Cognitive mechanisms underlying the spread of conventions in communities

该奖项是美国国家科学基金会社会、行为和经济科学博士后研究奖学金（SPRF）计划的一部分。SPRF计划的目标是为学术界、工业界或私营部门和政府的科学事业准备有前途的早期职业博士级科学家。SPRF奖励包括在知名科学家的赞助下进行为期两年的培训，并鼓励博士后进行独立研究。美国国家科学基金会寻求促进科学界各阶层的科学家，包括那些未被充分代表的群体的科学家，参与其研究项目和活动；博士后阶段是实现这一目标的重要专业发展阶段。每个博士后必须解决各自学科领域的重要科学问题。在普林斯顿大学的Thomas Griffiths和Adele Goldberg的赞助下，这个博士后奖学金奖支持一位研究语言惯例如何在社区中传播的早期职业科学家。为了成功地使用语言进行交流，说话者必须对单词的含义有一致的约定。然而，正如现代传播媒体所揭示的那样，这些语言惯例经常处于变化之中：新的含义迅速在社交网络中传播并被广泛采用。与此同时，不同子群体的讲话者在他们认为有意义的惯例上可能存在很大差异（例如，科学期刊中的缩写词）。这个项目研究了适应性认知机制，使说话者能够驾驭这种复杂的意义景观，以及这些机制如何反过来解释习俗如何在社区内部和跨社区传播。提出的社会期望如何表现和概括的模型为应用正式模型来更广泛地校准对社会规范和习俗的干预奠定了基础。这些干预措施有望解决从吸烟等公共卫生问题到参与投票等集体行动等大规模社会问题。此外，这项工作可能有助于理解语言习俗在种族或基于ses的社区中是如何变化的。通过对社交网络上的参照交流进行一系列大规模的行为研究，以及最先进的语言使用计算认知模型，本研究提出了一个新的假设，即说话者如何在不同的语境中理性地更新和部署他们对语言惯例的心理表征。这一假设在一个层次贝叶斯模型中得到了正式的精确，并使用新的行为数据与以前的模型进行了严格的比较。首先，该项目将研究支持新语言惯例在社区内传播的泛化机制。其次，该项目将调查发言者如何学习和灵活地在不同社区的不同惯例之间转换。第三，层次模型将使用深度神经网络架构进行扩展，以产生关于自然语言描述的更细粒度预测，从而实现进一步的理论发展。这个项目使用计算认知科学的思想，在两个跨学科传统（心理语言学和文化进化）之间建立新的联系，加深我们对大群体如何协调他们的信仰以更好地相互理解的理解。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。

This award was provided as part of NSF's Social, Behavioral and Economic Sciences Postdoctoral Research Fellowships (SPRF) program. The goal of the SPRF program is to prepare promising, early career doctoral-level scientists for scientific careers in academia, industry or private sector, and government. SPRF awards involve two years of training under the sponsorship of established scientists and encourage Postdoctoral Fellows to perform independent research. NSF seeks to promote the participation of scientists from all segments of the scientific community, including those from underrepresented groups, in its research programs and activities; the postdoctoral period is an important level of professional development in attaining this goal. Each Postdoctoral Fellow must address important scientific questions that advance their respective disciplinary fields. Under the sponsorship of Thomas Griffiths and Adele Goldberg at Princeton University, this postdoctoral fellowship award supports an early career scientist investigating how linguistic conventions spread through communities. To successfully communicate using language, speakers must share consistent conventions about the meanings of words. As modern communication media has revealed, however, these linguistic conventions are often in flux: novel meanings rapidly spread across social networks and become widely adopted. At the same time, speakers across different sub-communities may vary substantially in which conventions they find meaningful (e.g. the acronyms filling scientific journals). This project investigates the adaptive cognitive mechanisms that allow speakers to navigate this complex landscape of meaning, and how these mechanisms may in turn explain how conventions spread within and across communities. The proposed model of how social expectations are represented and generalized lays the groundwork for applying formal models to calibrate interventions on social norms and conventions more broadly. These interventions hold promise as solutions to large-scale social problems ranging from public health concerns like smoking cigarettes to participation in collective actions like voting. Additionally, this work could contribute to understanding how linguistic conventions vary across ethnic or SES-based communities. Using a series of large-scale behavioral studies of referential communication on social networks and state-of-the-art computational cognitive models of language use, the proposed work tests a novel hypothesis about how speakers rationally update and deploy their mental representations of linguistic conventions in different contexts. This hypothesis is made formally precise in a hierarchical Bayesian model and rigorously compared against previous models using new behavioral data. First, the project will examine the mechanisms of generalization supporting the spread of new linguistic conventions within a community. Second, the project will investigate how speakers learn and flexibly switch between different sets of conventions in different communities. Third, the hierarchical model will be extended using a deep neural network architecture to produce finer-grained predictions about natural language descriptions, thus enabling further theory development. This project uses ideas from computational cognitive science to make new connections between two interdisciplinary traditions (psycholinguistics and cultural evolution) that deepen our understanding of how large groups of people coordinate their beliefs to better understand one another.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.