Search Abstracts | Symposia | Slide Sessions | Poster Sessions
When words denoting entities and/or events become subjects: an EEG study
Poster Session D, Saturday, October 26, 10:30 am - 12:00 pm, Great Hall 3 and 4
Yingyi Luo1, Xiaoqian Zhang1; 1Institute of Linguistics, Chinese Academy of Social Sciences
To reveal the neurocognitive mechanism of the use of word category information (such as the division of nouns and verbs), researchers have attended to the representations when words of different categories are presented in isolation, and also the online processing when words are understood in a sentence context that constrains the word category. However, little neural evidence exists regarding how word category information modulates the build-up of syntactic structure when words serve identical syntactic functions in sentences with limited contextual cues. This study addressed this gap with Chinese, where nouns and verbs hardly dissociate from a morphological perspective. Specifically, words denoting events (“ju4jue2”, refuse/refusal) can function as either predicates (“ju4jue2 chi1fan4”, refuse to eat) or as arguments (“ju4jue2 hen3shang1ren2”, rejection hurts) without morphological change. Moreover, there is a significant group of Chinese words that are ambiguous between entities and events (AM), e.g., “bao4gao4” (“yi1fen4 bao4gao4”, a report; ”bao4gao4 lao3shi1”, report to teachers). Based on the Lexical Semantic Component Classification Framework, we selected 47 AM, 42 words exclusively denoting events (EV), and 42 words exclusively denoting entities (EN) from the high-frequency bisyllabic Chinese words, all denoting non-natural things and related to human cognition. Thus, this study particularly focused on how these three types of words influence the Subject Effect, which compares the neural activities induced by words when presented as standalone units and as sentence-initial subjects (maintaining identical word form). The aim was to investigate the role that word category plays in lexical access during the early stage of sentence construction. EEG data (128 channels) were collected from 21 native speakers. The experiment adopted a 3 (word type) x 2 (task) design. In the lexical decision task, target words visually appeared randomly in the word list (400ms + 400ms blank screen), and participants responded to pseudowords (filler) by pressing a button. In the sentence comprehension task, target words were embedded in a sentence framework "If + Target + BE/BE-NOT..., ...". Each sentence was visually presented word by word (400ms + 300ms blank screen). Participants judged the logical coherence of each sentence. Half of the participants completed the lexical decision task first. Data processing and 2-way ANOVA statistics (word type * task) were conducted using the EEGlab. ERP analysis showed that the divergence in Subject Effect between AM and EN was primarily observed in the early 0-100ms, with AM exhibiting a more negative effect in posterior regions under the sentence comprehension task. The difference between AM and EV in Subject Effect emerged after 300ms. Compared to the lexical decision task, the sentence comprehension task elicited more positive responses in posterior regions, particularly for EV. The results suggest a similar processing time course for AM and EN when functioning as subjects. The early increased negativity for AM may be due to interference from ambiguity. EV, compared to AM, require greater cognitive demand when serving as the sentence-initial subjects, likely due to the relatively low frequency of starting a sentence with (unspecified) events.
Topic Areas: Meaning: Lexical Semantics, Reading