Department of Psychology, University of Notre Dame, USA.
J Child Lang. 2020 Mar;47(2):483-507. doi: 10.1017/S0305000919000230. Epub 2019 Jun 13.
There is considerable controversy over the factors that shape infants' developing knowledge of grammar. Work with artificial languages suggests that infants' ability to track statistical regularities within the speech they hear could, in principle, support grammatical development. However, little work has tested whether infants' performance on laboratory tasks reflects factors that are relevant in real-world language learning. Here we tested whether the language that infants hear at home, and their receptive language skills, predict their performance on tasks assessing the ability to learn non-adjacent statistical dependencies (NADs) at 15 months, and whether that in turn predicts sensitivity to native-language NADs at 18 months. We found evidence for some (though not all) of these relations, and primarily for females. The results suggest that performance on the artificial language-learning task reveals something about the mechanisms of grammatical development, and that females and males may be learning NADs differently.
关于影响婴儿语法发展的因素存在着大量争议。针对人工语言的研究表明,婴儿在听到的语音中追踪统计规律的能力原则上可以支持语法发展。然而,很少有研究检验过婴儿在实验室任务中的表现是否反映了在现实世界语言学习中相关的因素。在这里,我们测试了婴儿在家中听到的语言以及他们的接受性语言技能是否可以预测他们在 15 个月大时评估学习非相邻统计关系(NADs)的能力的任务表现,以及这种表现是否反过来可以预测他们在 18 个月大时对母语 NADs 的敏感性。我们发现了这些关系的一些证据(尽管不是全部),主要是针对女性的。结果表明,在人工语言学习任务中的表现揭示了语法发展机制的一些情况,并且女性和男性可能以不同的方式学习 NADs。