Page 142 - Deep Learning

Creative Insight: The Redistribution Theory 125

Edward P. Chronicle, James N. MacGregor and Thomas C. Ormerod have
advanced an explanation for unwarranted impasses that contributes some-
59
thing new. They propose that subjects in insight experiments, and problem
solvers generally, translate their understanding of the problem at hand into
a sense of what it means to make progress. What does a productive step look
like? In the N-Balls Problem there are, for example, 9 indistinguishable balls,
8 weighing the same and one weighing slightly more, and the task is to find
the heavier one with two, and only two, uses of a balance scale. A plausible
60
notion of progress is to weigh as many balls as possible in the first round, to
have as few left to process with that single remaining round. In the 9-Balls
Problem, this will cause an impasse, because a 4-against-4 weighing could
leave the problem solver with 4 balls to choose among, a situation that can-
not be resolved with the single remaining weighing. There is no a priori rea-
son to doubt the progress criterion principle, and the experiments reported
by Chronicle, MacGregor and Ormerod provide strong support. The maxi-
mal progress explanation for impasses emphasizes decision heutistics while
the unhelpful knowledge explanation emphasizes problem perception and
memory retrieval. The relative prevalence of situations in which an impasse
is rooted in the activation of unhelpful prior knowledge and situations in
which it is better described as rooted in an inappropriate progress criterion
is unknown.

Resolving impasses: A role for forgetting?
There are multiple alternative explanations for the resolution of impasses. A
popular idea is that it helps to “sleep on it.” More precisely, the claim is that
the probability of finding the solution is higher after setting the problem aside
for a period of time, as compared with continued efforts to solve the problem
for the same amount of time. This notion is often ascribed to the Gestalt psy-
chologists, but it was first named by Graham Wallas, who in 1926 suggested
that creative problem solving follows a sequence of four phases called prepa-
ration (studying the problem to grasp its requirements and the goal), incuba-
tion (temporarily setting the problem aside), illumination (the spontaneous
transition into consciousness of the problem and the solution, typically in
some unrelated context) and verification (executing the solution). Wallas’s
61
1926 theory is sometimes cited as if it were the latest word on creativity.
62
His four-stage formulation has become so firmly entrenched in the culture of
creativity studies that it has become common sense, and the concept of incu-
bation and even the four stages are sometimes mentioned without acknowl-
edging Wallas.