Perspective-taking in Conversation

目前的语言使用理论认为，通过在对话中分享想法，人们形成了共享的知识。然而，这一假设在很大程度上未经检验，有充分的理由认为它是错误的。调查人员的初步研究提出了一个具有根本意义的基本问题：谈话者和听众在结束谈话时对所说内容的看法是否相同？答案是否定的。说话者比听者更容易记住自己说了什么。本研究以对话记忆中的这些不对称为出发点，以期对对话中记忆和语言使用之间的相互作用产生新的预测。预期的研究结果将有助于深入了解对话伙伴对所说内容的记忆何时相似，以及何时可能不同。由此产生的发现可能会改变关于如何在高风险环境（商业谈判、纠纷调解、第一次约会）以及日常生活中进行成功对话的基本建议。研究结果还可以指导人们通过与“Alexa”等设备的对话来预测人类可能记住的内容，从而为设备设计和人机交互提供见解。对教学法的潜在影响包括改善师生互动和小组项目中的沟通的建议。这一建议发展并测试了共享会话知识的理论观点的预测，该观点提供了与当前观点的根本背离，并做出了竞争性的预测。一系列的10个实验验证了这样一个假设，即在对话之后，合作伙伴会保留不同的对话记忆。因此，尽管他们有共同的谈话经历，但他们对所说的话会有不同的看法，以可预测的方式影响未来的决策和互动。这项工作还将为生态有效环境下的语言处理机制提供见解。更广泛的影响包括研究生，本科生和高中生进行独立研究项目的机会。为从代表性不足的群体中招收学生作出了特别努力。这项研究产生的所有去识别数据将在互联网上公开提供。总之，这项工作有可能促进对心理和语言过程的理解，推进计算机对话系统，并为教学研究提供见解。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

Current theories of language use assume that by sharing ideas in conversation, people develop shared knowledge. This assumption, however, is largely untested and there is good reason to think that it is wrong. The investigator's preliminary studies asked a basic question with fundamental implications: Do speakers and listeners each walk away from a conversation with the same idea about what was said? The answer is no. Speakers are much more likely to remember what was said than the person who was listening. The proposed work takes as a starting point these asymmetries in memory for conversations in order to generate novel predictions regarding the interplay between memory and language use in conversation. Anticipated findings will provide insight into when conversational partners are likely to have similar memories for what was said, and when their memories are likely to differ. The resulting discoveries could change basic recommendations for how to engage in a successful conversation, both in high stakes settings (business negotiations, dispute mediation, first dates), but also in day-to-day life. The findings may also guide expectations about what humans are likely to remember based on conversations with devices such as "Alexa," in turn providing insights for device design and human-technology interactions. Potential implications for pedagogy include recommendations for improving communication in teacher-student interactions and in group projects. This proposal develops and tests the predictions of a theoretical view of shared conversational knowledge that offers a radical departure from current views and makes competing predictions. A series of 10 experiments tests the hypothesis that following conversation, partners will retain predictably different memories of the conversation. As a result, despite their shared experience of conversing, partners will have a different perspective on what was said, influencing future decisions and interactions in predictable ways. The work will also provide insights into mechanisms of language processing in ecologically valid settings. Broader impacts include opportunities for graduate students, undergraduate students, and high school students to conduct independent research projects. Special efforts are made to recruit students from under-represented groups. All de-identified data generated by this research will be made publicly available on the internet. In sum, the work has the potential to advance understanding of psychological and linguistic processes, advance computer dialog systems, and offer insights into pedagogical research.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.