[Corpora-List] Corpus vs Intuition

[Gill Philip]

dear all,
I fully agree that "you can fool yourself using all discovery methods"
- which is why cross-checking is always sensible (though seldom, it
seems, ever done). Falsification of course can be done with the same
methods on different data, with different methods on the same data, or
with different methods on different data.

For example Yannick's example on semantic prosody was useful. You see,
while the term (intended as part of Sinclair's Unit of Meaning) comes
out of corpus analysis, the same concept was being mulled over over in
metaphor studies, where scholars had worked out that
metaphorically-used language [aka idiom-principled language] usually
seemed to be performing evaluatively in text. The problem was
understanding if that evaluative element was used consistently, and
expressing the same general meaning. Now, isn't it interesting that
many metaphor scholars are now using corpora, which lets them look
into this phenomenon (the "metaphoreme") using different data and
different methods. They don't call it semantic prosody, which is a
theory-bound term, but what they're looking at is essentially the
same.

Whichever methods and data any linguist uses, they will always be
incomplete. But it is important to acknowledge that incompleteness
(and not just as "more research needs to be done..." at the end of our
papers), and also acknowledge that we need to take account of what our
colleagues in other fields of linguistics are doing. Disagreements
fuel alternative approaches and accounts (ideally, at least), which in
turn cause disagreements etc etc. In the end, we're all just providing
building materials for those who come after us: there never will be
any permanent, definitive accout of any language phenomenon, although
we all give it our best shot.

just a little more food for thought.

best,
Gill


-- 
*********************************
Dr. Gill Philip
CILTA
Università degli Studi di Bologna
Piazza San Giovanni in Monte, 4
40124 Bologna
Italy

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