Request for Information
-----------------------
We are presently engaged in an international EU-sponsored
project called EAGLES (= Expert Advisory Group on Language
Engineering Standards). The work package in which we are
involved is concerned with provisionally defining standards
for the encoding and annotation of dialogues as machine-readable
texts. This task is to be undertaken primarily with reference
to the languages of the European Union and with reference to
language engineering applications.
In order to extend the database of information on which we
and our colleagues shall base our assessment of current
preferences in dialogue encoding and annotation, we are
sending out this circular to ask for your help.
If you are involved in a project that entails the encoding or
annotation (at any level) of dialogues in machine-readable form,
we should be very grateful for information on such representation
and annotation issues as those below. However, if you don't have
time to give details, a reference to web pages or other information
sources would be equally useful.
- the language(s) on which you are working
- the nature of the dialogues on which you are working
- the general thrust of your research
- how you transcribe your dialogues: e.g. do you use staves
or some other form of representation (e.g. SGML); do
you standardize filled pauses (e.g. um, ah); how do
you mark speaker changes/overlapping speech; do you
make any attempt at phonetic/phonemic transcription;
and so on.
- what levels of language you annotate: e.g. parts of speech;
syntactic structures; word stress; intonation contours;
semantic fields; and so on.
- the format(s) in which you annotate (e.g. TEI/SGML; TOBI;
F0 traces; etc.)
- the software (if any) that you use to accomplish the tasks of
transcription, encoding and annotation.
- what hardware you use
Copies of transcription, encoding and annotation guidelines
used on your project(s), together with references to background
publications and web pages, would especially be most welcome.
Thank you very much in advance for your help.
Geoffrey Leech UCREL
Martin Weisser Lancaster University
Andrew Wilson Lancaster LA1 4YT, UK