210 A CREATIVE COGNITIONTo illustrate the prob lem, consider cognitive theories of thinking. It is
prob ably agreed that thinking is really the core of cognition. IQ test con-
structors and education theorists claim to be in close touch with it. As
with nearly all cognitive theory, though, we are not told exactly what
thinking is. We are given lists of prob lems that thinking seems to solve,
but not how it does it. For example, in their leading textbook, Cognitive
Psy chol ogy, Michael Eysenck and Martin Keane tell us that thinking is
a set of functions, including prob lem solving, decision making, judg-
ment, reasoning, and so on. A vast variety of fl owchart models of these
pro cesses has arisen from an impressive array of ingenious studies on
animals and humans. Some of these distinguish between diff er ent kinds
of thinking, such as declarative (used in speaking) and procedural (used in
doing things) knowledge. But they are described by their ends, not their
means, which tend to be represented only by boxes and arrows.
Th e most popu lar recent notion— refl ecting the dominant computa-
tional model—is that thinking consists of operations on internal cognitive
symbols, or repre sen ta tions, using some kind of rules. Th e diffi culty lies
in specifying the nature of those repre sen ta tions, how they are set up and
manipulated, and by what kind of rules. Th e tendency has been to break
inputs down into ever- smaller if- then associations. On encountering a
prob lem, these associations are thought to be confi gured into appropri-
ate sets and sequences by the rules built into the system or learned in
some way. According to theorists like Steven Pinker, any prob lem can
be broken down into a series of steps like those forming the basis of
a machine “that thinks,” and so explain the vast repertoire of human
thought and action. So- called artifi cial intelligence grew out of the com-
putational assumption that machines can be made to think like humans
in the same way.
Th e view has been increasingly criticized for the implausibility of many
of those assumptions. Real inputs to real cognitive systems do not come
in the form of discrete ele ments, as already mentioned. Th ey are continu-
ous and coordinated, and usually novel, fuzzy and imperfect. Moreover,
although internal repre sen ta tions are usually described by analogy with
tangible images, like pictures, maps, three- dimensional models, or even
linguistic propositions, they cannot literally be like that in the soft tissue
This content downloaded from 139.184.14.159 on Tue, 17 Oct 2017 13:54:24 UTChttp://www.ebook3000.com