The Amateur Philosopher

Occasionally I’m just going to post about random ideas that are buzzing around in the back of my head. I’ve got three of them going right now: One related to technology, one related to economics, and one–the one I’m going to post here–related to my old field, philosophy, in particular, metaphysics and the philosophy of language.

In particular, I want to look at the premise set forth by Nathan Salmon’s Reference and Essence. Salmon was looking at a pair of theories often taken to be related (though Salmon’s argument was that they weren’t; not really)–the baptismal theory of names, and essentialism. Kripke, famously, proposed the baptismal theory of names in his Naming and Necessity, and had drawn essentialism as a conclusion from it; this step was accepted by much of the philosophy of language community. Salmon wasn’t interested in whether the baptismal theory of names (or, for that matter, essentialism) was accurate or not; he was simply trying to argue that the latter did not follow from the former.

The baptismal theory of names is as follows: Both proper nouns and natural kind terms (such as “tiger”) get their reference non-connotatively: That is, not via those referents satisfying predicates or collections of predicates that are analytically associated with the name. Rather, they get their referents via a connection (identity in the case of proper nouns, an unspecified connection in the case of kind terms) between those referents and the “baptismal source” of the name, the initial object the name was used to refer to. The only uniquely identifying analytic truth involving these terms is that they refer to whatever was used in this baptismal event; any other uniquely identifying fact about them is not part of the meaning of the term. I have lots of problems with this theory, which was what about half of my dissertation was about, but let’s ignore that for now.

The theory of essentialism is as follows: For a given object, there are properties that that object must have, independently of any analytic consequences of the way the object is described. So, for example, there are properties such that Barack Obama *must* have them, properties without which he would not be Barack Obama, independently of any analytic consequences of the actual term “Barack Obama.”

Salmon’s quick intuitive appeal (before he goes through actual argumentation) is, at first glance, a very reasonable one. The baptismal theory is essentially a linguistic theory, a theory about how words relate to their referents. Essentialism is very pointedly not a linguistic theory (that’s what all the “independent of analytic consequences” is about); it’s supposed to be a theory about the nature of objects. And you shouldn’t be able to deduce a metaphysical consequence from a linguistic theory; to put it in Salmon’s much more colorful language, “You can’t pull a metaphysical rabbit out of a linguistic hat.”

Salmon’s more or less right here, but the glibness of this is a bit dangerous. What I’m going to argue is not quite that essentialism follows from the baptismal theory by itself, but that it follows from the baptismal theory plus a relatively uncontraversial metaphysical theory. You can’t pull a metaphysical rabbit out of a linguistic hat, but you can pull a metaphysical rabbit out of a linguistic hat with a subtle and tasteful metaphysical hatband.

The hatband in this case is the denial of a truly bizarre metaphysical idea: Haecceitism. Haecceitism is the idea that objects have an unanalyzable “thisness”–a fact of just being them, which does not follow from any other property they may have, and is not excluded by any property they may lack. To believe in haecceitism is to believe that, for example, it might have been the case that I had all properties attached to my computer, while my computer had all the properties attached to me. To deny haecceitism is, at the very least, to assert the following:

I) For any distinct objects x and y, in any counterfactual situation C, there is some property p, not belonging to y, such that x has p both actually and in C.

And, by letting P be the disjunction of the above properties accross counterfactual situations, we get,

II) For any distinct objects x and y, there is a property P not belonging to y, such that it is necessarily true that x has P.

And letting Q be the conjunction of every P above across values of y, we get,

III) For any object x, there is a property Q such that it is necessarily true that x has Q, and in actuality (at least), Q is uniquely identifying of x.

Now, the baptismal theory of names claims that there are are no uniquely identifying analytic truths about names beyond the baptismal event. And I’ll point out (as Kripke does himself), that the baptismal event is not, in fact, a *necessary* truth about the object or kind in question. Certainly the fact that Barack Obama is named “Barack Obama” is not a necessary truth about him; nor is the fact that tigers are called “tigers” a necessary truth about them. So, if you believe in the baptismal theory of names, you believe that there are no analytic, uniquely identifying, necessary truths about them.

So, suppose x is Barack Obama, and Q is the property satisfying III for x=Barack Obama. Then the following claim is necessarily true and uniquely identifying:

IV.  Barack Obama has the property Q.

By the baptismal theory, there are no uniquely identifying truths involving the name “Barack Obama” that are both necessary and analytic. So, IV is not analytic, so Barack Obama has at least one property (namely Q) necessarily, independently of any analytic consequences of the name “Barack Obama.”

The only way to get the rabbit back into the hat is to accept haecceitism.