On direct and parallel modification in Japanese: an experimental study

Abstract

This study examines the empirical plausibility of two hypotheses on attributive modification in Japanese proposed in a previous study: (i) material and nationality/origin modifiers in Japanese are direct modifiers if they are not involved in parallel modification, and (ii) Japanese nominals containing two attributive modifiers with no pause between them can involve parallel modification when the first modifier is long, but not otherwise. To investigate whether the observation underlying these hypotheses is reproducible, we conducted an online survey experiment. The results indicate that the observation is not fully reproduced. Additionally, a prediction made by the hypotheses is not borne out, suggesting the possibility that at least one of them is implausible. Based on additional observations, we show that at least Hypothesis (ii) is untenable and suggest that the nominals described therein can involve parallel modification regardless of the length of the modifiers.