LINGUIST List 4.964

Sat 20 Nov 1993

Qs: Computational morphology, Jakobson quote, Sasse, MAC

Editor for this issue: <>

Directory

1. Flann Horgan, Q: Morphological Analysis Systems
2. Becky Passonneau, Latin phrase attributed to Jakobson
3. , query
4. Susan Ervin-Tripp, mac software

I am about to embark on a study of Computational Morphology [English language].
I would be interested to hear any opinions which may be out there concerning the
relative merits of different approaches. I am especially interested in the
building of morphological lexicons.

I have been looking at Koskenniemi's two-level approach and the PC-KIMMO
implementation of it. I also have the University of Pennsylvania morphological
lexicon and X-based management tools. This is, to a large extent, based on the
PC-KIMMO system.

I would like to know what people think of the above systems and what alternative
systems/models are available.

Thanks in advance.

Flann Horgan.

============================================================================
Flann Horgan <8805482ul.ie>

Undergraduate, Department of Computer Science & Information Systems
University of Limerick,
Limerick, Ireland.

============================================================================

I had a query from a student about the meaning of a phrase attributed
to Roman Jakobson. Can anyone tell me the translation, and anything
regading the accuracy of the attribution?

"Linguista sum: linguistici nihil a me alienum puto."

Becky Passonneau

Hello --
Does anyone know of e-mail address of 'Hans-J|rgen Sasse'?
He published two papers on Galab (Geleb or Dasenech, a Cushtic
language spoken in Ethiopia according to his description) in
the early seventies. Any information about him, or any other
published/unpublished source of Galab, would be very
appreciated. Please reply to 'hongshccit.arizona.edu'.
Thanks a lot.

hong

Does there exist a MAC search program which has the following
features necessary for the study of natural conversations?

-- can work on any ASCII text.
-- preserves the source text format, i.e. does not line up the target
 forms in a column unlike keyword-in-context concordances.
-- allows some option in window size for the number of lines
 above and below the target form which will be retained in the
 result file to allow context analysis.
-- allows an option in restricting or not the upper or lower case
 of the target form, or in reading or overlooking the diacritics
 or punctuation in a search.
-- allows the option of searching for discontinuous patterns,
 e.g. `look X up' could be retrieved with `look up' .

Susan Ervin-Tripp
Psychology
University of California, Berkeley