Annotating the meaning of discourse marker so in TED talks

Abstract

The purpose of this paper is to report on the results of crosslinguistic annotation of the English spoken discourse marker so and its counterparts in Lithuanian using the multilingual open translation project TED Talks as a resource of data. The purpose is achieved by investigating the domains and functions of the discourse marker so in English and Lithuanian as well as analysing its translations into Lithuanian. The present study was conducted in two stages. First, the meanings of the discourse marker so were annotated and compared to the meanings of their counterparts in Lithuanian. Later, translations of the discourse marker so into Lithuanian were analysed. In the present investigation, the taxonomy of functional annotation for spoken discourse (Crible and Degand 2017) and manual translation spotting were employed. The findings showed that the discourse marker so and its counterparts in Lithuanian in most cases express rhetorical consequence, followed by rhetorical structuring. It was also established that the most frequent variants of translation of the discourse marker so were those provided by bilingual English-Lithuanian dictionaries, whereas the least frequent translations were expressed by a particle or a verb. On the other hand, translation by omission was also frequently used.