Direct Reference in Thought and Speech


Kirk Ludwig




In this paper, I argue for the following twelve theses:


  1. Directly referring terms are a commonplace in language.
  2. Nothing follows directly about whether a direct reference theory for thought is correct from a direct reference theory being correct for some terms of natural languages.
  3. A direct reference theory for thought is forced on us by our holding four things: that the universe may be spatio-temporally symmetrical, that we do not know whether it is or not, and that we both uniquely refer and know we uniquely refer to spatio-temporal particulars.
  4. The most conservative solution to the problem of how unique reference to spatio-temporal particulars is possible is that we can refer directly to the self-at-a-time.  This is sufficient in principle for referring to spatio-temporal particulars around us, and if we are to know that we refer sometimes to the self, it is necessary.
  5. A theory of direct reference for thought requires that we have maximally direct knowledge that we refer to the self on some occasions.
  6. Both direct reference to the self and the direct knowledge that we refer to the self must be take as primitive relations.  We know that we bear these relations to the self because they are required to explain the possibility of our knowingly uniquely referring to spatio-temporal particulars.
  7. If we refer to spatio-temporal particulars by means of referring directly to the self-at-a-time, then any language we speak which includes terms that refer to spatio-temporal particulars must contain directly referring terms.
  8. A language which contains directly referring terms does not have to contain a first person pronoun or any other specialized indexicals.
  9. It is in principle possible for a language to contain no indexical expressions at all, but it is a practical impossibility for finite beings like ourselves who need to be able to uniquely refer to an indefinitely large number of particulars.
  10. Indexicals should not be thought of as reducible to a canonical form or to be intertranslatable: it is not required for their linguistic function that they be strictly synonymous with any other expressions in a language, and sentences in which we substitute the rules that can be given to characterize how to determine the referents of indexical expressions for those expressions do not have the same entailment relations as the original.
  11. It is a mistake to read off from the semantics of a sentence in a that-clause of a propositional attitude attribution the content of the thought of the person to whom an attitude is attributed.
  12. Consequently, there is no reason found in ordinary usage to think that there are any singular thoughts other than thoughts about the self.*


* Note that I would now qualify this claim.  See my “Singular Thought and the Cartesian Theory of Mind” (abstract).