Share on facebook
Save on delicious
RSS Feed
Driver Project 2. Grid-based Structuring of Assembled Records
This project seeks to explore the core research themes in the context of corpus linguistics. The aim here is to move beyond text-mining to analyze and unpack the relationship between language and gesture and thereby develop a much richer understanding of human communication. The research focuses on the analysis of �back-channels' - i.e., the ways in which hearers register with speakers that they hear what they are saying. It seeks to exploit techniques in computer vision and image analysis to support the identification of distinct gestural counterparts to linguistic back-channels. Researchers involved in the project have built on their achievements in the NCeSS Small Grants Project HeadTalk , rated as �outstanding� by ESRC reviewers.
Beyond pursuing it's own immediate ambitions, this driver project also aims to inform the development of digital records by articulating what is involved in structuring assembled records. It draws on the core skills of corpus linguists to articulate the work involved in organizing and analyzing large corpora of heterogeneous data : transcribing large amounts of video data, preserving the relationship between transcript and video, marking the data up to enable interrogation of the corpus, using vision recognition logs to identify gestural counterparts to linguistic back-channels, etc.
The results have been reported at the International Conference on e-Social Science series, major corpus linguistic conferences and journals, and in a number of new texts advancing the state of the art, reflexively informing research practice and shaping the development of computer support in the Node.