During the last years corpus based approaches have gained significant importance in the field of natural language processing. Currently, large corpora for many different languages are being collected all over the world. In order for these data to be useful for training and for testing implemented systems, the corpora must be annotated in various ways. The Discourse Resource Initiative (DRI) is an effort to assemble discourse resources in support of discourse research and applications. This seminar was the second in a series of DRI workshops with the goal to develop a standard to annotate corpora for semantic/pragmatic and discourse features.

As for the previous workshop in Philadelphia, PA., the participating 43 researchers had to do some homework, namely annotating different texts with the schema defined in the first workshop, before coming to the workshop.

During the workshop, the audience split in five groups to discuss different aspects of the annotations, namely coreference, forward looking functions, backward looking functions, segmentation, and information level and information status. At the end of the workshop the results were presented and discussed. This report contains the summaries from these five groups. During the workshop, additional groups met, e.g. to discuss tools for annotation and to discuss interactions between forward/backward looking functions and coreference.

There were also five plenary talks by

• James Allen: Towards a Standart for Annotating the Structure of Dialogue
• Hans Dybkjær: Dialogue Annotation in Europe
• Masato Ishizaki & Syun Tutiya: An Attempt to Standardize Discourse Tags in Japanese
• Lynette Hirschman & Rebecca Passoneau: Coreference Annotation
• Marc Vilain: A Model-Theoretic Coreference Scoring Scheme

Creative Commons BY 3.0 Unported license