Loanwords Adaptation: A Corpus Study of Sino–Japanese kan-on Initials

抄録

This study investigates the adaptation of Late Middle Chinese loanwords in Early Middle Japanese (i.e. Sino–Japanese kan-on). It focuses on the segmental adaptation of the Chinese initial consonants in Japanese. A corpus of 2136 kanji is assembled and the analysis is based on the Featurally Underspecified Lexicon model (Lahiri, 2018). The empirical data shows that the adaptation of the acoustic input from the source language is driven by the phonology of the borrowing language (Lahiri & Kennard, 2019; Kim, 2008, 2009). The result reveals that instead of matching a foreign segment directly to a native phoneme that most closely approximates the acoustic input, Japanese borrowers extracted phonological features from the acoustic input and mapped them onto the features in their native language. The adaptation is guided by two universal hierarchies: (1) PLACE features take precedent over manner features, (2) within PLACE, ARTICULATOR features rank higher than TONGUE HEIGHT features; and one language specific hierarchy: in Late Middle Chinese, [SPREAD GLOTTIS] dominates [VOICE]. Perceptual similarities are guided by feature specifications and hierarchies.