This is my evening for responding to all the causality messages. Here's
one more ...
Ronald J. Allen wrote:
> One should be cautious about bringing the legal treatment of causation
> into discussions of probability. "Causation" in the law is virtually
> never about causation; it is instead about responsibility. Conclusions
> in terms of causation usually are merely the means of carrying out those
> determinations of responsibility. In the hypothetical of A paying B to
> kill C, no one would claim that A's act is not what is called a cause in
> fact of C's death ("but for" A's act, C would not have died as he did).
> The pertinent legal question is whether A's conduct is within the
> definition of the elements relevant to the case, which is not advanced
> one whit by either philosophical or probabilistic analyses of
> causation. It depends instead on the purpose of the statute. Some
> might conclude that A should not be held liable for murder (but he would
> then be held liable for solicitation, conspiracy [and thus vicariously
> liable for the murder, interestingly], and express the conclusion in
> terms of A not causing B's death, but what they would mean is that A
> didn't cause B's death for the purposes of the statute.
I certainly can't argue with you about how the law actually works, but I
think it's important to distinguish the causality questions and the
policy questions (i.e., the purpose of the statute). They're both
important, but they should be dealt with separately. My impression
(from admittedly rather little basis) is that they are often confounded
in the law.
> When there are difficult questions of causation, such as simultaneous
> sufficient acts for a particular result (e.g., two people shoot the same
> person in the heart at precisely the same moment), the matter is
> resolved by definitional fiat. In this case, all US jurisdictions
> conclude that both acts "caused" the death for the purposes of the
> murder statutes. This definitional move, though, is fueled by the
> inability to see any moral or practical basis for distinguishing the
> two; they are equally reprehensible.
For what it's worth, our formal definition of causality reaches the same
conclusion: both shooters are the cause. The definition of
responsibility gives them each degree of responsibility of 1/2 though.
> What if A shoots C and C would die in one hour, but B comes along and
> also shoots C, and he dies in 30 minutes, but if A had not shot C, C
> would have lived an hour after being shot by B? Again, definition to
> the rescue driven by moral and practical intuitions. And again all US
> states agree: murder is the shortening of a life by any degree--both are
> guilty.
Our definition says that B is the cause of death, not A, in what is the
most reasonable causal model. However, I can imagine, as a matter of
policy, that the law might decie to punish A too. This is a case where
it seems to me useful to separate policy from causality.
> There is an infinite supply of interesting questions about who should be
> responsible (A poisons B's water. B goes off into the desert; C steals
> B's water and B dies of thirst). Puzzling over probability theory
> doesn't much help to resolve these, because as I say in most cases the
> puzzle is not really over causation.
>
> When there is a gap, again definitions to the rescue. In all states A
> in the poisoning hypo would be guilty of murder or attempted murder.
> Why the difference? Become some state legislatures disagree about the
> moral reprehensibility of A in such a case, but I can assure you that
> the ground for disagreement would not be dissolved as a result of any
> version of probability theory of which I am aware, at any rate.
I agree that probability won't help here at all. As I indicated to
Lotfi, our definition is not probabilistic. For what it's worth, in
what is the most obvious causal model, C (the thief) is the cause, not A
(the poisoner). The story is essentially isomorphic to the two
shooters. But, again, as a matter of policy, it does not seem
unreasonable to me to punish both. Before I go on, let me make clear
that Judea and I are far from the first to consider problems like this.
Philoosphers and legal theorists have been looking at them for years.
And there are certainly many (somewhat) formal models of causality in
the literature that try to deal with these issues. Of course, with
typical modesty, I think ours does a much better job :-).
> In any event, if any of this is of interest to anyone, I could
> elaborate. Also, work on causation in the law has advanced considerably
> since the publication 44 years ago of Hart and Honore's book. If anyone
> is interested, I'd be happy to provide citations.
I'm certainly interested. I'm reading Wright's "Causation in Tort Law"
and have read a few other random papers, besides Hart and Honore.
-- Joe
