[ Pobierz całość w formacie PDF ]

Consequently, we need not be overly precise here about the notion of a lean screener.
All we need is to characterize an event that secures the neutralization of would-be causal
relevance. If there is one, there are many, and for establishing causal irrelevance it does
not matter which one we pick. Hence we may characterize causal-relevance neutralizers
in a way that brings out the fact that they serve that role and yet do not include ingredi-
ents that are immaterial to that role.
21 Since the world is Markovian, any intermediate full world state is a stable screener for
any event that precedes it vis-�-vis any event that occurs after it. We can consider such
states global screeners.
22 Assume that all the parameters of the bullet at t that are pertinent to its future course are
1
specified in E .
1
23 Accordingly, there are infinitely many other such exhaustive channellers along this trajec-
tory. Such exhaustive channellers can also be specified in cases where there is more than
one causal route from A to C, for example, as conjunctions of such local stable screeners
vis-�-vis a given route. For further details, see Kvart 2001a: esp. section 5.
24 Note 13 rules out that the dog s charging forward triggered the avalanche. The very
release of the dog, we assume, yielded no vibrations of significance. In any case, there
are alternative neutralizers, other than the above E, such as E': the dog didn t come close
to the target (until t ). Implicit in Example 3 on p. 139 about the pipe and the tap was that
C
A was causally irrelevant to E there. In earlier writings I used the term  crn as an abbre-
viation for  causal relevance neutralizer .
184 Igal Kvart
25 And if there is one, there are many. One central feature of this analysis of cause is that in
preemption cases (early or late) the preempted cause is not a cause since it is causally
irrelevant to the effect, a feature that is secured by the presence of a causal-relevance
neutralizer. For a detailed analysis of how this approach handles such cases, see Kvart
(forthcoming d: sections 8 and 9, and this volume: section 6).
Note, however, a caveat to this analysis. Causal relevance and the presence of a stable
increaser are necessary, but not quite sufficient, for being a cause. It must be further ascer-
tained, when checking for causes, that there is no purely negative causal relevance despite
the presence of a stable increaser, which can be present if the route of the would-be posi-
tive causal relevance (indicated by a stable increaser) is neutralized by a positive rele-
vance neutralizer. This case can be analysed with the notions used here for analysing
causal relevance; see Kvart (forthcoming b). I ignore this complication below.
26 See Kvart 2001a: section 8, or, in a brief version, Kvart 2003: section 5. Note, however,
that the same pattern holds not just under the assumption that time is continuous (which
seems essential for physics as we know it), but also under the assumption that time is
merely dense, both of which allow for infinite regress. If time is discrete, then still
causal relevance amounts to the absence of a neutralizer (given probabilistic relevance);
but since all such sequences are finite, causal relevance is no longer exhibited by the
presence of infinite regress. In such a case one must be especially attentive to the
constraint on neutralizers that pertain to the upper end of the occurrence time of C, and
in particular in a case of E that occurs one temporal unit after t . Such a constraint is
n A
introduced on p. 172 in section 5, and in section 7 we consider in detail its application to
cases of this sort regarding radioactive decay (albeit fictional ones).
27 For a more detailed presentation of the above analysis of cause, see Kvart 1997: sections
6 12. For a more detailed presentation of the above analysis of causal relevance, see
Kvart 2001a: 59 90. For a related treatment using the above account for an analysis of
the thirsty traveller puzzle, see Kvart 2002. All of these papers are available also in my
home page: http://socrates.huji.ac.il/Prof_Igal_Kvart.htm.
28 For more on purely negative causal relevance, see Kvart 2001b: section 4, 1986: section
VIII. Recall also that we allow ourselves, here and elsewhere, to use A, B, C and so on
ambiguously as names of sentences as well as names of the events specified by these
sentences (narrowly individuated).
29 Kvart 1997: section 6.
30 This example is taken from Kvart 2001b: section 4.
31 See Kvart forthcoming a: section 8.
32 Under narrow event individuation, the framework within which we operate here, an
event is paradigmatically specified by an object, a property (or a predicate) and a speci-
fied occurrence time  that is, a temporal interval qualified by a temporal quantifier
indicationg the fragility of the event in question. Lewis s notion of fragility focused on
modal aspects, which are important for a counterfactual account of cause. But for the
probabilistic approach to cause advanced here, the modal aspect is not as important.
33 Kvart 2002: section 8.
34 I take this to be the brunt of a challenge posed by Murali Ramachandran.
35 Another neutralizer for B and C is: G  Billy s rock did not hit the window until the time
it was shattered (if it was shattered). G screens B off from C, and stably, and B lowers the
chance of G (see also note 37), with no stable increaser, and thus B is not a cause of G. [ Pobierz całość w formacie PDF ]