Mechanism, Dynamics, and Visual Experience
William Bechtel
The mechanistic program of
decomposing and localizing functions has proven very successful in discovering
the basic structure of the mechanisms underlying vision. Neuroscientists
have discovered a hierarchy of brain areas, each of which processes different
components of visual input. The strategy of developing such mechanistic models
relies in part of what McCauley and I refer to as the heuristic identity
theory--brain areas are identified with processing specific functional processes
on the basis of suggestive preliminary evidence. If an hypothesized identity
proves fruitful in directing further research, the identity claim becomes
incorporated into the foundations of the science and does not have to be
independently justified. This suggests that insofar as we want to understand
visual experience the strategy ought to involve tentatively identifying an
area in the brain in which what is processed corresponds to features of the
world of which we are visually aware. Once plausible tentative proposals
are advanced one can convert Leibniz's law into a discovery tool to expand
both our knowledge of neural
processing and the characteristics of visual experience. I will discuss a
number of pieces of evidence, such as that stemming from visual agnosia,
that support Jesse Prinz's hypothesis that areas in extrastriate cortex constitute
the loci of visual experience. The identification of brain areas with functional
roles is just a first step in understanding brain processes. The visual
system is a complex integrated system with feedback and collateral processing
as well as feedforward. Understanding the interactive processing in this
system requires a richer set of tools than have been brought to bear so far.
Dynamical systems theory offers such tools, and I will end with suggestions
about how a dynamical perspective might be brought to bear in understanding
extrastriate visual processing.