How Are New Words Integrated into Language Learners’ Semantic Memory? Is The Integration of New Words Related to How Well the Learners Remember them? A Theoretical Reframing of Lexical Integration and ERP Evidence from Visual Word Learning

Learning new words that one has never encountered before involves remembering (memory) the words (e.g., novid, ‘someone who has never had COVID’) and integrating (integration) them into existing knowledge. How are new words integrated into language learners’ semantic memory? Is integration related to how well the learners remember the words? The Complementary Learning System (CLS) account suggests that word learning experience is first stored as episodic memories through an initial fast learning hippocampal system. A subsequent lexical integration involves abstracting the stored episodic memories and gradually transferring them into semantic memory through the neocortical learning system, particularly during sleep.

This dissertation proposes a theoretical reframing of lexical integration that distinguishes three aspects of lexical integration (word form integration, development of lexicality and meaning integration), allowing the examination of different phases of integration. In support of this reframing, the dissertation reports three studies that taught native English speakers novel words (e.g., banara, ‘a dog that eats fruits and vegetables only’) and examined the remembering of novel words and their integration across a period of a week or so. Study 1 assessed memory of new word form (e.g., banara) and their form integration with the lexical competition effect on existing visually similar words (e.g., banana). This study found significant lexical competition effects immediately after learning and 7 days later. Interestingly, these effects were associated with how well the learners remembered the spelling of the novel word. Study 2 recorded the brain potentials to newly learned novel words and found that the N400 was modulated by how well the learners remembered the meaning of novel words. Study 3 comprehensively assessed memory of novel words across an 8-day period, and recorded brain potentials in response to novel words (e.g., banara, ‘a dog that eats fruits and vegetables only’) that were preceded by a prime that was either semantically related (e.g., cat) or unrelated (e.g., cup). As in Study 2, on day 2, performance of memory tests (on either word form or meaning) modulated the N400 to novel words. On day 8, an N400 semantic priming effect emerged, only in items that were correctly remembered on Day-2 memory tests, but not in items that were incorrect. Taken these results together, this dissertation demonstrates that (1) lexical integration involves multiple components, allowing the detection of different phases of integration; (2) gradual lexical integration leads to quality changes of novel words’ memory representations; and (3) memory and integration of novel words are positively associated: Better memory of novel words on an early test day leads to better integration later –The rich get richer phenomenon in the context of learning new words.

Committee Chair: Dr. Charles Perfetti, Ph.D.