Terminological Differences

Given the interest generated by Tamler's recent post on free will skepticism, I thought I would try to keep the discussion going by posting something about a related issue that I have often found puzzling.  On the surface, it is obvious enough what distinguishes libertarians, on the one hand, from compatibilists, semi-compatibilists, and revisionists, on the other hand.  The later, unlike the former, believe that we could be free and/or morally responsible even if determinism were true.  Similarly, it is obvious enough what distinguishes libertarians from free will skeptics.  The former, unlike the later, believe that we are both free and desert-based responsible. 

However, it is not always as clear to me what distinguishes compatibilists, semi-compatibilists, and revisionists from free will skeptics.  As far as I can tell, proponents of each of these views generally believe the following:

1. Humans do not have the kind of robust free will that libertarians think we have. 
2. Humans are nevertheless typically reasons responsive creatures with the ability to consciously regulate their own behavior, form second-order desires, etc.
3. Given the social nature of humans, the creation and enforcement of social norms is necessary and justified.
4. Given the importance of social norms, punishment may be necessary for encouraging compliance--especially when the wrong-doers have the capacities mentioned in 2.

So, wherein lies the disagreement?  It appears the two main disagreements are as follows:

• Should we call the kind of reasons responsiveness and conscious control that are emphasized by compatibilists, semi-compatibilists, and revisionists "free will" rather than something else?
• Regardless of what we call compatibilist control, is it enough for full blown desert-based moral responsibility (rather than the merely forward-looking stuff that presumably everyone agrees is compatible with determinism)?

The first of these two disagreements is entirely terminological.  Settling this dispute requires the parties to the debate to spell out why they believe using the term "free will" is beneficial or harmful.  For instance, Manuel thinks we should keep the term even if we have to (radically?) change the meaning, whereas I think we should jetison it entirely if we now realize we don't have the kind of causal powers we once thought we did.

The second disagreement does not appear to be merely terminological.  But it is nevertheless important for compatibilists, semi-compatibilists, and revisions to spell out very carefully how and why the kind of circumscribed free will we purportedly have is enough to ground desert-based responsibility.  Keep in mind that the issue is not whether humans ought to be held responsible for their actions.  As 4 above makes clear, this is something about which all parties to the debate agree.  The issue is whether the kind of free will compatibilists allege we have is enough for moral desert. 

In arguing that it is, one needs an argument that does not trade on any of the compensatory benefits of holding people responsible since these are benefits about which even the free will skeptic can agree.  No, what the compatibilists, semi-compatibilists, and revisionists need to show is that the kind of free will they are offering is enough for the kind of moral desert that is otherwise easily accommodated by the libertarian view.  Of course, you could be a revisionist about the meaning of "moral responsibility" as well such that it is forward rather than backward looking.  But then it makes the second disagreement mentioned above look merely terminological as well.

At the end of the day, it seems to me that when it comes to the distinction between compatibilists, semi-compatibilists, and revisionists, on the one hand, and free will skeptics, on the other, it all comes down to (a) whether we ought to call compatibilist control "free will," and (b) whether compatibilist control is enough for desert in addition to being enough to ground other forward looking notions of responsibility.  I, for one, think the answer to both (a) and (b) is no.  But I am curious to hear what others think...

Comments

[Neil]

This free will skeptic denies that punishment is justified. So does Pereboom. Constraint may be justified, but punishment is not mere constraint (constraint for non-blameworthy should be as comfortable as possible).

[tnadelhoffer]

Neil,

Right. I am technically anti-punishment as well. What I was trying to convey is that being anti-punishment doesn't mean one needs to be anti-accountability--which is the way skeptics often get portrayed. Perhaps I was inadvertently being revisionist about the meaning of "punishment." Oh, the irony :)

That being said, in light of your comment, I hereby officially change 4 to:

4*: Given the importance of social norms, state intervention may be necessary for encouraging compliance. The kind of intervention required will depend on the degree to which the wrong-doer has the capacities mentioned in 2.

Perhaps I should have had more coffee and/or more sleep before posting this early in the a.m.!

That being said, I am interested to know why a skeptic couldn't be pro-punishment for forward looking reasons. For instance, one might think it is appropriate to punish a dog for tearing up the furniture without thinking the dog is blameworthy. In this respect, couldn't a moral desert skeptic nevertheless be for certain forms of hard treatment at least in principle?

[Manuel Vargas]

Thomas-

You've identified a really important issue, I think, and one that will probably get more prominent as we continue to see a proliferation of views that involve various concessions on things like "a kind of responsibility" a "kind of blame" and so on.

And for the record, I'd love to hear more from folks about what general principles might be available to us to settle whether we jettison or retain terms when the world doesn't work out as we hope. Although I've got a horse in this race, I can recommend Susan Hurley's excellent "Is Responsibility Essentially Impossible?" Phil Studies (2000). It is something of a reply to G. Strawson's work, but it does a nice job of working through some of the issues involved in eliminativism vs. revisionism.

Neil:
"the free will skeptic denies that punishment is justified. So does Pereboom"

I love the (unintended?) implication that Pereboom isn't really a free will skeptic.

More seriously, I would have thought that free will skeptics deny the existence of free will, and that at least in principle, it is a further matter whether this denial entails that no punishment is justified. Now, as a matter of course, on most accounts of punishment it will turn out to be true that no free will = no punishment, but it seems undesirable to build in this view of punishment into the very denial of free will.

One easy way to get at it is via something like a semicompatibilist position (although, more accurately, a skeptical semicompatibilist, since John defines semicompatibilism in a way that makes it neutral on the matter of free will). That is, one could think that punishment requires moral responsibility but not free will. But this isn't the only way, I would imagine. Of course, some of this turns on what one means by punishment. You could define punishment in such a way as to rule out any justificatory ground that does not appeal to free will. But that seems like an unappealing account of punishment.

All that said, I take your point that one natural implication of free will skepticism is that one is unlikely to have much truck with punishment.

[Neil]

Implication about Pereboom unintended. A philosopher over here in the UK (who I won't name) once said that Derk is a compatibilist who calls himself a sceptic for "advertising purposes".

[Tamler Sommers]

"And for the record, I'd love to hear more from folks about what general principles might be available to us to settle whether we jettison or retain terms when the world doesn't work out as we hope."

I would too. We have clear cases on both sides, but very little guidance on how to settle the controversial ones...Harmon has some interesting things to say about this too. Once we learned about the relativity of motion, we didn't stop saying that things move or are at rest. So why should stop saying things are right and wrong, just because we learn that they can only be right and wrong relative to a culture or individual? I wonder how this might translate to the MR debate...

[Tamler Sommers]

"All that said, I take your point that one natural implication of free will skepticism is that one is unlikely to have much truck with punishment."

I don't agree. Bentham is a clear counterexample. He denies desert, endorses punishment (for utilitarian reasons, of course).

[Neil]

Tamler, Bentham wasn't a clear counterexample to the claim. He was wrong . Well, you needn't agree, but you should agree that it would take work to show that you can have punishment (as opposed, say, to tellishment) consistent with desert-scepticism.

[Tamler Sommers]

Neil,

Manuel seemed to be making a descriptive claim about how much truck free will skeptics are likely to have with punishment. Not how much truck they should have with punishment. In that sense, Bentham is a counterexample.

As for the normative question, I don't find the "tellishment"/"punishment distinction all that helpful. If you define punishment as essentially involving desert, then obviously the desert nihilist can't be in favor of punishment. But they can still favor the death penalty, long prison sentences, whacking people with batton canes etc if they're utilitarians and they think these actions will increase net happiness.

I guess I'd ask you this: How does playing around with definitions bear on substative question of whether free will skeptics could think it's justified to, say, imprison someone for ten years because they committed armed robbery...

[Matt King]

Thomas,
I think libertarians and skeptics are in the same boat here. Neither side gives an independent account of why on their view agents “deserve”, say, blame. So I think it presumptuous to suppose that compatibilists have a special burden to discharge here. And it isn't obvious (especially to compatibilists, I take it) that the libertarian or skeptic will have an easier time with the task.

For example, the following dialogue seems familiar:

Compatibilist (C): All that’s required for an agent S's being morally responsible for some action A is that - S was appropriately reasons-responsive; S did A intentionally, knowingly, and without mistake; S acted on a reflectively-endorsed desire, etc.

Libertarian or skeptic (LS): Here’s an example where S satisfies X, but surely S doesn’t deserve blame. After all, she isn’t – the uncaused cause of A; the ultimate originator of A; able to avoid doing A; etc. So in what sense can you say S *deserves* blame?

C: I’m not sure I understand the question. She deserves blame *because* she satisfied X.

LS: But that’s not sufficient to deserve blame. You have to satisfy Y.

So my conjecture is that disagreement about what amounts to deserving blame is just disagreement on when agents are morally responsible, which is captured in a number of disparate disagreements about particular principles and contested cases.

We can of course add details to the dialogue above. The LS can argue that blame in certain conditions wouldn't be fair, or warranted, or merited - but these smack of a certain rephrasing of 'deserved'.

So I grant that we might want to think (and argue) more about what desert-entailing responsibility amounts to. But I'd deny the compatibilist is in a different position from the LS with respect to that goal.

[Neil]

Tamler, if you're a consequentialist, you might well think that it is justifiable to imprison someone for 10 years for armed robbery. But you're committed to claiming that there is no essential link between the facts that S is causally responsible for the armed robbery and that S should be imprisoned. Even a rule utilitarian should countenance the possibility that the best policy might turn to entail that S should not be impriosoned, or that someone else entirely should be. In general, I guess I deny your claim that free will sceptics might think it is justifiable to imprison someone because they committed an armed robbery. Non retributive justifications of 'punishment' are forward-looking; the 'because' is backward looking.

[Paul Torek]

Manuel,

You've got a great question about what general principles might be available to us to settle whether we jettison or retain terms when the world doesn't work out as we hope. Unfortunately, the likely answer looks to be "none". The human mind is wonderful at recognizing further instances of categories, and absolutely awful at coming up with principled reasons for those classifications. Such feats of metacognition provide little evolutionary advantage, it seems, and we are woefully (or maybe it is mercifully) devoid of them.

All we can do is go through the examples and try our hands at abduction. Since the subject seems to be relatively new to historians of science, much less linguists and psychologists, there are dim odds of getting substantial theory handed to philosophers on a silver platter. But maybe that's just my ignorance of history of science, talking. I'll try to find out, because if there are any such theories, that would be intriguing.

[Kip]

Manuel,

With respect to revisionism, some of your favorite examples are whales, marriage, magicians, etc. But I don't think we jettisoned any of these terms. We did jettison, arguably, certain meanings attached to these terms. Is this what you mean?

If not, the closest example I can think of jettisoning an actual term is phlogiston. But it's still not jettisoned (I'm using it right now). We just hardly refer to it any more.

If I'm understanding the idea of "jettison a term" correctly, it raises interesting questions about free speech. What if someone wants to use a jettisoned term? Should we stop them? How could it ever be wrong, ethically, for them to use the term after it's been jettisoned?

[Note: using a term does *not* imply that the concept refers to something actual. We use the terms 'unicorn' and 'leprechaun' without committing ourselves to the existence of these things.]

Whether we are jettisoning terms or meanings, I don't think there are rules for how we jettison these things. For example, the meanings we attach to symbols seems to be arbitrary. Japanese is not ethically superior to English in mapping concepts to symbols. Or vise versa.

I can imagine one way in which there could be an ethical imperative to revise/jettison a term's meaning: if doing so caused some confusion, albeit an ethically desirable confusion. For example, if Hitler intends to bomb Poland, it might be ethically desirable to revise "Poland" to mean "Antarctica." I'm pretty sure, however, that this is not what you mean, with respect to free will. (But if it is, your revisionism seems to be surprisingly similar to Smilansky's Illusionism).

If everyone agrees, however, and no confusion is caused, it seems to me that there are no rules governing revision or jettisoning. Whales, marriage and magicians evolved to mean one thing, but they could have kept their old meanings, or evolved different ones, and there was no right or wrong to this. (For examples, "whales" could still mean fish; we would just say that whales don't exist.) Instead, we seem to have almost limitless freedom in mapping concepts to terms.

[Jonathan Farrell]

Just a small point: it could be that a (semi-)compatibilist/revisionist and a skeptic might both answer the substantive question 'no'. But the difference between their positions might be seen in their disagreement about other spheres and activities which (seem to) require that we have free will ('living' one's life rather than merely observing it unfold, creativity, love, etc.).

[I tried to post this earlier today, but it doesn't seem to have worked. Apologies if it appears twice.]

[tnadelhoffer]

Kip,

It's not so much a function of there being some kind of absolute right or wrong when it comes to revising terms rather than jettisoning them. The issue, however, is normative. The question is whether there are pragmatic/moral reasons for continuing to talk about free will and moral responsibility even if we now believe that the traditional meanings of the terms are no longer defensible in light of what we have learned about the nature and limitations of human agency and cognition. A bad analogy: If what we have now is merely Coca Cola Light, should we continue calling it Coca Cola? In this case, not much hangs on it (except for sales!). But in the case of free will and moral responsibility, it seems like more is at stake. Compatibilists and Revisionists alike have often been criticized for merely providing us with free will and moral responsibility light. The response is often that they provide us with all of the fw and mr "worth wanting"--but that is an open question. I, for one, think that if we don't have the kind of free will and moral responsibility the majority of people (at least in the west) have assumed that we have, we should not continue talking as if we did.

My goal is not to be unduly contrarian. Nor am I engaging in mere semantic quibbling. Rather, I think that what we call things matters. And I also think the terms “free will” and “moral responsibility” carry an awful lot of both metaphysical and historical baggage. From the story about free will’s role in our purported fall from grace in the Garden of Eden to the Cartesian dualism that aimed to make room for free will in an otherwise mechanistic universe, in the Western tradition the notion of free will has often been aligned with that which is supernatural within us—that ephemeral ghost that so curiously haunts our lowly bodily machine. Moreover, it has historically been conceptualized as that part of us that separates us from “the beasts,” makes us morally responsible for our behavior, and determines whether we end up with eternal damnation (or salvation). Given this web of historical associations I do not think that we should revise the terms “free will” and “moral responsibility.” If we don’t have the kind of agency and responsibility that people have traditionally thought we had, we invite confusion by continuing to use the old terms to talk about what we actually do have—especially when we could simply use other terms which are less loaded.

[Derk Pereboom]

I agree with Thomas’s last post. I’d like to add that there may come a time – perhaps it’s already happened – when in the philosophical debate the use of terms like ‘compatibilist’ loses its value. I think that core issue in this debate has always been whether determinism (or indeterminism) is compatible with desert that’s basic in the sense that the agent would deserve blame or credit, punishment or reward, just because she has performed an action, given an understanding of its moral status, and not, for example, by virtue of consequentialist considerations, or solely by way of a contractualist account. Philosophers who have denied this compatibility, or else have remained resolutely silent on the issue, have called themselves compatibilists. But then, first, in order to know what they mean by ‘compatibilist,’ further specification is needed, and if this is typically required, the term ‘compatibilist’ is arguably no longer doing any work. A second danger is that if it becomes legitimate to call yourself a compatibilist even though you deny that basic desert is compatible with determinism, or have no opinion about this issue, then perhaps almost everyone counts as a compatibilist, and that’s another way the term could become valueless in the philosophical discussion. In general, a function in philosophy of terms of this sort is to demarcate lines of philosophical disagreement, and when ‘compatibilist’ can no longer have this role, it might be better to opt for a different vocabulary.

[Eddy Nahmias]

So much to say, but I'll try to keep it brief for now. Thomas is right that it can be hard to discern the substantive differences between compatibilists, etc. and hard determinists, etc. (how do Pereboom and G. Strawson really differ from JJC Smart and Dennett?).

But there's more to the free will debate than just the question of what capacities are required to deserve retributive punishment and whether we have those capacities (and I also think there's more to the question of whether we deserve retributive punishment, whatever that is, than whether there is "intrisic value" in wrongdoer's suffering).

There are questions about how we conceive of ourselves and our powers to control our own lives and to change our habits and situation; about the important ways we are different from other animals and how those differences impact our social, legal, and interpersonal interactions; about the role moral and prudential reasoning can and do play in our decision-making; about the types of punishments that are justified and (a different question) most useful; about what is inevitable and what is changeable; and so on.

If skeptics about the libertarian sort of "free will" that is supposed to justify eternal damnation want to say that's the *only* proper referent of the term, so we should give it up, they can try to make that case, but I don't see the argument or data for this claim--the word is used (and I think *properly* used) to mark important distinctions in the issues I mention above. Unless we discover otherwise (and I don't think discovering determinism is true matters here), we do have psychological capacities (developed beyond any other animals') to control our lives in significant ways--at least if those capacities are properly developed--to reason about our futures and causally influence the way the future goes based on such reasoning, such that we are responsible in important ways that ground our interpersonal relationships, etc.

And to tell people (as scientists and the science media have been doing) that free will is an illusion will not be interpreted to mean we have everything we think except for some magical power to initiate new causal chains such that it would be justified for us to spend eternity in heaven or hell. Even religious people tend to focus on more earthly responsibility and control most of the time. Or so it seems to me.

[Kip]

Thomas and Derk,

Let's distinguish between:

1. We shouldn't change the meanings of terms if doing so would invite confusion (unless it's desirable confusion).
2. There is nothing wrong with changing the meanings of terms, such as free will and moral responsibility, provided that there is no confusion (everyone knows that "free will" used to mean X, but now it means Y).

With my moral realist hat on, I tend to agree with both 1 and 2. I think both of you are focusing on the situation in 1: the danger of confusion, as a practical matter, if people start using "free will" to only mean compatibilist will. I agree that this would have the problem of confusing the roughly half of the population reporting incompatibilist intuitions.

To my knowledge, Manuel has not addressed the question of whether his moderate revisionism would involve any confusion, and (if it did), whether this would be a bug or a feature.

[tnadelhoffer]

Kip,

It's not just that I think we invite confusion by revising concepts--it's that I think it is unduly concessive to the dualist/retributivist/libertarian/free-willers of the world when we continue using their terms even though we no longer mean what they mean by these terms. In short, I believe that just as evolution is hostile to religion (a la Dawkins), so I think the gathering data from the sciences of the mind are hostile to traditional conceptions of agency and desert-based responsibility. I simply don't want to hide this hostility under the banner of revisionism. Since I think the traditional conceptions of agency and desert-based responsible are bad for our species--even if they used to be adaptive--I would like to do my part to make these concepts go the way of phlogiston. What I don't want to do is call the caloric theory of combustion "phlogiston" when it would make more sense to just drop the term altogether from our vernacular so that it can wear its status as a by-gone relic on its sleeve.

[Kip]

Some more thoughts on this fascinating subject:

1. Like the Monty Hall problem, it might be (probably is) the case that the free will problem, as first stated, can't be solved, because free will is not properly defined. In that case, we can admit that, technically, the problem is undefined, while also emphasizing the most interesting interpretation of the problem.

Consider the Monty Hall problem. The original statement of the problem is unsolvable, because poorly defined. But once you grant certain intuitive assumptions, you get an unintuitive result: 2/3. This answer is so unintuitive that "approximately 10,000 readers, including nearly 1,000 with Ph.D.s, wrote to the magazine claiming the published solution was wrong."

In my opinion, the fact that so many people thought that that solution, based on those assumptions, was wrong, is a reason for granting the assumptions. In other words, we should define the problem in a way that gives the interesting or counter-intuitive result. Let's call this "Monty-izing" the problem.

2. I think *both* compatibilists and anti-realists are Monty-izing the problem. There is a genuine, counter-intuitive result at the heart of *both* compatibilism and anti-realism that the respective proponents feel is important to convey.

The compatibilist message (CM): even in a clockwork universe (similar to Conway's Game of Life), we can still have the fascinating array of freedoms and capacities that we enjoy every day. In other words, it's entirely possible that our world actually is deterministic. [Note that some compatibilists don't even think the above is counter-intuitive; and therefore want to spread the important message of its intuitiveness!]

The anti-realist message (AM): the anti-realist really has two messages. First, ultimately, in the end, it's all luck. Luck swallows everything. Second, humans are famously bad in making attributions about freedom, control and responsibility. We're plagued by a bunch of cognitive biases that that affect our reasoning about freedom and responsibility. And in particular, these biases tend to make us think that, at bottom, it's not all luck.

The closest we might come to solving the compatibilist-antireailist debate might be similar to the Monty Hall problem: granting that the original statement was incomplete, but asserting that one of the two above messages is more interesting and important, and therefore defining the problem to spread that message.

Of course, I tend to think the anti-realist's message is more counter-intuitive and interesting. Here are some factors to consider in deciding:

A. To what extent will the message accommodate or imply the other;
B. To what extent will the message affect our self-image;
C. To what extent will the message change or justify our moral practices;
C. To what extent is one message differentially more counter-intuitive than the other (on a scale from 1-10 of counter-intuitiveness, is CM a 9 and AM a 7)?

[Matt King]

Thomas,

Do you really think that the "dualist/retributivist/libertarian/free-willers of the world" are using a different concept of 'desert-based responsibility'?

It seems to me, agreeing here with Derk, that the debate is about whether or not we could be responsible for what we do (in the desert-based sense). Compatibilists have denied, principally, that determinism would rule out such responsibility. They have often argued this in virtue of defending a species of causation, freedom, or control sufficient to secure such responsibility. (Of course, not *ALL* compatibilists have tried to secure such responsibility, but there's an important sense in which they are not party to the debate, at least not in the same way.)

If this is right, then both sides agree on the concept of responsibility, at least to a large degree. What they disagree about is the notion of freedom/control necessary for such responsibility. (In this respect I think a lot of compatibilists are revisionists of a sort.)

And while we can note the data from the sciences that cast doubt on us having a more "robust" freedom/control, I don't know of any that cast doubt directly on our being responsible (in the shared desert-based sense). The most they can show is indirect: only on the assumption that such robust control is required for responsibility could such data show that we aren't responsible.

So skeptics think robust control is necessary; compatiblists don't. (Or: skeptics think the conditions are stronger than compatibilists do.) But they are still engaging the same concept (desert-based responsibility) - they're just disagreeing about the necessary and sufficient conditions.

[tnadelhoffer]

Matt,

My whole point is that when you, or anyone else, states that the issue is "whether or not we could be responsible for what we do (in the desert based sense)", it would be very helpful if you cash out "responsible in the desert based sense" in a way that neither trades on non-desert based forward looking considerations nor conflates the difference between being blameworthy (i.e., deserving of suffering/punishment) and being merely accountable (i.e., being an apt target of intervention, education, and rehabilitation).

The point of my original post was that compatibilists in particular are not always clear about these two kinds of different responsibility. They talk about desert--but in some cases, it is clear they are talking about the backward looking kind of retributivistic responsibility. In other cases, they talk about desert even though what they seem to have in mind is a forward-looking kind of accountability. My goal was to suggest that compatibilists who fall into the latter camp are not very different from free will skeptics. Both groups allow that (a) we have compatiblist control/reasons responsiveness, and (b) these are enough to ground some pared down kind of forward looking accountability. But just to say this is not to deliver the goods on desert.

By my lights, if compatibilists can't deliver the goods on desert, then they are all revisionists in the end. At the end of the day, there's a difference between merely giving us all of the varieties of free will and moral responsibility "worth wanting" (a la Dennett) and giving us all of the varieties of free will and moral responsibility that people ordinarily take us to have. Since I think (a) most people think we have full blown desert based heaven/hell style moral responsibility, and (b) compatibilist accounts of so-called free will are insufficiently robust to ground this kind of responsibility, I believe that many compatibilists are simply revisionists in disguise. And, as I tried to argue above, it is hard to see how the difference between the fw/mr revisionist and the fw/mr skeptic isn't simply terminological. The same can not be said for the difference between the libertarian and the skeptic. There, the disagreement is substantive. The libertarian says that sometimes making people suffer for what they do is intrinsically valuable given the nature of free will and desert-based responsibility. The free will skeptic says that this is never the case. In my own estimation, this is what was at stake in the original debate about free will. Perhaps this is why Kant and others thought the compatibilist was guilty of creating "a wretched subterfuge" or relying on "petty word jugglery."

[tnadelhoffer]

Eddy,

You state the following:

"There are questions about how we conceive of ourselves and our powers to control our own lives and to change our habits and situation; about the important ways we are different from other animals and how those differences impact our social, legal, and interpersonal interactions; about the role moral and prudential reasoning can and do play in our decision-making; about the types of punishments that are justified and (a different question) most useful; about what is inevitable and what is changeable; and so on."

No one denies that these are important questions. The issue is whether we ought to frame these questions in terms of free will and moral responsibility. Now, you say:

"If skeptics about the libertarian sort of "free will" that is supposed to justify eternal damnation want to say that's the *only* proper referent of the term, so we should give it up, they can try to make that case, but I don't see the argument or data for this claim--the word is used (and I think *properly* used) to mark important distinctions in the issues I mention above."

Here's a very simple argument. There is a core meaning of free will and moral responsibility--let's call this the radical core--that is arguably bad for humanity in that it tends to fuel a certain kind lust for revenge that lies at the root of both bad domestic policy (e.g., our criminal justice system) and foreign policy (e.g., our war on everyone who deserves to suffer for hating our freedom). Let's assume for the sake of argument that the assumptions about human agency that underlie this radical core are not consistent with the gathering data about human psychology. In this context, by continuing to talk about free will and moral responsibility in the less radical sense, we don't make it clear to people that the radical core they in fact believe in is total bullshit.

In short, by continuing to frame things in the old free will/moral responsibility/desert lingo, you occlude the fact that science really does threaten the radical core. I think this is a mistake--especially when you could simply use terms like conscious control, reasons responsiveness, accountability, etc. Why go the extra step and claim these are constitutive of free will rather than just saying these are the important features of our psychology that we have instead of or in place of free will? If you, too, think that what I am calling the radical core is bad for humanity, then it is you who needs an argument for continuing to use the very same terminology preferred by those for whom the radical core is the foundation of their beliefs about both themselves and others.

[tnadelhoffer]

Eddy,

It's worth pointing out that by "very simple argument," I meant an over-simplified-under-motivated-early-morning-argument designed just to keep the conversation going. I didn't mean "simply correct" (although, you know that's what I think)! :)

[Eddy Nahmias]

Much turns on whether you (Thomas) are correct that the magical self-creation (SC) powers we agree humans don't (couldn't?) have *do* in fact form the "radical core" of our notion of free will. It's funny you put it that way, because I've recently been toying with two pictures, one that shows these SC powers at the heart of a circle surrounded by the compatibilist powers of rational self-control, etc. The other picture has these compatibilist powers in the circle and the SC powers form a tangential circle touching the core circle like an epicycle.

I take the second picture to be more accurate. Many in religious Western cultures have internalized the philosophical/religious *theory* of SC powers that were developed, I would argue, to deal with the problem of evil and the problem of justifying eternal damnation. But this *theory* is: (a) not particularly intuitive on its own (not a product of our phenomenology or ordinary interpersonal interactions), (b) not required to ground our ordinary (earthly) practices of holding ourselves and others accountable (and *deserving* of credit and blame), and hence (c) not really at the core of our understanding of free will, but just an epicycle tacked on by a contingent and somewhat bizarre philosophical and religious history (Manuel, isn't this roughly the story we discussed a while ago?). As such, stripping it off involves a *minimal* revision to a useful and viable conception of free will.

Of course, if I am wrong (and the last paragraph certainly is full of speculation) and self-causal powers form the *heart* of our conception of free will, then ripping it out would be a radical revision, such that elimination of "free will" may be the more appropriate way to proceed. If I am right, then the elimination you suggest would in fact represent a serious *revision* (it's be more like saying phlogiston theory is essential to our conception of "fire" so we should eliminate "fire" and talk about "oxygen burning" or something).

I should add that the words "free will" may in fact have emerged to fulfill the theoretical purposes I mentioned above, but that does not mean that the term has not come to be understood to refer to our more ordinary capacities to control our actions and make choices such that we often *deserve* credit and blame for what we do.

[tnadelhoffer]

Eddy,

Actually nothing hangs on whether it is the radical "core" or merely a "tangential circle." The point is that many people appeal to these notions of MSC in their interactions with others. So long as calling the the non-MSC notions of control and accountability "free will" and "moral desert" (respectively) encourages those who believe in MSC to go about business as usual, then the onus is on you to explain what counsels continuing to talking this way when it seems to lend legitimacy to the MSC-ers of the world. After all, as I have said repeatedly, we already have different terms for the capacities you mention. So, I repeat: Why call these capacities free will when doing so runs the risk of encouraging the MSC-ers to continue thinking humans are the kinds of creatures who could deserve eternal damnation for their choices and decisions?

That being said, I am still entirely unclear why you state that free will skeptics are being revisionists rather than eliminativists. But I am heading off to class, so that is a battle that will have to wait until later on this evening...

[Kip]

Thomas,

Some points.

1. I share your progressive view on these topics. You and Derk and myself, along with Brian Parks and Tom Clark, are probably the only Gardeners adopting such views (in contrast to conservative anti-realists like Nichols, Smilansky, Strawson, Nagel, Double, etc.). I don't want my comments above to suggest a lack of sympathy for your cause.

2. Regarding revisionism, a concrete example might help. Suppose everyone in the world, almost magically, agrees to start calling dogs "cats" and cats "dogs" beginning tomorrow. Is there anything wrong with this (or are there moral reasons for or against)? I don't see any.

And I don't see how revising "free will", along these lines, would be any different. You say that, even in the absence of confusion, such revision would be a concession. But a concession implies that we're committed to "free will" meaning X, where X doesn't exist. We're not. Anti-realists are committed to X not existing, where X happens today to be symbolized as "free will". But we don't care if X is called "free will", "New York City", "Abraham Lincoln" or "shimboombawop." We shouldn't campaign to fix the referrent of "free will", always and forever, to what it means today. That unduly burdens our view.
In other words, if everyone decides tomorrow to use "free will" to mean New York City, that won't mean we were wrong. Free will--what we meant by free will--will still not exist. It will just mean that funny things are happening with words.

3. While I also believe that self-creation is an important part of our self-image, and our understanding of free will, I agree with Eddy that evidence for this claim is lacking. Here is one possible way to get such evidence:

I suspect that if we asked people whether a superpowerful alien could have designed our universe to fix and predetermine our life stories, most would say no. And most people would say no because they think, instinctively, of themselves as creating (in a buck-stopping way) their own life stories, rather than being players in a story that someone else has written. But it just ain't so.

[I find it very telling that Eddy seems to grant that the idea of free will might have *originated* in response to the problem of evil, and only later evolved to have a more compatibilist meaning.]

[Tom Clark]

Thomas,

Thanks for drawing out in this thread what I agree is one of the central issues in the free will debate, and for pointing up its social importance. The advance of neuroscience is putting increasing pressure on traditional dualist notions of self and control that are at least latent and often explicit in people's conception of free will. Thus far the literature seems to show most folks are *conceptual* incompatibilists who believe we live in an indeterministic universe (at least with respect to human decision-making), even if in emotionally freighted situations they often make compatibilist judgments about moral responsibility.

To the extent (and I think it's a great extent) that the term "free will" carries the connotation of being ultimately self-caused in our culture, continuing to use it obviously helps to justify non-consequentialist, desert-based punishment. But of course Eddy, Fischer, Dennett and many other compatibilists think we needn't be SC in order to truly deserving. Even though desert-endorsing compatibilists use forward-looking capacities (e.g., reasons-responsiveness) to pick out moral agents, they explicitly rule out the need for forward-looking considerations in justifying punishment. Very strange!

In any case, the non-revisionist recommendation of continuing to say that we have free will suits these compatibilists just fine, since the libertarian, "radical core" sense of the term serves to justify desert. If instead we simply talked about accountability, conscious control, reasons-responsiveness, etc. then the problem would immediately become apparent: what *is* the basis for desert in a likely deterministic universe? Is it because people are fully determined to be bad, or in spite of their being determined, that they deserve punishment?

Eddy,

You say that the theory of having self-caused (SC) powers is "...not particularly intuitive on its own (not a product of our phenomenology or ordinary interpersonal interactions)." Seems to me the opposite, since as many have suggested, the lack of introspective access to the neural goings on that determine our occurrent thoughts and decisions *does* make it seem that we are SC with respect to them. Lack of such access makes it seem like we're a categorically mental agent that comes up with a decision without being determined to do so, in which the self becomes a miniature first cause, a little god.

[Eddy Nahmias]

Thomas, here's a possibility. Suppose people think the compatibilist capacities *are* sufficient to justify (genuine) desert for earthly reward and punishment (not to mention what keeps getting neglected, ordinary reactive attitudes of genuine gratitude, resentment, etc.). They also can come to recognize that it would take magical self-creation capacities to justify eternal suffering (or to protect God from the problem of evil), and many also accept that we have such capacities, perhaps for religious reasons or perhaps because these ideas permeate our culture. Suppose they use "free will" language to talk about both sorts of capacities (just like philosophers do!).

In this case, to propose that we eliminate all talk of "free will" is a serious *revision* of our practices. We interact in terms of earthly rewards and punishments (and reactive attitudes) quite often. The eliminativist suggests we revise the language we use to do so (and risks leading people to believe those practices are unjustified and should be revised).

Of course, we may discover that we don't have the compatibilist capacities either, in which case elimination might be the appropriate response (by the way, why do you suggest that modern science may show we don't have free will? On your view, it shouldn't take science to show that, since you define free will in terms of capacities that can be argued out of existence without data).

Kip, I agree with your first sentence above: if we asked people, they'd say we couldn't be controlled by super aliens in this way. I disagree with your second sentence suggesting that this is because we believe in self-creation of a sort that goes beyond compatibilist accounts. I suspect most people would also say superaliens couldn't be controlling the movement of every particle in our universe. There are lots of reasons people might disagree with such a weird scenario. (But I'll be checking in with the folk about an alien control scenario later this semester.)

[Eddy Nahmias]

Tom C., briefly, I agree with you that we experience ourselves as mental agents (I also don't think this experience is an illusion; we *are* mental agents). But I disagree that this experience requires or leads people to a substance dualist position rather than, say, a nonreductive physicalist postion. And I disagree that it leads most people to think of their decisions as literally *uncaused* by anything (even their own prior mental activity). People think of themselves as agents and not just physical mechanisms. But without theoretical baggage thrown in, this does not (certainly need not) lead people to think of themselves as magical self-caused agents.

[Tamler Sommers]

For elimitivism to win out it seems we need to show more than:

(1) People take themselves to have MSC.

We also need to show:

(2) People think that MSC powers are necessary for (backwards-looking) desert.

You could also argue (as Shaun and Manuel do, and I sort of do) that even if (1) and (2) are false, you'd still need to show:

(3) Were people to discover that they have no MSC powers, they would not revise their judgments about the conditions of desert so that they were in line with compatibilism.

As Matt notes above, the concept of desert is holding steady here--it's only our conception of agency and of the conditions for desert that are either revised or eliminated.

[tnadelhoffer]

Eddy,

Do me a favor and briefly explain what you take to be "genuine desert" and what it would entail to say that someone "genuinely deserves" to be punished. Set aside the issue of eternal damnation for now. Let's just focus on what you think it means for a wrong-doer who satisfies the basic compatibilist conditions to "genuinely deserve" to be imprisoned in, for instance, a supermax prison--which is the typical destination for the kind of folks I suspect you think "deserve" punishment in the full blown sense. At the beginning of this thread, I asked revisionists and compatibilists to provide an account of desert that didn't trade on (or simply collapse into) the sorts of forward looking reasons for holding people accountable for their wrong-doing. Neither you nor anyone else has done so. Instead, it seems like everyone is saying that all parties to the debate mean the same thing by desert-they just disagree about the necessary conditions for proper attributions of desert. I am still unconvinced by this.

Keep in mind that I could talk in terms of desert when it comes to the difference between how I would respond to my four year old dog's pissing in the house and how I would respond to my twelve year old dog doing the same. In some sense, the younger pup "deserves" to be reprimanded in a way that my older dog does not. But that's a shallow kind of desert that can be entirely cashed out in terms of compatibilist-grounded forward looking considerations. But the important point is that however shallow it might be, this is nevertheless a way we do ordinarily talk about desert. Presumably, however, there is a deeper notion of desert--a notion that I believe is at the heart of the free will debate. That's the one I am after. And that's the one I don't think compatibilists can justify with their cognitive capacities approach to free will and accountability. You can obviously and admittedly use this approach to distinguish those for whom talk of accountability, rehabilitation, education, and reintigration makes sense. But what I don't think you can do is explain how your cognitive capacities approach undergirds the deeper sense of desert I am after. Libertarian free will, on the other hand, fairs fine in this regard--I just don't think we have it.

To repeat (hopefully, for the last time): There is a notion of desert in circulation (both among the masses and among people working on the philosophy of punishment) such that it is intrinsically valuable to make wrong-doers suffer for the harm they cause. Now quit being evasive and answer this question: Do you think the cognitive capacities you and other compatibilists and revisionists focus upon are enough to undergird this kind of desert? If not, why not?

[Matt King]

I won't be evasive, Thomas: I *don't* think agents can deserve punishment such that it is intrinsically valuable to make wrong-doers suffer - but that's just because I'm not sure it'd ever be intrinsically valuable to make someone suffer, even if they did act with the most robust account of free will on offer.

Incomaptibilists (skeptics included) sometimes seem to think that it's obvious that their conditions on moral responsibility are of the sort that would make it intrinsically valuable for wrongdoers to suffer. As Thomas claims: while the cognitive capacities of compatibilists can't undergird the deeper sense of desert he's after, "Libertarian free will...fairs fine in this regard".

But I don't find this obvious at all. Frankly, I'd like to see the story for why S's doing wrong with robust libertarian free will would explain why it would be intrinsically valuable for S to suffer. Perhaps such a story can be told - perhaps even an argument can be given. If so, that'd be great, because then we could evaluate the question on its own terms. And this would get at Manuel's first expressed interest - that too little has been said about what desert amounts to. That seems to me a question that is actually independent of what's necessary to be morally responsible, and one that incompatibilists and compatibilists alike probably ought to answer.

[Eddy Nahmias]

I agree with Matt that it is entirely unclear whether libertarian free will can ground whatever type of desert Thomas seems to be talking about (more on this in a post I'll put up soon that goes back to Tamler's discussion of TNR). If we are taking eternal damnation off the table (which libertarian free will could not ground), then it is unclear why an agent with "libertarian free will" *justifiably* deserves "intrinsically valuable suffering" any more than an agent with compatibilist capacities (someone please explain). But like Matt, I'm also not clear what to make of this notion of "intrinsically valuable suffering," or more importantly (as I've emphasized in earlier comments), why that notion alone should serve as the central issue in discussions of desert or free will.

So, I will answer Thomas' question by saying: I don't know whether the "compatibilist cognitive capacities" (CCCs) are sufficient to ground "this type of desert" because I'm not sure what that type of desert is or whether anything could ground it (or whether it makes sense to say that one of our ordinary ideas of desert is one that *could* not be justified). But I do think that CCCs, if we have them, are sufficient to ground the type of desert that is central to our ordinary practices, including (most of) our punitive practices (I'm against the death penalty too, but not because I think no one has free will). You scoff at the shallow notion of desert as applied to your dogs, but we are quite different than dogs, so if MR is a graded notion, then perhaps the grade of MR normal adults *deserve* (because of their more advanced CCCs) is significantly deeper than the one you mention. Can it be anything more than forward-looking? Sure, why not?

[tnadelhoffer]

Eddy,

You state "Can it [i.e., desert] be anything more than forward looking? Sure, why not?" This is precisely the kind of evasive move I have been complaining about all along. If the best the compatibilist can do in addressing this issue is offer a rhetorical question, then I am even less impressed with their ability to ground desert than I had been in the past! :)

That being said, here is a reason to think the kind of "intrinsically-valuable-suffering" kind of desert I keep bringing up is important: the research by social psychologists such as Carlsmith and Darley suggests this is precisely what undergirds folk ascriptions of desert and punishment.

It turns out that despite what people say, forward looking considerations have little effect on their intuitions about whether and how much someone should suffer. The only thing that matters is the moral outrage generated by the offense. When enough moral outrage is produced, making the violator suffer is justified for its own sake. It's not that someone ought to be punished because doing so will bring about some good above and beyond the fact that the violator has been made to suffer proportionality for his wrong-doing. The suffering is an end in itself...

That being said, I never said that the notion of intrinsically valuable suffering "alone should serve as the central issue in discussions of desert or free will" as you suggest--I simply said that compatibilists and revisionists who make it sound like their CCC account can deliver the goods on desert would do well to think about this kind of desert if they are going to insist that their views actually capture folk intuitions.

My claim is that the CCC account is not enough to undergird these intuitions and practices. As such, this would mean that all compatibilists are actually revisionists about moral responsibility (if not about free will). That's fine as far as it goes--especially, I take it, by Manuel's lights!--but then you really ought to quit pretending that you have the theoretical machinery at your disposal to capture the deliverances of common sense.

[Sofia Jeppsson]

I think the original post points to an interesting question. I really think it's common in the free will debate that verbal disagreements come disguised as something deeper.

But really, wouldn't philosophy do better without talk about what "our ordinary concept of free will/desert" is etc?

Now I've only read a teeny-weeny bit of experimental philosophy, but my impression is that so far there seems to be no generally agreed-upon conclusion of exactly what "people" mean by free will. And the same goes for desert and moral responsibility. Just asking questions in a way that doesn't push people in either a compatibilist or incompatibilist direction is extremely difficult, and it's hard to reach agreement between different philosophical camps on which is the neutral way to put things.

The discussion on desert also easily becomes so either/or: Either desert is something that could ground eternal punishment in hell or eternal bliss in heaven. Or it's just something we've invented to train each other by punishment and reward, much as a behaviourist psychologist could train his rats. My GUESS is that most layfolk have intuitions that fall somewhere inbetween these extremes, but are inherently vague, and could go in different directions in different contexts. But that's just a guess, and I won't pretend it's anything more than that.

[Mark Smeltzer]

Tom,

I think we would have to dissect the term "punishment" before we could draw any substantive conclusions on that point.

For instance, I generally think that desert does not warrant other agents to inflict physical pain upon evil doers, but I do think that desert warrants all sorts of punishments, even for intrinsic reasons.

Must punishment be identified with pain? For that matter, must capital punishment be physically painful?

To be clear, I am not sure that there isn't any intrinsic value in such suffering. However, I am inclined to believe that there is an intrinsic loss of value in the act of causing harm to other people -- such that whatever intrinsic gain there may be in causing the guilty to experience physical pain, it is lost in the very process of causing it.

[Sofia Jeppsson]

Another thought: I have the idea that there are at least three different dimensions where people can differ in their view on desert...

1. Backward-looking vs forward-looking justification for blaming or praising - perhaps this distinction is the same as distinguishing between final vs instrumental value in blame and praise, punishment and reward.
2. Desert as a social construct or an evolved part of our way of thinking, vs desert as some kind of metaphysical lump that sticks to people.
3. Is there any limit to how much people can deserve? Is desert just about praise and blame, or is it possible to deserve (if you're evil enough) eternal torment in hell and the other way around?

I think these three are independent of each other... although it might be that certain views on 1 tend to go together with certain views on 2 and 3 etc. What do you think?

[Paul Torek]

Sofia,

I'd like to stand your point on its head. I think that deeper disagreements can come disguised as merely terminological ones.

If a term, like "mass", is retained despite our being seriously wrong about the subject, this reflects the fact that what we still believe about mass is more important than what we no longer believe. For example, in Newtonian physics:
(1) gravitational force = G*m1*m2/r^2;
(2) F = m * a;
(3) mass is a fixed property of an object independent of the observer's velocity.

We had to give up (3), but (1) and (2) - especially (2) - are far more important. And we get to keep those. So the nearest thing to Newtonian "mass" is still called "mass", and rightly so.

Important features of a physics concept derive their importance from their theoretical role and ultimately their predictive power. In contrast, while "free will" does have some predictive uses, most of its importance comes from its roles in our ethics. (I'm just asserting this, but I hope it's plausible.) So if we disagree about the relative importance of "intrinsically valuable suffering" versus "just and fair backward-looking rules of punishment", this is probably an ethical disagreement. As a result of this ethical disagreement, we may disagree terminologically about whether compatibilist cognitive capacities count as "free will".

[Sofia Jeppsson]

Okay Paul, so you say it goes something like this:

1. People have different opinions on morality.
2. This leads them to different conclusions on what kind of free will is interesting, necessary etc, if that kind of morality is to be justified.
3. Everyone thinks that what THEY deem an important feature of a moral agent is properly called "free will".

If that's what you're saying I think you have a good point.

[Paul Torek]

Thomas,

If Eddy's answer doesn't satisfy, I'll offer another. It probably won't satisfy either, but it goes beyond firing back with a "why not?"

From my discourse ethics type perspective, a person "genuinely deserves" to be punished iff among the most justifiable set of norms is a backward-looking norm linking certain types of wrongdoing to punishment, and the person did in fact do such a thing. What distinguishes punishment from accountability is tricky, but among the clear cases, anything purely aimed at restitution does not count as "punishment".

When we justify norms, forward-looking considerations can and do come into play. That is why compatibilist cognitive capacities are important. "Punishing" the cognitively incompetent would be pointless cruelty. But this in no way implies that the norms themselves are forward-looking. It's easy to see good reasons why not.

If deterring crime were all we cared about, or were enough to override respect for persons, then those who were widely believed to be guilty could be "punished". That would certainly work out horribly for certain members of the moral community, for example Jews. (Lenny Bruce should be worked into discussions whenever possible. :) ) Clearly, such norms of "punishment" cannot be legitimate on a discourse ethics type approach, since those who would be sacrificed "for the greater good" could never agree to them.

Sofia,

You understand me perfectly, and your recap is marvelous. Strangely, though, your point number 3 is arguably not correct as it stands. It seems that free will skeptics think that what SOMEONE ELSE deems an important feature of a moral agent, is properly called "free will". The Someone Else in question, to whom we are supposed to defer in our terminology, seems to be the fire-breathing fundamentalists on the opposite end of the spectrum. Now that's strange!

[Eddy Nahmias]

Nice comments Paul and Sofia.

I just want to clarify that when I said: "Can it be anything more than forward-looking? Sure, why not?", I was not trying to be evasive; rather, I was turning the question back to the skeptic. The "it" I was referring to at that point in the post was the "type of desert that is central to our ordinary practices, including (most of) our punitive practices," that is "shallower" than the one involved in eternal (or any afterlife) punishment or reward, but that is deeper than what dogs have (because our significantly higher "grade" of CCCs).

Now, focusing on this sort of desert, I was asking the question of skeptics: Why *can't* it be anything other than forward-looking? I don't feel the burden to explain why it can be.