BRANEN and BRANES Corpora

Abstract

This paper presents two learner corpora built to investigate anaphora across learning environments: the Brazilian Learners of Anaphora in English (BRANEN) and the Aprendices Brasileños de Anáfora en Español (BRANES). Texts were written by language undergraduate students during an online course on anaphora, offered at a Brazilian University in 2020. The corpora provide insights for the analysis of the learning process of anaphora in English and Spanish by Brazilian Portuguese native speakers with intermediate-advanced level in the foreign language. Informants are 30 English and 15 Spanish learners, who were randomly divided into three sub-groups: one group that had two synchronous lessons on anaphora; another that had two asynchronous lessons; and a control group that did not take any lessons. Each participant wrote 100-150 words as a conclusion of a short story. The exercise was performed in four moments: before the course started, after the first lesson, after the second lesson, and a month after the course ended. The texts are available on Sketch Engine, a corpus manager and text analysis software, and contain information about the participants' group and testing moment. The BRANEN corpus was automatically part-ofspeech tagged with the Modified English TreeTagger and has 120 documents, 1,069 sentences, and 1,678 lemmas. For BRANES corpus, the Spanish FreeLing tagset was used, and it consists of 60 documents, 543 sentences, and 1,299 lemmas. The Concordance tool was used to retrieve sentences with pronominal and zero anaphora, which were then manually and independently annotated by two anaphora experts.

The synchronous and asynchronous learning of anaphora: A corpus-based analysis with learners of English and Spanish

“…As shown elsewhere (Bruscato and Baptista 2021a, 2022b, 2022c, students wrote short narratives of 100-150 words in four different moments: before the course started (to check students' performance before the intervention), after the first lesson (to measure their progression during the course), after the second lesson (to check their progression when they completed the course), and a month after the course ended (to investigate if students still remembered what they studied). Texts were different but followed the same structure, they were all third-person narrative fictional texts with multiple antecedents.…”