The Role of Language in Infant Category Learning

认知及其发展研究的核心问题之一是语言对认知的影响以及这种影响的性质。人们提出了各种各样的影响，从促进学习和记忆到改变我们感知和解释事件的方式。此外，口语可能会影响人类智力的核心——从已知情况中归纳和获取知识的能力，包括婴幼儿学习一般类别并将这些知识用于未来学习的能力。尽管这个问题至关重要，但仍有许多未知之处。最重要的是，我们不知道语言是如何影响认知发展的，以及这些影响在发展过程中是如何变化的。词汇是否从发育早期起就作为认知发展的教学信号？或者它们是发展的结果而成为教学信号？这个为期三年的项目的主要目标是通过结合行为实验和眼动追踪分析，以及计算建模和模型选择来回答这些问题。在最初的研究中，将对8- 24个月大的婴儿进行多类别学习实验，在实验中，单词或其他声音会伴随着不同结构的类别。在第二项研究中，将使用两类计算模型来模拟数据，其中一类将单词视为输入的一部分，另一类将单词视为教学信号。正式的模型选择程序将有助于确定哪一种机制的影响下的词是更一致的整个数据体。这项研究的结果将促进我们对语言如何影响早期认知发展的理解，从而有助于更好地理解典型的认知发展。拟议的研究也将有助于发展研究的方法论工具包。研究者将通过会议报告和期刊文章传播研究成果。研究人员还将对儿童和教师进行推广和教育活动，包括高中生和实验室活动，以及在哥伦布科学与工业博物馆为儿童举办的展览。

One of the central issues in the study of cognition and its development concerns the impact of language on cognition and the nature of such effects. A variety of effects have been proposed ranging from facilitation of learning and memory to changing the ways we perceive and interpret events. In addition, the spoken word may affect the very core of human intelligence--the ability to generalize and take knowledge from known to novel situations, including the ability of infants and young children to learn general categories and use this knowledge for future learning. Despite the central importance of the problem, much remains unknown. Most importantly, it is not known how language affects the developing cognition and how these effects change in the course of development. Do words function as teaching signals for the developing cognition from early in development? Or do they become teaching signals as a result of development? The primary goal of this 3-year-long project is to answer these questions by using a combination of behavior experimentation and eye tracking analyses, with computational modeling and model selection. In the initial study, multiple category learning experiments with 8-to-24-month-olds will be conducted, in which words or other sounds will accompany categories of different structure. In a second study, two classes of computational models will be used to simulate the data, one, in which words are considered part of the input, and another in which words are considered a teaching signal. Formal model selection procedures will help determine which mechanism underlying the effects of words is more consistent with the entire body of data. Results of this research will advance our understanding of how language affects cognition early in development, thus contributing to better understanding of typical cognitive development. The proposed research will also contribute to the methodological toolkit of developmental research. The investigator will disseminate the work via conference presentations and journal articles. The researchers will also undertake outreach and education activities with children and teachers, involving high school students and lab activities, and an exposition for children at the Columbus Museum of Science and Industry.