Page 166 - Deep Learning

Creative Insight Writ Large 149

example, it does not matter at the 0.10–10 years’ time band whether memory
retrieval is accomplished by spreading activation or through some other mech-
anism. What matters is that retrieval failure is a possibility. Nor can it matter
exactly how activation is redistributed at a single choice point. The principle
of redistribution of activation in a single processing unit – the central contri-
bution of Chapter 4 – does not have any consequences at the significant time
band. There are no regularities in the production of novelty at the higher levels
of time or complexity that owe their existence and shape directly to the fact, if
it is a fact, that activation in a single processing unit is redistributed in the par-
ticular way specified by that principle. What matters is that activation is capa-
ble of being redistributed. Only the gross characteristics – problem perception
determines knowledge retrieval; knowledge constrains search; feedback affects
the distribution of activation – punch through to higher levels of scale.
Conversely, the micro-theory of insight does not explain all phenomena
and patterns at higher levels of time and complexity. New mechanisms emerge.
34
For example, consider problem finding. It is often claimed that creative people
have a superior ability to spot a fruitful problem or a promising opportunity.
Case studies provide examples. John J. Wild, the inventor of ultrasound diag-
nosis, accidentally discovered that a cancer growth gave a different ultrasound
echo than the surrounding healthy tissue, opening up the possibility of using
ultrasound to diagnose cancer. He recognized that this problem was more
35
significant than the one he was working on at the time, so he set it aside to
explore the ultrasound technique. Serendipity, to accidentally stumble on a
solution to a problem that one is not working on and nevertheless recognize
it as a solution, is a closely related and well-established phenomenon. Both
36
problem finding and serendipity rely on a special type of sensitivity to the
environment that allows a person to be simultaneously goal-driven and inter-
ruptible. This phenomenon is not captured in the theory of analytical problem
solving and it does not seem amenable to an explanation in terms of represen-
tational change.
The long duration of significant projects also exposes them to the impact
of externalities. The role of externalities is discussed in Chapter 4 in relation
to fortuitous reminding and incubation, but their importance increases at the
higher levels of time and complexity. Events in the environment that either
distract the problem solver or trigger cognitive change are not accounted for
in the micro-theory of insight. But even a quick scan of the history of some
creative projects indicates a strong causal impact for events that are indepen-
dent of the internal dynamics of the creative work. If Darwin had not been
invited, through family connections, to participate in the voyage of the Beagle,