[ Pobierz całość w formacie PDF ]

respect to alleged cases of chance-lowering causation is a plausible strategy. The
bullet-biting strategy simply denies that common-sense intuitions concerning the
alleged counterexamples deserve to be taken seriously enough to undermine (IC),
and I argue that the desire to believe in chance-decreasing causes can be explained
away. Finally, in Section 4, I argue that the alleged cost involved in this strategy 
the denial of the transitivity of causation  is no cost at all, since there are no good
arguments for the claim that causation is transitive. I also show that another objec-
tion of Dowe s  that the strategy makes causal facts depend on facts that are
extrinsic to the causal process in question  fails to hit home.
2 Dowe s chance-relativizing strategy
In this section I argue that Dowe s strategy for rescuing (IC) does not succeed. I
begin by showing that his analysis of a hindering process  the process from
which we need to abstract away in order to restore chance increase between c and
e  fails. However, the failure of that analysis does not entail that we should
abandon the general chance-relativizing strategy altogether, and I therefore
provide a more intuitive, and more successful, conception of the alternative
process from which we need to abstract away. I then show that given this concep-
tion of the alternative process, the chance-relativizing strategy can be applied to
chance-decreasing causal processes that are not intuitively cases of causation, for
example the defoliant case described on p. 39. In other words, all cases of chance-
decreasing causal processes can be characterized as mixed-path cases, and hence
all such cases are, according to Dowe s strategy, cases of causation. So Dowe s
strategy is too successful: it makes not just some but all chance decreasers come
out as causes. Hence the general strategy does not succeed because it fails to
discriminate between (alleged) chance-lowering causes and chance-lowering
non-causes.
At first sight, Dowe s strategy seems intuitively plausible in the bus brick case.
There, two identifiable and reasonably independent causal processes are going on:
the process that involves the bus speeding along the street, and the process that
involves the brick falling. My push in effect stops Edna interacting with the first
causal process, but at the same time forces her to interact with the second. So  at
first sight  it seems plausible to say that the push does two things: it promotes
Edna s death by involving her in the brick process, and hinders her death by getting
her out of the way of the bus process.
It also seems to be a straightforward matter to imagine the situation with one of
the processes removed: it s easy to imagine the situation minus the bus, where I
push Edna (for reasons unknown, or perhaps for no reason), the brick falls, and
Edna dies; and it s also easy to imagine the situation minus the brick  where I push
her out of the path of the bus and safely on to the (falling-brick-free) pavement.
44 Helen Beebee
Clearly in the former case, where the bus isn t on the scene, the push increases the
chance of Edna s death. Hence it seems plausible to say that the push initiates a
 mixed path to Edna s death, and that we can sensibly relativize the chance of
Edna s death to just one path: the brick process.
However, we need to be a bit clearer about the nature of the  paths ; in particular,
we need to be precise about the nature of the bus-avoiding  process from which
we are supposed to be abstracting away. I shall argue that the details of Dowe s
analysis do not yield the intuitive picture presented above, and hence that if we
want to try to save some form of the chance-relativizing strategy, we are going to
have to hang on to the intuitive picture rather than the details of the analysis.
Picture the scene. Before the push, the bus is hurtling down the street and
heading straight for Edna. This is a bona fide causal process. Then we have the
push. (The bus continues to hurtle down the street, but these later stages of the
process are no longer relevant to Edna s death.) The push stops Edna from inter-
acting with the genuine causal process of the bus s travelling down the street.
However, according to Dowe, when we abstract away from the bus-avoiding  pro-
cess , we are supposed to be abstracting away from the hindering  process initi-
ated by the push. We are not  or at least not directly  supposed to be abstracting
away from (the earlier stages of) the genuine causal process of the bus hurtling
down the street, since that process is neither a potential preventer (that is, hinderer)
of Edna s death, nor a process initiated by the push.
So in what sense does the push initiate a hindering process  a process that could
(but in fact does not) lead to Edna s survival? Well, the push prevents a particular
course of events  say, bus-being-a-foot-away-from-Edna (event b), bus-hitting-
Edna (d), & , Edna s death (e)  from occurring: without the push, that sequence of
events would have been very likely to occur. On Dowe s account, the  process
actually initiated by the push is, as it were, the negation of the earlier stages of that
merely possible process b d e. Call the bus s hurtling down the street prior to the
push a, and the push c. Without c, the process that would have been very likely to
occur would have been a b d e. With the push, the bus-avoiding  process that [ Pobierz całość w formacie PDF ]