Department of Psychology, University of Pennsylvania, Philadelphia, Pennsylvania 19104, USA; email:
Department of English, Ithaca College, Ithaca, New York 14850, USA.
Annu Rev Psychol. 2022 Jan 4;73:1-23. doi: 10.1146/annurev-psych-032921-053737. Epub 2021 Oct 8.
The mid-twentieth century brought a radical change in how the linguistics community formulated its major goal, moving from a largely taxonomic science to Chomsky's revolution, which conceptualized language as a higher-order cognitive function. This article reviews the paths (not always direct) that brought Lila Gleitman into contact with that revolution, her contributions to it, and the evolution in her thinking about how language is learned by every child, regardless of extreme variation in the input received. To understand how that occurs, we need to discover what must be learned by the child and what is already there to guide that learning-what must be, in Plato's terms, "recollected." The growing picture shows a learner equipped with information-processing mechanisms that extract evidence about word meanings using various evidential sources. Chief among these are the observational and linguistic-syntactic contexts in which words occur. The former is supported by a mechanism Gleitman and her collaborators call "propose but verify," and the latter by a mechanism known as "syntactic boot-strapping."
二十世纪中叶，语言学界制定其主要目标的方式发生了根本性的变化，从一门主要的分类科学转变为乔姆斯基的革命，将语言概念化为一种更高阶的认知功能。本文回顾了莱拉·格莱蒂曼（Lila Gleitman）接触到那场革命的途径（并非总是直接的）、她对这场革命的贡献，以及她对每个孩子如何学习语言的思考的演变，无论他们接收到的输入有多么极端的差异。为了理解这是如何发生的，我们需要发现孩子必须学习什么，以及有什么已经存在来指导这种学习——用柏拉图的话说，必须“回忆”起什么。不断发展的图景表明，学习者配备了信息处理机制，这些机制利用各种证据来源提取关于词义的证据。其中最重要的是单词出现的观察和语言句法语境。前者由格莱蒂曼及其合作者称为“提出但验证”的机制支持，后者则由被称为“句法自举”的机制支持。
