[hist-analytic] Davidson's Hume

Bruce Aune aune at philos.umass.edu
Tue May 26 09:51:24 EDT 2009

I am late to respond to Steve’s rejoinder to my observation because I  
have been occupied with non-philosophic matters. I have not read all  
the comments his original email provoked, so I may be covering old  
ground. But I would like to clarify my earlier remark a little, anyway.

When I spoke of Davidson’s error, I had in mind the fact that “the  
cause of b caused b” contains a definite description, which Davidson,  
thinking of Russell rather than Donnellan, should have taken to imply  
that b has one and only one cause. In a subsequent note Steve asked  
what happens if we deny “the cause of b caused b.”  From a Russellian  
point of view, the definite description in the denied sentence shows  
it to be equivalent to “There is one and only one cause of b and this  
one thing caused b: in symbols, “Ex[(y)(Cyb <-> y=x] & Cxb].” The  
denial of this quantified sentence, expressed in English, reduces to  
“Either there is more than one cause of b or b has no cause.”  The  
error I said Davidson made amounted to neglecting the logical  
possibility that b has more than one cause. In speaking of logical  
possibility here, I mean possibility in the narrow logical (or  
formal) sense, not possibility in some broader sense. It is not a  
logical truth that an occurrence has exactly one cause.  It is not a  
logical truth that an occurrence has any cause at all.

Steve did mention the possibility that the definite description “the  
cause of b” might be used purely referentially in the given sentence,  
not attributively.  When Donnellan introduced the notion of a purely  
referential description, he observed that such a description may pick  
out a referent that does not strictly satisfy it. He illustrated this  
by an example like this:  “The man over there drinking a martini is  
my thesis advisor.” The referent of the description is a certain man,  
one that the speaker believes to be drinking a martini.  But that man  
may not actually be drinking a martini; he may be drinking water in a  
martini glass.  But the "referential" description might pick him out  
just the same. Suppose that ”the cause of b” is used, in “The cause  
of b caused b,” in this referential sense. Although the speaker uses  
the description to refer to an occurrence he or she believes to be  
the cause of b, the speaker’s belief may be false: the actual cause  
of b may have been some other occurrence.  If this is so, the  
speaker’s statement is contingently false.  I think we should  
consider that it was contingent all along.


More information about the hist-analytic mailing list