Predicting the end: monolinguals, L2 learners and interpreters' use of prosody to predict word endings

Prediction is essential to human cognition and language is no exception. Native speakers anticipate upcoming linguistic information rapidly and easily, but some studies show that second language (L2) learners have difficulty making linguistic predictions, even at advanced proficiency levels. I investigate whether prior experience with linguistic anticipation acquired via simultaneous interpreting explains adult learners’ trouble making L2 predictions. Simultaneous interpreting requires constant and quick predictions to ease the cognitive load of simultaneous interpreting. Also, I examine the role of working memory (WM) on the anticipation of morphology to shed light on how cognitive resources support prediction. To address the role of anticipatory experience and WM on L1 and L2 prediction, adult Spanish monolinguals and adult English learners of Spanish with and without interpreting experience completed a WM test and a visual world paradigm eye-tracking task asking them to predict word endings based on prosodic cues. Study 1 examines the effects of prediction experience via simultaneous interpreting on L2 prediction. Spanish monolinguals, and advanced L2 learners of Spanish with and without interpreting experience performed an eye-tracking task in which they saw two verbs on the screen while hearing a sentence and could anticipate the target verb based on lexical stress (paroxytone, oxytone) and syllabic structure (CV, CVC). Data showed that native and non-native speakers use lexical stress and syllabic structure in the initial syllable of a verb to predict its suffix, although the learners did not predict suffixes preceded by CV stressed syllables, and interpreters predicted faster than non-interpreters and even monolinguals under some conditions. Hence, prosodic cues facilitate morphological prediction during oral word recognition, and anticipatory experience enhances L2 prediction. Study 2 explores prediction of non-morphological word endings (e.g., PApa – paPÁ, ‘potato-dad’) to determine whether the findings in Study 1 are limited to words with inflectional morphology. Participants and tasks were identical to Study 1. Data revealed that monolinguals and interpreters also use lexical stress and syllabic structure to predict non-morphological word endings, but non-interpreters display more difficulties making predictions than monolinguals and interpreters, and can only anticipate word endings preceded by CVC unstressed syllables (which minimize lexical competitors). Therefore, prosody in the first syllable is key for lexical access and prediction, L2 prediction is more vulnerable to semantic interference, and anticipatory experience via interpreting enhances L2 prediction. Finally, Study 3 investigates the effects of WM on morphological prediction during oral word recognition. Participants and tasks were identical to Study 1. In addition, participants completed a letter-number sequencing WM task. Data showed that, in the stressed condition, higher WM monolinguals predicted earlier and higher WM interpreters predicted faster; whereas, in the unstressed condition, lower WM non-interpreters predicted earlier. Importantly, interpreters’ use of WM is closer to that of monolinguals. Taken together, these studies inform prediction models, and support accessibility models and usage-based models. The findings from this dissertation advance our understanding of prediction by teasing apart L2 proficiency from prediction experience. Also, findings support accessibility accounts, by showing that L2 learners can process their L2 on par with native speakers, and usage-based models of adult L2 acquisition, indicating that repeated exposure to prediction is essential to the optimization of the cognitive resources to gain L2 fluency.Ph.D.Includes bibliographical reference