You are viewing the SNL 2017 Archive Website. For the latest information, see the Current Website.

Poster C8, Thursday, November 9, 10:00 – 11:15 am, Harborview and Loch Raven Ballrooms

An fMRI Study of Syntactic Complexity Effect of Chinese Relative Clauses

Yanyu Xiong1, Chunglin Yang1, Sharlene Newman1;1Indiana University

The previous neuroimaging research about neural substrates of syntactic processing of complex sentences has found that the complexity effect is influenced by both syntactic integration cost and working memory. However, earlier research paradigm of comparing subject-modifying object-gap relative clauses (SO) with object-modifying subject-gap relative clauses (OS) conflated the external working memory load arising from main clause context (center-embedded vs. right-branching) with the internal processing load within the relative clause region. The results of this group of PET studies almost exclusively reported more activation of Broca’s area for SO. Other fMRI studies using only center-embedded minimal contrast of relative clauses found a more distributed network of syntactic complexity effect with inconsistent involvement of the inferior frontal region across different languages. The present study expanded upon the previous research by introducing another pair of minimal contrast of left-branching relative clauses. The experiment used 2 by 2 design with filler-gap dependency (object-gap vs. subject-gap) and external working memory load (center-embedded vs. left-branching) as two conditions. By utilizing the linguistic feature of pre-modification in Chinese, we reversed the subject-modifying and object-modifying conditions of English studies, that is, object-modifying relative clauses for the center-embedded condition and subject-modifying relative clauses for the left-branching condition, and were able to test: (1) whether left-branching pair showed the same activation pattern as the center-embedded pair does; (2) whether Broca’s area was activated for processing more complex object-gap relative clauses across both pairs; (3) whether there is universal syntactic processing preference for subject-gap clauses over object-gap clauses. We used a slow event-related fMRI paradigm including four blocks of sentences with each block including eight sentences for each type of four relative clauses plus nine filler sentences. 22% sentences were followed by comprehension questions. Subjects were required to make the semantic congruency judgment based on the previous sentence. The results of our study showed that Chinese relative clauses exhibited more brain activation for subject-gap clauses compared with object-gap clauses for both center-embedded and left-branching conditions, indicating that the claim of universal preference for subject-gap clauses does not hold. This opposite scenario of English relative clauses has already been supported by other self-paced reading, eye-tracking and ERP studies. Our investigation further revealed that the neural substrates for this object-preference in Chinese were different. In the left-branching condition, the left posterior precuneus was more activated for subject-gap clauses relative to right-gap clauses, indicating more retrieval of contextual information, whereas in the center-embedded condition, two clusters in the bilateral prefrontal cortex were involved. One is the middle frontal gyrus including Brodman area 6 and 9 and the other is the superior inferior opercularis. This pattern of activation indicated that with equal working memory load from the preceding sentential context before relative clauses, the longer filler-gap distance in subject-gap relative clauses increased integration cost. The inconsistent involvement of the inferior frontal gyrus across the left-branching and center-embedded conditions showed that Broca's area was more modulated by the working memory load than the syntactic feature of clauses per se.

Topic Area: Grammar: Syntax

Back to Poster Schedule