LINGUIST List 5.788

Mon 11 Jul 1994

Sum: Morphological rules, Phonectics fonts for WordPerfect

Editor for this issue: <>

Directory

1. Trey Jones, Sum: Morphological Rules
2. "Robert D. Rachlin", Summary: Phonectics Fonts for WordPerfect

Here is a summary of the responses to my query for morphological spelling
rules (or software that implements them). Since there were only a few
responses, I thought it best to exerpt them rather than summarize the whole.
Thanks to everyone who responded!

:Trey Jones
 Dataware Technologies

********From: lynnecacogs.susx.ac.uk (Lynne Cahill)
Do you know about the Alvey Natural Language Tools Project? They developed a
set of tools for processing English, including a lexicon with morphological
rules which did this sort of thing based on Koskenniemi's two-level morphology
but for spelling. There are papers on it and a book, "Computational
Morphology: Practical Mechanisms for the English Lexicon" by G. Ritchie et al,
1992, MIT Press. The software described in the book is public domain for
research purposes, I believe.

********From: Eubank Lynn Alan <eubankjove.acs.unt.edu>
[...] A retired colleague of mine
has written up a book on spelling conventions and rules to make
them work. It's not going to be published, as far as I know.[...] I know,
however, that he'd be more than happy to send you a copy if you
think it might be worthwhile. Here's his name, snail-mail address:
 Silas Griggs
 Division of Linguistics/ Department of English
 University of North Texas
 Denton, TX 76203/ USA

********From: Ken.Beesleyxerox.fr (Ken Beesley)
I work for Xerox Corporation and use finite-state morphology, [.. which]
consists of a lexicon compiler, a rule compiler, and a complete set of
programs for intersecting, composing, and otherwise manipulating
finite-state machines. [...] The runtime code is completely
language-independent, and in fact no code or algorithms of any kind are
written for a particular language. Using the compilers, linguists specify
facts about the language, and these facts are compiled into efficient
programs that do automatic analysis and generation. [...]

We have large systems available for French and English [...]
For information contact Ms Daniella Russo at russoadoc.xerox.com

Significant finite-state morphologies have been written for Japanese, Korean,
Finnish, Arabic, Armenian, Russian, Swedish, and several other languages.

Not to be too partisan, let me mention a non-commercial system:

A very simple but usable implementation of finite-state morphology is called
"two-level" morphology, and one version called PC-KIMMO is rather freely
available from the Summer Institute of Linguistics. For information
contact Evan Antworth: evan.antworthsil.org

Antworth has used this system to write a large morphological analyzer
for English, including many of the derivational phenomena that you
seem interested in. It is called Englex and contains a large lexicon.
[...] Englex is freely distributed for research only.

********From: ddinfo.ucl.ac.be (Geoffroy de Dorlodot)
Igor Mel'cuk (University of Montreal) has included this kind of
morphological correspondences in his "Meaning-Text Theory", by the use
of some of the "Lexical Functions" (example : lexical function S0 :
verb -> noun : S0(authorize) = authorization). This work has been
(partialy) done for English, French, Russian, German, Hungarian,
Polish, Somali, Albanian, Persian, ...

Some references in English :

- I.A. MEL'CUK, "Lexical Functions in Lexicographic Description",
Proceedings of the 8th Annual Meeting of the Berkeley Linguistics
Society, 1982, pp. 427-444.

- I.A. MEL'CUK, "Meaning-Text Models : A Recent Trend in Soviet
Linguistics", Annual Review of Anthropology, vol. 10, 1981, pp. 27-62.

- I.A. MEL'CUK & A.K. ZOLKOVSKIJ, "Towards a Functioning Meaning-Text
Model of Language", Linguistics, vol. 57, 1970, pp. 10-47.

- I.A. MEL'CUK, "Meaning-Text Linguistic Models and the role of the
Dictionary in Linguistic Description", Proceedings of the XIIIth
International Congress of Linguists, Tokyo, 1983, pp. 412-416.

- I.A. MEL'CUK & A. POLGUERE, "A Formal Lexicon in The Meaning-Text
Theory", Computational Linguistics, vol 13:3-4, 1988, pp. 261-275.

- I.A. MEL'CUK & A.K. ZOLKOVSKIJ, "The Explanatory Combinatorial
Dictionary", in "Relational Models of the Lexicon", M. Evens ed.,
Cambridge University Press, 1988, pp. 41-74.

********From: "Dr. Klaus Wothke" <kwothkeVNET.IBM.COM>
Some years ago I worked on a similar problem for German, French, and English
A survey of this work is published:
 Wothke, K. (1986): Machine Learning of Morphological Rules by
 Generalization and Analogy. In: Proceedings of COLING '86.
 Bonn.

********From: Afrikanistik2uni-bayreuth.de (Ralf Grosserhode)
Th. Schadeberg (schabergrulcri.leidenuniv.nl) has developed a
morphological parser for Swahili, called AINII. Though it works the
other way round, it should have the same kind of problem to solve.

I recently posted a request for information about downloadable phonetics
fonts usable with WordPerfect and a laser printer. WP 6.0 DOS includes
a phonetics character set (Character Set 2). I assume WP 6.0 Windows
does also.

Suzanne Fleischman, John Mathewson, Ralf Grosserhode, Karen Mullen,
and Timothy Montler provided help in my quest, and here's a summary.

Timothy Montler (montlerjove.acs.unt.edu) has developed a set of
bitmap phonetics fonts which can be used with WP 5.1. The fonts and
associated WP driver is available by anonymous ftp to ftp.unt.edu. The
path is pub/micro/ibm/phonts.zip. He also has authorized me to mention
that he will send the fonts and drivers to anyone who sends a letter of
request, a self-addressed, stamped disk mailer, and one 3.5" or 5.25" disk
to

Timothy Montler
P.O. Box 13827
University of North Texas
Denton, TX 76203.

Ralf Grosserhode from the University of Bayreuth tells me that the
Summer Institute of Linguistics has developed a TrueType font set for use
with Windows. He didn't mention just how one gets it. His address is
afrikanistik2uni-bayreauth.de

Last but not least John Mathewson (snezhasiucvmb.siu.edu) is
enthusiastic about a font he has available for the MAC. Since I don't use
a MAC, I couldn't take him up on his kind offer to send me a disk with
the fonts.

Best wishes to all.

Bob Rachlin

________________________________________________________________________
| Robert D. Rachlin Downs Rachlin & Martin Burlington, VT 05402-0190 |
| rachlinpanix.com Compuserve: 72420,3350 MCI: 583-3818 |
|................Tel: 802 863 2375 Fax: 802 863 2573..................|