Blog Coordinator

« Thanks, and a brief introduction | Main | Some Recent Work on Agency and Responsibility »



Feed You can follow this conversation by subscribing to the comment feed for this post.

Compatibilists are Hard Determinist wannabes, lacking the courage of their convictions. They know darn well that a person determined to do things by forces beyond his control is not responsible for his actions. But they also believe in Determinism, wanting to appear scientifically respectable. (Or perhaps they embrace the Mind Objection and/or think Agent Causalism is medieval.) Now what? A world without responsibility is too horrible to contemplate. Hence Compatibilism. Anyway, that's how I rationalized it before I manned up and became an Agent Causalist.

Galileans are earth-skeptic wannabes, lacking the courage of their convictions. They know darn well that a planet not at the center of the universe cannot be Earth. But they want to be scientifically respectable (or perhaps believe the Ptolemeic system is medieval). Now what? The non-existence of Earth is too horrible to contemplate. Hence Galileans (redefine "Earth", oh wretched subterfuge). Anyway, that's how I rationalized it before I manned up and became a Ptolemite.

The analogy doesn't work that well, but Robert left me no choice but to try ;-)

Thanks Robert, that puts another suggestion on the table: compatibilist commitments might be rationalisations, stemming from of fear of a world without responsibility.

This is an interesting suggestion both in being based on personal experience and in potentially contradicting something I said in the post, namely that if some strict incompatibilist condition on responsibility were a central and obvious part of our concept of responsibility, more people would find compatibilism blatantly contrived or clearly revisionary. With enough rationalisation at play, that might not be true.

But I wonder why you say that it is *fear* that would push people towards compatibilism? To put Eddy’s riposte (as I understood it) in the form of a question: why not merely say that people are pushed towards compatibilism by the perceived *implausibility* of the implication that no one is ever responsible for anything?


Because of all the heavy duty real life ramifications of HD (laid out masterfully by Derk P. in LWFW) we cringe at the thought that it might be true. My students understand this concern with hardly any prompting. I would call a philosophy sans nihilistic overtones 'implausible' were it faced with some rather obvious counterexamples: I am not afraid of Coherentism. But HD violates the core of our very self-image, the notion that we are agents who make a difference in this world and, hence, are appropriate objects of the reactive attitudes. The problem here, then, goes well beyond our intellectual concerns. Let's cut to the chase- Derk's valiant attempt to make room for love and friendship notwithstanding, HD means that, as my students say, life is meaningless, and that, my friend, is the scariest of thoughts.

Hi Gunnar--what a privilege and delight to meet you at Manuel's wonderful conference* (yes--the best conference of any ilk I've ever attended, asterisk or no). Allow me to take the opportunity here to thank Manuel publicly for that.

And from that conference, I bet you can anticipate what I'm going to say. I don't know about fear, but given the small margins of argumentative separation between compatibilist/incompatibilist positions that genuinely underwrite claims like John's well-known points about dialectical stalemate, maybe pragmatic considerations win out to resolve such a stalemate. Now that does not say how such considerations will do this. That is one complicated affair, involving (probably) meta-questions of what constitutes pragmatically worthy options of personal and social practices, and too much to pursue here. But I do want to raise the possibility that pragmatism might triumph in the clash.


Here’s my quickest version: Humans, like many other highly social mammals, have a deeply ingrained “revenge/reactance” system that is easily activated, affectively charged, and difficult to suppress. It’s partly why you want to kick the object upon which you just stubbed your toe (and it’s why older siblings take things out on their younger siblings)! In addition to this system, we also have a folk psychological system that enables us to categorize some behaviors as intentional and others accidental, etc. Both of these two systems are deeply ingrained in us and both are easy to activate, etc.

In addition to these two systems, we have much more recently come to develop beliefs and theories about agency, responsibility, punishment, etc. These beliefs and theories are not as old as the aforementioned systems, so while we can be deeply committed to them they only have limited penetrability when it comes to my revenge/reactance response, for instance.

So, why do I feel the pull of compatibilism in affectively charged situations, because my revenge/reactance system has been activated along with my folk psychological system—that is, I want to harm (so I have a desire to discharge my disapprobation) and I view this person as an intentional agent who could serve as my target. But I feel all of this despite the fact that, theoretically speaking, I don’t think the person deserves to suffer in any deep, non instrumental sense.

p.s. I don’t think this is a standard dual track theory (although I think it clearly shares many affinities with those theories—which I admittedly tend to gravitate towards). But I obviously haven’t worked out the details.

Now that I have given it a bit more thought, I am not sure everyone needs an error theory after all. Or, at least it seems to me that I don't! Reconsider the account I mentioned above. It's not that I think the output of my revenge/reactance system is somehow in error. It's working perfectly well (given what it was "designed" to do)! The same goes for my folk psychological system. In this case, the violator really did do it intentionally rather than accidentally. So, that system is working just fine as well. In light of these two things I want the violator harmed at the basest and most brutish level. The primitive rule might be, Harm he who has harmed! But at a higher level of reflection, I don't think the violator *deserves* to be harmed even if I feel he should be harmed (for instrumental reasons). But just because I have theoretical commitments that are discordant with some of my basic, moral, psychological mechanisms (or systems), I don't see why I thereby need an error theory.

I only need an error theory if I feel like the violator *deserves* to suffer in the deeper sense of the word. But that's not the feeling I get. I simply want him to suffer harm plain and simple. I don't additionally think he deserves to suffer.

So, you present me with a case involving a violator and I have a prepotent desire to harm him (or even someone else). My desire to see the violator harmed is at a lower level than you seem to be assuming. At that level, desert can and should have nothing to do with it. The revenge/reactance system evolved before we could have had concepts like free will and desert.

As such, I don't see that I have anything that I need to explain away. I need only have some baseline attitudinal and behavioral responses to certain stimuli that I think can't be justified in desert-based terms. That's it. Why think I additionally need an error theory?

Robert: Thanks, I now have a clearer understanding of your suggestion, and it seems right to me that this is what is going on with some people. Moreover, a study by Roskies and Nichols suggests that subjects who are asked whether responsibility is possible in a deterministic universe are much more inclined to think that it is if the universe in question is described as the actual universe, rather than a fictive one. This suggests that a commitment to responsibility attributions might be at work, perhaps for the reason your say, as incompatibilism threatens that commitment most directly when the actual universe is deterministic.

But I wonder how much of compatibilist tendencies can be explained this way. In a variety of experiments by different research groups, a substantial proportion of subjects asked are willing to attribute responsibility to agents in fictive deterministic worlds, even subjects who say that our universe is unlike the deterministic universe when it comes to action. Perhaps it could be their fear of HD that explains this, but the explanation would have to be somewhat indirect.

Alan, it was a pleasure meeting you in SF.

A quick clarificatory question: your suggestion that we need to invoke pragmatic considerations to settle the debate about incompatibilism, did you intend it to speak to my question about how compatibilist and incompatibilist intuitions should be explained (away)?


Why would the explanation have to be indirect? Aren't the subjects 'guilty' of projecting the horror they feel towards the thought that the actual world exists sans responsibility onto the fictive deterministic worlds? The reaction here is perfectly general: any situation in which persons fail to be appropriate objects of the reactive attitudes is fearful- a scary possibility. If your question is how can anyone be afraid of a mere possibility, let me assure you that I have spent an inordinate amount of time fretting over scenarios that fortunately never came to pass. What's worse, it always takes me awhile to get over dodged bullets.


Thanks for mentioning my sketch of an error theory (involving the distinction between regulative and guidance control). I develop this error theory in greater detail in: "The Frankfurt Style Cases: Philosophical Lightning Rods", forthcoming in I. Haji and J Caouette, eds., *Free Will and Moral Responsibility* (Cambridge Scholars Press). Also, I'll be giving this paper as my lecture at the upcoming FSU graduate student conference on free will and moral responsibility, Oct. 11-12. I hope to seem many Flickerers there!!

I'm already enjoying your blogging, Gunnar!

I've sometimes wondered: why exactly do we all need an error theory (given that we do, which is also an interesting question). Is this just a general mandate accross all areas of philosophy? Presumably, in *any* interesting philosophical debate, there will be more than one "side" that is plausible. The great debates are about issues concerning which there are various plausible responses. Does one *always* then need to offer a psychological explanation of why others have adopted views different from one's own? Is this a general "obligation" whenever one argues for a position in any interesting philosophical debate? Might one wonder why we need to be in the business of offering explanations of this sort, although I do feel the pull of the notion that error theories seem to render one's own view more compelling.

Thomas, thanks for your suggestion. I’m tempted myself by something like the aetiology you sketch as part of a full answer, and I might say a little about this in my next post.

I think that your point about not needing an error theory is an interesting one. So it might be that you don’t need an error theory for you own case because you don’t have compatibilist inclinations in the sense of inclinations to think that people might deserve harm for doing bad things if determinism is true. While you have inclinations to harm, these are not inclinations to think that people deserve harm.

But perhaps you still think that *others* might have compatibilist inclinations proper because their judgments of desert are influenced by their revenge/reactance system? (Another possibility would be that (most) apparently compatibilist judgments are concerned with something other than responsibility in the relevant desert-entailing sense, perhaps with the appropriateness of some modest displays of revenge/reactance system.)


I would want the revenge/reactance system to be completely theory-neutral. It's more brutely motoric than anything else--it's like an explosive chain reaction is caused by a harm within a social hierarchy. One must harm he who harmed. If he who harmed is lower down the rung, then the harm will be easy to rectify (but notice the offender who has been harmed may very well harm another further down the ladder just to vent). And if harming he who harmed isn't possible--e.g.., because the violator is further up the hierarchy--then the harmed may simply discharge his anger and resentment by harming someone lower down the rungs of power, which in turns leads to more discharged anger on the "under-deserving."

Eventually peace is made somehow, but for a moment, before the group descends once again into grudge--filled chaos based on norm violations! Or that's how I see that particular system. It's like Newtonian clockwork--harms caused (whether intentional or accidental) are met with harm to someone, somehow. It's why they used to punish inanimate objects that had caused harm. It's a discharge of moral steam embedded in a social scene.

It was only later that folk psychology took shape and our interest in intentions, decisions, choices, reasons, etc.

And much later still when "free will" and "desert" were invented, or so it seems to me :)

But if this is right, then I don't need an error theory for any of the outputs of the revenge/reactance system. That we prepotently respond to norm violations with anger is just a brute fact about human psychology. Mutable, but brute nonetheless. While this is a feature of psychology I would very much like to understand and explain. I don' see it as something I ought to be trying to explain away. There are no errors here. Only mechanisms that may have lost their evolutionary niche. But, that is a story for another day! That said, thanks for the interesting post!

John, thanks for the reference! (Any chance I could have a sneak peak?)

When I said that everyone needs an error theory, I had this in mind: given how prominent both compatibilist and incompatibilist intuitions are, a case for the either sort of view would be considerably strengthened if it included a credible explanation of why the contrary intuitions are there, an explanation that makes it likely that these intuitions are not veridical. Why strengthened? Because contrary intuitions are prima facie evidence against one’s view, being cause for caution absent sufficient reason to ignore them. It might of course be that there is overwhelming evidence on one’s side already, evidence translating into evidence that the contrary intuitions are wrong (and so into evidence that *some* error theory is true of those intuitions). But I don’t quite see that sort of antecedently overwhelming evidence in the case of compatibilism or incompatibilism.

A quick question about you error theory, if that’s OK: Suppose that most incompatibilists (here: laymen who find moral responsibility undermined by determinism) understand that when prior events determine an agent’s actions, they do so by determining what the agent wants and believes, which in terms determines what the agent decides, which in turn determines how she acts. I wonder if this would be a problem for you account, suggesting that incompatibilist intuitions do not rely on thinking that determinism rules out guidance control? (The reason I ask is that I have some empirical evidence that this supposition might be true.)

Thanks Thomas, I see how I should have been clearer. I understood that the revenge/reactance system does not itself output judgments of desert on your view, and so nothing that would be in error in the absence of desert.

What I thought was that your proposal was compatible with the idea that the outputs of the r/r system *influence* judgments of desert. Perhaps when someone feels the urge to hurt someone, there will also be a need (through dissonance aversion, say) to think that hurting would be socially appropriate, which in some social contexts (after the invention of "free will" and "desert"!) will trigger the thought that the intended harm is deserved. The idea was that this could be an error theory for certain compatibilist intuitions or judgments, even if the r/r system is error free, and an error theory closely aligned with your initial suggestion.

Robert, I had assumed that the fear you mentioned was primarily one of *living* in a world without responsibility, but you are right of course that one might find a fictive universe scary. That's an interesting suggestion; I have to think more about it. Thanks!

You're right Gunnar on how you read me; the contrary intuitions and the relatively small margins of argumentative force between them (given pervasive dialectical stalemate--which some I take it would deny) show that they both probably are best explained away and replaced with pragmatic concerns. So maybe the motivation for error theory presupposes that one view might eventually win out, but my advocacy of pragmatism supposes the opposite, based on considerations like the evidence for ongoing dialectical stalemate.

Great question and thread.


Good question about my Error Theory. The Error Theory I propose has it that people in general either don't distinguish regulative from guidance control, or, if they do, suppose that the two necessarily go together. They thus make the error of supposing that moral responsibility requires regulative control (given that they hold that moral responsibility requires control).

But I agree with you that my error theory does not apply to those who are incompatibilist for source reasons. I guess I would have to argue that such individuals make the error of supposing that ultimate sourcehood requires indeterministic causation. That is, given that responsibilty requires ultimate sourcehood, there are (I have contended) different accounts of ultimate sourcehood, some of which are compatibilist-friendly. Again, the error would be either not distinguishing the different kinds of "ultimate sourcehood" or, if one does distinguish them, failing to see that we don't need the indeterministic notion.

Harry Frankfurt essentially proposed an "error theory", although he didn't call it that. He distinguished between doing something when you couldn't have done otherwise and doing something *because* you couldn't have done otherwise. I suppose that his suggestion essentially was that many folks make the error of not distinguishing these, or, if they do, falsely supposing that it is merely doing something in a context in which one can't do otherwise that makes one not morally responsible.

I guess the idiom I'm looking for here is 'Perish the thought'. Or 'God forbid'. A cliched way of putting it might be 'Let's not even go there' (where 'there' refers to some horrifying circumstance nearly avoided or yet to come to pass). These expressions clearly indicate that people often mentally approach certain non-actual possibilities with a significant amount of trepidation.

Robert, that sort of mechanism no doubt exists, and I have invoked it myself on occasion. I'll have to think more about its workings in this case, and about the grounds for thinking that such a reaction would constitute an error in this case (apart from other independent grounds for accepting incompatibilism). Thanks again for the suggestion.

John, thanks the clarification. In a way, I think that something like different ways of understanding sourcehood is doing a lot of work in explaining the dialectic. More on that in my next post. (As for Frankfurt's suggestion, I have found it interesting, but I am not sure what the mechanisms involved in the mistake would be, and have doubts about its scope.)

I’m attracted to what Thomas proposes as the first stage of the story, that it begins with a revenge/reactance system that doesn’t generally feature erroneous presuppositions or associated beliefs. My sense is to side with Gunnar’s suggestion about the next stage. We’re social creatures with an instinctive sense that our reactions to others need to be justified, in particular when they are harmful, and this gives rise to the thought that harmful reactions are deserved just because of the wrong done. This sort of desert provides a simple and general sort of justification (or rationalization) that’s not intellectually taxing, unlike, say consequentialist or contractualist justifications. And if this is the etiology of basic desert justification, it’s hard to see how it could have much ethical clout, and hence the error theory.

Hi Gunner,

Thanks for the nice question you ask here.

A quick idea about error theory. I take it to be similar to the following situation. A student comes to a teacher declaring he finds a way to trisect an angle with ruler and compass. Of course the teacher can just say that it has been proved impossible. But this may leave the student unsatisfied even if the student is convinced by the proof of impossibility of trisection. Another way could be to help the student realize where he goes wrong in his procedure of trisection, which could lead to a stronger intellectual satisfaction.

As for compatibilism/incompatibilism. I am not sure if I totally agree with the way you put it (although it is surely very plausible and most popular). Of course, there exists a line (maybe a little bit artificial) between these two theories. However, I wonder if there is a natural line or intuitive line between them. On my view, human (rational) agency is definitely unique and different from the rest of the world. But how unique? Libertarian has very optimistic answer. Compatibilist has moderately optimistic answer. So, I am a libertarian who believes in very optimistic view about human agency. It doesn't stop me from thinking that compatibilist view is not that bad if libertarian view fails to work out.

Why should we assume that there is only one universally valid or a priori notion of free will (or only one notion of moral responsibility)? It depends on how we understand human agency, doesn't it? And such an understanding could be flexible, right?

For example, if we dig up a kind of ancient machine which we know little. After observing it being *operated* for one time. We can certainly see part of it starts to move and does something. Finally, the process of "operation" causes an external change in the environment (e.g., making a noise). Based on this, can we determine if this is a well-functional machine or not? I don't think so. It seems that it can be a well-functional machine if we understand it as a noise maker. However, it may not be the way of how ancients understood it. Maybe it is a traffic machine for them, and it is no longer well-functional. It seems that no universally valid standard is available for us to determine what we mean by a well-functional machine, even if it seems to manifest a disposition of causing something to happen in a reliable way.

Maybe this example applies to human agency, maybe not. I am curious about how others would think about it.

Hi Jiajun,

Thanks for chiming in. I'm inclined to agree with you that there is more than one notion of free will and one notion of moral responsibility (though matters a tricky here, and depend on how we individuate notions). I also think that your points about different ways of understanding human agency might be relevant in relation to what I will say in the next post.

I'm inclined to accept something like the Eddy et. al. view (John seems to endorse it too, in addition to Dylan). But as I was reading this (wonderful) thread, especially reading your exchange with Thomas and Derk, I started to have my doubts. Here is one problem that might potentially favor the Thomas et. al. view over the Eddy et. al. view.

Whatever it is that is going on in normal cases has a parallel in what we might call "clear cases of nonresponsibility." Consider something like the Robert Harris case but personalize it. Suppose that you were, or someone you loved was, one of Harris' victims. Thus, it isn't just that Harris (we'll call him) has done something horrible but that he has done something horrible to YOU, either directly or indirectly (by horribly harming someone you love).

In this case, I'd find it very hard if not impossible to turn off the "revenge/reactance system" no matter how much philosophical speculation I do about the true meaning of "can" or "control," even well after rationally coming to the conclusion that Harris was not blameworthy given his life and circumstances. Personally, were I a victim (or loved one of a victim) of something truly horrific, I would find forgiveness not just difficult but completely unrelated to the kinds of rational thought needed to explain away our intuitions were the Eddy et. al. view correct. That seems to suggest that the revenge/reactance system plays some significant role in my beliefs and attitudes and that the Eddy et. al. view can't be the whole story (even though I'm committed to something like that story).

The comments to this entry are closed.

Books about Agency

3QD Prize 2014: Marcus Arvan