Compatibilism is the position that states that free-will is compatible with determinism. Incompatibilism is the position that states that free-will requires indeterminism. It is hard to give an argument for either compatibilism or incompatibilism that does not beg the question. A good argument for incompatibilism would be to show that the compatibilist is committed to a particular idea that cannot be made sense of within a compatibilist framework. This is what (I think) Saul Smilansky tries to do in his paper “Determinism and Prepunishment: The Radical Nature of Compatibilism”, which you can find and read on this page:
http://philo.haifa.ac.il/faculty_pages/smilansky/determinism.pdf
It’s only 3 pages, and like all Analysis articles is wonderfully clear. No excuse for posting in this thread if you haven’t read it. “Prepunishment” refers to punishing someone prior to their committing a particular action. This is what I take to be Smilansky’s argument:
1. If compatibilism is true, we can have no moral objection to prepunishment
2. We have a moral objection to prepunishment, and we don’t want a theory of free-will that does violence to that intuition
3. Therefore, compatibilism is false
That argument looks valid to me. The premises also seem true to me.
In defence of P1, which I take to be the controversial premise, Smilansky says this:
But what if there is only the sort of free will that is compatible with determinism? Given determinism and complete predictability, compatibilism seems to lack any principled way of resisting the temptation of prepunishment. The mere temporal matter, that in prepunishment there is still no crime, and no one is yet deserving and blameworthy, does not seem to go very deep. Compatibilism allows that (a) a person’s committing of a crime is completely determined, and yet (b) he commits it out of his own free will, hence he is morally responsible for it, and liable to blame and punishment. If we perfectly know now that it is completely determined that a person will commit a crime in (say) a week, out of his own (compatibilist) free will, the compatibilist does not have a strong principled objection to prepunishing this person now, before he has actually committed the crime. The common-sense objection, that we must allow him to change his mind, does not apply here; for according to compatibilism it is already determined that he will not change his mind. Thus, there seems to be no point, from a compatibilist perspective, for waiting.
I think that this is potentially a very good argument, because the compatibilist typically thinks that if determinism is true, it does not affect our moral practise in any significant way. Indeed, some compatibilists (I’m thinking especially of Hume here) seem to say that our ordinary moral practise is dependent upon compatibilism being true.
What do you think?
(as for my own view, I used to be an extremely staunch compatibilist. Now, not just because of Smilansky’s paper, I am not so sure)
[...] Prepunishment as an argument against Compatibilism « no comments yet « links for 2009-04-19 [...]
Without reading the paper – sorry: lazy – the defence of the crucial premise requires that determinism extends to our judgments of responsibility, that is, that if we know whether someone will do X, we know whether they will be responsible for having done X. Davidson would deny this, I think. He thinks we can hold that the physical world is a closed system governed by strict causal laws, that the mental interacts with the physical, and that the mental is autonomous by insisting that there is no way of making law-like statements connecting the mental and the physical. That we could predict every physical event ever would then not imply that we could predict any mental events.
There’s also the possibility that the right to punish doesn’t depend on whether some act was committed or not, but that we can know that it was committed. Kant may well think that, for example.
Finally, I don’t see what’s (wholly) wrong with the ‘no crime’ answer: we think it’s perfectly appropriate to act in certain ways to people with certain features – to avoid people whose appearance we find repulsive, for example – whether or not they deserve that treatment or not. The fact is what grounds the treatment. Why could the fact of having committed a crime rather than not having committed it ground that treatment? After all, it may often be luck that distinguishes whether we committed crime X or Y (murder or assault, for example).
Rob,
I’m not sure Davidson can help here. Doesn’t his anomalous monism concern the *language* in which we express predictions? Yes, we can’t formulate strict psychological or psychophysical laws, but this doesn’t mean we can’t make predictions – it just means that they must be couched in purely physical language. We could still (in theory) predict what people will do, we just couldn’t include any propositional attitudes in our description. We’d still know that Smith would pull a trigger, this would cause a bullet to be fired, the bullet would hit Jones, and Jones would die and so forth.
If we want to say that mental event M1 caused mental event M2. If Davidson’s right, we aren’t going to get a strict law of the form M1-> M2 etc. But M1 will be token identical with some physical event P1 and M2 will be token identical with P2, so we can formulate a strict law P1 -> P2. Is that not enough?
If responsibility were a judgment about physical events, then that would be enough. But I’m assuming it’s a judgment about mental events, which are neither subject to lawlike generalisations themselves nor subject to lawlike generalisations about their relationship to physical events (if Davidson’s right). In a way this is the Kantian thought that punishment’s about what we can show people were responsible for again.
Pete, that’s a crap argument. Premise 1 is silly. Prepunishment stands in conflict with the deterministic thesis. Presumably, an agent A is prepunished if they will deterministically commit crime C in the future. They are locked up pre-emptively, and so do not commit the crime, thus contradicting the deterministic prediction that they would commit C.
If determinism is true, what is going on here is that the prior states of the world and laws of nature bring it about that people make a prediction about A committing C, so they lock up A, preventing him from committing C. A was never going to commit C because it was determined that he would be locked up beforehand. So he’s being locked up unjustly for a crime he didn’t commit.
No one within the deterministic system can have divine foreknowledge. Only an external LaPlacian demon-type observer could have such knowledge. Using that knowledge to interfere with the system renders that knowledge false, such as in the case of locking up A. Presumably, knowledge has to be true, but it is not true that A will commit C, so the people who locked him up could not know this in the first place.
Also, why to we have to prepunish C? I don’t think compatibilists would necessarily be committed to prepunishment. Why can’t we just warn A about the crime he will commit, or something similar? Punishment is a tad extreme and unjustified. And no, the Minority Report is not a good intuition pump…
Go and read Hume’s Enquiry.
This was all settled ages ago.
Next question.
P.S. Kant agress with Hume, so long as you focus on the phenomenal realm. Of course, Kant needs something more operating in the noumenal realm to get his ethical picture off the ground…but sane people can just ignore that.
Oh, and stop mixing up questions of (ethical) responsibility with questions of metaphysical freedom.
Go and read P.F. Strawson’s “Freedom and Resentment” to get yourself clear on that relationship.
Am I missing something here? I don’t understand how determinism, the person committing the crime and prepunishment can coexist. I would have thought that determinism implies that all future events are predetermined. If the person committing the crime was the predetermined event then there can be no prepunishment since that would prevent the crime occuring which by definition has to happen.
If prepunishment were applied then the prediction of the person committing the crime was wrong, since they were prevented.
Prepunishment is by definition a violation of determinism.