Opening Up Video Databases to Collaborative Commentary

The TalkBank Project has constructed a web-accessible database for spoken language interactions with transcripts that are linked on the level of the sentence to both audio and video materials. This database includes several large corpora documenting learning in classrooms, tutorial sessions, meetings, and the home. Now that these data are publicly available, we can begin to build tools to facilitate a new process called collaborative commentary. We can define collaborative commentary as the involvement of a research community in the interpretive annotation of electronic records. The goal of this process is the evaluation of competing theoretical claims. The process requires commentators to link their comments and related evidentiary materials to specific segments of either transcripts or electronic media. We will examine current work in the construction of technical methods for facilitating collaborative commentary through browser technology. We will look at seven spoken language database projects that have reached a level of web-based publication that makes them promising as targets for collaborative commentary. For each database, we will consider how collaborative commentary can advance the relevant research agendas. The TalkBank Project (http://talkbank.org) is an international collaborative effort that is building a web-accessible database for spoken language interactions with transcripts linked to both audio and video materials. The project has targeted 12 research areas for development. Of these, one of the most important areas involves the study of learning in classrooms, tutorial sessions, meetings, and the home. For this area, we now have published 12 video corpora linked to transcripts. Now that these data are publicly available, we can begin to build tools to facilitate a new process called collaborative commentary. We can define collaborative commentary as the involvement of a research community in the interpretive annotation of electronic records. The goal of this process is the evaluation of competing theoretical claims. To achieve clear connections between data and theory, commentators need to link their comments and other evidentiary materials to specific segments of either transcripts or electronic media. This means that, in order to even think about engaging in collaborative commentary, researchers must first have web access to corpora of transcripts of learning sessions linked to video data. The TalkBank database is designed to provide the raw material that can stimulate and support collaborative commentary. In this paper, I will first discuss the ongoing construction of this database. Then, I will summarize current work in the construction of technical methods for facilitating collaborative commentary on the database. The focus of this analysis will be on the use of video data for the study of learning in the contexts of the class, tutorial sessions, meetings, and the home. However, it is important to understand the extent to which work in these areas is embedded in a large interdisciplinary context that includes the analysis of video in areas as diverse as field linguistics, emergency medicine, and psychotherapy.