Quine, Ontological Relativity

Quine and Carnap agree that what exists (or at least what we take to exist) has something to do with our language. In "Ontological Relativity," Quine discusses how we understand existence claims like, "There are numbers," or "There are physical objects," and so forth.

He actually discusses rabbit, water, sepia, ox, and "gavagai," but ultimately his target is the metaphysical sorts of existence claims I just mentioned. Life is quite hard enough with apparently straightforward examples like rabbits, and so he waits until late in the article to muddy the waters with things like numbers. (Quine discusses translating an unknown language, and "gavagai" is a made-up word in his pretend example that is used by speakers of the language whenever a rabbit hops by.)

How do we come to understand language? We learn language from others, both by hearing them (in context) and by speaking to them. We also learn language by reading, watching TV, and in lots of other ways, but Quine thinks that they aren't importantly different from the main ones I mentioned. We don't learn language by psychically absorbing meanings from the ectoplasm, but through behavior. As a result, Quine believes, there can be no features of language that aren't exhibited in linguistic behavior. For example, if two people use a language in the same way, it doesn't make sense to even entertain the possibility that one of them might know the language while the other one has it wrong, for example, because one of them has the right meanings for the words and the other doesn't.

There is a familiar and seductive picture of language that would make such a possibility real: the so-called myth of the museum, in which knowing a language is knowing the "meanings" of each of the words (and the grammar). If the myth of the museum were true, Quine claims, it would be possible for two people to use the same language in the same way, but for one of them have it wrong.

Two "real" examples: In French, you negate a verb by surrounding it with "ne" and "rien." One person might take the word "ne" to mean not while "rien" means "anything." Another might take the word "ne" to be a mere piece of grammar that just marks the beginning of the verb while "rien" means "nothing." Who is right? On the myth of the museum, the question makes sense. Since behavior would not be affected, Quine argues that there is no fact of the matter. Both know everything necessary to use "ne ... rien" correctly.

The other example involve a distinction that is very important to Quine, that between "mass terms" and "terms of divided reference" (now often called "sortals"). A mass term names an undivided mass of something ("water"). A sortal names a collection of many things of the same sort ("cows"). Some phenomena are usually described using mass terms; others, using sortals. There are, however, a few cases where we have both: "cow(s)" is a sortal, but "cattle" is a mass term. We have ways of going from mass terms to sortals: "head of cattle." (Also, wood->sticks of wood.) In Japanese, when you count things (5 cows, 10 books, ...), there are three words involved: a number, a kind, and a third word in between. What do we do with the third word?

(5 bovine) cows or 5 (head of cattle)

It is said in books about Japanese that there are different number systems to count different kinds of things. (That is the first version.)

In the second version, Japanese has lots of mass terms where we use sortals, and you need to convert them in order to count.

There is no fact of the matter. In one of them, though, the Japanese term for "cow" is a mass term, while in the other one the very same word is a sortal. So, if someone says in Japanese, "Cow exists," what is being claimed?

Quine claims that we have the same problem always, in all language. We ordinarily, most of the time, translate (that is, understand) each other using the "homophonic" translation, that is, I normally assume that you use the word "cow" in the same way I do, and so on for all other words. Note that that is optional, and that we don't always do it. Is it possible for me to make sense of you using a translation that is far from homophonic? Quine claims that the answer is yes. Two speakers of Japanese might take each other to be different in the way illustrated above: one might use many number systems and sortals (count nouns) and understand the other as using one number system and mass terms. To do that would be to understand the other speaker as having a radically different ontology.

So long as that is possible in principle, Quine's philosophical points go through, and his real argument is just that language is so complex and learned in such an indirect way, it seems inevitable that there are variant ways of taking some parts of language.

Toy examples: rabbit, undetached rabbit part, time slice of a rabbit, a foot up and to the left, mirror world.

-- ShaughanLavine - 28 Jan 2008