Page 43 - Deep Learning

26 Introduction

matter how accurate, does not suffice to explain the regularities associated with
7
retrieval failures, tip-of-the-tongue feelings and similar mental events. This and
many other observations force the conclusion that there are mental processes that
are not conscious. The weakness of the phenomenological approach does not lie
in its descriptions of conscious experiences but in the unwarranted add-on claim
that there is nothing but conscious experience for a psychologist to describe.
Other researchers deny mind by equating it with the brain. This approach
8
is the more attractive, the less a person knows about psychology. It is a peren-
nial favorite among computer scientists, members of the medical profession
and particle physicists with philosophical aspirations. According to this view,
all psychological phenomena are to be reduced to descriptions of neurons,
modules, pathways, transmitter substances, axons, dendrites and synapses –
in short, the material anatomy and physiology of the brain considered as a
cause-effect machine. Once the brain has been completely described, all the
questions of psychology will also have been answered, or so the claim goes.
This reductionist approach ignores the fact that we need multiple levels of
description to understand any complex system. A description of human cog-
nition at the level of individual brain cells would not only be impractical in its
overwhelming complexity but it would also be uninteresting. To see the need
for another level of description, consider questions like the following: Why did
person X succeed in solving task Z while person Y failed? Why did X take twice
as long as Y to solve problem Z? Why did Y fail at time t but succeed at time t ?
1
2
Why does a sample of performances on tasks of type Z exhibit such-and-such a
regularity? These are among the types of questions that cognitive psychologists
want to answer. But an answer in terms of brain cells and neural processes
(e.g., because neurons N , N , N , …, N were active in such-and-such a way) is
3
2
1
n
not explanatory. To understand why the person performed the task at hand in
the way he did, we need to know what he was thinking. That is, we need an
explanation in terms of his beliefs, capabilities, concepts, decisions, memories,
perceptions, skills, thoughts and so on. Neuroscience is a fascinating science
and it has made great progress in explaining the workings of the brain, a most
worthwhile goal. No value is added to this science by the unwarranted meta-
physical claim that a complete description of the brain will answer all ques-
tions about mind.
Perhaps the most persistent denial of mind locates the driving force
behind our behavior outside ourselves, in the material environment. In the
first half of the 20th century, adherents of behaviorism denied that any sig-
nificant processes occur in people’s heads and asserted that very simple pro-
cesses, in combination with the structure of the environment, suffice to explain