Recapitulative Discourse Markers in English, Spanish, and Spanish Translated from English

Abstract

This paper unveils the similarities and differences between original English and Spanish recapitulative discourse markers (DMs), in fiction (F) and non-fiction (NF), and between translated and non-translated Spanish. It also aims to show that combining parallel and monolingual corpora can produce more focused results when tracking trends across language boundaries. Original English and Spanish data show how recapitulative DMs are distributed in both languages. Then, DMs with a low mutual correspondence (MC) value are selected for further analysis. English triggers to be compared with original Spanish data are identified in the English originals via the Spanish translations. Also, translated and non-translated Spanish are compared to identify translation trends. Results indicate that recapitulation is signalled by DMs more frequently in Spanish than in English. DMs in Spanish translation most frequently derive from well, anyway, you know/see, in short, I mean and after all. Additionally, many DMs in the translations do not have a trigger in English. Translated and non-translated Spanish results show the following trends, according to register: a) normalization in F (en fin, en síntesis) and counter-normalization in NF (en síntesis, en definitiva), b) deflation in F (en fin, en definitiva, total, a fin de cuentas) and in NF (en fin, en definitiva), c) dilation in NF (definitivamente, en resumidas cuentas, en síntesis), and d) equalization in F (definitivamente, en resumidas cuentas, en síntesis) and in NF (a fin de cuentas). The results may help to enhance contextual learning in generative language technology, leading to improved crosslinguistic performance.