The “orthodox” theory about reference, fathered by Russell, held that it was some collection of properties that determined the reference of a word. Specifically, it was thought that the word referenced the object that was uniquely picked out by those properties. And there have been numerous variations on this theory, some that state that the object referenced is the one that best fits those properties (even if some are not applicable to the object), and some that hold that different properties are more or less important in determining what is referenced. In contrast to this the causal theory holds that the reference is determined (at least for some words) by the object that was the causal origin of the word (or some variation on this theme). My goal here is not to argue that one of these theories about reference is right and the other is wrong, as I have established previously there is no right answer to that question. What I do want to address is how proponents of the causal theory attempt to argue against the “orthodox” theory.
One way to argue against the “orthodox” theory is to attempt to show that no collection of properties can pick out objects referenced by the word. And causal theorists attempt to do this by appealing to possible worlds. Their argument essentially goes as follows: let us assume there is some property, X, that picks out the referenced objects (or at least contributes to picking them out) for the word “Q”. The causal theorist admits that perhaps all “Q”s in this world are X. But, he or she says, surely there are possible world in which “Q”s are not X (it is possible for a “Q” to not be X, although all “Q”s in this world may actually be X). But “Q” still picks out the “Q”s in this possible world, even though they do not have this property. So obviously “Q” references some objects that are not X. For example, let us consider the reference of “Shakespeare”. For any property we might define “Shakespeare” in terms of (author of …, son of …, the man called “Shakespeare”, ect) we can always assert that it is possible that in some world Shakespeare is none of those things (not the author of those plays, born to different parents, called by a different name, ect). Since this holds for all properties (and combination of properties) it is impossible for properties to determine reference. And therefore the “orthodox” theory must be wrong.
This argument is fairly robust, but it slips in at least one questionable assumption when it invokes possible worlds. Specifically it doesn’t say how we can know what “Q”s are in other possible worlds. More specifically, we know what is and is not a “Q” in this world (based on some definition, in terms of properties), but how do we know what is and is not “Q” in our possible worlds, given, as the argument requires, that there are some of them that do not meet any definition that we can construct in terms of properties?
What is going on here is that some kind of essentialism has been subtly introduced, without our being aware of it, probably because essentialism is a very intuitive way of thinking about the world. By essentialism I mean the idea that an object is of a certain kind because it shares some indefinable “essence” with others of its kind, and essence that cannot be captured by properties (unless we admit properties such as “being Q”, but these are strange properties, properties that cannot be detected in any way). For example, going back to our example of Shakespeare, we can legitimately ask the causal theorist “why is this other man, in this other possible world, Shakespeare?” (or “how do you know he is Shakespeare?”). And there is no answer the causal theorist can give besides “he just is”, or “because he is essentially Shakespeare”.
And I think this is a serious weakness in the argument against the “orthodox” theory. For starters very few believe in essentialism, and I wouldn’t want to attribute such a belief to the majority causal theorists. And there are several reasons to reject essentialism (besides its unpopularity). One is simply Occam’s razor, essences play no observable role, and therefore the simplest explanation is that they don’t exist. Another with them is that they are basically unknowable. Because essences aren’t tied to any observable properties there is no way to know if an object does or doesn’t have a particular essence, and thus it is hard to say why we would come to believe that an objects have essences to begin with (we certainly don’t have a special essence-detecting sense).
And so, because I think that we have good reason to reject essentialism, no matter how intuitive it might be, I think we should also reject this argument against the “orthodox” theory. Now this is not to say there might not be good arguments against the “orthodox” theory. As I alluded to earlier I think the only facts about reference are conventional (that is to say that there is no “right” theory of reference), but there is a fact of the matter about which theory of reference people actually subscribe to. Thus it would be possible to construct an argument against the “orthodox” theory by showing that people don’t actually subscribe to it, that it is unintuitive in some way. But this seems like a job better suited to cognitive scientists (who could design and run an experiment to test this hypothesis, by posing carefully structured questions to a large group of people).