|
OBJECTIVES
Objectives of the
research
a.
To complete a series of psycholinguistic experiments in order to
delineate the weighting of scalar and contextual factors across a wide
range of quantifiers.
b.
To complete the design and implementation of a computational model
for quantifiers which covers both scalar and contextual variable sets, and
to run simulations which deal with a wide range of quantifiers
c.
To complete a series of psycholinguistic experiments in order to
evaluate the predictive qualities of the computational model.
WORKPLAN
Phase 1 Design of computational model + preliminary
experiments (Expts. 1-3)
Main Landmarks: 1a: Analysis and design
specification for the new model + adaptation of vision module to deal with
multiple objects. 1b: Experimental results on sentence-only
experiments. 1c: Relationship between spatial language model and new
quantifier model (NCPW9 proceedings paper; awarded “best paper” in
cognitive category at Spatial Cognition 2004 conference).
Phase 2 Simulation of psychological counting and
model integration + Experiments on the interplay between contextual
variables and on number judgement (Expts. 4-7, 9-11)
Main Landmarks. 2a: Simulation results on the
neural network models for the processing of scalar (e.g. psychological
number) and contextual variables (2nd Language & Technology Conference
Proceedings); 2b: Experimental results on context effects with fish
stimuli (Cogsci05 paper).
Phase 3: Simulation of quantifier ratings + Additional
experiments on abstract objects, functional similarity, and cross
linguistic investigation (Expts. 8, 12-14, 15-16).
Main Landmarks: 3a: Simulation results on the
prediction of quantifier ratings (ICANN05 paper); 3b Experimental
results on crosslinguistic study (ESSCS05 workshop, Italian AIP conference);
3c Paper on the comparative study using the developed computational
models and experimental data (Cognitive
Psychology journal article, in preparation).
Deliverables listed in publication page.
|
