Page 178 - Deep Learning

Creative Insight Writ Large 161

removed. Art historians prefer the term school, a loosely connected popula-
tion of artists or writers who paint or write in a particular style, explore a par-
ticular technique or – more seldom – adhere to a shared manifesto about the
purpose of art. A different type of historical unit is the market, a population of
producers and consumers. Although usually regarded as an economic rather
than cognitive entity, a market can be a locus for the production of novelty.
The explosion of electronic consumer products in the period 1950–2000 is an
example. A popular unit among historians is the nation, a population with
a geographic location and a shared culture. Historians and social scientists
also think in terms of armies, religions, social classes and other types of his-
torical entities, each of which might be a locus for the production of novelty.
The important observation for present purposes is that these units of analy-
sis are larger or last longer than the collectives that we call groups, teams or
organizations.
What are the characteristics of the production of novelty in historical sys-
tems? Examples of alterations in mode and tempo are ready at hand. The rate
of technological invention in, for example, Western Europe from a.d. 500 to
the year 1000 was obviously lower than the corresponding rate in a.d. 1500 to
55
2000. The rate of innovation in the information technology industry in the
period 1975–2000 was higher than the rate of innovation in, say, automobiles,
buses and trains over the same period. This year’s car is different from the
1975 car, but the year 2000 laptop computer was not merely different from
the computer of 1975. The personal computer did not exist in 1975; we were
still working on time-sharing machines. In the words of technology historian
N. Rosenberg, “One of the central historical questions concerning technical
progress is its extreme variability over time and place.” 56
In science, we might compare the relative stability of physics in the period
1790–1830 with the turbulent emergence of relativity and quantum mechanics in
57
1890–1930. As for nations, the increase in economic growth that produced the
Industrial Revolution in Britain in the period 1780–1830 was more rapid than
during any preceding period. The economic policy of New Zealand changed
58
at a more rapid rate in the 1980–2000 period than in the 1960–1980 period.
59
Such differences in rate of change are difficult to quantify with any precision, but
when considered in conjunction they nevertheless convince. Alterations in the
mode and tempo of novelty production recurs at historical time scales.
Can impasses at the historical level – periods of stagnation as a histo-
rian might call them – be caused by the constraining effects of prior knowl-
edge? In principle, the constraining effect of unhelpful presuppositions scales
in a straightforward way: If every member of a historical system embraces a