Show-Me The Argument Against Compatibilism

Patrick Todd over at Show-Me The Argument has posted an interesting argument against compatibilism based on the scenario used in Mele's Zygote Argument.  Go check it out!

Comments

[John Fischer]

Patrick's argument is fascinating. It build's on Mele's Zygot Argument, which I've sought to address elsewhere (in my APA Pacific comments and a forthcoming MIND review). Here are some rough and choppy first thoughts. Thanks, Patrick!

I do think there are cases in which two (or more) individuals can each be fully responsible for some upshot, although not solely responsible. So if two individuals simultaneously and independently shoot the mayor, then each is fully but not solely responsible for the mayor's being shot, in my view at least. Now it is kind of interesting to think about whether (given that what I've said about the two independent shooters is true) the case of Diana and Ernie is like this. Perhaps cases of independent simultaneous overdetermination are relevantly different from the case of Diana and Ernie--I'm not sure.

Distinguish: (A) two independent causal sequences issuing in an upshot.
(B): two components of a single causal sequence issuing in an upshot

In (A), I want to say that there can be two individuals who are fully responsible for the upshot, but not (of course) solely responsible. Although they are not solely responsible, they are also not partially responsible.

In (B), it is not clear what to say. Sometimes, there are various contributors who are partially responsible. But perhaps in a case like Diana, it is like A: Diana might be fully responsible for Ernie's behavior, and Ernie might also be fully responsible for it. How could I argue that this is like (A)? Well, since Diana brings it about that Ernie meets the conditions for (say) guidance control, I would use the sort of analysis I have developed previously: Diana's actions don't in any way distort or impair Ernie's exercise of the distinctive human capacities for control...

Maybe responsibility comes in packages. The packages are sufficient causal conditions for upshots, perhaps together with the background conditions that are necessary. (This is rough.) When you open the packages, there are "presents"--moral responsibility. Each package (of the sort described above) contains a full measure of responsibility. So, when there is overdetermination of the sort in the (A) scenarios, there is more than one package, and thus more than one measure of full responsibility. When there is just one package, there is just one full measure, and it has to be divided (ala a zero sum game). This would go along with Patrick's picture about the B-scenarios,. I guess I'd have to distinguish B-1 type scenarios from B-2 scenarios:

B1: two individuals put in parts of a total package that is causally sufficient for an upshot, but each part is not in itself sufficient, and each part does not cause the other.

B2: two individuals are along a single causal sequence to an upshot, where each individual part is necessary but not sufficient for the upshot, but one part causes (against a background of other conditions) the other.

B2 seems more like the case of Diana and Ernie. Now the question becomes: can one say what I want to about A about B2?

Well, how about this? In the B-1 scenarios, there's just one causally sufficient package, so just one full measure of responsibility. Happy birthday, but the present must be divided.

In the B-2 scenarios, although there's just one causal sequences, one can analytically divide it into two packages that are causally sufficient in the relevant way: the Diana package and the Ernie package. So, again, Happy Birthday, but there are more presents, and each can have a full one.

Sorry that the above is a big epigramatic.

[Neil]

We reject the ''I was only obeying orders" defence when it comes to war crimes. That is, we hold the foot soldiers who carried out the orders fully responsible for the crimes. But we don't conclude from their full responsibility that their commanders are not also fully responsible. So the principle that when an agent who is proximally causally responsible for an action is also fully morally responsible for it, agents upstream cannot also be fully morally responsible for it is false.

[Cihan Baran]

Neil,

I don't quite get the analogy. Of course, two different parties can be fully responsible for an action. That is to say, if two thieves co-operate in the robbing of a house, they are both fully responsible. Just because there is more than one individual, we don't divide the responsibility. In fact, I think (but not sure) punishments for organized are even heavier than individually committed crimes. This much seems elementary.

Where I think your analogy deviates from the case of Diane and Ernie is that Diane has way more power over Ernie than the commander does over the soldier. Perhaps, a more appropriate analogy might be this: Imagine that a slave owned by vicious fascists is forced to carry out crimes against humanity. Now in this case, it seems harder, intuitively, to blame the slave but it also depends on slave's moral character. Does the slave identify with her acts or does she feel compelled to do them?

Of course this analogy also falls short, since we would think that if the slave identified with her acts, the source of her identification would be herself.

A better analogy: Now imagine that the slave is being raised by fascists and she gets all her values and education from these fascists. I think in this case, even if she did identify with her acts, we wouldn't blame her - hence why we shouldn't blame Ernie.

Prof. Fischer,

I don't really think you can divide the "packages" that way. Or I am not really sure that your conceptual analysis accurately captures what's going on. Here is a different way to look at it.

Now, in one of your papers, you define "Transfer of Non-responsibility". Inspired by this, I'd like to speak of another principle, "Transfer of Responsibility." Now, I don't quite know what the relevant conditions for Transfer of Responsibility is and I am not going to identify them in this post - since I haven't really thought this through.

However, intuitively, this much seems to be true: When we perceive that A has been a mere instrument for B's purposes, we seem to think that responsibility transfers from A to B. So, let's suppose that when A opens the door, this will activate an electrical signal that will detonate a building due to the bombs B has installed. In this case, even though what A did was causally sufficient to destroy the building, the responsibility transfers to B.

Now, we might think of more complicated cases. Suppose that A and B are terrorists. A is an assassin but he fears that he might have doubts and fail at the last moment. So, when the time comes to shooting the mayor, B, having previously consulted with A, activates the electrodes in A's brain such that A shoots the mayor as a result of the random spasm in her finger.

In this case, I think the intuition is to also blame A ,since he also identifies with the deed. So the responsibility does not transfer to B.

Now suppose, however that B not only brings about A's random finger spasm but also, due to childhood brainwashings and various false conspiratorial theories about the world, A's identification with the deed. Now again, the transfer seems to transfer to B.

Back to case of Ernie and Diane. Even if there are two packages, as you say the Ernie package and the Diane package, since Ernie is a mere instrument for Diane's purposes and Diane brings about that Ernie identifies with Diane's purposes, the responsibility of Ernie package transfers to Diane.

Of course, this principle, transfer of responsibility, requires origination of responsibility - something that a hard determinist is bound to reject. Nevertheless, my point is that if there were responsibility, it would just transfer into the vacuum of causal chains before our birth. Perhaps, if an omnipotent god originated responsibility, everything would transfer to him/her.

[Neil]

Cihan, your ''better analogy" begs the question at issue. In your example, the slave is not responsible on grounds that a compatibilist (of a certain stripe) can recognize: due to a causal history which bypasses reasons-responsive mechanisms. The original case wasn't like that. Patrick's case was designed to avoid begging the question, by assuming that relevant compatibilist conditions were satisfied. His point was that if these conditions were adequate that Diana would be able to avoid responsibility by pointing out that Ernie was fully responsible. My case (and Neal's hitman example) show that he is wrong, unless he can show a relevant difference between the Ernie case and both the soldier case and the hitman case such that in the first alone responsibility is zero sum.

[Cihan Baran]

Neil, I'll grant you that my "better analogy" begs the question. (I am not quite sure of this because I am not quite sure about the term "reasons-responsive". In my opinion, Eddy made a similar objection to Alfred Mele in the relevant reading group, which Kip and he convincingly replied. The slave in my example has reasons for why she acts and she identifies with them. Or set the case that she is forced acquire to certain reasons, if you wish.)

However, I am confident about my central point: that is to say, if you bring about that someone does X AND someone's identification (whether it is reasons-responsiveness or other conceptual conditions) with X, that person can't be responsible for X.

[Neil]

Ernie need not identify with his action. Diana might create him so he does it in fit of akrasia. BTW, your principle threatens to generalize to many cases in which parents raise their kids to (say) carry on the family business .

[Dennis Millarker]

I think Patrick might really be on to something, here. While I agree with John Martin Fischer that it's possible for two people to be fully responsible for an upshot, I think that what Neal has pointed out is more problematic for the compatibilist. The way I see it, a compatibilist reading of the zygote argument has two people fully responsible for one decision: namely, Ernie's decision to A.

It seems starkly counter-intuitive that anybody but Ernie could bear full responsibility for Ernie's free decision.

[Kip Werking]

1. I agree with Dennis about the relevance of the distinction between full responsibility for outcomes and full responsibility for decisions. This is a distinction I've tried to make before.

2. Neil, I have several points to make:

2a. Your counter-example relies upon current practices. Currently, we're willing to hold both commanders and their subordinates fully responsible. But, from this empirical claim, how far do we get to the normative claim "we should hold both fully responsible?" In other words, might this practice be mistaken?
2b. Your counter-example is also based more upon practices than beliefs. We hold both commanders and subordinates responsible for *pragmatic* reasons. But even the more skeptically inclined, like Tamler and myself, might be willing to hold people responsible, without believing, in our heart of hearts, that they actually are fully responsible (e.g. Rawls and tellishment). Indeed, I don't think commanders or anyone else really believes that anyone is fully responsible for what happens; rather, their thinking is something like "you aren't fully responsible for what happened, but because what happened is so costly, and we are so bent upon preventing these things in the future, we are adopting something like a strict liability policy in order to terrify you into being on your very best behavior. You *will* be held accountable, even if you shouldn't be, and so you better do your very best."
2c. The question here is compatibilism about free will, not moral responsibility. It might be (as some like Thomas Nadelhoffer seem to suspect) that moral responsibility admits more of a compatibilist analysis than free will does. It certainly seems bizarre to me to say that both Diana and Ernie freely willed the same decision. Instead, something like Smilansky's Principle of Sole Attribution seems to capture an important part of "free will"'s content.
2d. Is it not relevant that the interaction between Diana and Ernie is so much more manipulative and insidious than that between a commander and subordinate? Even if we might be sympathetic to the commander case (despite all of the worries I've raised above), it seems relevantly different: the soldiers are interacting as agents and Diana has infinitely more control over Ernie than the commander does over the subordinate. The commander still has the option to say "no, I refuse to take your order" in a way that Ernie, designed by Diana for her specific purposes, does not. In other words, designing one's reasons-responsive mechanism to ensure a specific outcome seems just as disturbing as *bypassing* a preexisting mechanism. It's a mistake to think "just because there was no bypassing of a preexisting mechanism here, there is nothing to worry about."

3. I've always found the Zygote Argument (and others like it) to pack quite a punch. This is why hard compatibilism is called "hard" and even compatibilists like Gary Watson describe the idea of total-life-design to be "vertiginous". Something like the following empirical claim seems true:

THINK: We do not think of ourselves as the kind of agents who could have had their entire life stories written and designed by another (more powerful or intelligent) agent. We write the stories of our lives in a "real time", "buck stopping" way that is incompatible with others having written our entire life story for us.

FACT: But, we actually are such agents.

THINK, in combination with FACT, should generate a feeling of *surprise*. It should generate a remarkable emotion of *discovery*, like discovering that the earth actually circles around the sun. This is why Watson describes the idea as "vertiginous."

Now, THINK and FACT alone are not enough to generate the denial of free will. It might be we think we have super-duper amount of control and unpredictability, we're wrong about that, but "free will" only requires something short of this---something we do actually have. We could call this the PHEW view ("phew, that was a close one"). This seems to me very unlikely, because it seems that "free will" and "moral responsibility" would naturally track onto the powers and responsibility humans think we have---to whatever degree those may be.

But the compatibilist might also just deny THINK. When people call hard compatibilism "hard", and when Watson says the idea is "vertiginous", the compatibilist might say "hard!? vertiginous!? I've always thought that my life story was completely fixed by the state of the universe before I was born. This doesn't *surprise* me. And because there is no surprise, there is no threat to my beliefs about free will and moral responsibility that I need to address. I don't even empathize with, or understand, people who feel surprised at the discovery of FACT." We can call this view SCOFF ("I scoff at the idea that this should be surprising.")

What's interesting to me is that I'm not sure whether compatibilists tends to adopt (i) revisionism (either explicitly, like Vargas, or less explicitly, like Dennett, or perhaps even unwittingly), (ii) PHEW, or (iii) SCOFF (or perhaps even some combination of these). I would be very interested to hear what any compatibilists have to say about these possible responses.

[Roman Altshuler]

Maybe this has already come up in one of the (so far) 44 comments on the two sites, but I've only noticed an oblique reference to it (in Cihan Baran's post, above). I'm not sure that the argument, if it is meant to be a libertarian argument against compatibilism (rather than just an argument against currently dominant compatibilist theories) is legitimate, because so described it appears completely question-begging.

The set-up of the argument seems to assume that Diana has libertarian free will. The compatibilist can make decent arguments for insisting that Ernie is both free and fully responsible. And I agree that the claim that two individuals cannot both have full responsibility is not an argument so much as a questionable intuition--especially questionable since the cases that could test the intuition seem to be imaginary ones; by the way, weren't philosophers supposed to get rid of completely imaginary thought experiments on the grounds that the intuitions they generate are entirely unreliable?

In any case, the compatibilist is in trouble the moment he or she admits that the set-up of the argument is plausible. The point of the argument seems to be that Diana is fully responsible and therefore Ernie cannot be. This assumes that Diana has libertarian freedom. As soon as we agree to that possibility, the compatibilist has conceded to the libertarian. To admit that libertarian freedom is possible, I think, simply is to admit that compatibilist freedom is not "true" or "genuine" or "complete" freedom, because there is a greater kind of freedom.

The correct answer is for the compatibilist to say: Ok, this argument works IF you can explain to me how Diana happens to have libertarian freedom and what this freedom is like (and appealing to her divine status isn't enough; there is a reason we separate philosophy from theology). If there is a reasonable answer to this, then the compatibilist will need to take up the other alternatives.

What if Diana does not have libertarian freedom? What if her freedom, too, is compatibilist and some sort of greater power shapes her behavior? Then the question of whether or not Ernie is responsible just gets moved a step back to Diana, and the libertarian gets nothing out of the argument.

Am I missing the point? The libertarian can just say: right, but this is only an analogy for the situation we are really all supposed to be in, according to the compatibilist, where all our actions are really determined by natural factors. But that doesn't work: the argument relies for its entire force on our intuitions about responsibility, and we can't assign those intuitions to nature. In fact, the argument only appears to create a dilemma for compatibilism because it places a fully (libertarian) free agent at the beginning of the causal chain. This allows us to take the responsibility off of Ernie and place it in Diana's lap. That doesn't work in any kind of normal case, where the causes of a person's agency are not themselves agential.

Moreover, the argument seems to assume that there can be different degrees of freedom within a causal chain (Diana's libertarian and Ernie's compatibilist freedom), but not different degrees of responsibility (only one can be fully responsible). Why is that? Apparently, because the argument already assumes that Ernie cannot be free because there is a more free agent in the system. If we assume that there is a completely free Goddess (or God) who determines all our actions through a determination of sufficient conditions for those actions, then our freedom is indeed highly questionable. But its questionability here depends on an illegitimate comparison.

As I hinted at the start, I do not think the argument is question-begging if it is not meant to vindicate libertarians against compatibilists, but simply to attack current compatibilist theories. But if it is attempting the latter, then I think it fails for other reasons. The compatibilist already concedes that free agents lack the power to shape all causal factors relevant to their actions. That's the point of compatibilism. Simply pointing that out changes nothing.