The effects of familiarity and semantics on early acquisition of Japanese numeral classifiers

要旨

Japanese is one of the languages that has a numeral classifier system (e.g., Aikhenvald, 2000). The use of numeral classifiers is obligatory when enumerating, and numeral classifiers occur as a bound morphemes attached to a number. The selection of numeral classifiers is determined by the semantic properties of the referent being counted. Cross-linguistic studies of numeral classifiers indicate that animacy, shape, and function are often the crucial semantic properties on which the selection of numeral classifiers is based (e.g. Adam & Conklin, 1973). In addition, the primary function of classifiers in general is not referential, unlike nouns that identify referents; rather they denote the membership of referents in different categories based on the ways we perceive and interact with them (Denny, 1976). As such, numeral classifiers are considered a linguistic system that reflects conceptual categorization (e.g., Aikhenvald, 2000; Yamamoto, 2000, 2005). Therefore, investigations into how children acquire numeral classifiers could serve as a window onto the processes by which categorization develops over a period of time. There are about 150 Japanese numeral classifiers, of which approximately 30 are used regularly and daily (Downing, 1996). When selecting appropriate numeral classifiers, Japanese uses the crucial semantic properties observed cross-linguistically. First, the distinction between animate and inanimate referents is respected. For example, when counting people, a numeral classifier -ri for one or two people or -nin for more than two people is used as in (1), whereas counting non-human animates requires another numeral classifier -hiki as in (2).

主題