This paper examines popular 'conventionalist' explanations of why philosophers need not back up their claims about how 'we' use our words with empirical studies of actual usage. It argues that such explanations are incompatible with a number of currently popular and plausible assumptions about language's 'social' character. Alternate explanations of the philosopher's purported entitlement to make a priori claims about 'our' usage are then suggested. While these alternate explanations would, unlike the conventionalist ones, be compatible with the more social picture (...) of language, they are each shown to face serious problems of their own. (shrink)