Monroe spoken dialogue transcripts and annotations

Home | People | Projects | Publications | Resources

General | Participants | Tasks | Setup | Dialogue stats | Processing | Transcripts and annotations

Dialogue annotation

If you have not already done so, please email me to let me know you want to use this corpus.

We have annotated selected dialogues from the Monroe corpus. The behaviors annotated include:

Taken together, this work comprises a complete annotation for conversation acts. These annotations were designed to elicit three types of information: Other corpora have been annotated for some of these behaviors. For example, the Switchboard corpus has been annotated for speech acts using a variant of DAMSL [#!jurafskyetal98!#]; the Maptask corpus has been annotated for speech acts using the MATE workbench [#!mate!#]; and parts of the TRAINS-93 corpus have been annotated for speech acts using DAMSL [#!core97!#] and for initiative cues [#!chuandbrown97!#,#!chuandbrown98!#]. However, no other corpus has been annotated for all of these behaviors. The eight dialogues annotated were chosen at random, although care was taken to try to ensure that a variety of speakers were included. The dialogues chosen are: s2, s4, s6, s12, s14, s16, s17 and s18. Table 4 shows the annotators involved in different stages of this project.
Table 4: Annotator involvement
Annotator Initial Turn-taking Grounding and Argumentation Initiative
  processing behavior Speech acts Acts  
A yes no yes yes yes
B no no yes initial no
C no no no no yes
D yes no no no no
E yes yes no yes yes
F no yes no yes no
G no no no yes no

The following is a list of all of the dialogs in the corpus, with links to transcripts; reference/reconciled DAMSL and initiative annotations. In the future we may add our other annotations and the audio files.

This page describes how the transcripts were constructed. To summarize, we:

If you notice errors in the transcripts (missing words, extra words, incorrect words, non-language sounds that are missing or incorrectly marked), please email me.

For more information about the annotations, please see this thesis.

Dialog Transcript DAMSL Initiative
s1 s1.transcript
s2 s2.transcript s2rec.sgml s2.init
s3 s3.transcript
s4 s4.transcript s4rec.sgml s4.init
s5 s5.transcript
s6 s6.transcript s6rec.sgml s6.init
s7 s7.transcript
s8 s8.transcript
s9 s9.transcript
s10 s10.transcript
s11 s11.transcript
s12 s12.transcript s12rec.sgml s12.init
s13 s13.transcript
s14 s14.transcript s14rec.sgml s14.init
s15 s15.transcript
s16 s16.transcript s16rec.sgml s16.init
s17 s17.transcript s17rec.sgml s17.init
s18 s18.transcript s18rec.sgml s18.init
s19 s19.transcript
s20 s20.transcript

Last change: 1 Dec 2000 by stent