1.1 The Romans built this bridge.
What does it take to make 1.1 true? Surely we do not require that every single Roman participated in the building of the bridge. In other words, we do not view 1.1. as equivalent to 1.2. It seems more appropriate to translate The Romans in 1.1 with the existential quantifier, i.e. as 1.3. Still, sentences with definite plural noun phrases have traditionally been represented with the universal quantifier. This treatment is appropriate for some sentences such as 1.4 below, but apparently not for all cases.
1.2.All the Romans built this bridge.1.3 Some Romans built this bridge.1.4 The soldiers slept.
This thesis is concerned with representativity, the phenomenon exemplified by 1.1.We have a representative interpretation of a sentence when: -- it contains a definite plural noun phrase-- some member or subset of the noun phrase denotation satisfies the predicate-- the rest of the individuals in the noun phrase denotation do not satisfy the predicate-- the predicate is used to characterize the whole group referred to by the noun phrase
Representative predication is quite common in natural language use, but has not received much attention in the semantic and pragmatic literature. The first attempt to give a comprehensive account of the phenomenon is the Ph. D. thesis of Young Eun Yoon Kang, Weak and Strong Interpretations of Quantifiers and Definite Plural Noun Phrases (1994), which was followed up by an article in Natural Language Semantics (1996). Although Kang’s work has been commented on by a number of linguists (cf. Krifka (1996), Brisson (1998), Bech and Sauerland (2000) and Geurts (2002)), only Brisson (1998) undertakes a thorough treatment of the issues Kang raise.
The present thesis is an attempt to account for the semantic and pragmatic factors which make representativity possible, and how these factors can be incorporated into a formal framework.
In the first part of the thesis, we place representativity in the semantic landscape of related notions such as collectivity, cumulativity and genericity. Further, we review the theory presented in Kang (1994, 1996), and compare it to a number of other theories which have approached representativity from different angles. In particular, we compare Kang’s framework with that of Brisson (1998). Brisson claims that representativity is just a sub-phenomenon of a broader notion, nonmaximality. In the present thesis we show that representativity and nonmaximality are essentially different phenomena, and that Brisson’s framework cannot capture representativity.
Then we pass on to specifying our own semantic and pragmatic prerequisites for the representative interpretation. We conclude that this can be best be done by distinguishing three types of representativity which involve different semantic and pragmatic factors.
Finally, we give a formal account of the representative interpretation in Discourse Representation Theory (DRT).