Suppr超能文献

通过多重约束解决依恋模糊性问题。

Resolving attachment ambiguities with multiple constraints.

作者信息

Spivey-Knowlton M, Sedivy J C

机构信息

Department of Psychology, University of Rochester, NY 14627, USA.

出版信息

Cognition. 1995 Jun;55(3):227-67. doi: 10.1016/0010-0277(94)00647-4.

Abstract

Different theories of syntactic ambiguity resolution argue for different sources of information determining initial parsing decisions (e.g., structurally defined parsing principles, lexically specific biases, or referential pragmatics). However, a "constraint-based" approach to syntactic ambiguity resolution proposes that both lexically specific biases and referential pragmatics are used in parallel by the comprehender. Analyses of text corpora, sentence fragment completions, and self-paced reading experiments were conducted to demonstrate that both local information (lexically specific biases) and contextual information (referential presupposition) contribute to the on-line resolution of prepositional phrase attachment ambiguities. There does not appear to be a role for purely structurally defined parsing principles (i.e., minimal attachment). Present and previous evidence is consistent with a developing framework in which multiple constraints (bottom-up and top-down) interact immediately to determine initial syntactic commitments.

摘要

不同的句法歧义消解理论主张不同的信息来源决定初始解析决策(例如,结构定义的解析原则、词汇特定偏向或指称语用学)。然而,一种“基于约束”的句法歧义消解方法提出,理解者会并行使用词汇特定偏向和指称语用学。对文本语料库、句子片段完成情况以及自定步速阅读实验进行了分析,以证明局部信息(词汇特定偏向)和语境信息(指称预设)都有助于在线消解介词短语附着歧义。纯粹由结构定义的解析原则(即最小附着)似乎不起作用。当前和先前的证据与一个不断发展的框架一致,在该框架中,多种约束(自下而上和自上而下)会立即相互作用以确定初始句法承诺。

文献AI研究员

20分钟写一篇综述,助力文献阅读效率提升50倍。

立即体验

用中文搜PubMed

大模型驱动的PubMed中文搜索引擎

马上搜索

文档翻译

学术文献翻译模型,支持多种主流文档格式。

立即体验