CAREER: The Role of Good-Enough Processing in Language Comprehension

该奖项是根据2009年美国复苏和再投资法案（公法111-5）资助的。语言是一种极其复杂的信号，充满了多种不同类型的信息，包括单词（词汇）及其较小的部分（形态学）、句法结构（语法）、输入和前一个输入之间的联系（上下文）、输入和我们的一般世界知识之间的联系（合理性）。尽管信息很复杂，但人类的大脑处理传入的语音和文本的速度非常快，而且通常非常有效。然而，在处理过程中确实会出现错误，也会产生误解。然而，直到最近，很少有人注意到当人们误解语言输入时会发生什么。这个项目将极大地扩展所谓“足够好”的语言处理的研究。“足够好”的处理描述的是当一串语言输入中的一个或多个信息源没有被人类的句子处理机制整合到最终的解释中时发生的情况。目的是确定在各种语言理解任务中使用哪些信息源，并探讨手头任务的作用和处理能力的个体差异。这个项目的方法将结合在线测量听和看场景时的眼球运动、阅读时的眼球运动、自定节奏阅读、失语语言理解的损伤研究以及与离线理解任务的脑电波（erp）。实验将以英语和其他几种语言进行，参与者包括以英语为母语的人和以非英语为母语的人。我们的主要目标是理解“足够好”的处理是如何运作的，在大多数情况下，它实际上已经足够好了，以及它如何以及为什么会失败。理解语言是人类一切活动的基础。为了完全理解理解，理解误解是如何产生的是很重要的。这个项目试图揭示产生可预测的、系统的理解错误的机制。通过这样做，预计该项目将在开发更好的评估和测试语言理解的方法方面有价值，无论是在教育环境内还是在教育环境外。“足够好”的框架不仅对成年人的语言理解有帮助，而且对非母语人士、儿童和有语言疾病的人也有帮助。因此，本项目中各种研究的结果有可能以非常具体的方式影响语言教学和干预实践。就更近期的目标而言，该项目将培养博士生的实验设计和方法，包括眼动追踪和ERP数据收集，以及复杂的数据分析技术。该项目还将为本科生提供这些方法的经验，并为研究生学习做好准备。在该项目的最后一年，将举行一个关于“足够好”处理的讲习班，汇集研究人员和结果。

This award is funded under the American Recovery and Reinvestment Act of 2009 (Public Law 111-5). Language consists of an extremely complex signal, replete with information of several different types, including words (lexicon) and their smaller parts (morphology), syntactic structure (grammar), connections between input and previous input (context), and connections between input and our general world knowledge (plausibility). Despite the complexity of the information, the human mind processes incoming speech and text extremely fast and, generally, very efficiently. Errors in processing do occur, however, and misinterpretations arise. Until quite recently, though, little attention has been paid to what happens when people do misinterpret language input. This project will greatly expand the growing body of research on so-called "good enough" language processing. "Good enough" processing describes what happens when one or more information sources in a string of language input is not integrated by the human sentence processing mechanism into a final interpretation. The goal is to identify which information sources are exploited in various language comprehension tasks, and to explore the role of the task at hand and individual differences in processing capacity. The methodology for this project will combine online measures of eye-movements while listening and while viewing scenes, eye-movements while reading, self-paced reading, lesion studies of aphasic language comprehension, and brain waves (ERPs) with offline comprehension tasks. Experiments will be run in English and several other languages, with both native-speaking and non-native-speaking participants. The main goal is to understand how "good enough" processing operates such that, most of the time, it is in fact good enough and also how and why it can fail.Language comprehension is fundamental to all human endeavors. In order to completely understand comprehension, it is important to understand how miscomprehension arises. This project seeks to lay bare the mechanisms by which predictable, systematic comprehension errors are generated. By doing so, it is expected that this project will be valuable in developing better methods of evaluating and testing language comprehension, both inside and outside of educational settings. The "good enough" framework is informative to language comprehension not only by adult native-speakers, but also non-native speakers, children, and people with language pathologies. As such, the results of the various studies within this project have the potential to influence language teaching and intervention practices in very concrete ways. With respect to more near-term goals, the project will train doctoral students in experimental design and methodologies, including eye-tracking and ERP data collection, and in sophisticated data analysis techniques. The project will also provide undergraduate students with experience in these methods, and prepare them for graduate study. In the last year of the project a workshop on "good enough" processing will be held to bring together researchers and results.