Page 173 - Deep Learning

156 Creativity

by companies like Apple, Data Corporation, Hewlett and Packard, IBM and
Microsoft. In general, pronouncements by authority, explicit codification or
inclusion in an incentive system might serve to entrench unhelpful assump-
tions and create impasses that are unwarranted in the strong sense that the
collective contains individuals who are not at impasse, or have already over-
come an impasse, but who are not allowed by the authority structure of the
organization to act on their insights. From this point of view, communication
within an organization can act as an additional cause of impasses. That this
possibility is more than idle speculation is shown by the extensive literature
by business thinkers about how to create an institutional atmosphere that
fosters and sustains creativity.
If groups have unwarranted impasses, do they also have insights? The
scaling mechanism once again appears simple in principle: It is enough that
a single member of a collective has an insight for that collective to have that
insight, or “breakthrough” as they are often called in this context. It required
Mercator’s genius to invent the cartographic technique that bears his name,
but almost every map that we look at utilizes that technique. From the day
that Watson and Crick published their description of the double helix struc-
ture of DNA, the entire disciplines of biochemistry and genetics possessed that
insight. Once Walter Reed and his co-workers and associates had described
48
the mosquito path for the dissemination of yellow fever, the entire medical
profession knew. 49
Such cases suggest a straightforward application of the mechanisms
behind individual insight to collectives. However, they mislead by being
too clear-cut. Parallelism and communication introduce processes and phe-
nomena that have no analogues in the individual mind and complicate the
picture
The parallelism inherent in collectives can be utilized to conduct parallel
searches – that is, to launch multiple search processes that explore different
regions in a search space, or different search spaces, in the service of the same
subgoal. This is a different mechanism from division of labor as ordinarily
conceived. The latter pertains to different members of a team attacking dif-
ferent problems, or different parts of a problem; in the present case, the focus
is on different individuals working on one and the same problem, simultane-
ously and independently of each other.
The discovery of the structure of DNA provides an example. There were
50
at least three different research teams who took an interest in the problem
in the early 1950s. In addition to Watson and Crick at Cavendish Laboratory
in Cambridge, UK, and Maurice Wilkins and Rosalind Franklin at King’s