Corpora: register and genre

[Geoffrey Williams]

My intention was not to give an extended reply as we all know that text
categorisation is a complex problem However,
I tend to reject the term sublanguage as used in artificial intelligence and
generative linguistics as it implies a neat classification of language
rather than a continuum where one typology can blend into another. I prefer
to refer to discourse communites and their use of genre, particularly when
dealing with academic text. Register can be seen in two ways, both of which
I see as fairly broad categories; very general categories such as a
newspaper register or a more precise classification through the situational
features of field, tenor, mode as in Halliday & Hasan 1976. The former is
that adopted in language teaching and is eminently fuzzy.
Hence if we wish to go beyond more or less fuzzy registers then we should
turn to
genre for greater precision, in this I follow Swales 1990. To my mind
commerce/electronics and natural science/history are far too broad to be
significant unless you have a vast corpus, this is why I would like some
precision so as to compare something a bit more specific. Using a DC as in
Swales allows to tie down a more precise community and to justify content,
something I have argued for in several publications (notably Williams 1998 &
1999).
Now let us say we still wish to compare the two broad subject domains
mentioned, we must now turn to genre. Here I accept Biber's (1988) broad
genre of academic prose but require more precision. Academic prose covers a
multitude of sins being addressed to different audiences for different
reasons. One of the most studied genres is that of the research article
hence the IMRD common in science, in this case we can specify an audience
type which makes comparison between research articles in say biology,
mathematics and linguistics an interesting proposition. I would guess that
we would find 3 subgenre here with only biology using the IMRD model.
"Style" I naughtily used in a very vague way when wishing to speak of
features that might be used to demonstrate subgenre.
In summary thus I retain my initial analysis where reports etc form genre,
particular lexicogrammatic features adopted for use with a particular
audience.
Hope this helps
best
Geoffrey
refs:
Biber D. 1988. Variation across speech and writing. Cambridge : CUP.
Halliday M.A.K. & Hasan R. 1976. Cohesion in English London: Longman
Swales J. 1990. Genre Analysis. Cambridge : CUP
Williams G. 1998. Collocational Networks: Interlocking Patterns of Lexis in
a Corpus of Plant Biology Research Articles. International Journal of Corpus
Linguistics. Vol 3/1: 151-171
Williams G. 1999. Looking in before looking out: Internal selection criteria
in a corpus of plant biology. Papers in Computational Lexicography. Complex
'99. Hungary: Budapest : 195-204


********************************************************************
Dr G. C. Williams
Université de Bretagne Sud
4 rue Jean Zay
56100 Lorient
personal e-mail: geoffrey.williams at wanadoo.fr
*****************************************************************

-----Message d'origine-----
De : David Lee <david_lee00 at hotmail.com>
À : corpora at hd.uib.no <corpora at hd.uib.no>
Date : vendredi 1 septembre 2000 17:44
Objet : Re: Corpora: register and genre


>> I would say that you are dealing with two different registers, the
>>genre is
>> academic prose, but that is probably a very high level genre as the
>>styles are probably very different.
>
>[Beverley has now received apparently contradictory advice. (I suggested
>calling her groups of texts 'genres' in a previous mail.)]
>
>Geoff: Is there any reason why these two can't be different "genres" (as
>well)? You've used 'register', 'genre',  and 'style' in one sentence
>(plus 'discourse communities' and 'fields' later on), but have given no
>indication of how you distinguish among them.
>
>
>> If you are dealing with research publications
>> your sciences texts will probably follow variations on Swales's IMRD
>> model, that would not be the case in Commerce/Economics.
>
>Presumably you are working within the systemic-functional tradition and
>have the 'unfolding of stages' as a key criterion for 'register'. If
>commerce/economics as a  (subject) 'field' ('domain'? Terminology
>galore!) has not yet been established as having a particular GSP or
>field-mode-tenor configuration, why assume commerce/economics texts
>constitute a (single?) 'register'? My approach would be to view such
>texts as forming a more or less coherent genre (on the basis of having a
>fairly discernible 'discourse community' with shared interests), with
>different kinds of texts within the genre *perhaps* having distinct
>'registers' in the SF sense (e.g. financial reports, commercial law
>professional guidebooks) and other kinds being essentially formless and
>too varied to be 'register-typed' (e.g. economics textbooks).
>
>
>> One of your first tasks
>> will be to define what is prototypical of the genre you are studying
>> and also define the discourse communities you intend to look at as
>> both the fields you give are very wide
>
>
>Perhaps what Beverley was looking for was precisely a term which was
>capable of describing these very wide fields in some way. I have
>suggested 'genre' as an elastic term which assumes nothing (and makes no
>claims) about the internal or textual characteristics of the texts, but
>merely characterises them in socio-cultural, text-external terms.
>('Register', on the other hand, doesn't seem (to me) to have this
>elasticity or fuzziness: it tends towards technicality and specificity
>because it is defined in terms of specifiable textual features
>empirically established (e.g. Halliday & Hasan's Generic Structure
>Potential (GSP).) Granted, there are exceptions and problematic genres,
>and the more you look into this the messier it gets, but as a working
>definition, I think it works.
>
>
>
>David Lee
>