Team: Nadine Bade (ObTrEx, U Tübingen) and Florian Schwarz (U Pennsylvania)
The aim of this collaboration is to assess whether inferences resulting from violating the principle Maximize Presupposition behave differently from presuppositions and implicatures, thus testing predictions of theories which separate those inferences out from these more well-studied aspects of meaning (Sauerland 2008, Percus 2006). For that purpose, we want to run a series of eyetracking studies on definites/ indefinites and other items whose competition is based on Maximize Presupposition. This aspect of meaning has not been studied systemically using psycho-linguistic methods, with few exceptions.
It has been observed that presupposition triggers have to be used if their presupposition (PSP) is fulfilled in the context. Heim (1991) proposed an account based on the principle Maximize Presupposition (MP). According to theories working with MP, PSP triggers are ordered on a scale of a presuppositional strength with their non-presuppositional counterparts (Percus 2006, Chemla 2008). One of these scales orders the definite and indefinite determiner. The indefinite yields the inference that the PSP of the definite is false (“anti-uniqueness”) due to this competition, which is why it is infelicitous in (1).
(1) {The / # A} father of the victim came.
Anti-presuppositions have been argued to be distinct from both PSPs and implicatures since they are epistemically weak and project out of negation (Sauerland 2008). The weak status of anti-uniqueness is reflected by the fact that while the use of an indefinite in (1) allows for the inference that it is not certain that there is exactly one father, this cannot be strengthened to mean that it is certain there is not exactly one father. Alternative views on the competition between the definite and indefinite are that they both come with their own context restrictions, i.e. that the indefinite comes with a novelty condition (Heim 1982) or its own PSP of anti-uniqueness (Kratzer 2004). These make different predictions for the processing profiles associated with anti-uniqueness.