Understanding semantic competition in complex phrase comprehension and production

This thesis investigated the relationship between complex phrase comprehension and production. The work aimed to identify whether parallel competition effects exist between the two tasks; and importantly, the extent to which these effects drive from common processes and knowledge bases. In three studies, participants viewed pictures of various entities doing different actions. For each picture, they comprehended a recorded description (e.g. the teddy bear/man that the girl is hugging) about the entity being acted upon, and were asked to describe it in a production task. They also completed a number of cognitive assessments measuring vocabulary, inhibition, etc. Study 1 found that phrases containing highly similar and reversible nouns are more difficult to comprehend and produce in adults. Importantly, this difficulty varied as a function of individual inhibition skills over above vocabulary in both tasks, and production additionally recruited task-specific motor inhibition processes. Study 2 replicated the reversibility-based effects with children and adolescents. But young children differed from older participants as they experienced greater production interference and are less skilled in using certain production options to alleviate interference. Unlike adults, their language performance was predicted by variance on working memory capacity. Study 3 used eye-tracking to examine the time course of production competition. The results showed reversibility-based competition manifest at verb position, and is particularly relevant to individual’s semantic inhibition skill. This parallels previous comprehension findings, thus suggests shared competition resolution processes across tasks. Together, these findings suggest common reversibility-based competition processes underlying comprehension and production, and across development. Current models arguing for shared prediction processes in adults can potentially incorporate common inhibition mechanisms; however, our data imply that non-shared processes should also be considered. On the other hand, unlike adults, our children’s data supports the capacity-constraint account in language processing, thus suggesting a discontinuity of cognitive functioning in language development