Thursday, October 04, 2007

Is Water H20?

Hallq, over at Uncredible Hallq, is reading one of my favorite pieces by Hilary Putnam and failing to see what is so wonderful about it, so I thought I'd pipe in because it's a puzzle and an argument that I didn't get at first either, but have come to adore.

The whole thing starts when we ask, "What is water?" Seems an easy enough question, it's any substance made up of molecules that are a combination of two hydrogen atoms and one oxygen atom. That's just what water is. Water is defined that way and definitions can't be wrong.

As Humpty Dumpty says in Through the Looking Glass, words mean what I want them to mean, nothing more nothing less. If I take the string of letters "xix" and posit the definition purple navel fuzz that has been encased in Jello, then that is what "xix" means. Is my definition true? Of course, it's a definition, it has to be true.

This is important in the history of philosophy because philosophers of the generation before Putnam (like his adviser Hans Reichenbach and his wife's adviser Rudolf Carnap) worked hard to divide all sentences into two groups, those that are true of things in the world and whose truth was determinable by observation, and those whose truth is necessary. Sentences in this last group came in two flavors, they argued, those that are truths of logic and definitions.

And so they sought to understand the act of defining. Defining is a mental activity. When I know what a word means, I understand its definition, I grasp it intellectually. The definition is an idea, something in my head. When I use a word I understand its sense, its meaning, its conceptual definition.

So, now imagine if you will that there is a planet somewhere in space that we can call Twin Earth. On Twin Earth, everything is identical. Everyone has a Doppelganger, an exact twin, who right now is reading a post on my Doppelganger's blog which says exactly this. The only difference is that the substance on Twin Earth that rains from the sky, that they drink and brush their teeth with, that they refer to by the word "water" is a chemically different substance, call it XYZ. It turns out that XYZ has many superficial observable properties in common with H2O, it just has a radically different chemical make up. If we take Twin Earth water and Earth water the "same liquid" relation does not hold between them.

If we were to go to Twin Earth, we would report back that Twin Earth English is the same as Earth English, except that "water" means XYZ instead of H20. No problem, the word is ambiguous, they have a different meaning for the same signal. Nothing weird there. We'd get fooled at first, but eventually we'd figure out that Twin Earth water is not the same liquid as Earth water.

So, let's come back to Earth for a minute, only go back in time. When Archimedes used the Greek word for water, what did he mean what we mean by water? Well, there would be no cases in which a modern user who understands atomic theory and an ancient Greek would disagree. We would think that the "same liquid" relation held between all the same liquids.

Does that mean that Archimedes means what we mean by "water"? No, because if meaning is in the head, then we mean different things by water since his idea of what water is is different.

So, (here's the punchline) if we send Archimedes to Twin Earth and he sees Twin Earth water, will he think it is water? Yes. It fits his definition, the conception he has in his head. BUT it is not the same liquid. his definition, it turns out, is wrong.

His definition is wrong? How can a definition be wrong? It's a definition. It's something in your head that determines what a word means. It can't be wrong.

What Putnam is showing is that certain definitions, those connected with natural kind terms -- things of a kind that appear in nature from the same species -- are defeasible, that is, they can be wrong. This means that some definitions are not in the head, but in the world. It turns out that the way we've been thinking about definitions has been wrong.

So what? The reason we had been working on definitions was to be able to justify rational belief in science. Scientific results are based on empirical confirmation of propositions that get their meanings from theoretical definitions, definitions that we could depend on because they had to be right -- they are definitions, after all. These folks, the logical positivists, thought that we could take scientific theories and make axiomatic systems out of them, figure out what are the basic assumptions, what are the empirical claims and test to see if the theory is a good one. It required the common sense view of definitions that they are in the head.

What Putnam shows us is that the way we construct and justify rational belief has to be more complex than the nice neat rigid picture of Reichenbach and Carnap. We need to radically rethink why we should believe science. Turns out, Putnam contends, you might need something more naturalistic, something more like a web of beliefs, something more in line what the man with an office down the hallway in Emerson Hall with the name Quine on his door was saying...but that's a whole other post. When you see the argument in its context and why some unbelievably smart people thought the opposite, it turns out that it is not as trivial as it may seem in a vacuum.