Individual Differences in Verb Bias Sensitivity in Children and Adults With Developmental Language Disorder
Individual Differences in Verb Bias Sensitivity in Children and Adults With Developmental Language Disorder
复制标题
DOI：
10.3389/fnhum.2019.00402
复制
发表时间：
2019-11-19
影响因子：
2.9
通讯作者：
Farmer, Thomas A.
中科院分区：
文献类型：
--
作者：
Hall, Jessica E.;Van Horne, Amanda Owen;Farmer, Thomas A.
许多实验支持统计信息对儿童和成人语言学习和处理的假设效用。然而，这些研究的任务通常是非常笼统的，只有少数包括患有发育性语言障碍(DLD)的人群。我们想要确定，当统计学习的衡量标准与语言任务的相关性被选择时，当包括大量DLD参与者时，是否会显示出更强的关系。我们测量的语言能力是对动词偏误的敏感度--动词出现的可能性与特定的论点或解释有关。之前的一项研究表明，与典型的同龄人相比，患有DLD的成年人对动词偏见不那么敏感。尚未对DLD儿童进行动词偏向敏感性测试。在研究1中，49名7-9岁的儿童，其中17名被归类为DLD，完成了一项旨在通过内隐和外显测量来测量对动词偏见的敏感性的任务。我们发现，有DLD的儿童和没有DLD的儿童在内隐而不是外显的测量中对动词偏向表现出敏感，不同组之间没有差异。在研究2中，我们使用多元宇宙方法来调查统计学习中的个体差异是否预测了这些参与者以及成年参与者的数据集中的动词偏见敏感性。我们的分析没有发现统计学习与儿童动词偏向敏感性之间存在联系的证据，这并不出人意料，因为我们在研究1中没有发现组内差异。统计学习预测成人对动词偏向的敏感度是通过显性测量测量的，尽管结果并不可靠。这些发现表明，学龄儿童的动词偏向可能仍然相对不稳定，因此在儿童的句子加工中可能不会发挥与成人相同的作用。DLD患者在加工过程中使用的机制也可能与他们成年后典型的发育(TD)同龄人不同。因此，统计信息在有无DLD的个体中与语言处理的相关性可能不同。
A number of experiments support the hypothetical utility of statistical information for language learning and processing among both children and adults. However, tasks in these studies are often very general, and only a few include populations with developmental language disorder (DLD). We wanted to determine whether a stronger relationship might be shown when the measure of statistical learning is chosen for its relevance to the language task when including a substantial number of participants with DLD. The language ability we measured was sensitivity to verb bias - the likelihood of a verb to appear with a certain argument or interpretation. A previous study showed adults with DLD were less sensitive to verb bias than their typical peers. Verb bias sensitivity had not yet been tested in children with DLD. In Study 1, 49 children, ages 7-9 years, 17 of whom were classified as having DLD, completed a task designed to measure sensitivity to verb bias through implicit and explicit measures. We found children with and without DLD showed sensitivity to verb bias in implicit but not explicit measures, with no differences between groups. In Study 2, we used a multiverse approach to investigate whether individual differences in statistical learning predicted verb bias sensitivity in these participants as well as in a dataset of adult participants. Our analysis revealed no evidence of a relationship between statistical learning and verb bias sensitivity in children, which was not unexpected given we found no group differences in Study 1. Statistical learning predicted sensitivity to verb bias as measured through explicit measures in adults, though results were not robust. These findings suggest that verb bias may still be relatively unstable in school age children, and thus may not play the same role in sentence processing in children as in adults. It would also seem that individuals with DLD may not be using the same mechanisms during processing as their typically developing (TD) peers in adulthood. Thus, statistical information may differ in relevance for language processing in individuals with and without DLD.