Re: CHAT: Morphology with Kimmo
|From:||David G. Durand <dgd@...>|
|Date:||Monday, October 26, 1998, 21:52|
At 5:20 PM -0400 10/27/98, Carlos Thompson wrote:
>Due to my work I'm working with a tool for morphological parsing called
>PC-Kimmo (by a Fin called Kimmo Koskenniemi). In orther to implement a
>parsing of the Spanish language. (This is a word level morphological
To be precise, Kimmo Koskoniemmi devised the theory behind PC-Kimmo. Evan
Antworth of SIL developed the PC-Kimmo program.
The theory of phonology implemented by PC-Kimmo is not the standard ordered
rule model that they taught when I learned phonology (typical generative
phonology) -- but it's computationally tractable enough (in practice).
I use the kgen program to write symbolic rules, as hand-compiling FSA
tables is not my idea of fun! It's a bit buggy, however, and won't handle
some kinds of expression that you would expect that it could (for instance,
the same symbol set can only have on repetition specification in a given
rule). This makes patterns like C*VC* tricky, depending on what you want it
Xerox has some lexical compilation technology based on the same theoretical
model as PC-KIMMO, (Koskonienni works there now) and they used to offer a
web-based compilation service that produced PC-KIMMO driver files. Their
pages are at http://www.rxrc.xerox.com/research/mltt/fsHead/home.html --
worth checking out, as there's lots of info. I don't think that there's a
publicly available version of the tools (without paying big bucks, anyway).
David Durand email@example.com \ david@dynamicDiagrams.com
Boston University Computer Science \ Sr. Analyst
http://www.cs.bu.edu/students/grads/dgd/ \ Dynamic Diagrams
MAPA: mapping for the WWW \__________________________
>I wonder if someone else has use this tool. I think is very convinient to
>recognize forms in very flexional languages either natlangs or conlangs.
>I'm planing to use in my conlang projects. Kimmo can be found at Summer
>Institute of Linguistics http://www.sil.org/pckimmo