sj> To do an experiment, I need a
sj> HUGE (200 Mo) amount of bilingual parallel corpus, in english
sj> and french...
the canadian hansards are the classic source for this sort of
information. i note however, that they are not available from the LDC
(i could have sworn they were on the ACL/DCI disk).
but, researchers can look at:
http://www.parl.gc.ca/english/senate/deb-e/deb-e.htm
and
http://www.parl.gc.ca/english/senate/deb-e/deb-f.htm
and thereupon draw their own conclusions about the availability of the
data.