Topics Language

Question My question concerns analyticity. I'm a Danish undergraduate student of classics, so I don't have any formal education in philosophy. Anyway, here goes: How do you determine whether or not a proposition is analytic? I believe that the traditional definition is something like this: for a proposition to be analytic the predicate has to be contained in the subject (in the sense that the truth of the proposition can be determined purely on the basis of the semantics of the concepts used and of an understanding of the logical form of the proposition). But this does not seem to be enough. Consider this example: "Wolves live in packs". This would normally qualify as a synthetic proposition, but why exactly? Imagine that a person sitting in her favourite armchair uses her semantic mastery of the concept of wolves and determines the truth of the proposition without lifting a finger. Would that make the proposition "analytic". There seems to be something wrong here. One could say that every proposition that is not a member of the group of propositions sufficient to identify unequivocally the subject in question, does not qualify as a synthetic proposition. And since "living in packs" is not a member of the group of propositions sufficient to identify unequivocally the subject "wolves", it does not count as an analytic proposition. Is there some sense in this, or should I look at "analyticity" in another way?

Accepted:
September 14, 2008

Comments

Thanks for your excellent question. You're right to be dissatisfied with the "containment" characterization of analyticity. The reason is that containment is a topological or set-theoretic idea, and it's hard to know how to cash it out for the case of language. After all, subjects and predicates don't take up space, and it takes some doing to explain them in terms of sets also.

Your suggested definition in terms of 'sufficient to identify unequivocally the subject in question,' is a good idea, but I also think it won't work without a lot of modification: After all, 'containing interior angles' is true analytically of triangles (triangles have interior angles' is plausibly analytic), but that predicate doesn't identify unequivocally the concept of a triangle, since lots of other things have that feature as well.

A perhaps better characterization of analyticity is 'true by definition'. It's true by definition that triangles have three interior angles, and that bachelors are unarried. However, it can be pointed out plausibly that it isn't true by definition that wolves live in packs: We can coherently imagine lone wolves. If there's a difficulty with this, it's that it is often hard to know just by inspecting the concepts what is part of their definition. Is it true by definition that I'm a material being, that persons have rights, that time is linear? "Armchair reflection on the definition of expressions like "time", and so on is probably not going to take us very far in answering these questions.

In fact, frustration with issues like these is what led the philosopher W.V.O Quine to attack the "analytic/synthetic" distinction about a half-century ago. He argued that no non-circular characterization is workable, and indeed that the difference between analytic and synthetic is at best one of degree. For instance, if asked about the case of the linearity of time, I would venture that he'd answer that insofar as we take time to be necessarily linear, that's just an expression of the fact that its linearity is a deeply entrenched part of our best scientific theories. As such, it is in principle a doctrine we could find ourselves giving up if the scientific facts make it expedient to posit some entities (such as certain subatomic particles) has having unusual temporal properties.

Some philosophers still adhere to the traditional analytic/synthetic distinction, but many others follow Quine in thinking that it marks a difference of degree only, and doesn't bear anything like the philosophical weight it once did.

For more on these issues, please see George Rey's article in the Stanford Online Encyclopedia:

http://plato.stanford.edu/entries/analytic-synthetic/

Mitch Green