[Corpora-List] Bootcamp: 'Quantitative Corpus LinguisticswithR'--re Louw's endorsement

[Joybrato Mukherjee]

Dear all,

I have followed the discussion with great interest (as probaly most
Corpora subscribers), but I have to admit that I am not only surprised
but also increasingly worried to see how dogmatic corpus linguists can
be. Specifically, I don't think (and I can't believe that we are
seriously discussing these issues in a community that hasa lways been
proud of being open to new descriptions of language and new
models/concepts evolving out of the analysis of actual language use...):

- that anyone has hijacked the label "corpus linguistics" (is this
really possible?),
- that cognitive approaches to language and linguistics can be dismissed
as "flawed" in their entirety,
- that colleagues who work within a cognitively-oriented paradigm and/or
with sophisticated statistical methods have nothing to say about
"meaning",
- that it makes sense to tell people how to use corpora and what kind of
corpus use (or corpus linguistics, for that matter) is the only right
way of doing it.

In particular, I have a niggling worry that too many people accept all
to readily that there could be a divide between the "old-fashioned"
(true?) corpus linguists and the ones who use corpora and in a seemingly
inappropriate, toolbox-like, inherently non-Sinclairian way. I would
like to believe that a discipline such as corpus linguistics (and, yes,
I would argue that CL is both a discipline and a methodology...):
- must embrace new methods (including, of course, statistical methods,
given the importance of frequency in our field),
- must be open to new models and concepts (including
cognitive-linguistic frameworks) which can be linked to corpus-based
findings,
- must be open to a combination of corpus work and other methods
(including, if necessary and useful, introspection).

Whatever is useful to describe and analyse language in use and (if one
accepts that language is a mental phenomenon too...) the cognitive
dimension of language, must be allowed, should be done, discussed and
developed further.

The current discussion began with a flurry between Bill Louw and Stefan
Gries. There is unanimous agreement, I assume, that both have
contributed substantially, though in very different ways, to the
description and modelling of language in use: Both Louw's work on
semantic prosody and Gries's research on collostructions are - in my
view at least - very valuable contributions to (corpus) linguistics, and
both concepts help us immensely in understanding essential facts about
language in use. To me, it thus seems awkward being confronted with this
kind of (hostile?) exchange, developing into a strange discussion about
what the right kind of corpus linguistics is or should be...  

I firmly believe that Stefan Gries is right when he writes (in a
previous posting): "There should be room enough for all of us esp. given
that we share more beliefs with each other than with people who dream up
acceptability judgments of contextually isolated and unattested
sentences."

I couldn't agree more.

Best wishes

J. Mukherjee   


_______________________________________________
Corpora mailing list
Corpora at uib.no
http://mailman.uib.no/listinfo/corpora