Mandy Simons

The phenomenon we now know as projection was first observed by Frege in his brief remarks about presupposition in “Sense and Reference.” Frege observes there that the assertion that Kepler died in misery gives rise to the implication that the name Kepler has a referent; but that so too does the assertion that Kepler did not die in misery. Here we have the source of the observation that if p is a presupposition of S, then p is implied by (utterances of) S and by (utterances of) the negation of S. Since Frege, it has been observed that those implications which are shared by a sentence S and by its negation are also typically shared by a variety of other entailment-canceling embeddings of S: in questions, in the antecedents of conditionals, and under epistemic possibility modals. This observation has entered the canon in the form of the “family of sentences” test (Chierchia and McConnell-Ginet 2000). This test is 1 standardly used to demonstrate that some particular element of content projects. An application of the test is demonstrated below

We define a notion of projective meaning which encompasses both classical presuppositions and phenomena which are usually regarded as non-presuppositional but which also display projection behavior—Horn’s assertorically inert entailments, conventional implicatures (both Grice’s and Potts’) and some conversational implicatures. We argue that the central feature of all projective meanings is that they are not-at-issue, defined as a relation to the question under discussion. Other properties differentiate various sub-classes of projective meanings, one of them the class of presuppositions according to Stalnaker. This principled taxonomy predicts differences in behavior unexpected on other models among the various conventional triggers and conversational implicatures, while holding promise for a general, explanatory account of projection which applies to all the types of meanings considered.

Recall Grice’s well-worn example from Logic and Conversation about Smith, his girlfriend, and his trips to New York: (1) A: Smith doesn’t seem to have a girlfriend these days. B: He has been paying a lot of visits to NY recently. Grice says that in this dialogue, B implicates that Smith has, or may have, a girlfriend in New York. But in saying this, Grice under-describes his own example. For this proposition alone does not suffice to satisfy the requirements of Relation, the maxim presumed to be operative in this case. Grice says that “[B] implicates that which he must be assumed to believe in order to preserve the assumption that he is observing the maxim of Relation.” But the assumption that B thinks that Smith might have a girlfriend in NY is not in itself sufficient to render B’s utterance relevant. An additional assumption is required, one which explicitly links the issue of having girlfriends to the issue of travel to NY: perhaps, the proposition that a person who has a girlfriend somewhere travels there frequently; or that many people have long-distance relationships, and these involve frequent trips to the same place. If A can work out that B is making this supposition, then she can immediately see the relevance of B’s response to her remark. Without it, relevance cannot be established. So this general background assumption must be implicated by the utterance.1 Now, this background assumption is not enough by itself to guarantee relevance. Suppose that B believes the background assumption, but does not believe that Smith might be traveling to NY to visit a girlfriend. Then his utterance is still in violation of Relation. B should invoke the background assumption only if it is relevant itself

Grice observes that the primary discourse function of disjunction is the presentation of alternatives, each of which is relevant in the same way to a given topic. After a brief introduction , I offer in Chapter Two an account of Grice's observation and of further felicity conditions on disjunction, for example, the constraint against disjunctions in which one disjunct entails another. Using an enriched version of the Stalnakerian model of assertion, I define two constraints on information update--Relevant Informativity and Simplicity--and show that these suffice to account for the observations. ;I then argue that the requirement to abide by Relevant Informativity and Simplicity provides a basis for an account of two further properties of disjunctive sentences: their presupposition projection properties and the possibility of anaphora between disjuncts. I argue that presuppositions project in disjunction except where projection would lead to infelicity of the kind characterized in Chapter Two. To state this account, a treatment of presupposition which allows projection to be sensitive to considerations of felicity is required. I show that the Stalnakerian view of presupposition, combined with Heim's notion of local accommodation, provides the necessary framework. ;Similarly, I argue that anaphora across disjunction is disallowed only when it leads to an infelicitous disjunction. In certain configurations, anaphora between disjuncts leads to entailment between disjuncts, which is ruled out independently. The account is given using a version of Neale's E-type analysis of unbound anaphora. I show that the analysis has further desirable consequences with respect to more complex examples of inter-disjunct anaphora. ;Throughout Chapters Three and Four, I contrast my own proposals with accounts given in the Dynamic Semantic literature, offering a critique of Dynamic Semantic approaches to the phenomena discussed. ;In Chapter Five, I examine anaphora between NPs contained inside a disjunction and a pronoun in a following sentence. I argue that these pronouns are also E-type, but show that these cases require a reformulation of the E-type analysis. I develop a new account in which the interpretation of E-type pronouns is derived compositionally from the content of the antecedent clause