Anaphoric reference in L1 Spanish-L2 English: A corpus-based study of oral and written narratives

This dissertation investigates possible mode of production effects on the occurrence of topic continuity contexts and on Referring Expressions (RE) selection in this discourse configuration. A corpus study was conducted a corpus-based study to establish the following: i) How does the mode of production affect the selection of REs as zeros, overt pronouns and noun phrases (NP) in the subject position?; ii) How does the mode of production affect the syntactic (i.e., coordination) and discourse factors constraining the selection of REs across proficiency groups?; iii) How does the selection of REs occur developmentally in L2 English learners? By looking at spoken and written performance across proficiency levels, this investigation explores whether deficits persist (or not) in one of the two modes (or in both) in the narratives under study. To this end, this study looks at third-person singular grammatical subjects across different proficiency levels (from beginner to advanced learners) in L1 Spanish-L2 English, in contrast with L1 English. The data are spoken and written narratives from the COREFL corpus produced by the same participant and under the same task conditions. The analysis comprises 82 participants, 164 texts (82 spoken and 82 written texts, respectively) and 4.178 grammatical subjects. For data analysis, a specific fine-grained tagset was designed and used on the XML annotator UAM Corpus Tool (O’Donnell, 2009). Statistical analysis of the tagged data was performed using the same tool. The present corpus findings provide a comprehensive picture of the effects of the mode of production on the selection of REs and provide important insights as to learner’s performance in spoken and written performance. In particular, a highly marked effect of the mode of production in L1 Spanish-L2 English was found, while no such effect was found in L1 English performance. Overall, the results revealed that L1 English-L2 Spanish learners show greater and more persistent deficits in spoken production, while those in written production tend to disappear at advanced levels. In topic continuity coordination, the discourse configuration in learners is comparable to that of native speakers. However, the spoken production by beginner and intermediate learners still shows a higher amount of fuller REs than that by advanced and native participants. This study highlights the practical implications of considering the mode of production in RE selection, revealing distinct patterns in spoken vs. written discourse for L1 Spanish-L2 English learners and L1 English. The marked differences observed between L1 Spanish-L2 English and the control group of English native speakers in mode-specific performance emphasize the need for further comparative studies across different language pairs.