Entwicklung eines Modells der menschlichen Satzverarbeitung aufbauend auf Konzepten der Optimalitätstheorie und der Rational Analysis. Umsetzung des Models durch Computersimulationen mit Hilfe von corpusbasierten Techniken aus der Computerlinguistik; Test

基于最优理论和理性分析的概念开发人类句子处理模型。

• 批准号: 5322502
• 负责人: Dr. Frank Keller
• 金额: --
• 依托单位: School of Informatics
• 依托单位国家: 德国
• 项目类别: Emmy Noether International Fellowships
• 财政年份: 2001
• 资助国家: 德国
• 起止时间: 2000-12-31 至 2003-12-31
• 项目状态: 已结题
• 来源: https://gepris.dfg.de/gepris/projekt/5322502?language=en
• 关键词: Entwicklung; eines; Modells; der; menschlichen

Human sentence processing is highly automatic and takes place with great robustness and accuracy. Current models of human parsing typically focus on the processing of sentences that are carefully constructed to investigate how the parser reacts to ambiguities in the input. While successful in predicting certain structural and lexical preferences of the human parser, these models have difficulties explanining why the parser is generally robust. The proposed research program assumes that these limitations can be overcome by applying the concept of optimization to human sentence processing. Optimization is at the heart of recent developments in cognitive psychology and theoretical linguistics. I will develop an optimization-based account of human parsing using corpus-based and probabilistic techniques, providing a fully implemented parsing model that explains the robustness and broad coverage of the human sentence processor. By using concepts from Optimality Theory, this model will also offer a systematic account of crosslinguistic variation in human parsing. Its predictions will be tested against a broad range of experimental data from both English and German.

人类句子处理是高度自动化的，并且具有很高的鲁棒性和准确性。目前的人类解析模型通常集中在仔细构造的句子的处理上，以研究解析器如何对输入中的歧义做出反应。虽然成功地预测人类解析器的某些结构和词汇偏好，但这些模型很难解释为什么解析器通常是鲁棒的。拟议的研究计划假设，这些限制可以通过应用优化的概念来克服人类的句子处理。优化是认知心理学和理论语言学最新发展的核心。我将使用基于语料库和概率的技术开发一个基于优化的人类解析账户，提供一个完全实现的解析模型，解释人类句子处理器的鲁棒性和广泛的覆盖范围。通过使用优选论的概念，该模型还将提供人类句法分析中跨语言变异的系统解释。它的预测将根据来自英语和德语的广泛实验数据进行测试。