The Processing of the Second Syllable in Recognizing Chinese Disyllabic Spoken Words: Evidence From Eye Tracking.

Youxi Wang,Wei Shen,Hua Zhang,Xuelian Zang

doi:10.3389/fpsyg.2021.681337

Abstract

In the current study, two experiments were conducted to investigate the processing of the second syllable (which was considered as the rhyme at the word level) during Chinese disyllabic spoken word recognition using a printed-word paradigm. In Experiment 1, participants heard a spoken target word and were simultaneously presented with a visual display of four printed words: a target word, a phonological competitor, and two unrelated distractors. The phonological competitors were manipulated to share either full phonemic overlap of the second syllable with targets (the syllabic overlap condition; e.g., 小篆, xiao3zhuan4, “calligraphy” vs. 公转, gong1zhuan4, “revolution”) or the initial phonemic overlap of the second syllable (the sub-syllabic overlap condition; e.g., 圆柱, yuan2zhu4, “cylinder” vs. 公转, gong1zhuan4, “revolution”) with targets. Participants were asked to select the target words and their eye movements were simultaneously recorded. The results did not show any phonological competition effect in either the syllabic overlap condition or the sub-syllabic overlap condition. In Experiment 2, to maximize the likelihood of observing the phonological competition effect, a target-absent version of the printed-word paradigm was adopted, in which target words were removed from the visual display. The results of Experiment 2 showed significant phonological competition effects in both conditions, i.e., more fixations were made to the phonological competitors than to the distractors. Moreover, the phonological competition effect was found to be larger in the syllabic overlap condition than in the sub-syllabic overlap condition. These findings shed light on the effect of the second syllable competition at the word level during spoken word recognition and, more importantly, showed that the initial phonemes of the second syllable at the syllabic level are also accessed during Chinese disyllabic spoken word recognition.

Highlights

Humans can understand speech quickly and almost without effort
Data Coding A square region of 10◦ × 10◦centered around each printed word was designated as the region of interest (ROI) for eye tracking
Fixations that fell into the ROI were defined as “fixating on the current word” and those fixations that did not fall on the printed words were considered as falling on the background

Summary

Introduction

Humans can understand speech quickly and almost without effort. Successful speech comprehension requires the segmentation of the continuous speech stream into discrete spoken words, and the mapping of spoken words onto the corresponding lexical representations in the mental lexicon. Current spoken word recognition models have reached a common consent that a set of phonologically similar word candidates compete for activation as the speech signal unfolds. The cohort model (Marslen-Wilson and Tyler, 1980) posits that only competitors sharing similar word-initial phonemes with targets (e.g., “cloud” vs “clothes”) are activated for competition in the earlier phrase of speech perception. Other types of phonologically similar words such as rhymes (i.e., words that share ending phonemes such as “cloud” vs “proud”) are not involved in the competition process. The TRACE model makes an explicit assumption that cohorts and rhymes both participate and compete in the process of spoken word recognition. TRACE predicts that the cohort competitors are activated strongly and earlier while rhyme competitors are activated weakly and later due to the temporal properties of speech information

Objectives

Methods

Results

Conclusion