Abstract

A pair of interpreters of a speaker's sentences can disagree by assigning different truth-values to sentences in the speaker's language that the speaker neither accepts nor rejects. Alternately, they can assign different truth-values to some sentence that the speaker accepts as true. Neither source of disagreement is open to the speaker: on pain of inconsistency in the latter case, and ex hypothesis, the speaker neither accepts nor rejects the contested sentence in the former case. Arguably, interpreters possibilities of disagreement do not undercut first-person semantic authority.