ABSTRACT

The acoustic realization of phrasal prominence is proposed to correlate with the order of V(erbs) and O(bjects) in natural languages. The present production study with 15 talkers of Japanese (OV) and English (VO) investigates whether the speech signal contains coverbal visual information that covaries with auditory prosody, in Infant- and Adult-Directed Speech (IDS and ADS). Acoustic analysis revealed that phrasal prominence is carried by different acoustic cues in the two languages and speech styles, while analyses of motion showed that this acoustic prominence is not accompanied by coverbal gestures. Instead, the talkers of both languages produced eyebrow movements to mark the boundaries of target phrases within elicited utterances in combination with head nods. These results suggest that the signal might contain multimodal information to phrase boundaries, which could help listeners chunk phrases from the input.

摘要

提出了短语突出的声学实现，以与自然语言中的V（erbs）和O（bjects）的顺序相关。这项由15名日语（OV）和英语（VO）说话者进行的生产研究以婴儿和成人定向语音（IDS和ADS）调查了语音信号是否包含与听觉韵律相关的隐蔽视觉信息。声学分析表明，短语突出是由两种语言和语音风格的不同声学提示所承载的，而运动分析表明，这种声学突出并没有掩盖手势。取而代之的是，两种语言的说话者都产生了眉毛运动，以在与头部点头相结合的情况下标记所引出的话语中目标短语的边界。