ASTRACT. I argue that (a) the phenomenon characteristic of pragmatic presupposition, is distinct from (b) the phenomenon characteristic of semantic presupposition, and that there are sentences exhibiting (a) alone. I apply this to Stalnaker's defense of van Fraassen's theory of semantic presupposition against Karttunen. I show that, since Stalmaker fails to distinguish (a) from (b), this defense amounts to an unsuccessful attempt to explain pragmatically the supposed instances of (b) in Karttunen's counter-examples. I observe that, given the distinction between (a) and (b), we can salvage from Stalnaker the strategy of discounting certain of Karttunen's counter-examples which exhibit not (b) but merely (a).
- Traditionally, presupposition has been taken to be a semantic notion. Letting 'P' and 'Q' be metavariables ranging over sentences, we may define this with (l).* 01 P semantically presupposes C? iff p's being either true or false necessitates that Q is true. However, in a set of recent papers, Robert Stalnaker introduces a second, prugmutic notion of presupposition, definable with (2).3 ca P pragmatically presupposes Q iff whenever the utterance of P is conversationally acceptable, the speaker of P assumes Q and believes his audience to assume Q as well. ' An earlier draft of this paper was presented at the meetings of the Eastern Division APA in New York, December 27-30,1975. In preparing the present version, I have benefitted from the comments of Jay Atlas, Asa Kasher, and my anonymous referee. I would also like to thank Lauri Karttunen and Robert Stalnaker, both of whom have been very generous in their willingness to discuss their views with me. I hope that I do not misrepresent them in what follows. I should add that work on this paper was supported in part by grants from the Mount Holyoke Faculty Grants Committee. * van Fraassen, [ 101, p. 68. I observe van Fraassen's distinction between 'necessitation' and 'semantic entailment'; cf. his [9], esp. p. 138. ' Stalnaker, [6], pp. 200-203. Stalnaker provides a number of alternative definitions for his notion of pragmatic presupposition. (2) represents my selection of one from among these; Stalnaker himself pointedly declines to choose a favorite here. It can be verified that nothing in what follows turns upon this selection; all of my contentions can be recast to fit any of Stalnaker's alternatives.
It is worth adding that while Stalnaker holds pragmatic presupposition to be invariant under denial, he appears to exclude this invariance from his actual definition of this notion. It will be noted that my attack on his commitment to (3) (cf. infru) depends on the supposition of this exclusion (cf. i&r, sect. 2). If this point of interpretation is contested, then I might have to allow (3). However, I would then argue a corresponding point, niz. that in view of sentences ( 7),( 9), (ll), (13),and(l5) (' f rn ru, sect. 2) -and also sentence (19) (in@, sect. 3) -we need a notion of pragmatic presupposition whose definition does exclude this invariance. Thus, the point of interpretation determines only the form -but not the substance, of what I have to say.