Book Review
“We will not understand life and living organisms until we understand emergence.”
His basic definition of emergence:  “much coming from little” (Holland, 1998, p. 1).

“Emergence is a ubiquitous feature of the world around us . . . .  There are deep
questions about the human condition that depend on understanding the emergent
properties of living systems:  How do living systems emerge from the laws of physics and
chemistry?  Can we explain consciousness as an emergent property of certain kinds of
physical systems?  We are everywhere confronted with emergence in complex adaptive
systems—ant colonies, networks or neurons, the immune system, the Internet, and the
global economy, to name just a few—where the behavior of the whole is much more
complex than the behavior of the parts . . . .  Emergent phenomena also occur in
domains for which we presently have few accepted rules: ethical systems, the evolution
of the nation-state, and the worldwide spread of ideas are just a few that come to mind”
(Holland, 1998, pp. 2-3).

In science, many who describe themselves as reductionists object to the notion of
emergence—they insist that understanding comes only by breaking down the whole to
each of its finite parts.  However, emergence as we best understand it, “depends on
reduction.  Complicated systems are best described in terms of interactions of simpler
systems.  I emphasize interactions because there is a common misperception about
reduction: to understand the whole, you analyze a process into atomic parts, and then
study these parts in isolation.  Such analysis works when the whole can be treated as the
sum of its parts, but it does not work when the parts interact in less simple ways.  When
the parts interact in less simple ways (as when ants in a colony encounter each other),
knowing the behaviors of the isolated parts leaves us a long way from understand the
whole of the colony.  Reduction does not work in such a situation—we have to study the
interactions of the whole as well as the parts.  Emergence, as we use the term here,
occurs only when the activities of the parts do not simply sum to give coherent activity to
the whole.  Emergence means the whole is indeed more than the sum of its parts”
(Holland, 1998, p. 14).

Holland has a long treatise on game theory to demonstrate the emergence of long term
strategy.  “Our simplifying assumption to this point has been that opponents employ fixed
strategies, but this simplification sidesteps most of what happens when games are played
repeatedly.  Opponents learn.  A more realistic view is that all players simultaneously are
trying to build models of what other players are doing.  Under this extension, the game
situation becomes much more complicated.  An observer who has an omniscient
overview of the situation, even when that observer knows the initial strategies and the
details of the individual learning procedures, it is next to impossible to predict the course
of the game.  Emergence and perpetual novelty are ever present in situations where the
opponents are perpetually adapting to each other” (Holland, 1998, p. 42).

Holland next cites chaos theory, where prediction is virtually impossible even after
extended observation of early states.  Meteorology provides the contextual reference,
where small changes in local conditions (e.g. the butterfly flapping its wings in Argentina)
has an effect on global, long-term conditions even in a “well-behaved” weather system.  
Holland’s point is that even when we control for the most minute detail and all known
relevant mechanisms, novelty is perpetual.  In the complex unfolding sequence, the
pattern of the weather in any future state is an emergent property (Holland, 1998, pp.

Holland next turns to the human brain and its partner the central nervous system.  “Like
the ant colony, the central nervous system (CNS) is composed of numerous interacting
individual components, called neurons.  Individual neurons, like individual ants, have a
behavioral repertoire that can be reasonably approximated with the help of a small
number of rules.  And, like the ant colony, the behaviors mediated by the CNS are much
more complex, in both time and space, than the behaviors of constituent neurons.  Still,
the central mystery is much the same in both cases:  how does a persistent, flexible
organization emerge from relatively inflexible components?  The simile to the ant colony
is helpful because it demonstrates the plausibility that a neural network can acquire a
repertoire of behaviors that far exceeds the individual capacity of the individual
neurons.”  . . . With every input layer, each neuron responds to some small element of
the environment (the scene or waveform being presented by the eye for recognition, for
example), which is a function of feedback.

This performance is impressive, because the person does have to position objects
carefully (centering them and putting them into a standard orientation).  But to
accomplish pattern recognition over time, the neural network is set up in multiple layers:  
an input layer, several intermediate layers, and an output layer.  The result is a
feedforward network, in which neurons in one layer stimulate neurons in the next layer
to fire.  Moreover, the cycling process within the neural network creates an indefinite
memory, that is a much more advanced phenomenon than a simple feedforward
network, a phenomenon that cannot be achieved by the feedforward process alone.  The
result is “three observable emergent phenomena:  synchrony (groups of neurons entrain
themselves into synchronous firing), anticipation (groups of neurons prepare to respond
to an expected future stimulus), and hierarchy ( new groups of neurons form to respond
to groups already formed).”  The result: in combination we have the basis for speaking
about human agency: the ability of humans to take mental action.  “This is an emergent
phenomenon” not explainable through reductionistic processes:  “Emergence is, above
all, a product of coupled, context-dependent interactions that we call mind.  Technically,
these interactions, and the resulting system, are nonlinear.  The behavior of the overall
system cannot be obtained by summing the behaviors of its constituent parts” (Holland,
1998, p. 122).

In conclusion, Holland says “the notion of “getting more out than you put in goes against
intuition in the sciences.  Nevertheless, there is a real sense in which this occurs in
systems that exhibit properties of emergence” (Holland, 1998, p. 245).  “It may be that
the parts of the universe that we can understand in a scientific sense—the parts of the
universe that we can describe via laws, axioms, equations—constitute but a small fraction
of the whole.  If that is so, then there may be aspects of emergence that we cannot
understand scientifically” (Holland, 1998, p. 231).  “Because so many of the problems
that face us—ranging from the control of economies to understanding human
consciousness—involve emergent phenomena in a crucial way, one might causally infer
that this fact somehow signals an impassible barrier.  In such a view, we’ve gone as far
as science can take us” (Holland, 1998, p. 233).  “In scientific theory, the rigorous use of
prior models as sources for newer, more encompassing models provides a regular
succession.  Kepler’s insights have been succeeded and surpassed by Newton’s insights,
which in turn have been succeeded and surpassed by those of Einstein, Plank and
Heisenberg, and so it is likely to continue beyond the foreseeable future.  Yet this very
rigor restricts the scientist’s ability to deal with the broad, ill-defined domains that are so
much a part of human experience—domains characterized by words like “beauty,”
“justice,” “purpose,” and “meaning.”  The insights of poetry far surpass those of science
in these domains (Holland, 1998, p. 220).

Reviewed by E. Maynard Moore, Ph.D., August 2010
Holland, John H.  (1998.)  Emergence:  From Chaos to Order.  Reading, MA:
Addison-Wesley Publishing Company, Helix Books. [hardcover, 258 pp. ISBN: