Putnam’s views on issues of metaphysics and language have actually evolved extensively over time, and his main arguments for realism about natural kinds, which I’ll present here, do not accurately represent his views now, or for that matter, any time since about 1980. “Internal realism,” his current (last I checked) ontology, is extremely nuanced, and I’m not going to talk about it right here. Instead, the opponent whose arguments for natural kinds realism is, specifically, the Putnam of the 1970s. Putnam, in my opinion, deserves great respect for being one of the few philosophers to be completely forthright about changing his mind; he freely admits that his metaphysics has changed dramatically and that, in his opinion, his earlier incarnation was simply wrong.
Why investigate views whose own author has repudiated? The fact is, most of the philosophical community doesn’t agree with Putnam about his own earlier theories. I’d estimate that, currently, the hardcore scientific realist Putnam of the 1970s has more adherents than the internal realist of the 1980s and beyond.
Putnam’s Original Argument
Putnam had two very distinct arguments for realism about natural kinds. The famous “twin cases” that he discusses in “Meaning and Reference,” which I’ll talk about in this post, and a somewhat less famous (but much better, in my opinion, even though I don’t think they work either) argument from “Explanation and Reference,” which I’ll discuss in a separate post.
In “Meaning and Reference,” Putnam created a thought experiment he called Twin Earth. Imagine, he said, that in addition to our Planet Earth, there was an almost identical planet, many hundreds of light years away, called Twin Earth. The only difference, down to the finest detail, between Twin Earth and Earth is that on Twin Earth, the “water” (that is, the thing that fulfills all of the roles of water here on Earth, including being called “water” and so on), is not H2O, but rather a different compound which we (following Kripke; Putnam doesn’t name it) will call XYZ.
Now, let’s imagine a person (I’ll call him “Bob”) on Earth, and his counterpart, Twin Bob (called “Bob” by all his friends, of course–Twin Bob is just our name for him) on Twin Earth. Bob and Twin Bob have identical properties that they associate with the term “water.” But, argues Putnam, they don’t mean the same thing with the term “water,” because Bob’s “water” is H2O, whereas Twin Bob’s “water” is XYZ. So the extension of “water” can’t be a matter of any properties associated with the term; therefore, just as Kripke argued, the causal theory of reference must apply to the term “water,” which requires that water be a natural kind in a realist sense.
To start with, I want to make a distinction between two separate interpretations of Putnam’s example.
- In one interpretation, there are people on Earth and Twin Earth who can distinguish H2O from XYZ; Bob and Twin Bob just aren’t among them (perhaps only professional chemists are capable of making the distinction). Since Putnam saw himself as arguing primarily against internalism, this, I believe, was the interpretation he wanted.
- In a second interpretation, there are no people on Earth and Twin Earth who can distinguish H2O from XYZ.
I have a lot I might say about the first interpretation, but most of that isn’t really required here; I’ll talk about such things in a post that has nothing to do with natural kinds. What’s important for out purposes here is that linguistic externalism of this sort (that the meaning of words doesn’t depend entirely on properties associated with them by their speakers) does not entail natural kinds realism at all. Rather, as people like Tyler Burge have suggested, the meaning of these terms can be assigned by members of a small segment of society (professional chemists, in this case) to which other users of the term can defer. The extension of such terms can be determined by whether or not given samples satisfy properties associated with such terms; it’s just that not every individual speaker of the language is directly involved in this association.
What about the second interpretation? Essentially, the second interpretation suggests that both Bob and Twin Bob live in societies that are pre-modern chemistry. And I’ve already argued that, in such societies, the term “water” refers indiscriminately to H2O and XYZ. So Bob and Twin Bob do, indeed, refer to the same substance (all H2O and XYZ, plus any other chemicals that act sufficiently similar) with their terms.
So one way or the other, this argument of Putnam’s does not prove anything resembling realism about natural kinds.
To be continued…