Page 168 - Deep Learning

Creative Insight Writ Large 151

his revolutionary religious theses by nailing a single handwritten copy to
a church door. The contrast to attaching a document to an e-mail message
could not be greater. In general, between the fall of the Roman Empire in the
fifth century and the beginning of the first stirrings of the scientific revolu-
tion in the 13th, European science is better described as a scattering of nearly
unrelated geniuses than as a community of collaborators. The same is true of
art and technology. Yet, creative works appeared.
Even in collaborative enterprises, every new idea or concept arises in a
single mind for the very first time. As told in Chapter 3, there was a moment
in January 1935 when A. F. Wilkins formulated the thought that perhaps we
can detect enemy airplanes at a distance by bouncing radio waves off them. That
moment did not give us radar, but it was one event on the path to radar, and
a very interesting one. There is no contradiction between saying that, on the
one hand, the idea of detecting airplanes with radio waves came to one person
at a specific moment in time, and, on the other hand, that the development of
a working radar took a good many people the better part of a decade. These
41
statements do not contradict each other because they pertain to different ques-
tions: How, by what mental processes, did this or that idea arise in such-and-such
a person’s mind? versus How did the members of such-and-such a group interact
to solve their problem? At the level of the individual, ideas, concepts, percep-
tions, hypotheses, schemas and other cognitive units interact within a single
mind to form the individual’s view of the problem at hand. At the level of the
collective, the individual persons are the units, and their interactions constitute
the problem-solving approach of the collective, the latter considered as a cog-
nitive agent in its own right. Collectives can in turn interact to form systems
of yet greater scope, as when multiple research groups pursue the same field of
research while interacting through conferences, peer review, personal contacts
and so on.
In short, groups consist of individuals, so the question of whether novelties
are produced by individuals or groups is misconstrued. The theoretical task is
not to choose between the individual and the group or to prove that one system
level is more real or interesting than the other, but to investigate how events at
one level affect events at the other. Scaling along the dimension of collectivity
has different consequences from scaling along time and complexity.

Stagnation and Breakthroughs in Collectives
3
A group, team, population or organization – any collective of 10 to 10 indi-
viduals – engaged in the production of novelty can be said to search and pose