Seeing the Whole Picture
35
To a large extent, however, the concept of mind in psychology and cognitive
science has been advancing in a very different direction. The dominant theory
today is called the ComputationalTheory of Mind, and advocates that the human
brain is an information processing system. In this theory, mental processes
are computations realized by neural circuitry in the brain. Computation is
understood as the manipulation of symbols according to rules, hence mental
activity is believed to consist of processing symbols in the brain. At no point
in this theory is the existence of Gestalts recognized. (It is not denied, simply
ignored.)
Computation is an activity that has been extensively studied and is well
understood. It is a process that can be carried out in a variety of different
ways: by hand, by machine, and indeed by different kinds of machines. For
that reason, we say that it is machine-independent. Thus, if mental processes
are computations, then the same processes can be carried out on mechanical
computing devices. However, the mind is able to represent Gestalts, and to
draw inferences from data presented in complete wholes. Is it possible that
computers have the capability of doing that? In order to answer that question,
a few words about computation are in order.
Computer memory is nothing like animal memory. Memory in computers
is called “addressable memory” because every stored number has an “address” in
memory, and the number can be accessed only by providing its address. Stored
information is available to the computer only when a program instruction calls
out the contents of a specific memory address and inserts the contents into
the “accumulator”. The accumulator is where the data is processed, that is, it
is read and an operation is executed on it. 5
In each operation, the computer takes into account only the one number
(or item of coded data) in the accumulator, and has no access to any other
information. It “sees” only the one number that it has been instructed to
withdraw from memory and put into the accumulator. It decides and acts on
the basis of that one item of data, while all else lies temporarily in dead storage.
In the scope of one operation, the computer is not able to survey data
panoramically, hence at the instant of making a decision it takes account of
just one item of data. In contrast, when a decision is made by a person on the
basis of a Gestalt, a fully structured whole is taken into account simultaneously.
All of it together is needed to make the decision. As a consequence of this,
a decision made in a mind is altogether different from a decision made by
operations on a computer.
There is an additional and decisive reason why computations cannot deliver
Gestalt wholes: Computation, by its very definition, is the manipulation of
information. Gestalts, however, are not information: They are cognitive struc-
35
To a large extent, however, the concept of mind in psychology and cognitive
science has been advancing in a very different direction. The dominant theory
today is called the ComputationalTheory of Mind, and advocates that the human
brain is an information processing system. In this theory, mental processes
are computations realized by neural circuitry in the brain. Computation is
understood as the manipulation of symbols according to rules, hence mental
activity is believed to consist of processing symbols in the brain. At no point
in this theory is the existence of Gestalts recognized. (It is not denied, simply
ignored.)
Computation is an activity that has been extensively studied and is well
understood. It is a process that can be carried out in a variety of different
ways: by hand, by machine, and indeed by different kinds of machines. For
that reason, we say that it is machine-independent. Thus, if mental processes
are computations, then the same processes can be carried out on mechanical
computing devices. However, the mind is able to represent Gestalts, and to
draw inferences from data presented in complete wholes. Is it possible that
computers have the capability of doing that? In order to answer that question,
a few words about computation are in order.
Computer memory is nothing like animal memory. Memory in computers
is called “addressable memory” because every stored number has an “address” in
memory, and the number can be accessed only by providing its address. Stored
information is available to the computer only when a program instruction calls
out the contents of a specific memory address and inserts the contents into
the “accumulator”. The accumulator is where the data is processed, that is, it
is read and an operation is executed on it. 5
In each operation, the computer takes into account only the one number
(or item of coded data) in the accumulator, and has no access to any other
information. It “sees” only the one number that it has been instructed to
withdraw from memory and put into the accumulator. It decides and acts on
the basis of that one item of data, while all else lies temporarily in dead storage.
In the scope of one operation, the computer is not able to survey data
panoramically, hence at the instant of making a decision it takes account of
just one item of data. In contrast, when a decision is made by a person on the
basis of a Gestalt, a fully structured whole is taken into account simultaneously.
All of it together is needed to make the decision. As a consequence of this,
a decision made in a mind is altogether different from a decision made by
operations on a computer.
There is an additional and decisive reason why computations cannot deliver
Gestalt wholes: Computation, by its very definition, is the manipulation of
information. Gestalts, however, are not information: They are cognitive struc-