2008年6月27日: 中西公子, Scalarity of -Mo; 相原昌彦, Does Japanese Allow Degree Abstraction? Evidence from Superlative Constructions

The additive particle -mo `also, too' in Japanese gives rise to a scalar interpretation when it follows a numeral. In positive contexts, the relevant quantity is considered to be large (e.g. go-nin-mo ki-ta), while ambiguity arises in negative contexts (e.g. go-nin-mo ko-nakat-ta). I argue that -mo with a numeral triggers a scalar presupposition analogous to the one introduced by English even and that the scalar presupposition of -mo is responsible for "large" and "small" readings. I further show that the "large" reading in negative contexts serves as an argument for the scope theory of even (Karttunen and Peters 1979, among others). I then extend the proposed analysis to the cases where -mois used with `one' and show that the analysis is capable of explaining the following observations: while an unaccented `one' + -mo sequence (e.g. pan-o iti-mai-mo tabe-nakat-ta `(I) didn't eat any bread') serves as a negative polarity item, an accented sequence yields a similar ambiguity as the case considered above (e.g. pan-o iti¬-mai-mo tabe-nakat-ta, where ¬ marks the fall accent).

It has been observed since Ross (1964) that the superlative construction such as "John climbed the highest mountain" has two readings. One is called the absolute reading, where the heights of the relevant mountains in a relevant context are compared; the other is called the comparative reading, where relevant climbers' achievements of mountain climbing are compared (Szabolcsi 1986). Two theories have been proposed regarding this ambiguity. One theory holds that the ambiguity of this kind is derived by moving the superlative morpheme -est (i.e., movement theory) (Heim 1985, 1999; Szabolcsi 1986). The other theory holds that the ambiguity is derived by assignment of different values to the context variable C, keeping a single LF structure where -est stays in-situ (i.e., in-situ theory) (Farkas & Kiss 2000, Sharvit & Stateva 2002). As is pointed out by Heim (1999), decision between these theories is hardly made based solely on English. In this talk, through an investigation of Japanese superlative constructions, I argue that movement theory is required at least in Japanese. If the analysis of the superlative constructions proposed in this talk is on the right track, it argues for the view that Japanese comparative constructions involve degree abstraction (contra Beck, Oda, and Sugisaki 2004).