FrameNet Brasil Meets Multimodality: changes, consequences and open questions

FrameNet Brasil has been advancing towards including communicative modes other than written text into its annotation practice. In order to achieve this goal, new software infrastructure had to be built, so as to reflect the new analytical practices arising from the effort of extending a language resource originally tailored for computational lexicography into the domain of analyzing multimodal documents. In this talk, I present such an infrastructure and, by discussing sample analyses multimodal documents, I discuss the consequences of such an extension to the frame-based model, as well as point to open questions and new research possibilities.