SPEECH ACT, DIALOG GAME, and DIALOG ACTIVITY
TAGGING MANUAL
for Spanish Conversational Speech
Version 1.0
Project CLARITY
Language Technologies Institute
Carnegie Mellon University
Ann Thymé-Gobbel
Natural Speech Technologies, Inc.
Lori
Levin
Carnegie Mellon University
0. INTRODUCTION
This manual describes a 3-level manual discourse coding scheme developed
and used for manual tagging of the CallHome Spanish and CallFriend Spanish
databases used in the CLARITY project at the Language Technologies Institute
at Carnegie Mellon University.
The goal of CLARITY is to explore the use of discourse structure in
understanding conversational speech. The project combines empirical methods
for dialogue processing with state-of-the art LVCSR (using the JANUS recognizer).
The three levels of the coding scheme are (1) a speech act level consisting
of a tag set extended from DAMSL and Switchboard; (2) a dialogue game level
defined by initiative and speaker intention; and (3) an activity level
defined within topic units. Each of the three levels of tagging are discussed
in the following sections. The manually tagged dialogues are used to train
automatic classifiers.
INDEX
1. Speech
Act Level Tagging: Background Information
2. Speech
Act Level Tagging: Descriptions And Examples
2a. Semantic
Feature Dimension
2b. List
of Speech Act Tags
Questions
Answers
Agreement/Disagreement
Discourse
Markers
Forward
Functions
Control
Acts
Statements
Other
Tags
3. Speech
Act Level Tagging: Decision Trees
4. Dialog
Game Level Tagging: Background Information
5. Dialog
Game Level Tagging: Descriptions and Examples
6. Dialog
Game Segmentation
7. Activity
Level Tagging: Descriptions and Examples
8. Putting
It All Together
9. References
Contact: lsl+@nl.cs.cmu.edu
manualintro.html
last updated: January26, 1999