This started as an off-the cuff comment over at the blog for Philosophy Talk, a radio show run by two Stanford philosophers, John Perry and Ken Taylor. But I decided I liked this idea too much to leave it as an unpolished comment.
Reductionism, if you don’t know, is the thesis that claims from one “science” (this might be a true science in the technical sense, or simply any realm of human experience/endeavor) can be “reduced” to claims from another science. Psychology, for example, may be reducable to biology, biology may be reducable to chemistry, chemistry may be reducable to physics, and so on. What this means is that, for example, if we really had a completed biology (and, potentially, a lot of time to state what would count as simple facts from a psychological perspective), we could get rid of psychological language entirely: All psychological claims could be replaced with (admittedly rather more complex) biological claims.
I’m going to argue that reductionism is false–but not (unless there’s an extant argument I’m unaware of) on any of the grounds that people before have argued that reductionism as false. I’m not going to claim that there’s anything wrong with the metaphysics or science of reductionism, but rather that reductionism fails because of problems that have to do with language. I’m not talking about the practical problems here either; I’m not just claiming that it would be pragmatically impossible for us to use the language of even a completed physics to describe all of chemisty, or even that chemical facts might require things like infinitely-long sentences to express in the language of physics. I really mean that it would be theoretically impossible to express all of chemistry using *any* statements of physics, no matter what their length (even if we allowed for things like infinite disjunctions).
I’m actually going to take a much simpler case than the reduction of chemistry to physics: I’m going to talk about reducing thermodynamics to dynamics (and no, this has nothing to do with the unidirectionality of thermodynamics vs. the bidirectionality of dynamics or any other such phil sci chestnuts). Can you reduce all claims about, say, heat to claims about, say, molecular motion?
Well, at first, I might be inclined to say, “yes,” because I’m an identity theorist about heat and molecular motion–that is, I believe that heat is molecular motion. Sounds pretty reductionistic, right? But, in fact, identity theory–even identity theory about generalized phenomena rather than particular instances (this is generally called type identity theory as opposed to token identity theory), is actually not the same as reductionism.
Let’s look at the very claim I agreed to above: “Heat is molecular motion.” Now, is this statement analytic? That is, does its truth follow from the meaning of the terms “heat,” “is,” “molecular,” and ”motion,” plus basic rules of logic? Of course not. It’s certainly possible to understand all of those terms perfectly well, even if you’re a perfect logical reasoner, without beibg able to deduce that heat is molecular motion. Indeed, I’d venture to say that when molecules were first discovered, and the term “molecular” was coined, the discoverer already understood the terms “heat,” “is,” and “motion” (or their equivalents in some other language) and could certainly now understand the term “molecular” (or its equivalent), since he or she (he, at a guess, based on my admittedly limited knowledge of history of science) had *coined* it. And yet, he probably didn’t know that heat was molecular motion, and no amount of purely logical reasoning could have let him infer that heat was molecular motion.
So, if “Heat is molecular motion” isn’t analytic, it must be synthetic–that is, it must express a real fact about the world. I’d say, and I think most would be willing to agree with me, that it expresses this fact in the language of thermodynamics.
Now, can this fact be expressed purely as a fact of dynamics, where you still get to use terms like “molecular” and “motion,” but not the term “heat?” Definitely not. If we try to do so, we end up with a sentence like, “Molecular motion is molecular motion.” And that sentence clearly is analytic–in fact, anyone who understands what “is” means can deduce its truth from that alone. So it really doesn’t express any facts about the world–in particular, it doesn’t express the fact that heat is molecular motion.
Laws like this (laws of identity that span sciences), sometimes called bridge laws, all exhibit this phenomenon. They’re synthetic claims that cannot be expressed purely in the language of the reducing science. So reductionism is, in fact, false.
(And yes, to be truly accurate, the title of this piece should be “The Unreducability of Identity Theory,” not “The Unreducability of Reductionism.” Euphony carries the day.)
Post a Comment