[ Pobierz całość w formacie PDF ]
world is the watery stuff of our acquaintance, and also a priori that the C-extension of water at some w differs from
the watery stuff of our acquaintance at w.
Quine Revisited
I mentioned our theoretical disagreement with Quine earlier, noting that often in practice it does not matter. We are
now, though, in a position to say something directly to the theoretical disagreement.
The Quinean position that denies the possibility of full-blown analyticity in the sense of sentences that are genuinely a
63
priori was seen as radical when it was first propounded. Nowadays it is close to orthodoxy. The idea is that we cannot
make a clear distinction between what is a priori and what, for instance, is almost certainly
63
At least in America. As Hilary Putnam observes, the situation is different in Britain: Pragmatism , Proceedings of the Aristotelian Society, 95 (1995): 291 306, at 299.
THE ROLE OF CONCEPTUAL ANALYSIS 53
true or very obviously true between the encyclopedia and the dictionary, as it is sometimes put. I think, however, that
it is insufficiently realized just how radical the Quinean position is.
We use language to tell our community and our later selves how things are. Telling how things are requires
representational devices, structures that somehow effect a partition in the possibilities. For we say how things are by
saying what is ruled in and what is ruled out. The metaphysics of these possibilities is a controversial matter, as we
noted earlier, but anyone who reads a map is in the business of ruling in and ruling out some of the possibilities
concerning, say, where the source of the river is, or where the nearest town is.
Now suppose that it is impossible to effect a partition among the possibilities independently of how things actually are.
No mental state, no linguistic item, no diagram, no system of semaphore, divides the possibilities, except relative to how
things actually are. Then we can never say, diagram, depict, semaphore, think, . . . how things are. All we can do is say
(depict, think, etc.) how they are if . . . . We are always in the position of one who only ever tells you what to do if you
have high blood pressure, never what to do simpliciter. We can say how things are conditional on. . . , but cannever make
an unconditional claim about how things are. We cannot detach. This is a very radical doctrine. It is not that we cannot
say with complete precision how things are. We really cannot say how things are at all.
Thus, the Moorean insists that we can effect partitions in how things are in language, in thought, in pictures . . . ,
independently of how things actually are; independently, that is, of which world is the actual. But now it follows that
whenever we have two thoughts, propositions, sentences, sets of flags, or whatever devices of representation they may
be, call them R and R , such that the actual-world independent partition effected by R and the actual-world
1 2 1
independent partition effected by R is such that the set associated with R is a subset of the set associated with R , then
2 1 2
If R then R is a priori, because it is independent of which world is actual that whenever R is true, R is true.
1 2 1 2
What then is the difference between, on the one hand, being very sure that cats are animals, but regarding it as a
posteriori, and, on the other, regarding it as a priori that cats are animals? In the first case you have possibilities to
which probability might be
54 THE ROLE OF CONCEPTUAL ANALYSIS
moved, that you would describe as ones where cats are not animals; in the second you would not describe them that
way. So there is a difference in principle. In practice you may not have made the decision about how you will use
language that settles whether you'd use Cats are not animals to pick out the possibilities to which probability should
be moved should some very surprising discoveries be made, or whether you'd use There are some things that would
count as cats except that they are not animals . The nature of the possibility is the same in either case; what is unsettled
is how you'd pick it out in language and the latter may be unsettled simply because the possibility is so exotic there is
no point in expending energy in deciding ahead of time which way to jump should the need arise. We can, that is, agree
with the point emphasized by Hilary Putnam: in practice it is hard to find sentences in a natural language that are
64
determinately a priori. Our failure to decide in advance how we would jump in fantastical, remote cases gives
philosophers, with their notorious ability to think up fantastical, remote cases, plenty of scope to come up with a case
for which it is undecided whether, as it just might be, cat and animal apply, and so is a case where we can be induced,
without going against anything determinate in the meanings of the terms, to apply, say, cat and not apply, say, animal .
Thus, the case becomes one where cats are not animals. But the right conclusion is not that Cats are animals was
determinately not a priori. It is that Cats are animals is determinately not a priori after the story-telling, but before the
story-telling began, it was indeterminate whether or not it was a priori.
Although often we have not decided how to describe some incredibly unlikely, fantastical happening ahead of time,
sometimes we have. We have no trouble understanding stories about fantastically unlikely sequences of coin tosses,
about long series of bridge hands consisting of only one suit, or about Davidson's swampman. And, to make sense of
these stories, we distinguish between what is very confidently believed and what defines that which our very
64
See e.g. Hilary Putnam, It Ain't Necessarily So , Journal of Philosophy, 59 (1962): 658 71. As many have observed, post-Kripke this paper should be thought of as It Ain't A
Priori So , and for our purposes the issue about analyticity can be thought of as the issue about the a priori. I am indebted here to discussions with Michael Smith.
THE ROLE OF CONCEPTUAL ANALYSIS 55
confident belief is about which is not to say that the boundary cannot move with time, or is a sharp one.
The next chapter is concerned with two popular objections to conceptual analysis and, most especially, the distinction
between metaphysical and conceptual necessity, and its bearing on our enterprise.
Chapter 3 Conceptual Analysis and Metaphysical
Necessity
[ Pobierz całość w formacie PDF ]
