Hwangbo, Hyun Jin2018-12-142018-12-142018http://udspace.udel.edu/handle/19716/23991This dissertation studies the nature of lexical representation by conducting behavioral experiments in the artificial grammar learning paradigm. ☐ It begins with an important question in the phonological and morphological traditions: How are the pronunciations of morphemes represented in long-term memory? When there are multiple pronunciations of a morpheme (e.g., the English plural [-s], [-z], [-ɪz]), is each of the forms stored separately, or is only one form stored, with the other pronunciations derived by phonological transformations? In this dissertation, I refer to the former view as Morpheme Alternant Theory (MAT) and the latter view as Generative Phonology (GP), following Kenstowicz and Kisseberth (1979). ☐ This question is fundamental to phonology because the role phonology plays in grammar is significantly different according to each view. MAT argues that phonology plays a minimal role, and thus predicts that the pronunciations of the allomorphs are memorized or listed as their own forms as the underlying representations. GP argues that phonology is crucial in organizing the mental lexicon and in language, and thus phonology has greater priority than morphological listing. ☐ This dissertation brings a new kind of evidence to bear on the question. Artificial grammar learning experiments were conducted which compared how well learners could learn allomorphy patterns that were more or less amenable to a phonological analysis under the GP view. The results consistently reject the view that speakers just memorize which allomorphs are used in which phonological environments. A more careful examination of the results also shows they are inconsistent with the views that the subjects either prioritized phonological rule learning or that they infer phonological rules only after memorizing which allomorphs are used in which phonological environments. Instead, the results point to a more complex interplay between phonological and morphological learning. These results provide a foundation and direction for future research. Behavioral data can provide further evidence for how the mental representations of allomorphs are structured in our minds.Language, literature and linguisticsBiological sciencesPsychologyAllomorphy patternsMental representationsPhonologyThesis1078921814https://doi.org/10.58088/2624-tv862018-10-17en