Audiodescription under the Frame Semantics perspective

An exploratory study

Abstract

This paper presents an exploratory study based on Frame Semantics (FILLMORE, 1982, 1985), which analyzes the relations between frames annotated in the text of the audio description and the original audio of a short film transcribed for the purposes of our analysis, with a view to contributing to research in the field of audio description, audiovisual translation and Frame Semantics studies. Our aim was to investigate how the relations between the frames on FrameNet, as well as the categories designated by the top frames contribute to understand the process of articulation and/or semantic complementation of the analyzed cut of the short film. For this purpose, a monolingual comparable corpus was compiled with an excerpt of the text of the audio description and the original audio of the Brazilian short film "Eu não quero voltar sozinho" (I don’t want to go back alone). The corpus was analyzed using spreadsheets in an online editor and semantic annotations made on the Web Annotation Tool platform, belonging to FrameNet Brazil. The results show that the Inheritance relation was the most found among the frames and that in general, the same types of top level frames are evoked in the text of the audio description and in the original audio of the short film. On the other side, it was observed that in the text of the audio description, there is a reduced amount of lexical units that evoke inherited frames of State and are related to the emotions of the characters, showing a tendency for a more neutral language.