Abstract

I suggest a revised meta-linguistic account that distinguishes between the language used to talk about a particular language -- the meta-language -- from direct speech reports made within a language -- the object language. Making this distinction leads to a kind of meta-linguistic analysis of competence with names that is not simply tautologous, so long as competence with names is not construed as knowing this: 'Tyler' is whatever is called 'Tyler'. Rather, it should be this: the name 'Tyler' is whatever speakers of an object language call "Tyler," where the double quotes represent direct speech acts. This makes it possible that the schema, 'X' is whatever speakers of an object language call "X," could very well be informative, and serve as an explanation of what a speaker understands when competent with a name. Well, it would insofar as a speaker could learn something about the meaning of the the name 'Tyler' by examining its utterances. In fact, something like this must have to have some truth to it. After all, even non-fluent children learn to speak a language, in part, by attending to utterances of certain expressions in particular situations. The major difference is that when we are finally competent with the expression 'duck' it is not because we have grasped the previous schema. It is because we will engage in duck-awareness behaviors when our environment has something to with ducks at a certain time. In contrast, simply understanding the previous schema is exactly what competence with names consists in.