Ser vs. estar: A translation corpus study

This project analyzes previously proposed hypotheses for a linguistic phenomenon—the distribution of the Spanish copulae ser and estar—using a translation corpus with data from two languages: English and Spanish. Two meaning-based hypotheses for the distribution ser/estar are considered: the ‘Characteristics vs. Condition’ hypothesis–which relies on the semantic properties of the complement of ser/estar to predict which form of the copula is licensed—and a discourse-based hypothesis proposed in Maienborn (2005)—which builds in specific considerations pertaining to the discourse context in which each copula must be used. Data was extracted from chapters 1 to 4 of the first book in the Harry Potter series (English original text and Spanish translation) and used to build a corpus, which was then employed to analyze the hypotheses. Through comparing the existing hypotheses with naturally produced data, both the accuracy and the limitations of these hypotheses were assessed. Overall results showed significant overprediction of estar, and a discrepancy between how well the hypotheses fared in narrative vs. dialogue contexts.