Philosophy | Philosophical Psychology
P570 | 14468 | Weinberg


Models & Explanations in the Cognitive Sciences
Instructors: Jonathan Weinberg and Colin Allen

Modeling has become an absolutely essential tool for research in the
cognitive sciences, which include cognitive psychology, linguistics,
cognitive neuroscience, artificial intelligence, and robotics. But
the literature seems to contain a great many different kinds of
models, and a great many different uses for models. This course will
survey philosophical issues that arise in the consideration of this
great diversity of approaches to models and modeling.

Topics include:
--the methodological value of models: What advantages and/or
disadvantages does modeling have over other analytic or explanatory
activities?
--the ontological commitment of models: When (if ever) is it
appropriate to treat something that we appeal to in making a model as
reflecting some real aspect of the world? What is the relationship of
models to natural or scientific laws?
--debates over different fundamental approaches to modeling in
cognitive science, including mathematical, statistical,
computational, and process models, as well as connectionist,
dynamical, multi-agent, evolutionary, and robotic (both virtual and
incarnate) models.

We will also spend some time in the course looking at particular
models of interest such as generativist syntax, "mental model" models
of human reasoning, "theory" vs. "simulation" models of mental state
attribution, neurocomputational models of vision, dynamical models of
infant perseverative reaching, prototype and exemplar models of
concepts, and structural models of concept learning and
categorization.

Students do not need to have any particular scientific background in
order to take the course, but should have some degree of comfort in
reading scientific papers. Students will have an opportunity to
propose specific models for discussion. Requirements will include
frequent brief "reaction" papers, plus the student's choice between a
series of three medium-length papers or one large paper.

The course is cross-listed with HPSC X755 and COGS Q700.