Speech, Language and Hearing Sciences, Purdue University, West Lafayette, Indiana, United States of America.
PLoS One. 2019 Jul 11;14(7):e0219290. doi: 10.1371/journal.pone.0219290. eCollection 2019.
The mature lexicon encodes semantic relations between words, and these connections can alternately facilitate and interfere with language processing. We explore the emergence of these processing dynamics in 18-month-olds (N = 79) using a novel approach that calculates individualized semantic structure at multiple granularities in participants' productive vocabularies. Participants completed two interleaved eye-tracked word recognition tasks involving semantically unrelated and related picture contexts, which sought to measure the impact of lexical facilitation and interference on processing, respectively. Semantic structure and vocabulary size differentially impacted processing in each task. Category level structure facilitated word recognition in 18-month-olds with smaller productive vocabularies, while overall lexical connectivity interfered with word recognition for toddlers with relatively larger vocabularies. The results suggest that, while semantic structure at multiple granularities is measurable even in small lexicons, mechanisms of semantic interference and facilitation are driven by the development of structure at different granularities. We consider these findings in light of accounts of adult word recognition that posits that different levels of structure index strong and weak activation from nearby and distant semantic neighbors. We also consider further directions for developmental change in these patterns.
成熟的词汇系统编码了单词之间的语义关系,这些联系可以交替促进和干扰语言处理。我们使用一种新的方法,探索了这些处理动态在 18 个月大的婴儿(N=79)中的出现,该方法在参与者的生产词汇中以多个粒度计算个体化的语义结构。参与者完成了两个交替的眼动追踪单词识别任务,涉及语义上不相关和相关的图片上下文,分别旨在衡量词汇促进和干扰对处理的影响。在每个任务中,语义结构和词汇量大小都对处理有不同的影响。类别水平结构促进了词汇量较小的 18 个月大的婴儿的单词识别,而整体词汇连通性则干扰了词汇量相对较大的幼儿的单词识别。结果表明,虽然在小词汇量中也可以测量多个粒度的语义结构,但语义干扰和促进的机制是由不同粒度的结构发展驱动的。我们根据假设不同层次的结构索引来自附近和遥远语义邻居的强和弱激活的成人单词识别理论来考虑这些发现。我们还考虑了这些模式在发展变化方面的进一步方向。