Re: Term for "class family" wanted!

From: Richard Ulrich (Rich.Ulrich_at_comcast.net)
Date: 06/21/04 

Date: Mon, 21 Jun 2004 11:01:40 -0400

On Sun, 20 Jun 2004 20:44:23 +0200, "Dirk Schmidt"
<dirkschmidt999@yahoo.com> wrote:

> Hello,
> i write my doctorate and I have a problem.
> On the basis of verb alternations I have got some classes which can be
> grouped together, e.g. the classes
> Class 1a/b
> Class 1?a/b (doubt about a variant)
> Class 1a/*b (no b variant)
> Class 1*a/b (no a variant)
> constitute the 1a/b "class family." (1 is the number of the alternation, a
> and b the alternation variants)
> What is the term for these "class families" in statistics?
>
Statistics makes use of logical categories from the
underlying sciences.
For two 'factors', we can classify each as Yes/No.
If there is a 'questionable' category, there might
be Yes/maybe/no, to be treated as ordered. But since
you probably don't have a 'verb' if it is the No-No
category, this comment may be less than useful.

"Verb alternations" sounds like a topic in linguistics,
so you might pose the question to linguists. 

-- 
Rich Ulrich, wpilib@pitt.edu
http://www.pitt.edu/~wpilib/index.html