The previous post is what some people might call a "teachable moment". Everyone knows what "me" and "I" mean, after all -- they refer back to the speaker/writer of the utterance/sentence. It's not terribly difficult to follow. But how exactly do we do it?
There are a number of theoretical approaches to the study of meaning, aka semantics. Most of them -- but by no means all -- fall into a general category called "model-theoretic semantics", which, translated back out of jargon, simply means that they rely on a formal model of the world. These models are generally mathematical sets plus some facts about the world expressed in a formal language. Each set is associated with some label, which usually denotes a common property of the items inside the set.
Anyways, in many models, you will find a set labeled "individuals". Inside that set are a whole bunch of "me"s and "I"s. "But wait!", you say. Isn't it redundant to have the same item inside the set over and over again? And yes, it is -- or it would be, if it wasn't for the semantic notion of an index. Each one of those "me"s and "I"s carries a numerical index. So there's actually "me-1", "me-2", "me-3", etc. The link between the model and the real world comes from the fact (or at least, a claim) that you -- yes, you-2 -- associate each of those indices with a real object. Well, not necessarily a real object (or else we couldn't name fictional/imaginary objects), but at least with something else that concretely identifies a discrete individual in your mind.
The above is, of course, not the whole story. In a fuller model than the one above, there would be a unique context for each "me", and the indices would belong to the contexts. Your own personal "me" would refer to the context in your head that was specifically yours. But contexts are tricky things, as this guy could tell you (read his paper "A Plea For Monsters", when you're ready to have your context rocked).
So if you want a name which is context free, you need to get yourself a rigid designator. Saul Kripke, a philosopher of language, coined this term to refer to a label which refers to the same thing in all possible worlds -- a possible world being just what it sounds like, one way that the world could be arranged, but is not necessarily at present. I don't know if contexts were part of the models Kripke discussed in his original paper on the subject, but the notion certainly generalizes to them as well. "Semantic Compositions" refers to just one blog, and that's true in both my context and yours (and Google's -- I checked before I started this).
Comments