Page 137 - Contribution To Phenomenology

130 TOMNENON

paper, Connectionism provides a model for explaining active representa-
tions in terms not only of other active representations stored within the
system (remembered or sedimented beliefs), but rather in terms of
weightings as dispositions to produce certain representation under the
proper circumstances. The "experience" of a cognitive system then would
consist not in a practically infinite number of stored beliefs, but in the
weightings it has adopted to bring about appropriate beliefs under the
specific conditions. The upshot of all of this is simply that one should
not dismiss things like tendencies and dispositions within the mental realm
simply because they do not seem to fit neatly into the region either of
the mental, the physical, or the ideal. One must also be careful to avoid
the distortion of the phenomena by all too hastily concluding that they
must be kinds of stored beliefs that are kinds of "remembered"
judgments upon which one no longer focusses directly. Connectionism
suggests that there may be sedimentations of experience that are not now
and never were judgments in any sense, but rather exist only as a
tendency to process new information in certain ways, as functions within
mental life whose existence we become aware of only in their functioning.
Admittedly here, these very sketchy remarks leave out much that
would be necessary for this to become a serious hypothesis about the
nature of dispositions. But thinking about the problem in these connec-
tionist terms does point a way out of the dilemma I outlined above. It
would provide Husserl with a way of asserting both the accessibility of
all genuinely mental events in a paradigmatic sense, while at the same
time being able to posit other states that may not pass this test, even
though they do play a role in mental life, proto-nfiental states perhaps,
that would not have to be thought of simply as stored beUefs (which, I
might add, present an even greater problem for the traditional approach
if we extend our focus beyond the strictly theoretical realm and include
dispositions for actions as well).

IV

In the preceding section, I have indicated in a general way how the
connectionist model might provide resources for reconceiving a set of
problems within Phenomenology. In conclusion, I would like to suggest
that cognitive scientists could learn something from Transcendental
Phenomenology about how to conceive of representation more adequately,
something that might suggest a better way to construct systems that