Corpora: genre/register - separate realities

[Vladimir Rykov, PhD in Computational Linguistics, MOCKBA]

     1. When  I  studied BUC with computer for my PhD I never read its
texts.  My professor told me that I  have  not  Phylological  love  to
words.  But  once  I  stumbled at one phenomenon.  There was text with
fiction "stuff" in the "reportage" genre.  I read it.  The reason  was
obvious.  The text was about Inauguration of Kennedy. It was the first
time I saw that the text which is  "fiction"  itself  appears  as  the
reportage one.

     2. Last  Summer  I read a two-volume "The Don Juan Papers".  They
proved that famous Castaneda writings were softly "not true". But they
split the  problem  in  two parts - authenticity (A) and validity (V).
As if two dimensions of truth:  X-axis  and  Y-axis.  Castaneda  lied.
There was  no  Don Juan etc.  It was A.  But the concepts he described
were true (V).

     3. More simply.  So that you could critisize me easier. The girls
I met  often  told me that I would not recognise them if they'd change
their dress.  Then I realised that the girl is one thing and her dress
is smth different :-).

     4. Back  to our genres.  I think we should tag (as our No 1 David
Lee teaches us) our texts with two dimensions (sets) of tags.  One for
a  text itself.  David calls it register.  And another set of tags for
the text membership to the genre as it appeared in the publication.

     So we could be free to describe the  genre  pattern  of  specific
country, time etc. with our set of tags.

     And with another set of tags we could describe the various  kinds
of texts as they are.  So - girls and their dresses with different set
of tags.  As if X- and Y-axis. And - moreofit. It could be possible to
"catch" smth interesting on this X-Y plane.  I mean comparison - which
text registers are used in which genres (which dresses girls  use  for
which  reasons).  I  mean - the relationship between inner text nature
and its outer use.  Instead of smashing and squeezing the matter  into
one point.

     Vladimir Rykov



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/corpora/attachments/20000904/3a438c72/attachment.htm>