其他摘要 |
Discourse contextual learning is an important word learning approach in which the variability of context plays an important role in novel word learning. However, the question whether and how the variability of context influences the integration of novel word into the preexisting semantic network is still unclear. The present study used ERPs technique in semantic priming paradigm to explore these questions. We focused on two types of semantic relationships: taxonomic relation and thematic relation.
In Experiment 1, we explored whether and how the variability of episodic context influences the acquisition of thematic relation of novel words. The novel words were learned from unitary or variable descriptive episodic discourses. In a lexical decision task, the novel words served as primes while the corresponding concepts (CC), words thematically related to the shared episodes across the unitary and variable episodic conditions (THS-1), words only thematically related to the episodes of the variable episodic condition (THS-2), thematically related words in a different episode (THD), or unrelated words (NR) served as targets. Event related potentials (ERPs) revealed that the CC and thematically related targets which were related to the learning episodes (THS-1 targets for the unitary episodic condition and THS-1 and THS-2 targets for the variable episodic condition, respectively) elicited smaller N400s compared to the NR targets. Most importantly, the THD targets elicited a smaller N400 than the NR targets in the variable episodic condition but not in the unitary episodic condition. In addition, the thematic associations between the novel words and the thematically related words were also recognized in early processing stage in the variable episodic condition. These results suggested that the variability of episodic context facilitates the novel words form a rich thematic association in semantic network.
In Experiment 2, we further explored whether the variability of episodic context would also influence the acquisition of taxonomic relation and the depth of semantic representation of novel words. Similar to Experiment 1, the novel words were learned from different episodic conditions and then the lexical decision task was conducted. The novel words served as primes while the corresponding concepts (CC), taxonomic-related words (TR), feature-related words (FR), or unrelated words (NR) served as targets. Event related potentials (ERPs) revealed that the CC and TR targets elicited smaller N400s compared to the NR targets. Most importantly, the FR targets elicited a smaller N400 than the NR targets in the variable episodic condition but not in the unitary episodic condition. These results suggested that the variability of episodic context also facilitates a deeper understanding of the novel word.
Our findings suggested that the variable episodes result in a deeper understanding of a word’s meaning and a more stable and richer semantic representation in semantic network. |
修改评论