Abstract：
People learn language variation through exposure to linguistic interactions. The way we take part in these interactions is shaped by our lexical representations, the mechanisms of language processing, and the social context. Existing work has looked at how we learn and store variation in the ambient language. How this is mediated by the social context is less understood. We report on the results of an innovative experimental battery designed to test how learning variation is affected by a variable\'s social indexicality. Hungarian native speakers played a co-operative game involving verb nonwords. These were built on existing inflectional variation in Hungarian. Participant behavior shifted in response to an automated co-player\'s preferences, and this reflected a change in the overall lexical patterns of the players, affected by the particular verbs introduced by the co-player. Patterns persisted in subsequent testing. Learning was similar for variables with or without social meaning. Results show that participants can learn and retain a range of variable morphological patterns in a simulated interaction. Participants seem to have equal capacity to pick up variables with and without social meaning. This suggests that the social meaning of a pattern does not clearly constrain learning morphological variation and becomes relevant downstream in learning.
摘要：
人们通过接触语言互动来学习语言变异。我们参与这些互动的方式是由我们的词汇表示形成的，语言处理的机制，和社会背景。现有的工作已经研究了我们如何学习和存储环境语言的变化。人们对这是如何由社会背景介导的，人们对此知之甚少。我们报告了一种创新的实验电池的结果，该电池旨在测试学习变化如何受到变量的社会索引性的影响。匈牙利母语人士参加了涉及动词非单词的合作游戏。这些是建立在匈牙利现有的屈折变化基础上的。参与者的行为根据自动玩家的偏好而改变，这反映了玩家整体词汇模式的变化，受共同玩家引入的特定动词的影响。模式在随后的测试中持续存在。对于有或没有社会意义的变量，学习都是相似的。结果表明，参与者可以在模拟的相互作用中学习和保留一系列可变的形态模式。参与者似乎具有同等的能力来拾取具有和不具有社会意义的变量。这表明模式的社会意义并没有明显限制学习形态变异，而是与学习的下游相关。
