Suppr超能文献

语言处理中的扩展视角转换与话语经济性

Extended Perspective Shift and Discourse Economy in Language Processing.

作者信息

Harris Jesse A

机构信息

Department of Linguistics, University of California, Los Angeles, Los Angeles, CA, United States.

出版信息

Front Psychol. 2021 Mar 31;12:613357. doi: 10.3389/fpsyg.2021.613357. eCollection 2021.

Abstract

Research spanning linguistics, psychology, and philosophy suggests that speakers and hearers are finely attuned to perspectives and viewpoints that are not their own, even though perspectival information is not encoded directly in the morphosyntax of languages like English. While some terms seem to require a perspective or a judge for interpretation (e.g., epithets, evaluative adjectives, locational PPs, etc.), perspective may also be determined on the basis of subtle information spanning multiple sentences, especially in vivid styles of narrative reporting. In this paper, I develop an account of the cues that are involved in evaluating and maintaining non-speaker perspectives, and present an economy-based discourse processing model of perspective that embodies two core principles. First, perspectives are subject to a "speaker-default," but may shift to a non-speaker perspective if sufficient contextual cues are provided. Second, the processor follows the path of least resistance to maintaining perspective, opting to maintain the current perspective across sentences as long as the shifted perspective continues to be coherent. The predictions of the model are tested in a series of offline and online studies, manipulating the form of an attitude report and the tense of the sentence that follows. Implications for processing perspective and viewpoint in speech and narrative forms are explored.

摘要

跨越语言学、心理学和哲学领域的研究表明,尽管视角信息并未直接编码在像英语这样的语言的形态句法中,但说话者和听话者却能敏锐地察觉到并非他们自己的视角和观点。虽然有些术语似乎需要一个视角或评判者来进行解读(例如,绰号、评价性形容词、方位介词短语等),但视角也可能基于跨越多个句子的细微信息来确定,尤其是在生动的叙事报道风格中。在本文中,我阐述了与评估和维持非说话者视角相关的线索,并提出了一种基于经济性的视角话语处理模型,该模型体现了两个核心原则。首先,视角遵循“说话者默认”原则,但如果提供了足够的上下文线索,视角可能会转向非说话者视角。其次,处理器遵循维持视角的阻力最小路径,只要转换后的视角保持连贯,就会选择在句子之间维持当前视角。该模型的预测在一系列离线和在线研究中得到了检验,这些研究对态度报告的形式以及后续句子的时态进行了操控。同时还探讨了该模型对处理言语和叙事形式中的视角和观点的启示。

https://cdn.ncbi.nlm.nih.gov/pmc/blobs/83d4/8044538/ca609855ed3f/fpsyg-12-613357-g0001.jpg

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