When philosophers pose theories of language, are they implicitly dealing with just human language, or are their theories meant to address all possible languages?