The Automatic Content Analysis of Spoken Discourse. Final Report on SERC-DTI Project Nr. GR-F/36385 IED4-1-1143.