This paper, originally published in 2001, deals with the question of the source of presuppositions, focusing on the question of whether presuppositions are conventional properties of linguistic expressions, or arise as inferences derivable from ordinary content in combination with some general conversational principles. I argue that at least some presuppositions should be analysed as conversational inferences, on the grounds that they show two of the hallmarks of such inferences: contextual defeasibility and nondetachability. I make this case for the presuppositions associated with change of state predicates and with factives. I argue further for the need for a general principle for deriving presuppositions as inferences by illustrating a variety of cases of presupposition-like inferences not clearly involving a lexical presupposition trigger. In the second half of the paper, I move towards the development of a general conversational account of the relevant presuppositions. Building on a brief comment in Stalnaker (1974), I develop the following pair of ideas: first, that an utterance embedding a proposition P may be seen as raising the question whether P; and second, that P may be related to a further proposition Q in such a way that it would make sense to raise the question whether P only if one already believed Q to be true. It is these required prior beliefs that constitute conversationally derived presuppositions. Although the account developed here is only a preliminary attempt, the relevance of contextually salient questions, or sets of alternatives, to an account of presupposition has been taken up in subsequent work, notably Abusch (2010) and Simons et al. 2010.

Republished in Capone, A. et al. (eds), Perspectives on Linguistic Pragmatics. Springer, 2013, 329-348.