Session 12, Spring 2003, 2 June 2003, 16-17:30
Abstract
Natural Language processing is important for Computer Scientists, both
for theoretical and practical reasons.
The advances in this field have been rather slow, compared to the development
of other applications of computers.
The reason is that the problem is much more difficult than what was
expected at the beginning, but perhaps also that
the main stream research is based on wrong premises.
The talk will show that a central tenet: reference-based semantics,
has serious inconveniencies, and will try
to explore another, inference-based, approach to semantics. Several
examples will be discussed.
