Monday, October 09, 2006

Deep Tautologies

I love language.

I had cause in the car last weekend to repeatedly utter the sentence, "We'll get there when we get there." Now if understood literally, the sentence is what we call a tautology, that is, it is a sentence that because of its form is always true. But like a meteorologist who boasts about his 100% accurate weather predictions because he simply repeats "Tomorrow it will rain or it won't," these sentences are always true because they contain no information -- they say nothing, so they can't be wrong. But yet, we use them all the time to convey information. How does this happen? This was a question Hanno and I pondered a while back and to prove that I, Hanno, or perhaps both of us are real philosophers, here's what we came up with -- There are at least three ways in which you could have a meaningful sentence that looks like a tautology:

Pseudo-tautologies

A sentence is a pseudo-tautology if the A's in a sentence of the form "A is A" are really different A's. My favorite is one that came out of Hanno's mouth years ago when we were discussing the history of logic and he said, "That was before Bertrand Russell was Bertrand Russell." Of course, there was never a time before Bertrand Russell was himself, but I understood what he meant because I realized that the first occurrence of the name referred to the man himself and the second to a definite description like "the famous logician/philosopher who came up with Russell's paradox and the theory of types." The name meant two different things -- things of the sort Russell himself helped to disambiguate -- and so the sentence was meaningful if not hilariously ironic on several levels.

Pointers

A term Hanno coined from computer programming lingo for sentences that really represent other sentences. "We'll be there when we're there" really means "Stop complaining. Be patient." Many pointers like this appear in form to be declarative sentences, but are really commands. The idea is that if you can disguise your imperative as a sentence that is always true, then your advice comes across as necessarily true, as something that must be followed. There is a raging debate amongst really geeky philosophers of language over whether we figure out what command goes with what seeming tautology based only on pragmatic aspects of language like context or whether the type of noun phrase used in the fake tautology is part of the inference, what H.P. Grice termed an "implicature."

Deep Tautologies

The third and most interesting ones are what we dubbed, "deep tautologies." The idea is that a tautology is only a tautology if the A's in a sentence of the form "A is A" are what we called, sharply delineated, that is, it is the sort of term that (a) do not admit of degrees and (b) have a set of clear criteria which draw a sharp line between those things that are and are not A. "Pregnant is pregnant" makes sense, but "beautiful is beautiful" doesn't because things are more or less beautiful and there is not a sharp line separating all things into the beautiful and the non-beautiful.

But some terms have both a well-delineated and a non-well-delineated sense. Take green. Green has shades. Something can be a little green, but not too green. But then there are cases where something is or is not green. If a waiter delivers a plate and you say "Take it back, this meat is green," and the waiter says, "But it is only a little green around the edges," you may indignantly respond, "Green is green." In other words, when I used the word "green" to describe the meat, I meant it in its well-delineated form. Shade and amount does not matter. There is green meat and not green meat. That is all. So when we use deep tautologies, what we are conveying to our listeners is that we mean the ambiguous word in the well-delineated sense.
Ain't language fun?