[ Pobierz całość w formacie PDF ]
be begged. What's more controversial, and more germane to the conceivability argument,
is the next alleged source of a priori knowledge, what I'll call semantic form, or, more
generally, information concerning the application conditions of concepts. So if I know
that all bachelors are unmarried a priori it's because I know that nothing counts as a
bachelor unless it's unmarried. Whether I do in fact have a priori knowledge of this sort,
and how much, is a serious matter of debate.
Kant thought there was a third type of a priori knowledge, the synthetic a priori. Both
logical and semantic form yield analytic a priori judgments, so if there is a synthetic a
priori, it must come from another source. The most compelling example is mathematics,
where it seems as if there is a direct apprehension of what must be true in some domain
and it isn't based obviously on either logical or semantic form. But of course the idea that
we could have epistemic access to such facts has seemed mysterious, and for that reason
many attempts have been made to ground mathematical knowledge in either logical or
semantic form. We needn't worry here about this problem, since it seems clear that the
argument with which we'll be concerned has nothing to do with the synthetic a priori.
I said above that if we judge a situation metaphysically necessary, then we must have
available a representation of that situation relative to which it is
end p.41
conceptually necessary. Another way to put the point is this: I deny the existence of
brute necessities, metaphysical necessities that transcend logic (where semantic
constraints on concepts are understood to be part of logic in the relevant sense). If, for
some alleged metaphysically necessary situation, there were no description of it relative
to which it was conceptually necessary no description of it (not just none that we can
think of, but none at all) that manifested either formal validity or semantic/conceptual
necessity it would be hard to understand what could ground the metaphysical necessity.
At the very least, I don't see how we could be in a position to judge that the situation in
question was metaphysically necessary.6
As will become evident below, a major issue dividing advocates from opponents of the
conceivability argument is the nature and extent of the second source of a priori
knowledge, that deriving from semantic constraints. Because I am very sympathetic to
the Quinean attack on the analytic/ synthetic distinction I would like as much as possible
to reduce the a priori to logical form. However, just how far I can go in this direction and
still hold on to my rejection of brute necessity is unclear. Let me explain.
To begin with, consider the following objection to my account of the relation between
metaphysical and conceptual necessity.7 I claimed above that if we do have a description
of a situation relative to which it is conceptually necessary, then the situation is
metaphysically necessary. But take the situation of Aristotle being a student of Plato.
Presumably this is not a metaphysically necessary situation, since Aristotle could have
decided not to study with Plato. Yet that situation does have a description relative to
which it appears to be formally valid (hence conceptually necessary), namely, The
greatest student of Plato was a student of Plato.
In reply, I think we have to recognize that the statement in question is ambiguous. On one
reading it describes a general situation and on the other a singular situation. The general
situation involves a relation between, roughly, the properties of being the greatest student
of Plato and being a student of Plato. That relation certainly does hold in every possible
world and therefore that situation is indeed metaphysically necessary. On the other
reading, where we take the phrase the greatest student of Plato to have an implicit
dthat operator (see Kaplan 1979), the statement represents the singular situation
containing Aristotle and the property of being a student of Plato. But on that reading the
statement isn't formally valid.
However, this reply immediately raises another objection, this time to the converse claim,
that every metaphysically necessary situation has an a priori description. Consider the
situation of Aristotle's being human. If one follows Kripke (1980), as I am inclined to,
then this is a metaphysically necessary situation. Aristotle, though he could have failed to
study with Plato, could not have failed to be human. Nothing that is not human counts as
the same individual as Aristotle. Now, relative to which description is this situation
conceptually necessary? One is tempted to offer something like The human being who
did such-and-such is human. That certainly seems to have a logically valid form. But
given my reply to the challenge above, it looks like this statement won't do the trick. If
we read the human being who . . . rigidly, with an implicit dthat operator, then it
isn't formally valid, no more than is the statement The greatest student of Plato was a
student of Plato. But it's only on that rigid reading that it represents the singular situation
involving Aristotle and the property of being human. So if Aristotle is a human being,
or Dthat human being is a human being is to count as conceptually necessary, it must
be by virtue of an a priori principle to the effect that human being is a privileged sortal.
That is, if the predicate applies at all to an object, then it applies to that object in every
possible world in which the object exists.
It appears, then, that one can't simultaneously hold onto the following three doctrines: (1)
there are de re necessities of the sort exemplified by Aristotle's being human; (2) there
are no brute necessities; and (3) conceptual necessity reduces to formal validity. To my
mind, giving up either (1) or (3) is preferable to giving up (2), and I suppose I'm most
inclined to give up (3). Whether I can reconcile giving up (3) with my sympathy with the
Quinean argument against analyticity depends on whether or not there is a principled
basis for restricting the range of conceptual connections that can ground a priori
knowledge to a relatively small set, one that includes the sort needed to ground our
knowledge of de re necessities. But whatever the answer to this question, I don't believe
it will have a bearing on the debate over the conceivability argument, so I will let the
matter rest here.8
2.3 The Conceivability Argument
If we ignore certain complications not relevant to this discussion, we can say that
materialism rules out the metaphysical possibility of a zombie. A zombie is a creature
that is physically identical to a conscious creature but lacks conscious experience. So my
zombie twin call him Zjoe is physically identical to me but there is nothing it is like
to be him. Since conscious experiences are realized in physical states, according to
materialism, and since realization entails bottom-up necessity, materialism entails that
Zjoe is not possible. So any argument that establishes the possibility of Zjoe refutes
materialism.
The conceivability argument begins with the premise that zombies are conceivable. Not
everyone would grant this premise, but it is quite plausible, and it's certainly something I
want to grant. Let's be clear what this means. It's not that I think that zombies could really
exist, given what I know about myself and other human beings. Since I know that I'm
conscious, and also that my conscious experience depends in some way on the physical
[ Pobierz całość w formacie PDF ]