Discourse and spoken language treebanks
- The network teams were invited to contribute to a common "pool" of
spoken language transcriptions. The transcriptions should be based on
about 30 min of talk and can be sent to firstname.lastname@example.org
- As many teams as possible should try to convert their own
trancsriptions to the Tiger XML format, to be paired with the
- On the basis of 1 and 2, the teams in the network were asked to
reflect on whether Tiger XML should be extended to cover the needs of
discourse and/or spoken language structure.
- Teams that are able are encouraged to try to parse their own
transcriptions (and if possibleof other teams) and make the result
- The teams were also encouraged to code at least 50 utterances
manually. Here there was a suggestion to try to capture both internal
utterance structure and "between utterance" structure.