Figure 7.1 Asymmetric causal dependence (the arrows represent
causation or causal dependence)term ‘S’ will refer to Ss, and only Ss, provided that the causal connections
between ‘S’ and any other (non-S) objects which may happen to cause
tokenings of ‘S’ are asymmetrically causally dependent upon the causal
connection between tokenings of ‘S’ and Ss. So if any other types of object
besides Ss cause tokenings of ‘S’, they will only do sobecauseSs cause
tokenings of ‘S’. The account, then, is this (also represented diagram-
matically inWgure 7.1 above):
mousewill meanmouse(and so shrews will bemisrepresented bymouse) if:
(i) mice cause tokenings ofmouse, and
(ii) if mice had not caused tokenings ofmouse, shrews would not have,
and
(iii) if shrews had not caused tokenings ofmouse, mice still would have.
Thus (i)–(iii) capture the idea that theshrew-to-mouseconnection is asym-
metrically causally dependent upon themouse-to-mouseconnection.
Note that what Fodor oVers is only asuYcientcondition formouseto
meanmouse, not anecessary andsuYcient condition (that is, what im-
mediately precedes clauses (i)–(iii) is an ‘if’, not an ‘if and only if’). This is
in line with his (1974) conception of what naturalisation should involve. In
general, he thinks, we cannot hope for areductionof some problematic
term T into purely natural terminology (which would involve a statement
of necessary and suYcient conditions for T to apply). For most higher-
level properties admit of multiple instantiation in lower-level facts. The
most that we can hope for is a statement, in natural terms, of one of the
realising conditionsfor T to apply. That is to say, the most we can hope for
is a statement ofsuYcientconditions for its application. So it is no good
looking to refute Fodor’s account byWnding cases wheremousemeans
mousebut conditions (i)–(iii) do not collectively apply. Rather, we should
need toWnd a case where conditions (i)–(iii) apply, butmousedoes not,
intuitively, meanmouse. Can this be done?
Here is a possibility to think about. Consider any term T withmarginal
cases(for example, ‘red’, or ‘sport’). Call the central cases where T applies
Cs, and the marginal cases Ms. (Note that the Msarecases of T, only
marginal ones.) Does it not seem plausible that the application of T to the
Informational semantics 165