Integration of Background Knowledge in Sentence Processing: A Unified
Theory of Metaphor Understanding, Semantic Illusions and Text Memory

Raluca Budiu and John R. Anderson

Abstract


One of the challenges of cognitive psychology is developing
general models that explain a wide range of empirical phenomena.  We
describe a unique language comprehension model that fits data from
several text comprehension domains: metaphor understanding, processing
of semantic illusions and text memory.  We show how background
knowledge plays a similar role in all these processes, helping or
hampering them. The model assumes that sentence processing at the
semantic level is incremental, nondeterministic and incomplete, and
uses background knowledge hints at each step. The model is implemented
in the ACT-R framework \cite{anderson:98}.  The empirical phenomena
that we model are: position effects on metaphor understanding,
influence of distortion ``quality'' on semantic illusions and memory
for related stories.