POTENTIAL BETWEEN BRAINS 241upon the dynamics of feed- back loops between interacting nestmates and
hence shape the collective response of the whole ant society.”^4
Such emergent systems of positive and negative feedback loops has
been studied in a number of colony be hav iors, including nest building,
vacating, and cleaning, as well as food foraging. As described by Detrain
and Deneubourg, “While excavating, the insect adds pheromone to the
cavity walls and the laid pheromone, in turn, stimulates other nestmates
to dig at that site. As the nest volume increases, the density of insects and
so the frequency of their visits to the digging sites decrease what ulti-
mately leads to a self- regulation of the excavated nest volume.”^5 Similar
self- organized dynamics have been described when dealing with dead
bodies, forming nest clusters, and defending against predators.
Individual ants are not entirely if- then machines; they can exhibit
some complexity of cognition and of individual learning, even with a
brain of only about a quarter of a million neurons. For example, at least
some ants do a “learning walk” to integrate positional coordinates of the
nest site. When foraging and exploring, they also integrate a variety of
cues, such as sun position, polarized light patterns, visual patterns, odor
gradients, wind direction, and step- counting, to navigate and compute
a “bee- line” home. By tuning their responses to the emergent patterns of
interactions in the group, they are doing more than following built-in
rules. And they are obtaining greater adaptability than they could
achieve alone.
Th us the cognitive systems, even of ant brains, are already co ali tions
of attractors (i.e., attractor landscapes), as explained in chapter 7. In ant
colonies the individual ants are now forming co ali tions at a higher level.
As well as the co ali tions of attractors in brains, there are “co ali tions of
co ali tions” across brains. Higher- level attractors emerge among brains,
just as they do among individual neurons within them. Th ese are new
statistical abstractions that go beyond the information given. Th ey do not,
as it were, hang in cyberspace. Th ey are dispersed among individual
brains— through learning—to regulate individual be hav iors.
Th e point is that this new level of attractors, based on deeper statisti-
cal patterns in the here and now, provides far greater adaptability than
built-in rules or individual learning, possibly could. And that applies to
the group as a whole, as well as to its individual members.
This content downloaded from 139.184.14.159 on Tue, 17 Oct 2017 13:54:53 UTC