Pragmatic Strengthening in Plural Predications and Donkey Sentences


Συγγραφέας: Manfred Krifka


Manfred Krifka: Pragmatic Strengthening in Plural Predications and Donkey Sentences (pdf, 55K)
The classical analysis of donkey sentences like (1.a,b) in Kamp (1981) and Heim (1982) assigns them truth conditions as given in (2.a). That is, they are treated as quantifications over farmer-donkey pairs. Partee (1984) and Kadmon (1987) have pointed out that the proper reading of (1.b), and a preferred reading of (1.a), is rather a quantification over farmers, as illustrated in (2.b). (1) a. lf a farmer owns a donkey, he usually beats it. b. Most farmers who own a donkey beat it. (2) a. MosT({(x,y)l FARMER(x) A DoN1<EY(y) A owN(x,y)}, {<><,y>l BEAT<><,y>}) b. MosT({xl E|y[ FARMER(X) A DoN1<EY(y) A owN(x,y)] }, {xl x BEATs(DoNKEY o1¤(x)) }) c. MOST(A, B) <:> card(ArB) > card(A—B) (2.a) is called the symmetric interpretation, and (2.b) the (subject-) asymmetric interpretation by Kadmon (1987). ln (2.b), the donkey variable y is called dependent. With asymmetric interpretations the question arises how the dependent variable is interpreted within the second argument, the nuclear scope. This issue is taken up in Rooth (1987). He distinguishes two cases, a "weak” or existential interpretation as in (3.a), and a “strong" or universal interpretation as in (3.b). (3) a. MosT({xl E|y[FARMER(x) A DoN1<EY(y) A owN(x,y)] }, {xl Ely[FARMER(x) A DoNKEY(y) A owN(x,y) A BEAT(x,y)] }) b. MosT({xl E|y[FARMER(x) A DoN1<EY(y) A owN(x,y)] }, {xl Vy[1¤ARMER(x)A DoN1<EY(y) A owN(x,y)] + BEAT(x,y)] }) (3.a) says that most farmers that own a donkey beat at least one of their donkeys. (3.b) says that most of those farmers beat each of their donkeys. The question is: What determines the choice of reading in asymmetric quantifications?