Time course of variability in spoken word recognition

口语单词识别的时间变化过程

基本信息

批准号：
7250061
负责人：
CONOR T MCLENNAN
金额：
$ 7.03万
依托单位：
CLEVELAND STATE UNIVERSITY
依托单位国家：
美国
项目类别：
财政年份：
2006
资助国家：
美国
起止时间：
2006-08-21 至 2010-06-30
项目状态：
已结题

项目摘要

DESCRIPTION (provided by applicant): The long-term objective of the proposed research is to gain a better understanding of the processes and representations underlying spoken word recognition. Substantial empirical evidence now exists in support of the hypothesis that surface variability in speech has both processing and representational implications. The processing work demonstrates that variability in speech, such as a change in the identity of a talker, hinders perception, resulting in slower and/or less accurate spoken word recognition. The representational work demonstrates that variability also has long-term consequences for the perception of spoken words, suggesting that the representations underlying spoken word recognition can include highly specific surface details. Furthermore, recent work also suggests that variability effects follow a time course, manifesting themselves at predictable points during perceptual processing. More specifically, some types of surface variability do not seem to affect early stages of spoken word recognition. Rather, abstract linguistic representations appear to drive the initial on-line perception of spoken words and more specific representations, containing surface details associated with variability, only affect latter stages of processing. The aim of the proposed work is to provide a more immediate and fine-grained analysis of the time course of variability effects. In particular, behavioral measures of the time course effects of two sources of variability, talker-identity and speaking rate, will be collected using the eye-tracking paradigm. The proposed research should ultimately lead to better theories and models of both normal and disordered spoken language perception.

描述（由申请人提供）：拟议的研究的长期目标是更好地了解口语识别的过程和表示形式。现在存在大量的经验证据，以支持言语表面变异性具有处理和代表性的含义的假设。处理工作表明，语音的可变性，例如说话者身份的变化，阻碍感知，从而导致较慢和/或不准确的口语识别。代表性工作表明，可变性对口语的感知也会产生长期的影响，这表明口语识别的表示形式可以包括高度特定的表面细节。此外，最近的工作还表明，可变性效应遵循时间过程，在感知处理过程中以可预测的点表现出来。更具体地说，某些类型的表面可变性似乎不会影响口语识别的早期阶段。相反，抽象的语言表示似乎可以推动对口语和更具体表示的最初在线感知，其中包含与变异性相关的表面细节，仅影响处理的后期阶段。拟议工作的目的是对变异性效应的时间过程进行更直接和细粒度的分析。特别是，将使用眼球轨道范式收集两个可变性，说话者身份和口语率的时间过程影响的行为度量。拟议的研究最终应导致正常语言感知的更好的理论和模型。