Part of
Usage-based Approaches to Japanese Grammar: Towards the understanding of human languageEdited by Kaori Kabata and Tsuyoshi Ono
[Studies in Language Companion Series 156] 2014
► pp. 171–192
Three different sets of empirical data are compared to examine the usage distributions of ni and e, two particles that have been traditionally considered to be interchangeable in marking allative relations. The data from the speaker’s judgment test are more consistent with the semantic analyses of the two particles, exhibiting interaction between semantic contexts and particle preference. However, the usage distributions in the spoken and written corpora exhibited a strong bias for ni regardless of the semantic environment. The discrepancy between what is generally considered to be ‘acceptable’ in Japanese grammar and the actual usage data stresses the impact of the empirical data in the further understanding of lexical representation.