Genes, Brains, and Human Potential The Science and Ideology of Intelligence

REAL GENES, REAL INTELLIGENCE 115

But as with the rules of the road or of grammar in speech, it is mostly
deep information in the sense of many variables interacting at many dif-
fer ent levels. Accordingly, the most impor tant aspect of evolution has not
been that of ge ne tic adaptations to recurring circumstances. It has been
that of intelligent systems able to deal with changing environments by
abstracting such information at increasing depths. Th at involved in driv-
ing or human speech is vastly more complex than the intelligence of the
cell. But even in cells, it can be diffi cult to describe. We can just about
envisage interaction among three variables; beyond that we require math-
ematical tools, especially where relationships are nonlinear and changes
are not uniform. Moreover, with increasing numbers of variables the
system tends to become dynamical, as described above.

INTELLIGENT SYSTEMS

As usually envisaged, a gene is sensitive to a very specifi c environmental
change and responds with a very specifi c response. An intelligent system
is sensitive to how one change is conditioned by other changes, and that
sensitivity shapes more adaptable responses.
Living things, then, need to be good at registering those statistical
patterns across everyday experience and then use them to shape the best
response, including (in the cell) what genes to recruit for desired prod-
ucts. Th is is what intelligence is, and its origins coincide with the origins
of life itself. Indeed, in an impor tant sense, intelligence is life, and life is
intelligence.
Accordingly, what is now being discovered is that even molecular net-
works can “learn” the statistical rules encountered in their environments.
Th e rules are assimilated in reconfi gured reaction networks, themselves
due to changed reactivities of molecular components. Such abilities are
being abundantly revealed in single cells and single- cell organisms. Even
bacteria, it turns out, are “dynamic predictors actively oriented toward
what comes next.”^6 Th is follows experiments by Ilias Tagkopoulos and
colleagues, showing how bacteria can adapt to changing environments by
learning statistical associations between variables. Th e bacterial biochem-
ical networks create internal models of the complex environment.^7

This content downloaded from 139.184.14.159 on Tue, 17 Oct 2017 13:52:50 UTC