Ninio Anat
The Hebrew University of Jerusalem,Israel.
J Child Lang. 2016 Sep;43(5):1103-30. doi: 10.1017/S030500091500046X. Epub 2015 Sep 2.
The environmental context of verbs addressed by adults to young children is claimed to be uninformative regarding the verbs' meaning, yielding the Syntactic Bootstrapping Hypothesis that, for verb learning, full sentences are needed to demonstrate the semantic arguments of verbs. However, reanalysis of Gleitman's (1990) original data regarding input to a blind child revealed the context of single-word parental verbs to be more transparent than that of sentences. We tested the hypothesis that English-speaking children learn their early verbs from parents' single-word utterances. Distribution of single-word transitive verbs produced by a large sample of young children was strongly predicted by the relative token frequency of verbs in parental single-word utterances, but multiword sentences had no predictive value. Analysis of the interactive context showed that objects of verbs are retrievable by pragmatic inference, as is the meaning of the verbs. Single-word input appears optimal for learning an initial vocabulary of verbs.
成人对幼儿说出的动词所处的环境被认为对于动词的意义并无信息价值，由此产生了句法引导假设，即对于动词学习而言，需要完整的句子来展示动词的语义论元。然而，对格利特曼（1990）关于盲童输入的原始数据进行重新分析后发现，父母说出的单个动词所处的环境比句子的环境更具透明度。我们测试了这样一个假设：说英语的儿童从父母的单个词话语中学习他们早期的动词。大量幼儿产出的单个及物动词的分布情况，能被父母单个词话语中动词的相对词频有力地预测，但多词句子则没有预测价值。对互动环境的分析表明，动词的宾语以及动词的意义都可以通过语用推理来获取。单个词输入似乎最有利于学习动词的初始词汇。
