演讲制作
生产(经济)
认知心理学
心理学
语言生产
音节
语音识别
理解力
预测(人工智能)
心理语言学
积极倾听
计算机科学
手指敲击
脑磁图
沟通
名词
语音处理
韵律
言语感知
句子处理
轮流
异步(计算机编程)
口语
作者
Liu Shun,Suiping Wang,Xiqin Liu
出处
期刊:Language and Speech
[SAGE Publishing]
日期:2026-04-29
卷期号:: 238309261441302-238309261441302
标识
DOI:10.1177/00238309261441302
摘要
Although the production system is thought to play a vital role in language prediction, direct evidence demonstrating both enhancement and suppression of real-time prediction within a consistent paradigm is scarce. To explore this problem, our study used a dual-task visual world eye-tracking paradigm across two experiments, investigating the production system's influence on prediction by both impairing and enhancing its availability. In Experiment 1, participants either performed articulatory suppression (silently repeating the syllable /pa/) or finger tapping while listening to sentences. Results showed that articulatory suppression attenuated anticipatory eye movements toward target objects, compared to finger tapping. This suggests that reducing the availability of the production system diminished predictive processing. In Experiment 2, participants listened to sentences and concurrently performed speech shadowing (covertly repeating the heard speech in real time). By comparing its results with previous data, we found that speech shadowing enhanced anticipatory fixations relative to listening, indicating that increasing production engagement facilitates prediction. In both experiments, prediction effects were tested before the target noun was presented, providing unambiguous evidence of anticipation. Together, these findings offer integrated, bidirectional causal evidence that engaging production processes during comprehension can both strengthen and weaken prediction, depending on whether production resources are augmented or constrained. This research advances our understanding of the mechanisms underlying prediction in language processing and underscores the importance of production in shaping anticipatory comprehension.
科研通智能强力驱动
Strongly Powered by AbleSci AI