Is our referential ability something always mediated by some sort of descriptions, senses, intensions, connotations, etc.? Does it have sense to say that, at least in some cases, we get to refer to the world in a direct, non-mediated way? An affirmative answer to the first question gives place to the so called “descriptivist theories of reference”. An affirmative answer to the second one gives place to “non-descriptivist theories of reference”, also called “theories of direct reference”. Frege is the paradigmatic example of descriptivism. Russell and Kripke are paradigmatic examples of non-descriptivism. Stuart Mill also defended a non-descriptivis position. For that reason, being non-descriptivist is “to maintain a Millean theory of reference”.

Orthogonal to the mentioned tension between descriptivist and non-descriptivist theories of reference, there are two main ways of explaining our referential ability. We can try to explain it as derived from some intentions or we can try to explain it as derived from some objective facts (for instance, causal facts, informational facts, etc.). The problem is that even if we were in the ideal situation of knowing all the possible truths about us, ourselves and about the world, the references of the terms of our languages, and the references of our own thought, would remain indetermined. References could change without any change in the truth values of the sentences.

The situation we have just described is a version of Quine’s thesis about the indetermination of reference. Truth value can be determined by the way things are. The world also can determine the references of our languages and thoughts. And references can determine truth values. There is, however, a radical indetermination of reference by truth values. Truth values do not determine references. Moreover, truth values do not determine that we get to refer. All the truth contained in an ideal description of the world would be compatible with the non-existence of such a world.