LINGUIST List 5.553

Mon 16 May 1994

Disc: E-mail language

Editor for this issue: <>

Directory

1. "Simeon J. Yates", Re: 5.526 Qs: E-mail language, Monographs, Agni, OCR for Cyrill

Message 1: Re: 5.526 Qs: E-mail language, Monographs, Agni, OCR for Cyrill

In response to the query on e-mail language I thought I
might make a 'public' rather than a 'private' resposnse.
I have spent the last few years conducting a corpus
based comaprison of computer-mediated communication with
speech and writing. Rather than attempt to summarise
the whole of the results in a short e-mail here are a
few of the results and some of the outstanding
questions. (N.B. the analysis used heavily a Hallidayan
approach)

Using various mesaures of the 'textual/modal' aspects of
CMC (e.g. Lexical Density/Type-Token ratio's) most CMC
genre were found to be closest, though statistically
different from written genre.

Using basic measures of the 'interpersonal/tenor'
aspects of CMC (e.g. pronoun use) the similarities to
speech and writing were dependednt upon the CMC genre.
One common feature was the lack of third person
reference in CMC, with comparable proportional use of
1st and Second person pronoun use between speech and
CMC.

Using basic measures of the 'ideational/field' aspects
of CMC (e.g. modal auxiliary use) CMC was found to
considerably different from both speech and writing with
 much greater occurance of such linguistic features.
Having said this the proportions of different forms of
modal auxiliaries in use were similar between speech and
writing.

Statistically identifiable genres were also uncovered in
CMC. These genres could be described in terms of their
use of spoken and/or written generic resources in the
construction of messages. This is too much of a
simplification as some of the genres found
demonstrated features found only in the CMC corpus.

Questions:

Why the higher levels of modal auxilary use? I have
some ideas but I'm fishing for more.

At what level do genres truly function? Are they simply
post hoc (possibly ideological) categorisations? Or do
they represent a resource to be used at the
lexico-gramatical level of text production? The results
from my CMC analyses would indicate the latter. Having
said this the 'genres' within LOB and London-Lund
corpora which I used for comparative purposes
demostrated much less comparability at the
lexcio-gramatical level (though this my be due to
limitiations in the data sets and statisitical
techniques)?

Simeon.


Dr. Simeon J. Yates
Institute of Communications Studies
University of Leeds
Leeds
West Yorkshire
LS2 9JT Phone: +44 532 335806
United Kingdom Fax: +44 532 335808

E-mail: SIMEONICS-SERVER.NOVELL.LEEDS.AC.UK