, do you think that this would actually affect most folk? I don't think many people can (just from that phrasing) appreciate the metaphysical subtleties you mention: "that is, if I do X, it is false that NECESSARILY I do X. (This would be a bad modal fallacy.) And everyone also agrees, that determinism DOES entail that if I do X, then this is true: NECESSARILY, given the actual past and laws, I do X."

So even if that phrase is question begging, I don't think it would have much effect on the folk. For someone without much knowledge of philosophy, "X has to happen" and "There is only future possible and it entails X" are probably identical.

One thing I wish I would see more in these metaphysical "can do otherwise" type of debates is Kadri Vihvelin's distinction between "ability" and "opportunity"(See Vihvelin "Libertarian Compatibilism"). A piano player who has the ability to play piano *can't* play the piano if there is not a piano nearby, because she lacks the *opportunity*. I think most people would agree that people have the ability for free choice but that the contentious issue is whether they'd have the opportunity if determinism were true.

(2) Also, in the paper you post, you defend the claim that most people have intuitions that support Mechanism Incompatibilism and not Pure Incompatibilism. However, I think that most people also believe that "garden of forking paths" kind of freedom is essential for the "intentional stance" - take away the garden of forking paths and for most people, the whole portrait of human agency collapses into a mechanistic picture.

One thing I am curious to know is whether people are more akin to take the "mechanistic stance" (as opposed to the "participant stance") towards deterministic agents. I am guessing that people are far more likely to take the mechanistic stance towards a deterministic agent as opposed to an indeterministic one.

My hunch is that yes, you are right - mechanism seems to be the real issue in the free will debate. However, I disagree in that mechanism and determinism are independent of each other (I am not entirely sure you make this independence claim but so sorry if I am misreading you) - I think determinism would (probably) support or (less probably) entail a mechanistic view of ourselves.

[jon kvanvig]

Hi Eddy, you wrote, "And everyone also agrees, that determinism DOES entail that if I do X, then this is true: NECESSARILY, given the actual past and laws, I do X."

I think that this is clearly the dominant view, but some theistic philosophers who adopt a non-Humean view of laws will want to leave room for miraculous exceptions to the principle. They may say, that is, that determinism is compatible with your not doing X in a world with the same past and same laws, because of a miraculous intervention. That doesn't affect your point here, however, since it will still be necessarily true that given actual past and laws and no miraculous intervention, I do X.

[Thomas Nadelhoffer]

Cihan writes:

"1) Assuming that the phrase "have/has to happen" is question begging, do you think that this would actually affect most folk? I don't think many people can (just from that phrasing) appreciate the metaphysical subtleties you mention: "that is, if I do X, it is false that NECESSARILY I do X. (This would be a bad modal fallacy.) And everyone also agrees, that determinism DOES entail that if I do X, then this is true: NECESSARILY, given the actual past and laws, I do X."

So even if that phrase is question begging, I don't think it would have much effect on the folk. For someone without much knowledge of philosophy, "X has to happen" and "There is only future possible and it entails X" are probably identical."

If we've learned anything from experimental philosophy, it's that we simply can't get very much argumentative mileage out of this sort of speculation. I think that the x-phi literature is replete with examples that show people are markedly more sensitive to the wording of scenarios than philosophers might have otherwise assumed.

Consider, for instance, the results of Eddy's experiments. I suspect most people would have assumed that people would not have responded differently to the two kinds of cases--especially given how nuanced the different cases were. And yet, the results were very robust!

The moral of the story is this: If you think people won't notice the sort of nuanced differences in wording that worry Eddy, you really ought to get your hands dirty and do the empirical spade work. Simply speculating will only get you so far--especially when the gathering data suggest that people *are* adept at noticing slight (yet important) difference in wording.

[Eddy Nahmias]

Just to clarify Thomas' last sentence: it's not that people are adept at noticing (consciously) these differences or that they could tell you what feature of the scenarios is driving their judgments. Rather, the point is that people respond to the (sometimes subtle) differences, so the differences are in fact influencing the way people think about the scenarios (whether they know it or not). One reason we use scenarios and ask for judgments about them is because we are (usually) interested in people's intuitions rather than their theories. We sometimes ask them to explain themselves (and occasionally we've even asked them to offer their theories), but that is different data interesting in different ways.

Cihan, you write, "For someone without much knowledge of philosophy, "X has to happen" and "There is only [one] future possible and it entails X" are probably identical." You may be right that there is no interesting difference between these two phrases, but neither is entailed by determinism, so both would be question begging to use to test folk intuitions about incompatibilism. Again, determinism does not entail that there is only one future possible. Rather, it entails that (and it's hard to say this in a folksy way): it is only possible, *given the actual past and laws*, for the future to occur in one way. (Even to say "given the past and laws, only one future is possible" or "given the past and laws, you couldn't do otherwise" suggests a scope fallacy.)

Now, the proper wording (or proper idea), if the folk could grasp it properly, may still suggest a "closed future" and no "garden of forking paths" in such a way that most of them would offer incompatibilist judgments. But so far, the data I've collected suggests otherwise.

Finally, you may be right that determinism suggests mechanism. But the big question is whether it *properly* suggests mechanism. In our paper, we point out that scientists tend to treat the two theses as equivalent, some philosophers present determinism to suggest mechanism, and some folk may take them to be equivalent (but they didn't seem to do this when we make the determinism explicitly "at the psychological level"). And we argue that the claim that incompatibilism is intuitive is driven largely by this conflation of the threatening theses of mechanism with determinism.

But I have not seen an argument that determinism properly entails mechanism, though perhaps much depends on how we understand mechanism. Personally, I see no conflict between determinism and a non-reductive metaphysics of mind (even property or substance dualism). And it seems obvious to me that mechanism does not entail determinism.

I would be very interested to hear from people whether they think I am wrong about this, since if determinism does properly entail mechanism, then not only is part of my research program screwed but I may become an incompatibilist!

[Cihan Baran]

Thomas and Eddy, yes, I agree that "has to happen" has a fatalistic tone to it that may sway people's intuitions/responses. On the other hand, I hope we also agree that it's not due to people's appreciation of subtle modal discourse. (I remember I was very much puzzled when I first learned about logical possibility. "What do you mean one *can* travel faster than the speed of light?")

Eddy writes that:Rather, the point is that people respond to the (sometimes subtle) differences, so the differences are in fact influencing the way people think about the scenarios (whether they know it or not). One reason we use scenarios and ask for judgments about them is because we are (usually) interested in people's intuitions rather than their theories.

Now this makes me skeptical about X-phil. If what X-phil is after is folk's unconscious and perhaps emotive responses under the umbrella term "intuition" to philosophical theses, is that really valuable? I thought what were valuable were people's reasoned reflections or theories about these issues. If slight re-wording of these scenarios (that leaves the meaning intact) can make statistically significant changes in people's responses, then maybe what they are really measuring is people's gut/emotive reactions. Would that really tell us what the folk concepts of MR and FW really are? Again, would that be valuable?

And lastly, I would be interested in reading about the interaction between mechanism (as understood here) and the free will debate. I believe that partly the reason why I am an incompatibilist is that determinism supports a mechanistic view of agency. (Once a friend mentioned to me how his math teacher seemed like an "automaton" while he was reflecting on determinism and I believe the intuition would be much more common.)

(Let me mention in passing. Determinism itself allows -in my opinion- unanswerable challenges to compatibilists in the form of manipulation arguments (Pereboom's 4-case argument, Mele's Zygote argument) . Still you'd soften me up to compatibilism if you could show me that in a deterministic world, the intentional stance was warranted.)

[Tom Clark]

Eddy, if under determinism (using your language) “it’s only possible *given the past events and laws* for the future to occur in one way,” and x resulted, then nothing but x could have resulted given the past events and laws, right? So again, it seems to me a fair and clear way to present determinism in your scenarios would be to say that “once specific thoughts, desires, and plans occur in the Ertan’s mind, they will definitely cause the Ertan to make the specific decision he or she makes, and the Ertan could have made no other decision given that those thoughts, desires and plans occurred.”

This doesn’t beg the question about whether an agent could have done otherwise *if* other thoughts, desires and plans had occurred, which of course the agent very likely could have (what Dennett calls the “wide” sense of possibility in chapter 3 of Freedom Evolves). But it does make clear that the contra-causal (“narrow”) sort of could have done otherwise is ruled out. If you don’t rule this out, then you’re not really testing people’s intuitions about whether people can have (say) desert-entailing FW/MR in a deterministic universe, since the contra-causal notion might still be operating in their responses. Unless the scenarios are clear about what determinism entails such that the folk grasp it properly, the results will be ambiguous.

[Joe Campbell]

Tom,

You're missing that the relevant 'can' might designate a general power or ability. Maybe you don't have a bike but you can still ride a bike. If we loose this 'can,' I think science is impossible. Isn't there a relevant sense in which an experiment CAN be replicated even if it is not in fact replicated and determinism is true?

I can ride a bike even though I have no bike. I can ride a bike but my nephew Jack cannot ride a bike even though he has one. Maybe what Frankfurt and your argument jointly show is that there is a sense of 'can' that no one should care about, since it isn't essential to moral responsibility and no one satisfies it anyway.

[Tom Clark]

Joe,

I was referring to the relevant "can" in saying that my amended description of determinism doesn't rule out the wide sense of possibility, such that if the past had been different, the agent likely could have done otherwise. Eddy is trying to control for determinism in seeking to tease out the role of mechanism in shaping people's judgments of FW/MR. But you can't control for it unless you make it maximally salient, and my amendment does that.

I agree that no one should care about the (impossible) contra-causal sense of can and that it isn't necessary for MR. The important issue as I see it is what *sorts* of MR and responsibility practices does the (real) compatibilist sense of can justifiably support. What the folk believe about this will certainly inform the debate, but won’t be dispositive. If people are wrong to think mechanism rules out being MR, as Eddy suggests in the paper, they might also be wrong to think that compatibilist MR supports backwards-looking, non-consequentialist desert and retribution, if in fact they think it does.

[Eddy Nahmias]

Tom, your updated wording is much better. I'll have to think about whether to try it.

I want to clarify one thing. You say, "If people are wrong to think mechanism rules out being MR, as Eddy suggests in the paper, they might also be wrong to think that compatibilist MR supports backwards-looking, non-consequentialist desert and retribution, if in fact they think it does." You may be right about this--I have certainly never argued that if the folk have compatibilist intuitions, then compatibilism is the correct view.

However, the reason I think people may be wrong to think mechanism rules out MR is because they seem to have a hard time comprehending how mechanism is consistent with the intentional stance. So, the issue here would be showing (if it can be shown) how, to put it simply, mental states can exist and be causally relevant even if dualism is false. I don't think folk intuitions are particularly relevant to that issue. Once it were shown, then I would predict most folk would not be Mechanism Incompatibilists. So, their view is false because it trades on a seemingly underdeveloped understanding of the mind-body problem.

Now, the folk may be wrong to think that compatibilist FW and MR supports backwards-looking desert and retribution. But it's harder for me to see how to demonstrate this, since to me it seems that our ordinary MR beliefs and practices are more relevant to the truth of these issues (than they are to the question of whether mechanism conflicts with the intentional stance, etc.). Perhaps good arguments would demonstrate the purported incompatibilist truth to them, but these arguments always seem to bottom out in some premise or principle that relies on our intuitions and practices...

But this is an issue I struggle with and would love to hear more about from Gardeners: what are the "truth-makers" for claims about MR? And what role do our ordinary beliefs and practices have in establishing these facts (if they be facts)?

[Anybody catch Jodie Foster on Daily Show last night talking about how good revenge feels? She reminds us how hard it would be to give up moral anger ...]

[Alan]

The remark about Jodie Foster is apt and ironic to too many layers. It's because of her--and more aptly John Hinckley's obsession with her--that gives us some insight into the folk's current thought about FW and responsibility as it's reflected in criminal law.

Read about the influence of the American Law Institute's Model Code on insanity law prior to the Hinckley case. That code instituted two prongs of exculpation: understanding of morality (McNaughten) and substantial capacity to conform one's actions to law (FW--by some understanding--to choose the good).

Then read about the Insanity defense Reform Act of 1984 as post-Hinckley reaction (in the Federal jurisdiction of the Hinckley trial). The FW prong is dropped in favor of a McNaughten rule alone.

Then by 1986 Utah, Montana, and Idaho dropped the insanity defense altogether. (Although these states incoherently still allow testimony to ameliorate mens rea--but FW is not a prominent part of that either.)

The folk--through legislation--are increasingly unconcerned with FW as detached from the general rational capacity to recognize the difference of right and wrong in classic McNaughten fashion. FW might still be important in individual actions for the folk and especially for philosophers in thinking about the folk, but in assessing the criminal liability of others in codified law, FW took a big hit from Hinckley.

[Kip Werking]

Quick response to Eddy:

I don't want to nitpick Eddy's comments too much, because I admire so much of his work, but this did raise my eyebrows:

"Again, determinism does not entail that there is only one future possible. Rather, it entails that (and it's hard to say this in a folksy way): it is only possible, *given the actual past and laws*, for the future to occur in one way. (Even to say "given the past and laws, only one future is possible" or "given the past and laws, you couldn't do otherwise" suggests a scope fallacy.)"

Of course determinism entails that there is only one future possible. That's the *definition* of determinism!

Eddy also asks:

"I would be very interested to hear from people whether they think I am wrong about this, since if determinism does properly entail mechanism, then not only is part of my research program screwed but I may become an incompatibilist!"

Personally, I think determinism->mechanism. But I suspect that you use the term "mechanism" in a narrower sense than I do. Your writing, it often seems to me, seems to imply that mechanism must be reductive or "bypassing". In my view, a mechanical system like a human being need not be reductive in this sense. But we would need to be precise about the details.

[Kip Werking]

Here are some comments on Nahmias et al.’s, new paper.

1. I’m no expert on experimental philosophy, but this strikes me as the gold standard for empirical work about FW/MR. The sheer abundance of data is staggering and extremely helpful.

2. Eddy repeats here an error theory for incompatibilism/non-realism that he has defended elsewhere. One might call it the “bypassing” threat, because it suggests that “our deliberations and conscious purposes are bypassed by forces that are out of our control.” In that way, I prefer to call it the Zombie threat: people hear about determinism and think it would render them into zombies—unconscious or conscious, but without their consciousness affecting anything. I think this error theory is too strong. Certainly, this is not what motivates most philosophers who deny the existence of free will: Pereboom and G. Strawson, and those who follow in Strawson’s tradition (Smilansky, Sommers, etc.). I think Double and Waller and Honderich and Nadelhoffer, etc., would also avoid the same elementary mistake of worrying about the Zombie threat.

So, at best, Nahmias et al. suggests that there is a serious and previously unmentioned divergence between folk or ordinary non-realists and more scholarly ones: they disbelieve in free will for completely different reasons. But I would go even further, I suspect (but cannot claim to know) that most ordinary people who deny the existence of free will, also fail to make this mistake of worrying about the Zombie threat. Yet, not only does Nahmias, et al., write:

“The threat to FW and MR here seems focused not on the fact that there are sufficient conditions, going back to before we were born, to ensure our decisions and actions, but rather on the purported implication that our deliberations and conscious purposes are bypassed by forces that are out of our control.”
But they even later write that:

“And we would need more evidence that incompatibilist intuitions
get their foothold in most people (perhaps even incompatibilist philosophers!)
because determinism is mis-presented in a way that leads people to take the mechanistic stance.”

3. Nahmias et al. also suggest that, contra Nichols and Knobe’s (primary?) hypothesis, affect has an enabling, and not a biasing, effect on human moral reasoning (at least in these contexts). Indeed, at one point they call such cool reasoning about fw/mr to be “suboptimal.” I think this is an extraordinary suggestion, although Nichols and Knobe consider it themselves. Certainly affect has a biasing influence on human decision making in many or most contexts. When I am hot under the collar is precisely when I am most likely to say what I don’t really mean or do something I will later come to regret. Furthermore, I find that I can properly use any sufficiently simple concept when I am “cool under the collar” so to the speak. When I am cool under the collar is precisely when I make the right judgments about calculus, geography, physics, psychology, investments, etc. I can grasp and delineate complicated ideas like “integral”, “compound interest” and “confirmation bias”. So, I am initially skeptical of the suggestion that, despite my ability to use “integral” and “compound interest” well, when I am cool under the collar, I am incapable of using “free will” or “moral responsibility” properly under those same conditions, when I am otherwise so competent. It remains most unclear why I should have this difficulty properly using “free will” and “moral responsibility” then (although that suggestion is convenient to (neurotic) compatibilists). And it is even more mysterious why affect and anger, which are otherwise so biasing and distorting, should have an enabling function then.

4. I would not place too much emphasis on the subjects’ unwillingness to check “I don’t know.” Most people, including myself, hate admitting we don’t know anything, even when we are most confused. Future research might develop more subtle or sophisticated (non-self-report) methods for measuring subject confusion.

5. Similarly, I would like to hear more about the manipulation check. For example, if it was a T/F question, then even 50% of confused subjects would get it right. In contrast, if subjects were supposed to write down, for example, the only prime number in a series of 50 numbers, it would be much less likely that they could guess properly.

6. Nahmis et al. include questions about Erta “because we thought it might help people reason counterfactually about the implications of determinism, especially if they do not think our universe is deterministic.” But, if that is true, I do not understand the importance of the real world cases, and I certainly would regard real world cases (which do not address the above concern) with more skepticism than the Erta cases.

7. To see how problematic this can be, consider the following quote from the article: “For instance, in support of prediction 1, we found that in the Psych Real scenario a significant majority of participants agreed that people would be free and responsible even if their decisions were determined by prior events, as long as those events included their thoughts, desires, and plans (see Table 1, column B).” But, unless I am mistaken, the comparable figure for “free will” in the Erta case is 53.1%--which is not much support at all, and certainly not “strong support” for the article’s first thesis. I am not sure why (other than convenience) the article would cite the Psych Real and not the Psych Erta number, when only the latter addresses the authors’ concerns about subjects’ prior commitments and understanding of determinism.

8. The data and discussion in the appendix is priceless. It’s good to see numbers supporting the idea that people find fw’s existence to be more dubious than that of moral responsibility (something I’ve noted before). It’s interesting to note that certain authors seem to avoid the term “free will” and argue instead for “moral responsibility” or “autonomy”, etc.

9. I think the “surprising” answers to the questions about criminals suffering and fw requiring a soul may just reflect the more liberal and more educated nature of undergraduate subjects in general. Nevertheless, it was good to hear people say that criminals don’t need to suffer (in a presumably retributivist sense).

10. Finally, although it might seem that I am somewhat disappointed with the article, this is only because I have focused on the areas of disagreement, in a spirit of constructive criticism. I find the article extremely impressive and I want to thank Eddy, D. Justin, and Trevor for sharing it with all of us.

[Tom Clark]

Eddy,

Just as people might be wrong to think mechanism rules out MR because they don’t understand how mechanism is compatible with the intentional stance (as I think it is), so they might be wrong to think MR entails retributivism because they don’t understand that human agents are likely determined (and that indeterminism wouldn’t help establish MR or retributivism anyway). As I see it, ordinary MR beliefs and practices related to retribution are founded on a fiction. Once people see that we don’t have contra-causal free will (as most philosophers and scientists agree is the case), then it will be difficult for them to justifiably believe that retribution is warranted. Why, after all, should people be punished, in ways that need not have any beneficial personal or social consequences, for actions that were necessitated by past events and laws? Why do people non-consequentially *deserve* to suffer for what they’re fully caused to be and do? Because of how good revenge feels? I doubt any Gardeners would reflectively endorse that justification. Moral anger gets tempered by determinism, seems to me, and that’s a good thing given what unchecked retributive appetites can lead to.

Compatibilists have to give a good reason why having forward-looking capacities for reasons-responsiveness, guidance control, etc. (the capacities that take advantage of the compatibilist “can”) make one deserve retributive punishment, which by definition *rules out* the need to have forward-looking, consequentialist justifications. Although I’ve asked many times, I haven’t yet heard a good reason from any compatibilist-retributivist out there. If there is, I might become one too; if there isn't, then retribution needs to be reconsidered.

Ordinary beliefs and practices, however widespread, don’t have a privileged status in establishing claims about MR since they might not have a factual basis and might not serve ends that we fully and reflectively endorse. The “truth-maker” about all this is simply (ok, complexly) the on-going argument about what sort of culture we want to be.

[quidnunc]

I'm only a student so correct me if I'm wrong:

Does retributivism really presuppose contra causal free will or is it enough to have some shared capacities regulating behavior? There's no reason to believe we're better at balancing flow of "disposition" and "situation" than we are decomposing elements in economics but a typical evolutionary explanation of anger, blame, punishment need not presuppose fair proportion when assigning blame to be forward looking because it's a heuristic. Our best reasoning and science can cast doubt on the way we mitigate blame and temper punishment (e.g. it's clear justice need not be quite as harsh in the Leviathan of modern society) but ultimately fairness is human relative and it seems reasonable to stipulate that these capacities are effective enough mechanisms to regulate behavior not a measure approaching some perfect standard, only the best we can manage with some degree of improvement possible in the future. I don't see it as an error theory any more than I see freedom being limited by having a well defined nervous system.

[Paul Torek]

Eddy,

Thanks for the paper. This is my favorite: "Rather, it may be that certain emotional responses should be considered enabling factors that engage the cognitive processes we employ from within the participant stance..." Gems like this are why I'm the self-appointed president of the Eddy Nahmias fan club. I think it's high time we stopped equating neutrality with objectivity and perceptiveness, especially when it comes to moral responsibility. Would anyone say that being devoid of empathy makes someone a better judge of moral matters? Emotion can be your friend. Emotions involve the quick unconscious processing and encapsulation of large amounts of information. Do not discard.

I think Tom's updated wording, on what a deterministic agent can't do, is still misleading. The scope of the modal operator is wrong. Just because the subjects can't formally treat modal inferences doesn't mean that the X-phi'er can afford to be casual about such things.

I also don't see why "definitely cause" is not sufficient.

Eddy, I'll take a stab at your excellent question about "truth-makers" for claims about MR, and the relation to our ordinary practices. MR is moral through and through, and metaphysical only secondarily at best. In broad agreement with Habermas and Scanlon, I'd say that means those truths are consequences of publicly justifiable norms, attitudes about virtues and vices, etc., where "publicly justifiable" does most of the meta-ethical work. But what is publicly justifiable does depend on our ordinary practices, at least in the sense that moral debate has to start somewhere, and it starts by proposing reforms to our current practices. That leaves plenty of space to challenge current practices, however. Other Gardeners have already pointed out the power of wide reflective equilibrium to move us far from the starting point.

[Tom Clark]

Paul,

Assuming Eddy's description of determinism is on the mark, namely that “it’s only possible *given the past events and laws* for the future to occur in one way,” how would you state its implications for the Ertan's decision with the proper scope of the modal operator? Seems to me we'd say something like "the decision couldn't have been other than it was, given that those thoughts, desires and plans occurred," but you apparently disagree.

[Paul Torek]

Tom,

Eddy's English keeps the modal operator operating on the whole conditional, most naturally formalized as:
N ((p & l) -> f).

Your English moves it inside the conditional, operating on the consequent, most naturally formalized as:
(p & l) -> N(f).

Yes, it's maddening. It must be very hard to write survey questions that don't sound stilted and awkward, and still convey the right ideas.

[Cihan Baran]

Paul,

Hume, too, wrote about "necessary connections". ("We have sought in vain for an idea of power or necessary connection, in all the sources from which we could suppose it to be derived.") Was he too getting his modal discourse wrong? Was he also not properly employing his modal operator? Should he have said something like "We have sought in vain for an idea of or necessary connection given the laws of nature and facts of the past..."?

I think not. So many of the philosophers' terms are convention. In Hume's time, there was no such convention. It would be foolish to blame Hume for not using the modal operator correctly.

Similarly, it's a desperate move to say that folk have some sort of intuitive access that allows them to differentiate how modal operator is used in these different wordings, since the use of the modal operator in English is determined by convention and one without the knowledge of that convention has no rational way of getting at it.

There is a clear, folk sense of "has to happen" - that is, the possibility of a single future- and if determinism is true, this sense describes the world as truthfully as 1+1=2.

If these surveys' results are sensitive to such wording differences, they detect people's emotive reactions more than their intuitions.

[Tom Clark]

Paul, many thanks and I take your point. Substituting for Eddy’s English, we get the following in the Ertan psych scenario: “It’s only possible that, given those thoughts, desires and plans occurred, the Ertan’s decision happened as it did.” That is, N((t,d,p) -> d), where the modal operator applies to the whole conditional.

I *think* N((t,d,p) -> d) is equivalent to N~((t,d,p) -> ~d), which translates to:

“It’s impossible that, given those thoughts, desires and plans occurred, for the Ertan’s decision to have happened other than it did.”

Or more colloquially: “It’s impossible that, given those thoughts, desires and plans occurred, for the Ertan to have made another decision.”

If this keeps the proper scope of the operator, then it might work as an ordinary language expression of determinism that would help make it more salient in the scenarios. So for instance the amended scenario would read (my 3rd and last revision, promise!):

"So, once specific earlier events have occurred in an Ertan’s life, these events will definitely cause specific later events to occur *and it’s impossible that, given those earlier events, any other later events could have occurred*. For instance, once specific thoughts, desires, and plans occur in the Ertan’s mind, they will definitely cause the Ertan to make the specific decision he or she makes *and it’s impossible that, given those thoughts, desires and plans occurred, for the Ertan to have made another decision*.”

If this doesn’t work, then I’d be interested to hear your (or Eddy’s or anybody’s for that matter) take on how to express in ordinary language the single possible future aspect of determinism in this scenario, which is what I think is missing in Eddy’s original. Unless that aspect is included, then I don’t think people’s responses to the scenarios will be a fair indication of their intuitions about ascriptions of FW/MR in a deterministic universe.

[Eddy Nahmias]

I was away at the wonderful IU free will conference (we really need someone at this blog to do the conference summaries some other blogs do... NOT IT!), so I've fallen behind on this discussion. I cannot respond to all the comments above--below are a few responses--but this discussion has inspired me to create a new post someday soon on an issue I've long wanted to publish on: the fact that, in the debate about the compatibility of free will and determinism, it seems that as much depends on how one defines and understands the concept of determinism as the concept of free will (how one understands "true MR or desert" seems to be the crucial third issue).

I will try to find examples from various sources (high-level philosophy, intro texts, scientists, exp phil surveys, etc.) that suggest proper, improper, and misleading definitions and/or implications of determinism, and see what people think. As I've said before, I suspect that some people adopt incompatibilism because determinism is presented in a false or misleading way--e.g., to suggest fatalism or epiphenomenalism or a sort of inevitability or inability to do otherwise that requires an argument rather than a stipulation...

Kip, the definition of determinism is not that it "entails that there is only one future possible." Some (e.g., van Inwagen) have defined it as the view that there is, at any given time, only one *physically* possible future. We have to see what people think of that definition (I think it's problematic). But usually people (including van Inwagen) define it to mean: "Necessarily, given that the actual past and laws of nature are such and such, then the actual future will be so and so." Or, as Kane puts it in his new intro book: "it must be the case that, *if* the earlier determining conditions obtain, then the determined event will occur... Determinism is thus a kind of necessity, but it is a conditional necessity" (2005, 6). (Notice no mention of causation.)

So, now looking at Paul's and Tom's logical forms applied to my psych cases, we have Nec.[(thoughts, desires, plans) -> (choice)], which is equivalent to ~Poss.~[(t,d,p) -> c], which is equivalent to ~Poss.[(t,d,p) & ~c]. That could be phrased pretty colloquially as "It is impossible that the Ertan has the specific thoughts, desires, and plans he has and then makes a different choice than he actually makes."

Notice, by the way, that this also means: Nec.[~c -> ~(t,d,p)] which could be phrased "So, it must be the case that if the Ertan had made a different choice, then he would have had different thoughts, desires, or plans leading up to the choice." That sounds entirely reasonable to me (though I wouldn't want to say it's required for free choice, since I'm OK with indeterminism too, in close call cases). Of course, at this point we're off and running into debates about whether this sort of choice means that in order to choose otherwise we have to have an ability to change the past or the laws.

My only point here is that it is misleading to assert baldly that determinism entails that you can't choose otherwise or that the future is necessary (or inevitable) or that nothing can occur except what does occur, etc. And it is not ridiculous to think that, even in a deterministic world, things could happen otherwise, in that if they did, the past or laws would have to be different than it actually was (but of course, few think the past or laws couldn't be different).

(Again, I ask, if anyone knows a way to say that, even if determinism is true, we could hold fixed the actual laws, and say things could happen otherwise without saying the *entire* past would have to be otherwise).

OK, no one should be posting this long, so I better shut up and I will perhaps return later to some of Kip's other points and Tom and quidnunc's interesting points about retributivism.

[Joe Campbell]

Tom,

I was away at the IU free will/agency conference – which was a blast! – and unable to respond. Here are a few comments.

1/ You write that you were “referring to the relevant ‘can’” but isn’t the whole debate an issue about which sense of ‘can’ is the relevant one? Also, you can’t deny that general abilities are necessary for moral responsibility. If the only way that I could have saved someone from drowning is if I could fly like Superman – in the general abilities sense – then I am not morally responsible for failing to save that person. Who would deny that? Thus, the general ability sense of ‘can’ is relevant. What is debatable is whether it is strong enough to ground free will or whether, as you and others suggest, a libertarian interpretation the all-in ability sense of ‘can’ is required.

By the way, to my knowledge, no one has even attempted to prove that what you call the relevant sense of ‘can’ is necessary for moral responsibility. What would a proof of this look like?

2/ Do you really want to say that the folk have the libertarian interpretation the all-in ability sense of ‘can’ in mind when they advocate principles like PAP rather than the general ability sense of ‘can’? The latter strikes me as intuitive and folkish whereas the former seems like something only a philosopher could understand. And what empirical evidence is there to support your claim?

3/ I still don’t understand why someone would insist that a kind of ability that all but a few philosophers think is impossible could be necessary for moral responsibility. And I really don’t see how anyone can think that the folk would consider this kind of ability to be necessary for moral responsibility, even though they think that moral responsibility is possible. Especially since there is another kind of ability that is unquestionably necessary for moral responsibility and – were it necessary for moral responsibility – would easily explain why some people are morally responsible for their actions.

4/ With regard to your last point, I think that you overestimate the folks abilities to understand the scope of operators. In my own logic classes, for instance, I have a hard time getting them to see the difference between Not-(A & B) and (Not-A & Not-B). I don’t see why you would expect them to be better with regard to modal operators than they are with regard to negations. This is why I would avoid the use of modal terms when expressing determinism in experiments intended to test the folks intuitions.

Cihan,

When Hume talked about a possible ‘necessary connection’ between causal events he was thinking of this in a wide-scope sense, not a narrow-scope sense. Also, he famously rejected our having any ideas about such a connection other than the psychological expectation of the effect upon seeing the cause.