Irene,
As far as experimental data is concerned, it might be interesting to
note responses in a gap-fill, and see in what way (if at all) they
correspond to corpus data: language users are very creative and can
usually justify oddities (such as those you provided in your examples)
by locating them in an imagined context, so what seems unusual is
therefore normalised into ironic/ situationally relevant/ humorous, etc.
Perhaps if you tried to elicit the key collocates, rather than responses
to them, you may find out more about perceptions of 'normality'.
e.g.
Engineering of these goods will ___________ affect the market.
She caused me a lot of __________________.
...and so on.
If you're lucky you should get a series of related answers which
fit into one or more groups of recognisable semantic preferences.
If they don't behave as you would like them to, however, you can
then start asking why your production data differs from the corpus
data, and (re-)open a can of worms for us all...
Just a suggestion, but worth a shot.
Best,
Gill
*************
Dr. Gill Philip
Universit? di Bologna
CILTA
Piazza S Giovanni in Monte, 4
40124
Bologna
Italy
tel (++)39 051 209 7763
fax (++)39 051 209 7751
email g.philip [at] cilta unibo it
This archive was generated by hypermail 2b29 : Thu Jan 27 2005 - 11:32:15 MET