April 20, 16-17:30, Erasmusbuilding 14.01
Henk Zeevat, UvA
Uniform Contexts for Cognition?Abstract:
Standard notions of context in dynamic context come close to providing
a uniform notion of context, but do not quite get there. From a
cognitive evolutionary perspective this is unsatisfactory: a context
of interpretation different from a global context should not diverge
in properties from the global context and the way information is
treated there: extra contexts are a late addition to the cognitive
scene. It is extremely implausible that assertion and
presupposition (anaphora) would not be exactly the same when an
auxiliary context is constructed. It is equally implausible to assume
--- for other than strategic reasons--- that basic global context can
only be updated. A visual scene is open to further refinement, but
also to correction.
My desiderata for contexts are therefore both uniformity under
assertion and presupposition (with anaphora as a special
case) and the possibility of corrections. I will try to show that
various notions of context that can be formulated from DRT or update
semantics have serious problems in these respects.
My own attempt is preliminary: I will provide a syntactic model of
contexts based on the work of Gazdar and Kamp and will go through the
technical details. Some consequences for the treatment of
presupposition and denial will be explored including a number of
empirical linguistic applications.