Anticipatory processing of novel compounds: Evidence from Japanese
Anticipatory processing of novel compounds: Evidence from Japanese
复制标题
新型化合物的预期加工：来自日本的证据
DOI：
10.1016/j.cognition.2014.11.033
复制
发表时间：
2015
期刊：
影响因子：
3.4
通讯作者：
Y. and R. Mazuka
中科院分区：
文献类型：
--
作者：
Hirose;Y. and R. Mazuka
我们的研究认为，头部前预期处理的水平低于句子的水平。视觉世界眼球跟踪研究表明，在处理日语新复合词时，如果前面的修饰成分的韵律信息表明正在应用复合重音规则 (CAR)，则可以在中心词之前构建复合结构。这种韵律提示排除了修饰名词的单头分析，否则这将是一个自然且经济的选择。一旦提前计算了中心部分的结构表示，解析器就会更快地识别复合含义。这对维持结构整合和单词识别是独立过程的模型提出了挑战。与此同时，我们的结果以及之前的发现表明，在理解复合名词的过程中，不同来源的信息的处理过程中可能存在某种程度的分期。
Our study argues that pre-head anticipatory processing operates at a level below the level of the sentence. A visual-world eye-tracking study demonstrated that, in processing of Japanese novel compounds, the compound structure can be constructed prior to the head if the prosodic information on the preceding modifier constituent signals that the Compound Accent Rule (CAR) is being applied. This prosodic cue rules out the single head analysis of the modifier noun, which would otherwise be a natural and economical choice. Once the structural representation for the head is computed in advance, the parser becomes faster in identifying the compound meaning. This poses a challenge to models maintaining that structural integration and word recognition are separate processes. At the same time, our results, together with previous findings, suggest the possibility that there is some degree of staging during the processing of different sources of information during the comprehension of compound nouns.