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

[Wolfgang Teubert]

Dear All,
 
A number of responses I have received via the list or in private suggest that the future will see the integration of corpus linguistics with cognitive approaches. I disagree.
 
I am glad to notice that the emergence of corpus linguistics had the effect that people now are actually working with real language data. They are using the toolbox that corpus linguists have developed over the last forty years, and other tools developed by other people. These tools provide results. To some extent, these results depend on the mark-up including the annotation carried out beforehand. A word tagged as an adjective will be counted as such. To some extent, we can also learn something new. If I am interested in synonyms and antonyms, I might look up the phrase 'friends and', and I may not know beforehand which nouns will come up as the next word on the right. This is what corpus technology delivers as 'brute facts'.
 
It is then up to me to interpret the findings. I can use existing concepts like 'antonym' and 'synonym' to pigeonhole the results. But it matters how I define these constructs. Should I see synonymy as something that plays a role in the mental mechanism that turns thought into language and language into thought? William of Ockham discusses at great length the question whether there are synonyms in the lingua mentis. If mental concepts are language-independent and universal, what do we do with synonymy? How synonymous are the words I find in WordNet's synsets? What indeed do we know about mental concepts? Are there more concepts than words (Sperber/Wilson), or are there about as many (Fodor), or are there only some fifty (non-synonymous) semantic primitives from which all other concepts are built (Wierzbicka)? How synonymous are 'chien' and 'dog'?
 
The problem is that the mind does not allow introspection. No one has ever presented evidence for a single mental concept. Everyone is working within their own model, or a model of a given school. But they are only models. What exactly do we gain from this modelling?
 
If someone can show me anything about the synonyms and antonyms of the plural 'friends' that is not already there in the discourse, I will have been convinced. If we really want to take such concepts seriously, we can only, I believe, develop them bottom-up, by interpreting the textual evidence. We will then find that relationships between words, as they are actually used, embedded in the unpredictability of their contexts, are a lot more complex and irreducible to the rules of the simple cognitive models fed to us are able to accommodate. To the extent that language is symbolic, it is contingent and not rule-based.
 
Language is symbolic. A sign is what has been negotiated between sign users. The meaning of a sign is not my (non-symbolic) experience of it. Meanings are not in the head, as Hilary Putnam never got tired of repeating. The meaning of a sign is the way in which the members of a discourse community are using it. It is what happens in the symbolic interactions between people, not in their minds.
 
This is why I find cognitive linguistics flawed.
 
Wolfgang Teubert


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