Page 157 - Deep Learning

140 Creativity

Identification of key steps on the basis of autobiographical memory or his-
torical materials is fraught with difficulty and uncertainty. Regan and I attempted
an experimental validation of our analysis of James D. Watson and Francis N. C.
Crick’s achievement into eight key conceptual steps by constructing a labora-
tory version of the DNA structure discovery problem. We compared the time
to task completion for experimental subjects who received hints on these eight
points with subjects who did not receive those hints. The effect on the solution
time was strong and directly related to the number of hints the subjects received;
see Figure 5.1. Our results support the hypothesis that the double helix concept
can be approached along a trajectory that is dotted with those eight conceptual
inventions, but they cannot prove that Watson and Crick did in fact move along
that particular trajectory. We can be certain that, whatever their trajectory, it
required multiple insights.
To conclude with an example from art, Herschel B. Chipp has analyzed
the painting of Guernica into 10 phases, each of which is marked by a distinct
change in the concept of the composition. At one point, Picasso repainted the
17
head of a screaming horse so as to turn it from upside down to right side up,
drastically changing the character of the horse’s scream from pitiful to defiant.
The three examples of the electron microscope, the structure of DNA and the
painting of Guernica illustrate the lower boundary on the number of insights
in significant creative projects.

2000

1500
Solution Time (secs) 1000

500

0
0 2 4 6 8
Number of Abstract Hints
Figure 5.1. The effect of priming eight key insights on the time to solve a simplified
version of the DNA structure problem.