ISSN 1671-3710
CN 11-4766/R
主办:中国科学院心理研究所
出版:科学出版社

• •    

语言经验对概率词切分的影响

于文勃1,王璐1,程幸悦1,王天琳2,张晶晶3,梁丹丹1   

  1. 1. 南京师范大学
    2. 纽约州立大学奥尔巴尼分校
    3. 南京师范大学心理学院
  • 收稿日期:2020-06-28 修回日期:2020-07-31 出版日期:2020-10-19 发布日期:2020-10-19
  • 通讯作者: 梁丹丹
  • 基金资助:
    音乐学习促进语言韵律加工的认知和神经机制;江苏省高校优势学科项目

  • Received:2020-06-28 Revised:2020-07-31 Online:2020-10-19 Published:2020-10-19

摘要: 概率词切分指个体利用音节间的转换概率切分语流、发现词语边界的过程。经典的概率词切分研究多采用“学习-测试”范式,首先要求被试切分一段无意义人工语言,随后对切分效果进行测试。近年来,研究者逐渐关注语言经验对概率词切分的影响,本文从被试的语音经验和被试掌握的语言知识两方面来阐述语言经验如何对概率词切分产生影响。今后的研究中,一方面可以在语言经验的分类上进行拓展,细分群体语言经验和个体语言经验的影响,一方面还可以更多地关注普通话母语者的语言经验如何作用于概率词切分过程。

关键词: 语言经验, 概率线索, 词切分

Abstract: Statistical word segmentation or statistical learning (SL) refers that individuals could segment continuous speech by tracking the likelihood of syllable co-occurrences (Transitional Probability, TP). The classic “learning -test” paradigm was widely used in this field, in which participants were firstly required to segment an artificial language and the learning effect would be test in the following forced choice task. However, the potentially experience-dependent aspect of SL has received more attention in recent years. In this paper, we discuss how linguistic experience place influence on the process of verbal SL from two aspects, one is the knowledge of phonological experience, the other is the prior knowledge of subjects. In future studies, investigation of how to separate two types of linguistic experience by test method and how Mandarin’s linguistic experience influences statistical word segmentation performance might be critical insight into better understanding of statistical word segmentation.