Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/3/84; site teddy.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!allegra!bellcore!decvax!genrad!panda!teddy!mjn
From: mjn@teddy.UUCP (Mark J. Norton)
Newsgroups: net.nlang
Subject: Re: collectible words
Message-ID: <474@teddy.UUCP>
Date: Thu, 11-Apr-85 09:26:19 EST
Article-I.D.: teddy.474
Posted: Thu Apr 11 09:26:19 1985
Date-Received: Sat, 13-Apr-85 05:02:37 EST
References: <7188@watdaisy.UUCP>
Distribution: net
Organization: GenRad, Inc., Concord, Mass.
Lines: 43

> 	- all words which end in <mumble> (eg. "-nymy" patronymy,
> toponymy, eponymy, pseudonymy etc.) and a description of the
> etymology of the class of words *as a whole* (i.e. as opposed to
> just looking it up in the O.E.D.).

I could respond to many of the others, but this is one I've had 
personal experience with.  I was involved with natural language
R&D at Wang Labs a few years back, and was assigned the task of
writing a program to find the root of a word.

This is by no means an original project, so I collected up info
from other similar programs (Winograd, etc), and attempted to
extend it.  To test for endings on words, I took a fair sized
word list (Random House Collegiate) and wrote a program which
inverts all the words in the list (net.nlang => gnaln.ten) and 
sorted the new list.

This gave me a list of words from which could be extracted, all
words which end in <ending>.  This was very useful to me, since
I then had many examples of how the word was modified to add the
ending (modify => modified).  This satifies the first part of
your request.

As a side project, I started collecting a list of word endings
(prefixes as well, but that's not of intrest here).  I happened
to be reading Webster's Unabridged one day (not uncommon in 
that line of work), when I noticed that many endings had
definitions associated with them.  Furthermore, some of these
had etymology explained as well.  This satisfies the second
part of your requirements.

It is unfortunately the case that such work has not be published
to any great extent, publishers would see little profit in it.
It is of intrest to a limited group of people.  A compendium of
similar works, perhaps in almanac form might be viable, though.
I'm afraid you actually want such a source, you will have to make
it yourself given the algorithms above.

	Mark J. Norton
	decvax!genrad!panda!mjn

PS. I am still very intrested in this subject, and have continued
	my researches on my own.  Correspondence on it is invited.