Phonological feedforward and feedback mechanism in reading aloud : a mixed-effect modelling analysis

Abstract

Feedforward mechanisms (print-to-sound) in Chinese word naming has been widely investigated, whilst feedback mechanisms (sound-to-print) are left unclear. Mixed results of feedback effects, including facilitative, inhibitory or null effect, were reported. The current study aims to examine the feedforward phonological regularity and consistency effect, and feedback mechanisms through homophone density using a Cantonese database. University students were recruited to name 4376 Chinese characters. Mixed effect modelling was used to analyze participants’ naming latency and accuracy. Results demonstrated that irregular words were named significantly slower and more errors than five other regularity conditions (i.e., non-phonetic compounds, bound radical, regular, semi-regular and multiple pronunciations radical). Feedforward consistency was significantly facilitated word naming in low-frequency word. Number of homophones alone did not affect naming. However, consistency effects were found for low-frequency words with more homophones, whereby highly-consistent words were named faster than words with low-consistency. The results indicated the role of feedback consistency which made facilitative effects in high consistent word. The study revealed the importance to include other regularity conditions, especially non-phonetic compounds, to provide a comprehensive overview of Chinese word naming. Furthermore, modification of current word recognition models was needed to explain feedback effects from phonology.