Behavioural and pupillary effects of homophone density on Chinese character naming

Abstract

Homophones (words with same pronunciation but different meanings) is an important metalinguistic component for Chinese reading. However, how people process them remains unclear. There are mixed findings on whether homophone density effect is facilitative or inhibitory and whether the locus of homophone density effect lies in semantics or phonology. This study aims to resolve the above debates and gives insight on the current Chinese word recognition models. Twenty university students were recruited to name 4,376 Chinese characters that were of varying homophone density, number of senses and phonological neighborhood density. Behavioral measures (reaction time and accuracy) and pupillary measures (mean peak pupil dilation and mean pupil size) were analyzed using mixed-effect modelling. Result revealed that high homophone density led to longer reaction time and inhibited naming. Importantly, homophone density and number of senses a character carries interacted, indicating semantics as locus for homophone competition. Phonology determined naming accuracy; accuracy decreased with increased phonological neighborhood density. Yet, the pupillary measures were insensitive to homophone size. These findings using Cantonese and numerous characters suggested that the inhibitory effect of homophone density arises from competition amongst meanings of co-activated homophone mates in semantics. Lastly, a feedback connection between phonology to semantics and inhibitory links in phonology were suggested in the current lexical constituency model.