Another standard method for detecting presuppositional expressions is con- sidering their inferential potentials in conditionals. It is widely agreed that a sentence which contains a presuppositional expression licenses a presuppositional inference even when it is embedded in the antecedent of a conditional. Again, the presuppositional inferences thereby contrast classical entailments and implicatures since the latter tend to disappear when embedded in such syntactic environments.
(174) a. Bertrand stopped smoking.
↝Bertrand used to smoke. (presuppositional inference) ↝Someone stopped smoking. (classical entailment)
b. If Bertrand stopped smoking, his boss gave him a raise. ↝Bertrand used to smoke.
↝̸Someone stopped smoking. (175) a. e Governor of Alaska resigned
↝ere is a Governor of Alaska. (presuppositional inference) ↝Someone resigned. (classical entailment) b. If the Governor of Alaska resigned, she’s running for president.
↝ere is a Governor of Alaska.
↝̸Someone resigned.
Now, to demonstrate that there is a difference between existential-there sentences and sentences with weak determiners in subject position, we care- fully consider which inferences these sentences license in relevant complex syntactic environments. And since the difference is quite subtle, I use a method devised by von Fintel (1998) to make it easier to detect. e method is to preface each conditional with a statement indicating that the speaker is agnostic about the existence ofFs (whereF is the restrictor for the determiner). is should have the following effect: If the speaker is openly agnostic about existence ofFs, the existence ofFs cannot be common ground. Hence, it should then be infelicitous for a speaker to use an expression which requires that the existence of a non-empty set of Fs is already common ground. Each sentence in (176a-176c) should therefore be read as immediately following the initial assertion in (176).
(176) I’m not sure whether there are any apples in the pantry.
a. If there are some apples in the pantry and they’re ripe, we should bake an apple pie.
b. If there are several apples in the pantry and they’re ripe, we should bake an apple pie.
c. If there is an apple in the pantry and it’s ripe, we should bake an apple pie.
e discourse continuations in (176a-176c) are felicitous even though the speaker has openly declared herself agnostic about the existence of apples in the pantry. But, now compare these to the continuations in (177a-177c). (177) I’m not sure whether there are any apples in the pantry.
a. # If some apples in the pantry are ripe, we should bake an apple pie.
b. # If several apples in the pantry are ripe, we should bake an apple pie.
160 | Projection Tests
c. # If an apple in the pantry is ripe, we should bake an apple pie. e discourse continuations in (177a-177c) should be equivalent to the continuations in (176a-176c). at is, there should be interpretations of (177a-177c) available where the weak determiners take scope inside the if-clauses.5 But these interpretations are simply unavailable as the con- tinuations in (177a-177c) are clearly infelicitous. And given the standard existential analysis, this cannot be explained.
In contrast, if these determiners are assumed to trigger existence presup- positions, the infelicity is to be expected. If the weak determiners require that an existential presupposition is common ground, it is no surprise that a speaker cannot rst declare herself agnostic about the status of some propositionpand then proceed to use a term that requires thatpis common ground.6
5 In fact, as already pointed out in Chapter III, the interpretation on which the determiner phrase is embedded inside the scope of the if-clause should not only becompatiblewith these sentences, it should bemandatorysince if-clauses are scope islands.
6 While certain discourse particles, e.g. ‘but’ or ‘however’, seem to improve felicity judgments about (177a-177c), these discourse particles are prototypicalcontrast markers. eir discursive function is to contrast information already given with information to be conveyed. To explicate precisely the semantic and pragmatic effects of discourse particles such as ‘but’ or ‘however’ is a complicated affair, but observe that these discourse particles also succeed in improving felicity judgments for sentences that contain standardly accepted presuppositional expressions, e.g. factives and aspectual verbs.
(1) I’m not sure whether Peter is having a party.
a. . . . but if John knows that Peter is having a party, we should go! b. # . . . and if John knows that Peter is having a party, we should go! c. # If John knows that Peter is having a party, we should go! (2) I’m not sure whether Sue used to smoke.
a. . . . but if she stopped, she’s probably drinking more coffee. b. # . . . and if she stopped, she’s probably drinking more coffee. c. # If she stopped, she’s probably drinking more coffee.
e contrast is perfectly clear. e (b) and (c)-sentences are downright infelicitous whereas the (a)-sentences are not. is, I take it, is not evidence that factive or aspectual verbs fail to trigger presuppositions. Rather, it is evidence that these discourse particles facilitate coherent interpretations. Explaining this data is a complicated task that I am unable to undertake here, but my tentative (albeit rather vague) suggestion is that these contrast markers shi modal salience; the contrast markers indicate to the interlocutors that the continuation should be interpreted at a possible world where the presupposition is satis ed. I.e. one could imagine that these discourse particles license