3.328 If a sign is not used then it is meaningless. That is the point [Sinn] of

Occam's razor. (If everything behaves as if a sign had meaning, then it has
meaning.)

I take it the "meaning" above means "reference." But this remains to me a
highly cryptic prop.. I do think that to the extent that one holds a theory of
simple names of the sort W expounds in TLP, you can tell what your ontology must

do to satisfy Occam's Razor. You will need to insist that the only things that
(really) exist are those that are named (by simple names). So, to give a
complete list of "what there is" one must supply a list of all the nameable
objects—and nothing else. The problem is that W doesn't give a single example
of an object.

It's a pretty theory, nevertheless. "There's a cat on the mat" will still make
sense, even if it's there's no cat on the mat in question. But, on W's view,
"Axelrod is on the mat" will be nonsense if "Axelrod" is a simple name but there

is no Axelrod. So, "the cat" will make the cut only if it's identical to some
Axelrod or other.

Compare Quine, who says in "On What There Is" that to the extent that the
made-up "socratizes" can be used to dispense with all sentences containing the
name "Socrates" we need not count Socrates as one of the indispensable elements
of the universe. The problem with this version of Quine's thesis, as Alston
pointed out in "Ontological Commitment," if the analyses and analysanda are true

in just the same cases, whatever one is committed to the other would seem to be
as well. But Alston's argument won't work against the tractarian view, I don't
think, because, for W, the sentence itself displays its commitments by showing
which of its elements are simple names (those that are senseless if they don't
refer). If "Socrates" is a real name, "Socrates is F" doesn't seem to be well
paraphrased by "There's an X such that X socratizes, X if F, and for all Y if Y
socratizes then Y = X" because it is simply a denial that "Socrates is F" is
atomic.

Put it this way: Let "S is F" stand for a prop. If it is paraphrasable in the
manner described above, it is NOT an atomic prop. But, using Quine's trick, any

prop may be so paraphrased. Therefore, either there are no atomic props or the
paraphrase does not work.. But W spends a good chunk of time and trouble
attempting to prove that there MUST be atomic props (though he can't or won't
give us any examples of them). If he has succeeded, then we know that Quine's
trick is defective.

What seems so paradoxical about W's ontological scheme here is that, rather than

pointing to science as the arbiter of what there is, he points to meaning. He
claims that the very possibility of determinate statements/understa nding
requires that there be atomic props. But how will we tell which props are
atomic?? As I've said before, one of the attractive things about the Tractatus
is that it is so audacious.

Walto