Center for Language Studies, Radboud University, Nijmegen, The Netherlands.
Max Planck Institute for Psycholinguistics, Nijmegen, The Netherlands.
Psychon Bull Rev. 2019 Dec;26(6):1850-1869. doi: 10.3758/s13423-019-01634-5.
The status of thematic roles such as Agent and Patient in cognitive science is highly controversial: To some they are universal components of core knowledge, to others they are scholarly fictions without psychological reality. We address this debate by posing two critical questions: to what extent do humans represent events in terms of abstract role categories, and to what extent are these categories shaped by universal cognitive biases? We review a range of literature that contributes answers to these questions: psycholinguistic and event cognition experiments with adults, children, and infants; typological studies grounded in cross-linguistic data; and studies of emerging sign languages. We pose these questions for a variety of roles and find that the answers depend on the role. For Agents and Patients, there is strong evidence for abstract role categories and a universal bias to distinguish the two roles. For Goals and Recipients, we find clear evidence for abstraction but mixed evidence as to whether there is a bias to encode Goals and Recipients as part of one or two distinct categories. Finally, we discuss the Instrumental role and do not find clear evidence for either abstraction or universal biases to structure instrumental categories.
在认知科学中，主题角色（如施事和受事）的地位存在很大争议：对一些人来说，它们是核心知识的普遍组成部分，而对另一些人来说，它们是没有心理现实的学术虚构。我们通过提出两个关键问题来解决这场争论：人类在多大程度上以抽象的角色类别来表示事件，以及这些类别在多大程度上受到普遍认知偏见的影响？我们回顾了一系列有助于回答这些问题的文献：针对成人、儿童和婴儿的心理语言学和事件认知实验；基于跨语言数据的类型学研究；以及新兴手语的研究。我们针对各种角色提出这些问题，并发现答案因角色而异。对于施事和受事，有强有力的证据表明存在抽象的角色类别，并且存在将这两个角色区分开来的普遍偏见。对于目标和接受者，我们发现有明确的证据表明存在抽象化，但对于是否存在将目标和接受者编码为一个或两个不同类别之一的偏见，证据并不一致。最后，我们讨论了工具角色，并没有发现明确的证据表明存在抽象化或普遍偏见来构建工具类别。
