Ganushchak Lesya Y, Konopka Agnieszka E, Chen Yiya
Leiden University Centre for Linguistics, Leiden UniversityLeiden, Netherlands; Education and Child Studies, Faculty of Social and Behavioral Sciences, Leiden UniversityLeiden, Netherlands; Department of Psychology, Education and Child Studies, Erasmus University RotterdamRotterdam, Netherlands.
School of Psychology, University of Aberdeen Aberdeen, UK.
Front Psychol. 2017 Feb 28;8:250. doi: 10.3389/fpsyg.2017.00250. eCollection 2017.
This study investigated the time-course of online sentence formulation (i.e., incrementality in sentence planning) as a function of the preceding discourse context. In two eye-tracking experiments, participants described pictures of transitive events (e.g., a frog catching a fly). The accessibility of the agent (Experiment 1) and patient (Experiment 2) was manipulated in the discourse preceding each picture. In the condition, participants heard a story where the agent or patient was mentioned explicitly (). In the condition, the agent or patient was not mentioned but was primed by the story (via semantically or associatively related words such as ). In the condition, the stories did not explicitly mention or prime either character. The target response was expected to have the same structure and content in all conditions (SVO sentences: ). The results showed that participants generally looked first at the agent, before speech onset, regardless of condition, and then at the patient around and after speech onset. Analyses of eye movements in time window associated with linguistic planning showed that formulation was sensitive mainly to whether the agent was literally mentioned in the context or not and to lesser extent to conceptual accessibility (Experiment 1). Furthermore, accessibility of the patient (be it literal mention of its name or only availability of the concept) showed no effect on the time-course of utterance planning (Experiment 2). Together, these results suggest that linguistic planning before speech onset was influenced only by the accessibility of the first character name in the sentence, providing further evidence for highly incremental planning in sentence production.
本研究调查了在线句子构建的时间进程（即句子规划中的增量性）作为前序语篇语境的函数。在两项眼动追踪实验中，参与者描述及物事件的图片（例如，一只青蛙捕捉一只苍蝇）。在每张图片之前的语篇中，对施事（实验1）和受事（实验2）的可及性进行了操控。在 条件下，参与者听到一个故事，其中施事或受事被明确提及（ ）。在 条件下，施事或受事未被提及，但由故事进行了启动（通过语义或联想相关的词，如 ）。在 条件下，故事既未明确提及也未启动任何一个角色。预期目标反应在所有条件下具有相同的结构和内容（主谓宾句子： ）。结果表明，无论条件如何，参与者通常在言语开始前首先看向施事，然后在言语开始时及之后看向受事。对与语言规划相关的时间窗口内的眼动分析表明，句子构建主要对施事在语境中是否被字面提及敏感，而对概念可及性的敏感程度较低（实验1）。此外，受事的可及性（无论是其名字的字面提及还是仅概念的可得性）对言语规划的时间进程没有影响（实验2）。总之，这些结果表明，言语开始前的语言规划仅受句子中第一个角色名字的可及性影响，为句子生成中的高度增量规划提供了进一步的证据。
