LINGUIST List 5.91

Wed 26 Jan 1994

Qs: Word analysis, Speech to text, Icons to lang, Brown corpus

Editor for this issue: <>

Directory

1. a mcelligott, Word Analsyis Approach
2. , T
3. Mark Heinicke, from icons to language
4. Gerald McMenamin, Access to Brown Corpus

Hi,

Has anybody come across a technique, algorithm or approach to achieving
the following:

 Given a word in a language have a program to derive its root form
 and state its attributes.

 e.g. given tries return try, verb; try, noun and possibly other
 details.

 PC-KIMMO achieves this to a point but a more general approach
 would be nice, as I do not wish to duplicate parts of the
 lexicon to achieve this.

Thanking you in advance,

AMcE.

__________________________________________________________________

 Annette McElligott, CSIS Dept., University of Limerick, Ireland.
 Tel: +353 61 333644 ext. 5024; Fax: +353 61 330876
 Email: mcelligoitdsrv1.ul.ie or mcelligottaul.ie

An academic friend is having increasing difficulties with his sight.
I would be grateful for any details on usable speech-to-text systems.

Noel Heather
Lecturer in Arts Computing
Dept.of English
Royal Holloway
University of London
Egham
Surrey
UK

e-mail: n.heatherrhbnc.ac.uk

I'm not a linguist, but I have a project that may involve some linguistic
assistance. I want to help a handicapped person to write. He has
severe cerebral palsy such that he cannot speak more than a few words,
and has minimal use of his hands. He can write with agonizing slowness,
and type very slowly with lots of errors. He's also not really a "left-
brain" thinker, so typewriting is cognitively difficult anyway. It is
an all-day task for him to write a 100-word letter.

However, he can operate a mouse/trackball (I'm trying to find some gadgets
to make that easier; he can't lift the mouse, for example, which makes
long moves difficult; the track ball presents other problems). So I'm
thinking along the lines of an icon-driven parser/assembler (whatever
the inverse of a parser is). Perhaps such software exists in the PC
world? If not, what English grammar software is available?

I'm looking for modifiable software, not hardwired stuff which I've
already seen some of. He should be able to create his own icons.

A sample problem for this software:

Given:

[icon for "self"; modified to possessive] [icon for "parents"] [icon for verb
TO BE; modified to future] [icon for "home"]

Produce the sentence:

"My parents will be home."

(Perhaps many words, particularly intransitive verbs, adjectives and adverbs,
 will not be icons, but simply words spelled out in cells that can be
 pointed to; also the modifiers such as tenses of verbs and possessives/objects
 of nouns if needed)

Any ideas or suggestions along these lines would be welcome. Probably a lot
would not be of interest on the list so please reply to me if you think
your comments would be clutter:

Mark Heinicke
mheinickGlue.umd.edu
Dept. of Geography
University of Maryland at College Park

p.s. please let me know if this is an inappropriate use of this list. It
 seemed like a decent linguistic challenge to me.

This is a request for help directed to anyone who has access to the
Brown Corpus of Edited American English.

This is a very narrowly defined request. I am interested in searching
the corpus or a concordance of it for a single word/phrase. I would
appreciate it if anyone can help with access or can actually look up my item.

My question is, what is the relative frequency of occurrence of the
two-part verb "hear out", as in, "Please hear me out," including other
forms of the verb (i.e., heard, hearing, etc.).

Thanks for the help.

Jerry McMenamin
Dept. of Linguistics
California State University, Fresno
Fresno, CA 93740