The Effect of Familiarity and Semantics on Early Acquisition of Japanese Numeral Classifiers

Hiromi Sumiya and Eliana Colunga

Abstract for paper presented at
Boston University Conference on Language Development
Boston, MA
November 5, 2005

This study investigated comprehension of six "basic" Japanese numeral classifiers that are divided into category-based (-nin/-ri for people, -hiki for small animals, -dai for vehicles and machineries) and shape-based (-ko for three-dimensional objects, -hon for one-dimensional objects, -mai for two-dimensional objects) with a special focus on young children's ability to generalize their knowledge about classifiers. Ninety monolingual Japanese children aged 3- to 5-years-old participated in the study employing a forced-choice comprehension task. The stimuli consisted of familiar or unfamiliar real objects presented in a between-participants design. The results showed that comprehension increased with age except comprehension of -ko, whose performance level remained at 50% even among older children. The results further indicated that comprehension of category-based classifiers was significantly better than that of shape-based classifiers. These results suggest that certain features, such as animacy and semantic transparency between classifiers and referents, have a significant impact on the acquisition of numeral classifiers.

Hiromi Sumiya is a doctoral student in the Department of Linguistics at the University of Colorado. She can be reached at hiromi.sumiya@colorado.edu.

Colorado Research in Linguistics - Volume 19, Issue 1 - June 2006