[Corpora-List] little favour

[Krishnamurthy, Ramesh]

Hi Diana and all

To me, the problem with non-corpus linguistic descriptions,
e.g. in dictionaries and grammars, is that they give very little indication
of 'importance/salience in the system'.

Most dictionaries list all their headwords with equal status. The clues about importance lie in a) which headwords they choose to include and exclude b)the length of the entry. Headword/subentry status depends on derivational relationships.

Similarly, grammars list various statements at various levels in their models of grammar, but the hierarchy again depends on the grammar model and its categories, not on actual occurrence/usage in daily life.

I remember in (c 1992) Michael Halliday saying something to me like (and I apologise to him if I have misinterpreted what he said): 'when I learned Chinese, I was given numerous grammar rules regarded as essential; over the past 50 years, I have realized that half of those rules are almost never needed, another quarter are needed occasionally, and only a quarter are needed on a regular basis. I wish I had known this 50 years ago'!

The potential of corpus linguistics is to highlight those features that we need to attend to 'on a regular basis', and then to try to account for the less regular or rarer features. Patrick Hanks uses the terms 'Norms' and 'Exploitations' to cover two major aspects of regular usage.

Elicitation exercises may be useful to understand the differences between actual usage and intuition (which may well include many 'folk-myths' about the language)... let us wait and see what Prof Magnusson's informants provided.

Best
Ramesh

-----Original Message-----
From: corpora-bounces at uib.no [mailto:corpora-bounces at uib.no] On Behalf Of Santos Diana
Sent: 09 September 2007 15:18
To: W.Louw; corpora at uib.no
Subject: Re: [Corpora-List] little favour

Dear all,

I think this is a very interesting question: what is the relationship
between corpus data (whose use Ramesh so well defended to give an
appropriate description of an English verb) and the (hypothesized poor)
correlation with made-up sentences from corpus linguists, which Geoffrey
Sampson and Ulf Magnusson consider such an interesting question.

Elicited made-up sentences are probably unrepeatable and members of
corpora-list, by being corpus linguists (and therefore linguists) may be
somehow suspect as average English speakers, so I still side with
Ramesh's initial posting that it is puzzling why one would want to
gather this data in the first place.

I am absolutely for using our own language skills and introspection to
interpret corpus data. However, in this particular case, I am quite
puzzled to see the two kinds of data above compared, and what points
this comparison might illuminate.

If the resulting descriptions are widely different, should that surprise
us? Even assuming that we are talking about the same number of speakers
in the BNC and those who answered the "little favour" (probably not),
are in any sense the two activities comparable?

I don't think so. In BNC, they are using the verb "slip" in normal
communication; in the "little favour" case, they are illustrating two
kinds of usage of "slip" with 3 sentences each...

So I am curious to hear what kind of arguments can be involved here;
which I am obviously missing...

Thanks in advance for your answers,
Diana

> -----Original Message-----
> From: corpora-bounces at uib.no [mailto:corpora-bounces at uib.no]
> On Behalf Of W. Louw
> Sent: 7. september 2007 17:55
> To: corpora at uib.no
> Subject: [Corpora-List] little favour
>
> Hello All
>
> It strikes me that little mentalist favours can turn into a
> disservice to scholarship. It was Gottlob Frege about 100
> years ago who first cautioned us NOT to (as Firth later said)
> 'fish in our own tank' for made up examples starting from a
> single word. Frege's logic demands that there be both a
> function and an argument. This latter 'argument' enabled him
> famously to state the difference between the morning star and
> the evening star, notwithstanding the fact that they refer to
> the same astral body. The answer lies in the collocates.
>
> We would be a lot better off dropping something as manifestly
> weak as grammar in favour of collocation, which is altogether
> more thorough at finding outer limits computationally. The
> outer limits of SLIP will be found in its different
> involvements in _Sachverhalten_ or Wittgensteinian 'states of
> affairs'. Once you have all of these, you can establish the
> line of best fit for your SYMBOL.
> Collocation is capable of re-assigning the categories of mere
> grammar. Meaning is what we ought to be after, unless we have
> come to language study in order to enjoy the solipsism of
> manipulating symbols. Geoffrey appears to be defending a
> practice that I once used disparagingly in an article on
> semantic prosody and that he insisted he be allowed to remove
> as a condition for reprinting the article. Ramesh, to his
> credit, includes in his recent re-print of the article my
> 'little favour' request to readers to come up with sentences
> involving the form 'without feeling'.
>
> Bill Louw
>
> Head
> English Department
> University of Zimbabwe
>
> _______________________________________________
> Corpora mailing list
> Corpora at uib.no
> http://mailman.uib.no/listinfo/corpora
>

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

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