Degree constructions in Japanese

Date of Completion

January 2008


Language, Linguistics|Language, Modern




This dissertation provides a study of a language that adds another dimension to the semantics of degree constructions in natural languages that has long been monopolized by analyses that are based on English and related languages. In doing so, I will be particularly concerned with the Japanese language to provide empirical data. ^ I will first adopt an unconventional approach that comparisons in Japanese are contextually made rather than compositionally. Among several possibilities of how such contextual comparisons are made, I will pursue an analysis, under which Japanese gradable adjectives emerge out of the lexicon as comparatives. Thus tall in Japanese denotes "x is d much taller than a contextually given degree c" instead of "x is d-tall." This assumption will be called the lexical analysis. ^ A set of predictions are made based on the lexical analysis. First, the direct degrees of gradable adjectives are bound inside the adjectives; thus, they are never overtly filled. Second, the lexical analysis implies that including a comparative morpheme in the syntax is redundant since the comparative semantics is already introduced in the lexical entries of adjectives. This explains the absence of comparative morphemes in Japanese. The lack of degree operator movement in Japanese comparatives directly follows from the lexical analysis. Third, each gradable adjective introduces a comparison under the lexical analysis. This implies that there are as many comparisons as the number of gradable adjectives in one sentence. This is in fact confirmed in multihead comparatives in Japanese. ^ The lexical analysis affects degree constructions other than comparatives, since the semantics of gradable adjectives is a building block of degree constructions. Thus, the analysis must be tested with various degree constructions in Japanese. I will discuss comparative conditionals and exclamatives. The lexical analysis is assumed to be one of the many parameters that govern large cross-linguistic variations of comparatives. It will be shown that the lexical analysis, coupled with other parameters, accounts for a certain range of Korean and Chinese data. ^