Word count

[Brian MacWhinney]

Dear Anne,
  I see no evidence on the archives (at LinguistList) that your message ever
got posted, at least not during May.  I am guessing that readers might be a
bit hesitant to nominate themselves as "authorities", particularly given the
underscoring of principled.  The best shot that I ever took at unearthing
principles for judging something to be a word in child language or anywhere
else was in a chapter back in 1982 called "Basic Syntactic Processes" in a
volume edited by Stan Kuczaj.  The core idea here involved distinguishing
rote, analogy, and combination as word and construction formation processes.
A more readable and articulate version of this position can be found in Ann
Peter's 1983 book on "The units of language acquisition".  In one way or
another, there are probably about 200 articles dealing with the issue of
trying to distinguish rote from combination.  There are probably another
couple of hundred papers trying to formalize these ideas into methods for
computing mean length of utterance and related measures.
  Regarding negative contractions, such as don't, I think most people would
view these as monomorphemic initially and perhaps forever.  In some of the
most frequent cases, the phonology alone is an indicator that they are not
simple combinations of auxiliary and negative.
  Why do you ask?  How can knowing this be of any forensic importance?

--Brian MacWhinney, CMU


On 5/17/04 9:50 PM, "Anne Graffam Walker" <agwalker at cox.net> wrote:

> Haven't heard from any of you experts out there in response to the message I
> sent (I hope!) on May 10th, and haven't seen it come in to my own Inbox.  I' m
> very new at this , so maybe I'm expecting something that doesn't happen (a
> sender getting a message he/she sent).  If so, would someone be kind enough to
> let me know at least that the message is/was floating out there in the
> internet ether?  Many thanks.
>
> Message repeated below, just in case.
>
> "Hello, all.  I'm looking for an authority or two or ten on a principled,
> underline "principled", method for counting words in a child's utterance.  I'm
> especially interested in how negative contractions are treated.
> Thanks to anyone who can help me on this."
>
> "Anne Graffam Walker, Ph.D.
> Forensic Linguist"
>
> [6404 Cavalier Corridor
> Falls Church VA 22044-1207
> Ph: 703-354-1796
> Fax: 703-256-2914]
>
>