Corpora: Re:What is a corpus

[Torgny Rasmark]

My 5c.
Words are containers. The meaning of a word is a more or
less
temporary assignment. This assignment is a result of
negotiation among
the participants of the communicative process. A set of such
assignments could be called a protocol. Communication events
of a
frequent nature tend to, more or less, reuse parts of
earlier protocol
'versions'. Frequently reused protocols become ingrained.
Then we tend
to regard them as facts. If our culture has systematic
language
education the protocols become norms. An achievement would
be to regard
the process itself rather than the result of it as the
object of
education.

That said: Of course must the corpora list discuss the
meaning of the
word corpora. This discussion illustrates possible
boundaries for the
notion itself and serves as ground for our using of it in
the
negotiations we must pursue to express our opinion
concerning corpora
and language.

Let Language be the written, sounded or signed (painted,
danced, ...?)
results of all human intercommunication hitherto. Then to me
a corpus
is a subset of Language.

This is a practical definition. As a systems designer of
corpora
acquisition and access software, it enables me to regard
results of an
access as an object for further inquiry along the exact same
procedure.
>From the systems point of view (if it has one) that is the
same as
regarding the output from a corpus inquiry as another
corpus.

Any selection from Language have some bias. The trick is to
know what
bias. Then one can (possibly) draw valuable scientific
conclusions from
the selection. Then it is relatively uninteresting if the
selection is
a subsubsub...subcorpus or merely a corpus.

Best Regards
Torgny
-------------- next part --------------
A non-text attachment was scrubbed...
Name: torgny.rasmark.vcf
Type: text/x-vcard
Size: 369 bytes
Desc: Card for Torgny Rasmark
URL: <http://listserv.linguistlist.org/pipermail/corpora/attachments/20000203/31a71fe7/attachment-0001.vcf>