DeLanda, Manuel. Nonorganic Life, in Incorporations. Edited by JonathanCrary and Sanford Kwinter.

New
$E,ANDA-ANUELh.ONORGANIC,IFEvIN)NCORPORATIONS%DITEDBY*ONATHAN#RARYAND3ANFORD+WINTER.EW
York: Zone Publishers, 1992.
9ORK:ONE0UBLISHERS

Nonorganic Life
Manuel DeLanda

According to Thomas Kuhns well-known theory, scientific revolutions

are triggered by a paradigm-induced gestalt switch. A traditional
way
of conducting scientific research is replaced by a new one (a new para-
digm is implemented), and scientists come to perceive phenomena that
previously were invisible. Kuhn gives the Copernican revolution as one
example of such a shift. Unlike Chinese astronomers, who had been able
to observe the occurrence of sunspots centuries before Galileo simply
because their cosmological beliefs did not preclude celestial change,
early Western astronomers were unable to see changes in the cosmos.
Sunspots, for example, remained invisible that is, insignificant
-

and anomalous until Copernicuss ideas changed the ways in which

European astronomers could look at the heavens.

The last thirty years have witnessed a similar paradigm shift in scien-
tific research. In particular, a centuries-old devotion to conservative
systems (physical systems that, for all practical purposes, are isolated
from their surroundings) is giving way to the realization that most sys-
tems in nature are subject to flows of matter and
energy that continu-
ously move through them. This apparentlysimple paradigm switch is, in
turn, allowing us to discern phenomena that, a few decades ago, were, if
they were noticed at all, dismissed as anomalies.
For example, when we approach systems as if they were conservative
that is, artificially isolate them (experimentally or analytically) from
ambient fluxes of energy and matter we are led to expect that

these systems will eventually reach a point of steady-state equilibrium.

However, when we acknowledgethat these fluxes necessarily flow
through the system, a new possibilityemerges a dynamic equilibrium.

One of the most striking examples of this is the spontaneous assembly

ofa chemical clock." In a normal chemical reaction, the interact-
ing molecules simply collide randomly, transforming one anotherwhen
the energy generated by their collisions passes a certain threshold.
lfwe imagine the substances involved as, say, red and blue, we

zone :29
 Nonorganic Life

would expect their random interactions to result in a violet mixture.

Indeed, a conservative model would almost certainly preclude any other
expectation.
Reality, however, is full of surprises: under certain conditions, some
chemical reactions behave in a most counterintuitive way. The reactants,
ratherthan reaching a steady-state equilibrium (a violet mixture), sud-

W
denly turn completely red, then blue, and back to red, according to a
perfectly regular rhythm. In order to perform such a feat, the billionsof
interacting molecules must somehow act in concert, since only by coor-
dinating their movements could they produce rhythmic motions with
such precision. According to the old paradigm, this spontaneous coop-
eration among molecules was so unlikely that, until very recently, it was
thought to defy the laws of classical thermodynamics. Indeed, while the
reign of the previous paradigm lasted, such chemical clocks were for all
practical purposes invisible:
[The chemical clock effect was] first reported in 1921 by William Bray, in
the decomposition of hydrogen peroxide into water and oxygen, with an
iodine catalyst. But chemists then believed wrongly thatthe laws of

thermodynamicsforbid oscillations. Instead of following up B1-ays discov-

ery, they concentrated on explaining it away, on the grounds that his exper-
imental method must have been at fault an attitude that set them back

nearly forty years. In 1958 the Russian chemist B. P. Belousov observed

periodic oscillations in the colour ofa [chemical reaction]. Ilya Prigoginc
had by then shown that far from thermodynamicequilibrium, the usual
laws of thermodynamics dont hold, and people were more prepared to
take the results seriously?
In a sense, we could say that the contributions of Prigogine and
others have made visible this and other self-organizing chemical reac-
tions. What were formerly rare phenomena, irritants to be explained
away, began, under the new paradigm, to appear everywhere and in
every form:
Only a few years ago a chemical reaction was supposed to have a decent
reaction order and not much more. But now, such a reaction is of low
standing if it does not include a few of the followingitems among its prop-
erties: bifurcations, bistability,
catastrophe, chaos, dissipative structures,
echo waves, multistabilityoscillations, superchaos, symmetry breaks, trig-
ger waves, etc. Our fascination with such phenomena in chemical systems
is due to the fact that we believe they are the fundamental elements of
dynamics which may integrate to form life.3
The spontaneous emergence of temporal patterns (chemical clocks)
or spatial patterns (chemical waves and spirals) has
many profound con-
sequences for theories of evolution. These phenomena could elucidate
how the right conditions for the emergence of life came about in the
so-called primeval soup. Prigogine, for instance, calls attention to the

130
 special thermodynamicconditions necessary for the emergence of order
out of chaos. At equilibrium, he says, matter is blind, but in these far-
fromequilibriumconditions, as he calls them, it becomes capable of
perceiving weak gravitational and magnetic fields.4
In other words, at the onset of a process of self-organization (when
a chemical clock begins to assemble, for example), the mechanisms
involved become extremely sensitive to minor fluctuations in the envi-
ronment. A small change in external conditions, one that in thermody-
namic equilibriumwould have had negligible consequences caused

perhaps by a relatively weak gravitational or magnetic field is ampli-

fied and directs the kind of chemical clock that is assembled (the period
of its oscillations, for example),thereby naturallyselecting one self-
assembly pattern over another. Thus, because of their extreme sensitivity
to initial conditions, spontaneously emerging chemical patterns in the
primeval soup could have become beneficiariesof the pruning process
of a prebiotic natural selection.
This idea has been developed by Peter Decker, a German researcher
of selforganizing phenomena. Decker calls any chemical system sensi-
tive to small fluctuations during its self-assembly a bioid," which he
defines as an open system capable of generalized Darwinian evolution.
Unlike the chemical reactions used by Prigogine as examples of self-
organization, which have been criticized for the ratherartificial condi-
tions under which they occur,5 Deckers could have actually occurred
in certain reactions that were present in the primeval atmospheric con-
ditions on Earth. The onset of one such reaction (the formal reaction)
and the self-organizing processes to which it might have given rise rep-
resent, according to Decker, the acquisition of the very first bit of evo-
lutionary information on the planet.6
Of course, the kind of information that may be generated by chem-
ical oscillations is rather meager and repetitive when compared with
even the most primitive organisms. Yet, according to other recent find-

ings, the same systems that pulsate regularly under some conditions are,
under other conditions, capable of oscillating in extremely intricate pat-
terns. Because these patterns make the behavior of the chemical system
essentially unpredictable, they have often been described as chaotic.
For our purposes, what matters is that these chaotic oscillations, under
the right conditions, could have generated the complexity we observe in
the living world. Prigogine, for example, invites us to imagine a chaotic
chemical reaction in which, at certain points, some of the chemicals
precipitate or diffuse outside the reaction space. If we also imagine that
these chemicals accumulate on a tape some kind of natural storage

device, like a self-replicating macromolecule the spatial patterns gen-

erated will be as asymmetric and informationrich as those we observe

in organic life.7

zone 131
 Nonorganic Life

Another class of self-organizing phenomena, very different from

those just discussed but also instrumental in reshaping how we under-
stand the richness of expression of the material world, is called solitary
waves. Differing from ordinary waves in that they do not quickly dissi-
pate as they move, they maintain their shape for a relatively long period
of time. When, as happens, a solitary wave maintains its exact identity
after interactingwith other waves, it becomes a soliton.3 As with
chemical oscillations, various forms of this phenomenon have been
observed for a long time over a century and a half but they were

relegated to the realm of anomalies. The first person to study these phe-
nomena was John Scott Russell, a Scottish engineer and ship designer.
In 1834, he witnessed the spontaneous emergence of a soliton in the
surface of a canal near Edinburgh, and after chasing the coherent mass
of water for several miles, he became convinced he had seen something
extremely important. But science was not ready for this discovery, even
though a mathematicalexplanation for it was available as early as 1892
(the Kortwegde Vries equation).9 As the components of a new scien-
tific paradigm began to consolidate in the 1960s, solitons were recog-
nized everywhere. In the ocean, for instance, they are called tsunamis:
Tsunamis are formed when a strong seismic shock occurs in the ocean
floor. The wave, only a few inches or feet high, can travel intact across the
ocean for many thousands of miles. The human problem begins when
. . .

the tsunami reaches the continental shelf. In shallower waters, nonlinear

effects at the sea bed act to shorten the wavelengthof the wave and increase
its height. The result is awesome. From a soliton a few inches or feet high,
the tsunami becomes a 100-foot mountain of water crashing into coasts and
harbors. The tsunami thatkilled thousands in Lisbon in 1775 caused many
writers in the age of Enlightenment to question the existence of a benevo-
lent God. In 1702 a tsunami in japan drowned over 100,000 people, and in
the seismic soliton created by the volcanic explosion of Krakatoa Island in
1882 thousandsdied.
Solitons can occur at much greater scales than tsunamis. In the
atmosphere of jupiter, for instance, there exists a soliton (the famous
Red Eye) with a diameter roughly equal to the distance from the Earth
to the Moon. On the other hand, solitons have also been found at ex-
tremely small scales: for example, this is the way in which electrons
traveling through solid objects form charge-density waves. Indeed,
anything that flows in regular waves (electricity, sound, heat, light) can
give rise to solitons. Not surprisingly, the laws of classical thermody-
namics would seem to forbid such coherent waves from arising spon-
taneously: if a given amount of energy is introduced into a flowing smoke.
medium, the energy should diffuse evenlythroughout the components
of the medium. And indeed, this does normallyhappen; but in special
oxide
circumstances, the energy can form into a pulselike wave and retain its Zinc

132
 coherent identity over time and in some cases, even after collisions
withother solitons.
Chemical clocks (periodic and chaotic) and solitons exist within our
bodies. An important chemical reaction in our own metabolism, which
serves to transform glucose into useful energy (glycolysis),has been
shown to generate spontaneouslyrhythmic oscillations.11 Chaotic oscil-
lations, on the other hand, occur in neural activity associated with the
control of the heartbeat and the secretion of some hormones. (Apparently,
nonperiodic oscillations are more adaptable and flexiblethan rigidly
periodic ones, lending them, perhaps, a functional advantage. 12) Solitons,
for their part, are helpful in understanding how energy is conveyed
throughout the human body: since regular waves dissipate their energy
too rapidly to be of any metabolic use, coherent pulses that maintain
their identity could explain how energy produced in one part of the
body fuels processes in another, distant part. Solitons traveling along
the backboneof certain proteins could be the answer to this puzzle.
Similarly,electrical signals travel through the nervous system too fast to
be accounted for by the traditional understanding of incoherent waves;
thus, solitons have been proposed as the most likely mechanism for
neural signal transmission, and in this regard they have been described
as constituting the elementaryparticles of thought.14 In short, it seems

that our bodies are inhabited as much by the phenomena of nonorganic

life" by the
as more familiar phenomena of organic life.
Chaos and solitons have become the yin and yang of the new physics. 15

At one extreme, there are relatively simple systems (a chemical clock,

a mechanical pendulum), which can generate bewilderinglycomplex

behavior; at the other, there are extremelycomplicated systems (oceanic

and atmospheric motion), which can generate simple and coherent
structures (pulselike solitons, or more generally, fronts, shock waves
and bubbles). When the two are combined by adding friction to a

solitonbearingequation, for instance the result is the formation and

competition of spatial patterns of great complexity and beauty. Indeed,

what the study of chaos and solitons wasto the 1980s, the study of pat-
tern formation may be to the 19905.15 What is important for our pur-
poses, though, is thatboth periodic and nonperiodic oscillations and
coherent structures are among the unexpected possible behaviors of
inert matter, behaviors of which we have onlyrecently become aware.
Matter, it turns out, can express itself in complex and creative ways,
and our awareness of this must be incorporated into any future materi-
alist philosophy.
One may well wonder why it has taken so long for these material
effects to be recognized. Of the many possible explanations, one un-
doubtedly deserves special mention: our mathematicaltechnology
was simply incapable of modeling self-organizing phenomena. The kind

Z0119 '33
 Nonorganic Life

of equation most abundant in science (both classical and quantum), the

linear equation, is extremely useful for modeling physical systems

not so much becausethis mathematicalform captures all the relevant

aspects of behavior, but because it provides a workable approximation,
easily solved, and thus adequate in its predictive ability. Nonlinear equa-
tions, on the other hand, were not so easily solved, at least not before
the advent of digital computers; even if they gave us a more realistic
representation of reality, they were useless for practical applications. As
a result, nonlinearities (like friction and air drag) were eliminated as
much as possible from mathematicalmodels, making nonlinear effects
like chaos and solitons invisible (that is, visible only as anomalies to be
eliminated). According to one mathematician:
So docile are linear equations, that classical mathematicianswere willingto
compromise their physics to get them. [Consequently,] the classical theory
deals with shallow waves, low-amplitude vibrations, small temperature
gra-
dients [that is, eliminates nonlinearities]. So ingrained became the linear
habit that by the 1940s and l950s many scientists and engineers knew lit-
tle else. Linearity is a trap. The behaviour of linear equations is far
. ..
. . .

from typical [in nature]. But if you decide that only linear equations are
worth thinkingabout, selfcensorship sets in... Your textbooks fill with
.

triumphs oflinear analysis, its failures buried so deep that the graves go
unmarked and the existence of the graves goes unremarked.'7
Thus, one of the main causes of the paradigm shift thathas allowed
us to see matter as capable of self-organization is an advance in the
technology that materially supports mathematics,and with it mathe-
matical technology. Needless to say, this will not be an overnight re-
placement, and much of science (classical and quantum) will remain
linear. But nonlinear science has begun to reveal new and startling facts
about matter, in particular, that the behavior of entirely different mate-
rial systems can exhibituniversal features. Physicists, for example,
have known for some time that phase transitions in matter (from liquid
to solid, from magnetic to nonmagnetic, from conductor to
supercon-
ductor) display a common mathematicalstructure despite the diversity
of their actual physical mechanisms. More recently, it was found that
the onset of turbulence in a flowingliquid (a self-organizing process) is
also closely related to such phase transitions, as well as to the onset of
coherence in laser light. In a very important sense, then, all these tran-
sitions may be said to be mechanism independent: one and the same
mathematicalmechanism can account for all these events that would
seem otherwise to be wholly unrelated. 3

Similarly,the onset of rhythmic behavior in a chemical reaction is

identical to many other processes in which equilibriumsystems suddenly
begin to oscillate a welldocumented example is the spontaneous

aggregation of slime mold amoebas, in which one individual amoeba

34
 functions as a clock that regulates the assembly of other separate cells
into one large multicellular organism. Nonlinear oscillations have been
observed in fields as diverse as electronics, economics and ecological
relations (such as predatorprey systems). Soliton phenomena, as we
saw, occur at every scale, from ]upiters Red Eye to tsunamis, to atomic
charge-density waves, and in every kind of energy flow (heat, light,
sound). This startling universality of the mathematicsof self-organiza:
tion is one of the most revolutionary elements of these new theories:
There is no better, there is no more open door by which you can enter
into the study of natural philosophy,Michael Faraday has written, than by
considering the physical phenomena of a candle. Although the wisdom of
this remark lay unrecognized for well over a century, it is now rather widely
understood that the study of nonlinear diffusion in a candle. is closely
. .

related to the dynamics of nerve impulse propagation, the spread of conta-

gious disease, and the beating of the human heart. The same equation
. ..

thatsolid state physicists and electrical engineers use to describe the prop-
agation of magnetic flux quanta (called fluxons) is also employed by
. . .

theoretical physicists as a model for elementary particles. Similarly,the

soliton on an optical fibre .is closely related to suggested mechanisms for
. .

the transport and storage of biological energy in protein. 19

In mathematicalterminology, the events at the onset of self-organi-
zation are called bifurcations.Bifurcations are mutations that occur
at critical points in the balance of power between physical forces

temperature, pressure, speed and so on when new configurations

become energetically possible, and matter spontaneouslyadopts them.

It is as though inert matter, confronted with a problem stated in
terms of a balance of forces, spontaneously generates a machinelike
solution by drawing from a reservoir of abstract mathematicalmech-
anisms. For instance, in the case of chemical self-organization, two dis-
tinct forces are at work: the rate at which the substances diffuse and the
rate at which they react with each other. When the balance of power is
dominated by diffusion, the result is a steadystate equilibrium,but the
moment reaction rates begin to dominate, the chemical substances are
suddenly confronted with a new problem, and they must respond to the
challenge by adopting a configuration that meets the new energetic
requirements. In this case, the new stable solution to the problem is to
enter into an oscillatoryequilibrium, and it is in this way that chemical
clocks are born. But, again, more important than the self-assembly of
these clocks is the factthat essentially the same solution is available to
other systems, like populations of electronic circuits or of economic
agents, which involve balances of power between forces of a completely
different kind.
Gilles Deleuze and Flix Guattari have suggested that this abstract
reservoir of machinelikcsolutions, common to physical systems as

Z008 135
 Nonorganic Life

diverse as clouds, flames, rivers and even the phylogeneticlineages of

living creatures, be called the machinicphylurn30 a term that would

indicate how nonlinear flows of matter and energy spontaneously

gener-
ate machinelikeassemblages when internal or external reach
pressures
a critical level, which only a
very few abstract mechanisms can account
for. In short, there is a single machinic phylum for all the different living
and nonliving phylogeneticlineages.
The machinicphylum remained largely invisible until the advent of
digital computers; or rather, considering how pervasive nonlinear be-
havior is throughout nature (it could hardly have escaped everyones
attention), we had to learn to recognize it. Bifurcating sequences lead-
ing to complex behavior had been observed by mathematicianssuch
as Henri Poincar as early as the 1890s
although those earlyglimpses
into the wild spaces of the machinic phylum horrified most who saw
them. Those passing glimpses have, with the proliferation of computers
in mathematicalinvestigations (giving rise to experimental mathemat-
ics),opened onto vast landscapes

computer screens becoming our

windows onto the machinicphylum in more than a figurative sense.
If computers have emerged as windows onto this world, it is because
the nonlinear mathematicalmodels of bifurcation processes can be given
a visual representation, a phase portrait. The first
step in creating a
phase portrait is to identify the relevant aspects of the behaviorof the
physical system to be modeled. It is impossible, for example, to model
an oven by considering each and
every atom of which it is composed,
but one can consider the single aspect of the oven that matters: its tem-
perature. Similarly,in modeling the behavior of a pendulum, only its
velocity and position are important. In technical terms, the oven has
one degree of freedom, its change in temperature; the pendulum,
in turn, has two degrees of freedom. A bicycle, on the other hand

taking into account the coordinated motion of its different parts (han-
dlebars, front and back wheels, right and left pedals) is a system with

approximately ten degrees of freedom.

The next step is to create an abstract space (called phase space)
that has as many dimensions as the system to be modeled has degrees
of freedom. In this way, everythingthat matters about the
system at
any given moment in time can be condensed into a single point: a point
in one-dimensional space (a line) for the oven, a point in two-dimen-
sional space (a plane) for the pendulum or a point in ten-dimensional
space for the bicycle. Moreover, as the system changes in time (the oven
heats up or the bicycle makes a turn), this point in phase space will also
change, thus describing a trajectory. This trajectory, in essence, will
contain all the information that matters about the history of the mod- crystal.
eled system. For example, if the system under study tends to oscillate
between two extremes, like a driven pendulum, its trajectory in phase Liquid
136
 space will form a closed loop, which represents a system whose move-
ment consists of a repetitive cycle. A free pendulum, on the other
hand, which eventually comes to a standstill, appears in a phase portrait
as a spiral. More complex systems willbe represented by more complex

trajectories in phase space.

Now, it is one thingto model a system with a set of equations, but
quite another to solve those equations. Sometimes, when the equations
modeling a system are so complex that they cannot be solved that is,

when they are nonlinear scientists can nevertheless learn something

about the systems behavior by looking at its phase portrait. They cannot
make precise quantitative predictions about the system, but they can
use the phase portrait to elicit qualitative insights about the general
traits governing the systems long-term tendencies. In particular, there
are certain special spots in phase space thattend to attract or repel all

nearby trajectories; that is, regardless of where a trajectory begins, it

will tend to drift toward certain points (called attractors),or to move
away from certain others (called repellors).
Because these trajectories represent the behavior of real physical sys-
tems, the attractors and repellors in a phase portrait represent the long-
term tendencies of a system. For instance, a ball rolling downhill will
always seek the lowest point. If it is pushed up a little, it will roll down
to this lowest point again. Its phase portrait will contain a point attrac-
tor: small fluctuations (the ball beingpushed up a little) will move the
trajectory (representing the ball) away from the attractor, but then the
trajectory will naturally return to it. Other systems will have two point
attractors or more. A wellmade light switch, for example, whose resting
states are on and off, has two attractors and if it is delicately set
between them, chances are good that it will spontaneously snap into
either position. Its phase portrait cant tell us when it will do so that

is, provide a precise quantitative prediction, which implies that the

equation has been solved but it can describe the systems long-term

tendency to snap into one of its points of equilibrium. In fact, the spe-
cific form of the nonlinear equation modeling the system will reveal this
before its specific trajectory has been calculated.
Attractors, though, do not necessarilyappear as points. For example,
an attractor with the shape of a closed loop (called a periodic attrac-

tor or limit cycle) will force all trajectories passing nearby to wrap
around it, to enter into an oscillating state, like a pendulum. If the phase
portrait of a physical system has a periodic attractor embedded in it, we
know that no matter how we manipulate the behavior of the system, it
will tend toward an oscillationbetween two extremes.
These two attractors, points and closed loops, were the only types
known before computer screens opened windows onto phase space.
What followed was the discovery of a much wilder array of creatures

ZOFIB 137
 Nonorganic Life

inhabitingphase space: among them were attractors with strangely

tangled shapes, called strange or chaotic attractors, representing
turbulent behavior in nature, and the incredibly complex, chaotic
boundaries separating basins of attraction (a simple attractors sphere
of influence in phase space). Far more startling, though, was the discov-
ery that attractor
an spontaneously mutate into a different attractor.
can

These spontaneous transformations are known as bifurcations. For

instance, a physical system that originally tended toward a steady-state
equilibrium (a point attractor) can suddenly begin to oscillate between
two extremes (a limit cycle); this would describe the self-assembly of a
chemical clock. Spontaneous selforganization can take other forms as
well: the onset of turbulence in a flowing liquid appears in phase space
as a cascade of bifurcations thattakes a circle (limit cycle) and, through

successive doublings, transforms it into a strange attractor. Roughly, we

could say that phenomena of self-organization occur whenever a bifur-
cation takes place: when a new attractor appears on the phase portrait
of a system, or when the systems attractors mutate in kind.
We have, then, three distinct entities inhabitingphase space: spe-
cific trajectories (corresponding to systems in the actual world), attrac-
tors (corresponding to the long-term tendencies of these systems) and
bifurcation events (corresponding to the emergence in these systems of
new structural tendencies). Bifurcation events are brought about by

changes in certain control parameters, which represent the more or

less constant conditions affecting physical objects; temperature, pres-
sure, gravity and so on. As the value of a parameter shifts through cer-
tain ranges, the attractors of the physical system represented will usually
change subtly; at certain critical points, though, a bifurcation will take
place, and the attractors will transform themselves.
As I said earlier, because essentially the same attractors and bifurca-
tions are available for many different physical systems, they may be seen
as virtual or abstract mechanisms that are incarnated in different con-

crete physical mechanisms. This is not to say that attractors and bifur-
cations exist in some platonic realm waiting to be realized rather,

they are intrinsicfeatures cfthe dlvnamics qfpb}/sical systems, and they have
no independent existence outside cjltbose physical systems. Yet, attractors

and bifurcationsdo constitute an abstract reservoir of resources avail-

able to nonlinear flows of matter and energy a condition that applies

as much to the beating of hearts as to earthquakes, flames and clouds,

tsunamis and amoebas. And it is in this respect that I introduce the

term machinicphylum to designate a single phylogeneticline cutting
through all matter, living or nonliving, a single source of sponta-
neous order for all of reality. More specifically, the attractors define
the more or less stable and permanent features of this reality (its long-
term tendencies), and bifurcations constitute its source of creativity

138
 and variability.Or to put it more philosophically,attractors are veritable
figures of destiny, for they define the future of many systems.
For instance, when the dynamics of a physical system are governed
by a periodic attractor, it is as if the very flows of matter and energy
rushing through the system were bindingit, or destining it, to an oscil-
latory future. And yet, for several reasons, this iron-clad determinism
should not be given too mechanistic an interpretation. For one thing,
the phase portraits of most systems usually contain more than one at-
tractor, which means that the system in question has a choice between
several destinies. Then, there is the freedom built in to chaotic attrac-
tors. When a systems dynamics are caught in a strange attractor (deter-
ministic chaos), that system is bound to be creative, that is, to explore
all the possibilitiesof a small region of phase space. Furthermore, even
if we are destined to follow the attractors guiding our dynamical behav-
ior, there are also bifurcations, critical points at which we may be able
to change our destiny (that is, modify our long-term tendencies). And
becauseminuscule fluctuations in the environment in which a bifurca-
tion occurs may decide the exact nature of the resulting attractors, one
can hardly conclude thatall actions we undertake
as individuals or

collectively are irrelevant in the face of these deterministic forces.

Bifurcations may not be a guarantee of freedom, but theycertainly

do provide a means of experimenting with and perhaps even modify-

ing our destinies:

From the physicists point of view this involves a distinction between states
of the system in which all individual initiative is doomed to insignificance
on one hand, and on the other, bifurcation regions in which an individual,

an idea, or a new behaviour can upset the global state. Even in those regions,

amplification obviously does not occur with just any individual, idea, or

behaviour, but only with those that are dangerous that is, those that
-

can exploit to their advantage the nonlinear relations guaranteeing the sta-
bilityof the preceding regime. Thus we are led to conclude thatthe same
nonlinearities may produce an order out of the chaos of elementary pro-
cesses and still, under different circumstances, be responsible for the de-

struction of the same order, eventually producing a new coherence beyond

another bifurcation.
Although attractors and bifurcations appear only in the phase por-
traits of systems that are dissipative (those in which energy is not
conserved), even conservative systems show a partitioning of phase
space into regions, some of which define systemic long-term tenden-
cies. Such a phase space may be seen as an energy landscape, with

peaks valleys that correspond to maxima and minima of free energy.

and
These valleys serve as attractors in the sense thatthese systems will
spontaneously adopt a physicalconfiguration that minimizes the amount
of free energy in other words, their trajectories will tend toward

zone 139
 Nonorganic Life

low-energy valleys. Despite the important distinction between dissipative

and conservative systems, in both cases a feature of their phase portraits
constitutes a figure of destiny and for that reason, I shall simply

refer to these as attractors.14

A similarlyimportant distinction exists between equilibriumand
nonequilibriumphase transitions. Although the former illustrated by

the transition from gaseous to liquid or to solid states of matter is a

bifurcationthat involves the generation of a new energy landscape in

phase space, the latter case the transition from Calm (laminar)to

turbulent flow in fluids involves the creation of a new set of attrac-

tors. Again, though, in both cases a critical point in the balance of power
between physical forces brings about the alteration of the systems long-
term tendencies and so I shall refer to both as bifurcations. A more
complete theory of the machinic phylumwould have to address these.
issues in detail, but this rough picture of the mathematicsof self-organi-
zation will suffice for our purposes here: defining the concept of nonor-
ganic life and the theoriesthat connect it to our own organic bodies.

Let us consider a few hypothetical scenarios of how reality as we

know it could emerge from the interplay of attractors and bifurcations
inother words, how more or less permanent physical structures
(from organisms to entire planets) could be created by the machinic
phylum,by the mere nonlinear flow of matter and energy. Lets begin
with an impossibly simple case, that of a crystal planet made entirely
of a single element. We can imagine this planet as it exists at the mo-
ment of its birth, that is, as it precipitates from a protostar, a ball of
superhot plasma (which at a very high temperature is an electrically
charged cloud of subatomic particles). As the plasma cools, it increas-
ingly acquires form: atomic nuclei appear, then the naked nuclei
acquire electrons, forming atoms, which then form molecules (precip-
itating a gas), which condenses into a liquid from which, eventually,

there crystallizes a solid, our planet.

The plasma, at its highest temperature, may be regarded as formless
there are not even atomic nuclei, only flows of subatomic matter and
energy. The possible bifurcations, on the other hand, the critical tem-
peratures at which new forms arise, are already there, in the sense
that for any given element these points are very precisely defined. As
the plasma crosses these temperature thresholds, these bifurcations are
actualized.The actualizationof a bifurcation, as noted, involves the
creation of new attractors (or of new energy landscapes), which define
the stable configurations that the components of the plasma may adopt.
These attractors, in turn, must be actualized for those configurations to
become real. Thus, the naked atomic nuclei of the cooling plasma pre-

14.0
 cipitate a whole set of new attractors, which specify the virtual orbits
that electrons may occupy. The passing electrons that become trapped
in them are, in a sense, actualizationsof these virtual orbits. When all
of these orbits are filled that is, when all the attractors have been

incarnated the result is a stable form: an atom. Thus, form emerges

out of formless matter, order emerges out of chaos, through the incar-
nation of bifurcations (critical temperatures) and the actualizationof
the resulting attractors.
Although this hypotheticalcrystal planet would hardly look like a
real planet, even at this stage, the simple cooling process in the plasma
could give rise to complex patterns. Control parameters, other than the
one I have described (temperature) as driving the bifurcations, would

complicate the emerging forms; pressure and volume are two obvious
factors. Because these would vary over the surface of the planet, they
would disrupt the formation of a homogeneous crystal, producing
instead regions of gas, liquid and solid in different parts of the planet
(and perhaps even areas where all three phases coexist). Moreover, the
possible paths to solidificationwould depend on which element we
chose: some elements (for example, silicon) have a choice of solidify-
ing either as crystal (quartz) or as a glass (opal, an amorphous solid).
The result that is actualized, the path taken, will depend on how fast
the bifurcation is crossed in any given area: if the solidification is rapid
wave. (a process called quenching),the result will be a glass, but if it is slow
(annealing),it will be more or less perfect crystal.
sound
a
a

Obviously,a more realistic scenario would involve a planet made of

more than a single element. This would allow a vast increase in complex-
by
ity, even if we assume only a limited numberof elements. (On Earth,
agit ed only ninety-two elements occur naturally; others, like plutonium, are
too unstable to occur spontaneously,although they can be created

plate under laboratory conditions.) With this addition, new phase transi-
tions are created, like those at the onset of chemical reactivity. In other
on words, the different elements will begin to interact and form new sub-
setling stances. Furthermore, the gas, liquid and solid phases of the materials
may enter into Complex mixtures called colloids examples of

sand which are a solid and a gas (an aerosol), a solid and a liquid (gel), or
any other, more complex combination.The colloidal state of matter
fine is indeed very common.
of
Clearly, if we admit enough such additions to our original scenario,
Grains the result will be a more realistic planet. Any extra details we add will
simply increase the numberof possible bifurcations available to matter
figure.
as, for instance, the novel phase transitions of the colloid state (like

Chladni
the bizarre mutations from solid to liquid we observe in quicksand a

bifurcation actualized through a small shock). Yet the planet thatwould

result from such a process would simply be a collage of states of matter

zone 141
 Nonorganic Life

interactingwith one another. A real planet, on the other hand, has moun-
tains, valleys and other geological formations. Were we to peer inside
the mountains, for example, we would find folded layers of different
types of stone. In short, we would observe a historical structure not

a collage of states of matter.

On Earth, the main historical process responsible for the formation

of mountains and other structures, is called the geological cycle.
Briefly put: Erosion and weatheringcreate the raw materials for the
cycle (stones, pebbles, grains), which rivers, through a process known
as hydraulic sorting, sort out and deposit as sediment at the bottom

of the ocean. The sediment already has a structure, since the sorting
process causes matter to deposit in distinct layers. When these layers
are buried under further deposits, they undergo a transformation: the

pebbles and grains that form these layers are cemented together into
sedimentary rock (for example, sandstone or limestone). When these
rock layers are folded under the pressure of movements of the Earths
crust, mountains emerge which are then sculpted by erosion, and so

on ad infinitum.3

Selforganizing processes drive the geological cycle. The tectonic

forces behind the burial and folding of sediment are driven by molten
rock flowing up from beneaththe crust in convection cells, a coherent
flow arising after a temperature bifurcation, constituting a kind of
self-assembled conveyor belt. The rivers that sort out the pebbles and
grains that make up the layers of sediment, can be seen as self-organized
hydraulic computers. The dynamics of a rivers flow at different
points in its course are governed by different attractors, turbulent flow
from strange attractors, and coherent pulses (solitons) from weaker
forms of nonlinear stabilization.It is these different regimes of flow
that give the river its capabilityof sorting raw materials by grain size,
shape and even composition. Finally, the processes of erosion and
weatheringalso act as filters or sieves, sorting out materials by their
degree of stability. Mountains are worked over by these processes,
removing those components that are not fully stable, and leaving behind
those that are strongly locked into their attractors. (For this reason,
granite, a highly stable crystalline configuration, is typically found on
high ground. 33)
Thus, adding these historical processes changes our simple picture
of the birth of a planet. Instead of a simple sequence of bifurcations
taking us from a plasma to crystal, here the formation of hardened
structures involves attractors and bifurcations engaged in more com-
plex interactions. The self-organized machinic assemblages they create
(conveyor belts, hydraulic computers) provide the labor needed to gen-
erate geological constructions in a two-step process: first, the raw mate-
rials are selected and sorted (sedimentation),then they are consolidated

:42
 into permanent structures (cementation or hardening). Because the
end result of these two operations is geological strata, this process of
double articulation has been given the name of stratif1cation.33
According to Deleuze and Guattari, processes of stratification occur
not only in theworld of geology but in every sphere of reality. In other
words, any sphere of reality the hydrosphere, the biosphere and so

on
can be defined in terms of flows of matter and energy and the
reservoirs driving those flows. At any given point in time, portions of
these flows will be involved in any numberof activelyself-organizing
processes; other portions of the flows, however, will have sedimented
or hardened into more or less stable structures. Thus, we can describe

a given region of the planet at a given moment by specifying which of

the three possible states of these flows predominates: freely self-orga-
nizing, loosely bound or rigidly bound. But becausethese states are
neither irreversible nor exclusive, we can speak of various components
in terms of the degrees of stratification they exhibit.
For instance, a rock may seem to us the archetypal example of per-
manence and stability,but when one takes the long view, even rocks flow:
their atoms migrate along grain borders (self-diffusion), dislocation
boundaries within grains move, cracks and fissures propagate. In this
sense, the flow of rocks is very viscous; they constantly change, but at
extremely slow speeds. Furthermore, under extreme heat and pressure,
rocks may undergo a bifurcation (limestone, a sedimentary rock, meta-
morphoses into marble).3"' Alternatively,rock may be melted into lava
and reincorporated into the convection flows driving plate tectonics.
Stratification,then, is in no way a terminal state: free matter and energy
stratify, and the stratified destratifies.
Moreover, hydrological strata (rain belts, ice caps, rivers), atmos-
pheric strata (pressure systems, wind patterns) and organic strata (food
chains, dominance hierarchies) can also be understood as complexes of
matter and energy flows that are stratified to different degrees.
What we have,then, is a kind ofwisdom of the rocks, a way oflis
tening creative, expressive flow of matter for guidance on how to
to a
workwith our own organic strata.

Organic stratification, which gives rise to our own bodies (among other
things),unfolds through chemical reactivity that is, the onset of a
-

reaction transforming one set of substances into another. This, too,

can be considered a phase transition becausethese reactions involve
activationthresholds, the minimum amounts of energy necessary for
chemical processes to begin. The control parameters determining these
processes, however, are by no means fixed: catalysts substances capa-

blc of manipulating these activation thresholdswithoutthemselves

ZONE 143
 Nonorganic Life

being affected raise the thresholds or lower them, and in this way
can

can either inhibit or enhance a reaction.

In a very literal sense, catalysts are the first form of matter that can
affect directly the control parameters driving bifurcations, either pre-
venting or acceleratingtheir actualization. (More precisely, catalysts
push systems away from their attractors and toward the border of their
basin of attraction, where small fluctuations can push them into the
domain of a different attraction.35) In some cases, a substance can cat-
alyze its own production (autocatalysis);in others, two or more sub-
stances can cooperate with one another by catalyzing each others
reactions (crosscatalysis). Chemical clocks arise only in these kinds of
reactions. Among all the elements, the most powerful catalysts are met-
als, and for this reason metals have been said to beara privileged status
in the machinic phylum:
[W]hat metal and metallurgy bring to light is a life proper to matter, a vital
state of matter as such, a material vitalism that doubtless exists everywhere
but is ordinarilyhidden or covered, rendered unrecognizable. [M]etal
. ..

is coextensive withthewhole of matter, and thewhole of matter to metal-

lurgy. Even the waters, the grasses and varieties of wood, the animals are
populated by salts or mineral elements. Not everythingis metal, but metal
is everywhere. Metal is the conductor of all matter. The machinic phylum
is metallurgical, or at least has a metallic head, as its itinerant probe-head
or guidance device.37
This special capabilityof catalysts to intervene in the dynamics of
other processes is the necessary precondition for life to begin. Indeed,
the catalysts involved in living processes, called enzymes, are far more
specific than metals in their dynamic effects, allowing for much more
detailed control of chemical reactions. The instructions for building
each enzyme are stored in DNA, and this allows particular enzyme
designs to be not only inheritable, but also capable of being fine-
tuned for specific functions through the action of natural selection.
Because the translation of sequences of DNA into enzymes is itself an
activity regulated by enzymes, we seem to have here a classic chicken-
or-egg dilemma. _

Recent theories of how a machine as complicated as DNA and its

translation operations came to be assembled begin with networksof
cooperating nucleic acids and enzymes. One suggested precursor, called patern.
a hypercycle, might have been a selfreplicating molecule (nucleic acid)
coded to produce an enzyme that enhanced the production of a second
nucleic acid, which, in turn, coded for an enzyme thatenhanced the ocilspe
production of the first a dynamical cross-catalyticsystem governed

by attractors and subject to bifurcations.

Such a cooperating network might have been the mechanism
through which the machinic phylumgave rise to organic life. Organic
Cathode-ry
I44
 strata, however, differ from geological strata in that they have enveloped
withintheir sphere of operations the control parameters driving bifur-
cations. In the case of our hypotheticalcrystal planet, the forces deter-
mining the control parameters (temperature, pressure, volume) had not
yet been integrated into, nor harnessed by, the mountains and rocks
they gave rise to. The latter simply emerged as the flows of matter and
energy underwent bifurcations, but these structures had little or no
effect on which bifurcationsoccurred. With organic strata, on the other
hand, the stratified structures in particular, DNA also drive the

control parameters, and are thereby able to determine which bifurca-

tions are actualized.
For our purposes, we can treat living systems as made of two kinds
of components: those that function more or less independently of the
rate and flow of matter and energy, and those that depend critically on
these rates of flow (that is, on degrees of viscosity or stratification). The
bifurcationsof the latter occur only at critical points in the rate of flow
of matter and energy, whereas the instructions stored as genetic infor-
mation for the assembly of enzymes are completely stratified and rate
independent. In this sense, DNA and its enzymes can be seen as a com-
plex parallel computer regulating the relative viscosities of different
flows so as to permit the actualizationof certain bifurcationsbut not
others. This is, in essence, the hypothesisproposed by philosopher
Howard Patee, who calls informationbased structures like DNA sym-
bol systems:
[S]ymbol systems exist as rate-independent (nonintegrable) constraints. . ..

[They do not] depend, within wide limits, on the rate of reading or writing,
or on the rate of energy or matter flow in the symbol-manipulatinghard-

ware. On the other hand, the effect. of symbols functioning as instruc-

. .

tions is exerted through the selective control of rates. For example, the rate
of reading or translating a gene does not affect the determination of which
protein is produced. However, the synthesis of the protein as instructed by
the gene is accomplished through the selective control, by enzymes, of the
rates of individual reactions.
These separate roles played by DNA and nonlinear, self-organizing
processes can perhaps be best illustrated by looking at the way in which
the genetic information contained in a fertilized egg is slowly converted
into a fully developed individual of a given species. This is the process
known as embryogenesis. Roughly speaking, during embryogenesis
two kinds of processes occur: those mediated by DNA, thus subject to
genetic control, and those governed by attractors and bifurcations, thus
constrained but not created by genetic information.
An egg consists of a nucleus (containing genetic information) and
cytoplasm.Traditionally,the latter was regarded simply as a source of
energy and nutrients for the developing embryo, but it is now known

ZODB 145
 Nonorganic Life

to play a much more fundamental role. As a rule of thumb, We should

look for manifestations of nonorganic life in those processes in the devel-
oping embryo where the nucleus is not involved as a source qf1'ry"orma-
tional constraints. This qualification is necessary because the nucleus may
be involved in the emergence of bifurcations, though not by virtue of its
genetic material. For example, its position in the egg can create an ini-
tial polarity between different zones of the egg. This initial polarity,
established during the formation of the egg (oogenesis), has been found
to constitute a fundamental source of asymmetry guiding the develop-
ment of the first few stages in embryogenesis. Even when the genetic
activity of the nucleus has been inhibited, an egg will undergo certain
early bifurcations due to its global dynamics:
Self-assembly during [the early stages of ] embryonic development is not
mediated by direct gene intervention. When all the transcriptions have
been prevented [through the use of an inhibitor] the regular cleavage pat-
terns are retained. However, the polarity of molecular organization of both
the eggs cytoplasm and its nucleus (chromatic) are essential for normal
development. Hence the main features of [early] embryogenesis cell

differentiation, induction, determination of pattern formation all stem

from the oogenetically originated, spatial distribution of preformed infor-

mational macromolecules. The initial condition of embryogenesis is oogen-
esis. The epigenetics of embryonic development is built on the topological
self-organization and orientation of macromoleculesof the total egg.4
In later stages of embryo development, the emergence of the pat-
terns forming various organs and structures
an eye, a leg and so on

similarlyinvolves events directly mediated by DNA and those resulting

from the global dynamics of interacting cells. Historically,the study
of embryogenesis has shifted in the importance attributed to each of
these factors. A century ago, there were two main theories of embryo-
genesis: one positing that organs are somehow preformed in the egg,
and another positing that organs arise out of the dynamics of develop-
ment. The first approach held that, as an egg cell divides, each cell
would contain half of the total organism and so on throughout the
course of embryogenesis. This has, of course, been proven false, for
the destiny of a cell is not rigidly predetermined but, rather, is regulated
during development:
The classical demonstration of [the phenomenon of regulation] was pro-
vided in the 1890s in H. Drieschs experiments on seaurchin embryos.
When one of the cells of a very young embryo at the two-celled stage was
killed, the remaining cell gave rise not to halfa seaurchin, but to a small
but complete sea-urchin. Similarly,small but complete organisms devel-
oped after the destruction of any one, two or three cells of embryos at the
four-celled stage. Conversely, the fusion of two young seaurchin embryos
resulted in the development of one giant sea-urchin.

I46
 Several hypotheses have been proposed to explain the regulation of
cell differentiation, the process through which a group of essentially
similarcells for example, an eg after the first few divisions gives

rise to incredibly diverse tissues. Following the discovery of DNA, it was

thought thatdifferentiationwas entirely genetic that as cells differenti-

ated, their DNA changed as well, guiding the development of various sets
of proteins and enzymes for various cell types. It was found, however,
that DNA remains essentially the same in cells of totally different types.
What, then, accounts for cell differentiation? One general hypothe
sis holds that during embryogenesis certain chemical patterns emerge
and guide cells to produce specific proteins at specific times. There are,
in fact, several theories about exactly how chemical patterns accomplish
this task, which differ mainly in the importance they attribute to genetic
information in the guidance of these processes. On the one hand, the
positional information hypothesis asserts that simple gradients of
chemical concentrations are interpreted by differentiating cells according
to a very complex genetic code. On the other, the nu ,.aa:terii hypothe-
sis asserts that chemical patterns are very complex, and tlirf genetic code
with which the cells interpret them is very simple. Clearly, there is a
trade-off here: the greater the complexity one attributes to a self-orga-
nizing chemical pattern, the less the genetic information one need pos-
tulate for its interpretation, and vice versa.
For a long time, the latter hypothesis was favored because a mecha-
nism for spontaneously forming simple chemical patterns seemed much
more likely than one for complex patterns. But, as I suggested at the be-

ginning of this essay, this situation is changing rapidly: chemical reactions

are now viewed as capable of undergoing several kinds of bifurcations

that result in the self-assembling Chemical clocks, traveling waves, spirals

and other complex patterns previouslythought impossible. So the hy-
pothesis of a prepattern is becomingaccepted more widely, and with it an
increased role for nonorganic life in the creation of organic structures.
Regulation and regeneration were what first prompted scientists
to postulate forces in embryogenesis other than genetic information.
At particular stages in the development of an embryo, groups of cells
interact and thereby develop along specific paths. These paths, though,
do not seem to be wholly determined by genetic control: if a group of
cells in a developingembryo is transplanted from one region to another,
they will develop, not into the structure emerging in the first area, but
into the structure corresponding to the second. These stable paths of
development that guide or canalize cell behaviorhave been called
chreods.44 Confronted with such self-organizing phenomena, one
scientist has felt the need to postulate the existence of physical forces
unknown to ordinary physics, which are supposed to act in a nonener-
getic way, independently of space and time. I see no need, however,

zone 14-7
 Nonorganic Life

for postulating mysterious agents to account for chreods: the attractors

guiding the global dynamics of interacting cells suffice to account for
their canalized behavior, and the bifurcations arising in those dynamical
interactions adequately account for the emergence of the attractors
themselves. More specifically, the enzymes produced by the expression
of genetic information enter into chemical reaction networkswhose
dynamics are guided by attractors. Each type of cell (bone, muscle,
nerve) is now thought to arise from a different attractor, with the DNA
simply pushing cells from one basin of attraction to another.This
description has the advantage of explaining the spontaneous emergence
of new patterns by invokingonly intrinsic features of the global dynam-
ics of cell populations, without postulating additional forces with mys-
terious properties.
Roughly speaking, then, the beliefthat all embryological events in-
volve direct genetic supervision is giving way to a new vision: the egg
begins as an undifferentiated field that undergoes a series of bifurca-
tions, each of which produces a new set of attractors thatform the
chreods guiding development until the next bifurcation. One scheme
for classifying attractors and bifurcations, known as catastrophe the-
ory, presents a taxonomy of seven elementary morphogenetic (thatis,
form-producing) events. To each of these sets of attractors and bifurca-
tions corresponds a different process of selforganization in cell groups:
in the right conditions, these will tend spontaneously to form a pleat or
a fold, a pocket, a spike or a furrow."
The more self-organizing processes scientists discover, the smaller
the domain of DNAs control appears to be. Rather than including a
complete blueprint for the construction of an entire individual, DNA
must include only enough constraints to harness self-organizing pro-
cesses in the creation of the structures that characterize a particular

species. Mathematicalanalysis has established (for now) thatthe attrac-

tors and bifurcations guiding these dynamics are relatively few in num-
ber, so organic processes of stratification are themselves constrained by
dynamical forces. In other words, morphogenesis for example, the

spontaneous emergence of form in a relatively formless egg is limited

by the available attractors and bifurcations. I Ience, DNA constraints can

only operate within these limits. In the words of one biologist:
There are several consequences ofthis view of morphogenesis. First, it is
evident that morphology is generated in a hierarchical manner, from simple
to Complex, as bifurcations result in spatially ordered asymmetries and

periodicities, and non-linearitiesgive rise to fine local detail. Since there is

a limited set of simple broken symmetries and patterns that are possible virus.
(e.g., radial, bilateral,periodic), and since developing organisms must start of
off laying down these elements of spatial order, it follows that these basic
forms will be most common among all species. On the other hand, the Crystal
148
 finer details of pattern will be most variable between species, since the
pattern-generating process results in a combinatorialrichness of terminal
detail, and specific gene products in different species stabilizetrajectories
leading to one or anotherof these. The fact that virtually all the basic
. ..

organismic body plans were discovered and established during an early

evolutionary period, the Cambrian, is often remarked with surprise,
but it is just what one would expect on the basis of the above argument.
Ancient and recent morphological characters are secondary conse-
quences of the hierarchical nature of morphogenesis and the exploration
of its potential in time.
The idea of an undifferentiated egg slowlydeveloping into a full
organism as it crosses bifurcations corresponds to our earlier image of
a formless plasma giving rise to a crystal planet as it cools down and

undergoes phase transitions. But as we saw before, we must also bring

into play the action of historical processes of stratification in order to
complete the picture. Like volcanoes in the lithosphere or hurricanes
in the atmosphere, creatures in the biosphere must meet strict matter
and energy budgets. The flow of flesh (biomass) through food chains
constitutes the main form of energy circulation in organic strata that
-

is, energy budgets are met by all living creatures by eating and avoiding
being eaten. In all ecosystems, planets are at the bottom of these food
chains, constituting a reservoir of solar energy stored chemically
(throughphotosynthesis). On top of this layer of primary productivity
come several layers of consumers (herbivores, carnivores). The specific
job that an animal or plant performs in an ecosystem (what is called its
niche) is defined by its position in one of these circuits for the circu-
lation of biomass. So in this sense, organic strata are, like any other
strata, composed of temporary coagulations of matter (the bodies of
plants and animals), themselves the product of the ceaseless flow of
energy through ecosystems. And yet, animals and plants are unique
in that they must also meet a second budget, a genetic budget that com-
pels them to try to spread their genes as far and wide as possible. For
these creatures, meeting their energymatter budget is, in a sense, of
secondary importance. They must eat and avoid being eaten only to
endure until the mating season, when the main objective of their lives is
revealed: the continuation of a genetic line, the preservation of the por-
tion of the gene pool theyencapsulate. The gene pool of a species may
be seen as a veritable reservoir driving flows of genetic material through
the bodies of its individual members. More specifically, a gene pool sup-
plies the raw materials for the pruning process of natural selection. It
contains the stored experience of the species over many generations,
the knowledge of how to survive successfully in a given environment.
Individual animals and plants are like temporary experiments with
which gene pools probe current environmental conditions to make sure

ZOTIB 149
 Nonorganic Life

that past successes are still viable (in other words, that energymatter
budgets can still be met as bcfore).50
Although gene pools are designed to replicate themselves very accu-
rately (and thereby preserve past experience intact) random copy-error
(mutations) and sexual shuffling of gene groups (recombination)gener-
ate enough variation so thatthese genetic reservoirs can respond to new
environmental challenges. In short, thanks to variation, gene pools can
evolve and be submitted to historical processes of stratification. Roughly
speaking, mutation and recombinationplay the role of erosion and
weathering that is, they provide the raw materials for the sorting pro-

cess of natural selection. In the simplest case, for example in the case
of selection by climatic conditions, this process merely sorts out the fit
from the unfit, or more generally, the stable from the unstable. And yet,
like the hydraulic computers mentioned above, selection pressures are
also patterned by nonlinearities. For instance, some pressures may add
some directionality to evolution, as in the case of natural arms races:

a thickeningof the armor in a prey species directly provokes a sharpen-

ing of the claws and teeth in its predatory counterpart, which in turn
puts pressure on armor designs to get even thicker. Similar catalytic
loops occur between parasites and hosts, or between male decorations
and female sexual choice. In other cases, when the fitness of a parti-
cular trait or behavior is nonlinear (that is, when selection pressures
depend on how frequent that behavior is exhibited in a population),
behavioral patterns come to be stabilizedby attractors (the socalled
evolutionary stable strategies)
Thus, selection pressures of many kinds play the same role that pro-
cesses of sedimentation perform in the case of rocks. They select the
stable from the unstable, and then sort out whats left into the layers of
a food chain. The analogy does not end there. Like patterns in the sedi-

mentary deposits that form at the bottom of the ocean, the accumu-
lated patterns of adaptive traits and behaviors brought about by natural
selection are very ephemeral. They slowly sediment over many genera-
tions but they can be wiped out by a single large-scale bifurcation, like
-
the onset of the ice age. 50 what then corresponds here to the process
of cementing togetherloosely accumulated pebbles into hardened sedi-
mentary rock? The answer, according to the discipline of macroevolu
tionary dynamics, is the process of speciation, that is, the birth of a
new species. When a portion of a population becomes reproductively
isolated from its parent group, the information contained in its gene
pool becomes permanently injected into the larger phylogeneticlineage
to which both groups belong. Speciation acts like a ratchet, preventing
accumulated adaptations from being eroded away. In this form, what
was a loosely bonded set of anatomical and behavioral traits is now hard-
ened into the more or less permanent structure of a particular species.
 Thus, the simple images of developing eggs and cooling plasmas,
thoughgood illustrationsof the action of self-organizing processes, must
always be complemented by a description of the actual historical pro-
cesses of stratification thathave determined how attractors and bifurca-
tions have been actualized. These images are also misleadingin another
sense. On the one hand, eggs and plasmas are good metaphors for the

machinic phylum, which I defined as a destratified, nonlinear flow of

matter-energy. In both cases, matter and energy originally
exist in a more
or less formless state in which actual forms and structures exist only
virtually as the various bifurcations that eggs and plasmasare capable
of
undergoing. On the other hand, these images could seduce us into think-
ing that the phylum something
machinic is that exists only initially,
and
graduallydisappears
that it as concrete forms and structures emerge.

example
As the extreme of limb regeneration indicates, though,
the
phylum does not vanish or become irrelevant in the aftermathof the
systems early development. On the contrary, as with the nonorganic
systems described above, the phylum remains immanent throughout
the organisms development: it can be found whenever a process of self-
organization occurs whenever a phase transition takes place or an
-

internal chemical reaction undergoes a bifurcation. Indeed, the dynamics

of the complex networks of chemical reactions thatmake up our own
metabolic processes are guided by bifurcations and attractors that guide
the system between stable, or homeostatic, states:
Homeostasis refers to the relative constancy of the internal environment
with respect to variables such as blood sugar, blood gases, electrolytes,
osmolarity,blood pressure, and pH. The physiological concept of home-
ostasis can be associated with the notion of steady states in mathematics.
Steady states refer to a constant solution of a mathematicalequation [a
point attractor]. Although the mean blood pressure is maintained rela-
tively constant, as we all know, the contractions of the heart are approxi-
mately periodic. Likewise...withthe rhythms of heartbeat, respiration,
reproduction, and the normal sleepawake cycle. Less obvious, but of equal
physiological importance, are oscillations in numerous other systems for

example, release of insulin and luteinizing hormone, peristaltic waves in the

intestine and ureters, electrical activity of the cortex and autonomic ner-
vous system, and constrictions in peripheral blood vessels and the pupil.
Physiological oscillations are associated with periodic solutions of mathe-
matical equations [periodic attractors].54
Perhaps the best example of these processes is the internal physiological
cycles involvingwaking, sleeping, temperature and so on, the so-called
circadian rhythms. Circadian, which literally means about a day,
refers to the fact thatthis internal rhythm has a period of roughly
twenty-five hours. Consequently, as our internal clock adapts to the
external daynight cycle, it undergoes a bifurcation every day. Because

Z0118
 Nonorganic Life

our circadian rhythms are dynamical processes they are guided by

periodic attractors they are capable of synchronizingor entraining

to an external rhythm. Although entrainment to the lightdarkCycle

of the regular Earth day, for example, can by desynchronized (as in jet
lag), the same dynamic flexibilitythatallows self-organizing clocks to
adjust to the external environment also allows them to undergo other
kinds of bifurcations. Thus, a flash of light at a critical moment (of
bifurcation) can send ones clock into a totally different period or stop
it altogether. A whole range of pathological phenomena has been asso-
ciated with the occurrence of bifurcations and has received the name
dynamical diseases.55
The phenomenon of entrainment the spontaneous phase synchro-

nization of different oscillating entities is common in nature. One

well-known example is the slime mold amoeba: in normal Circum-

stances, a group of these amoebas will behave as unrelated individuals,
but when the level of environmental nutrients reaches a critically low
value, they assemble themselves into a coherent colony with differen-
tiated organs. Entrainment also occurs in laser light, which differs
from ordinary light in thatthe photons oscillate in phase (that it,
they are synchronized),resulting in the emission of a coherent beam.
Entrainment takes place in so many physical systems that it is, in a
sense, a mechanismindependentprocess:
Populations of crickets entrain each other to chirp coherently. Populations
of fireflies come to coherence in flashing. Yeast cells display coherence in
glycolyticoscillation. Populations of insects show coherence in their cycles
of eclosion (emergence from the pupal to the adult form).. Populations
..

of women living together may show phase entrainment of their ovulation

cycles. Populations of secretory cells, such as the pituitary, pancreas, and
other organs, release their hormones in coherent pulses.
In phase space, entrainment appears as a torus-shaped attractor,
which the trajectories of different systems (the oscillating entities) come
to wrap around. The transition from nonsynchronized to synchronized

breakdown.
oscillations can be understood as a bifurcation in which a set of separate
limit cycles transform themselves into a single attractor. Entrainment is
common in living systems because periodic attractors are their main
form of organization. Indeed, one crucial difference between geological
and organic strata is the formers tendency to develop around points of
of static equilibrium (minima of energy, point attractors), as opposed
to the latters tendency to make use of forms of dynamic equilibrium
(periodic and even chaotic attractors). Were one to track the flow of
proces
the
matter around the planet, one would see how it becomes stratified

Dielctr
in
along these lines. For example, although the key element of life, carbon,
is for the most part locked in rocks, some of it notably in the form of

carbon dioxide moves freely through the biosphere; some of this, in

152
 turn, is trapped cyclicallyby plants (throughphotosynthesis),captured
in theirbiomassand released when their leaves fall in autumn.

There is a sense, then, in which we are all inhabited by processes of

nonorganic life. We carry in our bodies a multiplicity of self-organizing
processes ofa definite physical and mathematicalnature a set of

bifurcations and attractors thatcould be determined empirically,at

least in principle. Yet, is there any way to experience this nonorganic life
traversing us (for example, through the use of meditation techniques
or psychedelic chemicals to destratify ourselves)? As noted above,
there is a wisdom of the rocks from which we can derive an ethics
involving the notion that, ultimately,we too are flows of matter and
energy (sunlight, oxygen, water, protein and so on). At any moment in
these flows, we can distinguish some portions that are more viscous
(hardened, stratified) than others. An ethicsof everyday life, in these
terms, would involve finding the relative viscosities of our flows, and
giving some fluidity to hardened habits and making some fleeting ideas
more viscous in short, finding,through experimentation, the right
consistency for our flows (the right mixture of rigid structures, sup-
ple structures and selforganizing processes).
Clearly,though, it is impossible to derive ethical lessons from the
machinic phylumwithout considering processes of selforganization
and stratification at the level of society. Animal and human societies,
like individual organic bodies, are also crossed by the machinic phy-
lum. For example, a particular population of animals is regulated by a
simple equation concerning its birth and death rates and the reservoir
of its environmental resources. This simple equation has been shown to
undergo bifurcations: a population that for many years tended toward
a steadystate equilibrium may suddenly begin to oscillate between

two extremes. Similarly,the respective biomassesof two populations

(one prey, the other predator) are also linked by equations susceptible
Phalanx of soldiers. to bifurcation.
Elaborating on the observation that phase transitions and other bifur-
cations occur at the level of animal populations, Arthur Iberall has pro-
posed a model of human society in terms that emphasizeself-organizing
processes. In his view, societies function as an ensemble of flows and the
reservoirs driving those flows: water, metabolic energy, bonding pres-
sures, action modes, population, trade, technology and so on. In doing
so, Iberall is not out to replace standard accounts of human development
but to stress the role of flows and phase transitions in determining
social field stability. He goes on to say:
I view the discontinuous social change manifested by the appearance of
food-producing societies (e. g., from huntinggatheringto horticulture to

Z0118 I53
 Nonorganic Life

settled agriculture) as evidence of internal rearrangements, new associations

and configurations, and a new phase condensation as if a gaslike phase

of matter were becomingliquidlike or solid statclike. At his beginning,

. ..

modern man apparently lived in huntinggatheringgroups operating in a

range appropriate to human size and metabolism. If, as is appropriate
. ..

to his size, man had the typical mammalian metabolism and roaming range
of about 25 miles/day,cultures separated on the order of 50 miles would
have little interaction. The 70- to 100-mile separation of populations, as
. ..

empirically found, is highly suggestive of a system of weak force, gaslike

interactions. [D]ecreases in the levels of the required potentials (temper-
. ..

ature, water, food) cause condensation [liquification] of small bands on

fixed centers of population. The nature of the social phase Condensation,
. ..

however, depends on the amplifying capabilityof the technological poten-

tial. Associated withthose two chief potentials water supply and tech-

nology (tools) came changes in modes of living, improvement in the use

of water resources, and localized social development through the domesti-

cation of plants and animals.. .59 .

Eventually, these fluidlikesocial formations crystallized into strati-

fied civilizations. Based on the archaeological record, Ibcrall concludes,
civilizationsbegan when there was extensive trade (convective flow) among
population concentrations (condensations). The urban centers held cumu-
lative populations greater than 2500 and were composite groups. The thresh-
old size can be estimated from the absence of complex cultures of smaller
population.
This picture of social evolution corresponds in many ways to the
creation scenario of our hypothetical crystal planet. Clearly though, if a
simple liquid solution can harden into crystal or glass, ice or snowflake,
depending on the multiplicity of nonlinearities shaping the solidifica-
tion process, human societies which have a wider range of attractor

types have far more leeway in how they develop stable configurations.

And while there is much to be learned from analyzing in detail the

actual processes of stratificationand destratificationthathave occurred
in different societies at different times, even a picture as simple as this
already points to certain principles of a geological ethics. Notably,
society cannot be understood as climbinga ladder of progress, as
though huntergatherers, agriculturalists and civilizedcommunities
were stages on a path toward everincreasing perfection. Rather, under-

standing these transformations as phase transitions would imply that a

State apparatus is not essentially better than a primitive society, since
after all, there is nothingintrinsically better about a solid than a liquid.
From the viewpoint of a geological ethics, early societies may even
have achieved a better consistency among their flows, a viscosity more
in tune with their ecosystems than our own. Indeed, it seems that some
early societies may have sensed the approach of social solidificationand

154
 developed mechanisms to prevent actualizingsuch a bifurcation. This,
of course, is not to say that we should return to some lost paradise of a
savage state, a bygone era of greater innocence and harmony. Rather,
we must work on the society in which we find ourselves, tracking the
flows of matter and energy, destratifying hardened institutions, setting
into flux human practices that have sedimented in short, we must

find the right viscosity for our fluxes, the exact consistency that would
allow humanity to self-organize withoutthe need for coercion and war.
Iberall pictures the flow of goods along trade routes as an example
of convective flow, as thoughprecapitalist markets were indeed self-
organizing structures creating order out of chaos. We may have to dis-
cover in the history and present of human practices those institutions
that realize best the workings of the machinic phylum, or, more pre-
cisely, the degrees of stratificationof each of its components: those
flowing more or less freely, those sedimented into more or less supple
structures and those rigidified into permanent institutions (entrenched
political hierarchies, inflexiblebeliefsystems). The geological strata
teach us that even the seemingly most rigid structures can flow (how-
ever slowly), mutate (metamorphic rocks) or even be reincorporated

into self-organizing processes (convection flows of lava).

Part of such an effort would involve creating maps showing the
attractors that govern the dynamical behaviorof social flows, and more
important, the bifurcation regions where to paraphrase Prigogine

a dangerous idea can amplify itself by taking advantage of the nonlin-

earities that guarantee the stabilityofa given social system. The work of
Deleuze and Guattari is exemplary for the creation of such maps: they
show how our lives may be viewed as a composite of rigid structures
(family,school, military service, office, marriage), supple structures
(temporary alliances, transitory love affairs, loosely knit groups) and,
finally, lines of flight, the bifurcations that could allow us to change
our destinies as defined by those two types of structures. In every area
of human reality (art, politics, love), they attempt to measure the
degrees of stratification of the flows of matter and energy at work in
these domains. For example, after showing how music originates from
the expressive powers of matter itself (in particular, the self-organizing
processes of animal territories that give rise to bird songs), they argue
Marching formation that music has a greater capacity to destratify than does painting. It is
at Nazi rally. as though the modulated flows of air that we experience as music have a

greater capacity to set our emotions and thoughts into fluxthan do the
more viscous, spatial flows of form and color found in painting.

It follows, then, that the kind of stratometers that could perform

these measurements of relative viscosity in various material and energy
flows need not all be mathematicallybased. A great novel can, for a given
society, capture the coexistence and interactions of rigid structures and

ZOHB 155
 Nonorganic Life

sedimented habits, as well as the active and creative forces driving a

society to undergo bifurcations. That there is plenty of room for intu-
ition and experimentation in trackingthe machinic phylum in no way
obviates the need for, or possibilitiesof, stratometers of a more mathe-
matical kind. Some might be computer programs Capable of Creating
complexphase portraits of interacting dynamical systems, and of
detecting their various elements those operating linearlyin equilib-

rium conditions (rigid structures), those that are nonlinear in near

equilibriumconditions (supple structures) and, finally,those operating
nonlinearlyfar from equilibrium. (Only the latter allows the flows of
matter and energy to express themselves.)
In order to visualize better the uses to which we could put these
stratometers, consider as examples two kinds of human institution
operating at different degrees of stratification:the marketplaceand the
State. As noted, cities and towns may be seen as composed of a series of
reservoirs (water, protein, labor and so on) driving a variety of flows and,
in the process, giving rise to the institutions that channel, amplify or
control these flows. On the one hand, the more or less fluid structures
of markets arise from the spontaneous interaction of many agents, with-
out any central organ directing the process of assembly. On the other,
the more rigid hierarchical structure of State institutions comes to life
not spontaneously but through the goal-directed activities of an elite.
Mainstream economic science, however, fails to capture the fluid
nature of market dynamics: the operation of markets is treated as
though it resembled the growth of perfect crystalline structures, that is,
as though it were governed by a tendency to move toward a single point
of equilibrium (where supply meets demand) a point of optimal effi-

ciency, from the viewpoint of society as a whole (full employment equi-

librium). In order to obtain this wellbehaved dynamics from their
models, though, economists must rely on two unwieldy assumptions:
first, that there exists perfect competition among all economic agents
(thatis, no monopolies or oligopolies manipulating prices); and second,
that these agents have perfect informationabout market conditions as
well as an unbounded rationality to act on that information, so thateach
is able to negotiate the deals that maximize the agent's benefit (utility).54
But when these unrealistic assumptions are relaxed, the dynamics of
the market begin to resemble in many respects those of self-organizing
structures in liquids, from the controlled flow of a convective cell to the
wild patterns of turbulence or even solitons. If, alternatively, the rela-
tively long-lasting and stable structures of some markets are under-
stood as suggesting that a solid metaphor were appropriate, market
behaviorwould resemble, among all solids, that of glass which is
crystal.
Liquid

guided by multiple equilibria more than thatof a crystal with a single

equilibrium. The more amorphous structure of glass results precisely

156
 from the conflictingconstraints imposed by many attractors (local
energy minima) and from the possibilityof being trapped by an
attractor that prevents the optimal use of energy. Markets function like
glass in that, in the presence of imperfect conditions (imperfect deci-
sion making), they can be trapped in suboptimalequilibria like a

supplydemandequilibriumwith high unemployment as has been

well known since Keyness work in the 19305.55

As Joseph Schumpeter showed around the same time, moreover, the
equilibriumon which a marketsettles need not be a stationary state:
more likely than not it will be a cycle instead. The upswing of a cycle

will bring prosperity, and the downswing a period of recession and high
Cloverleaf intersection. unemployment. This behaviorhas been accounted for by models that
generate these nonlinear dynamics through a periodic attractor:
Kondratieffcycles, for example, in which prices and interest rates fol-
low a fifty-twoyearlong-wave motion which has been operating for

at least two centuries have been explained as resulting from positive

feedbackloops forming in specific areas of the economy (specifically,
in the capital goods sector, which produces the machinery needed to
generate consumer goods). By amplifying small fluctuations in the rele-
vant sector, these positive feedback loops drive the economy away from
equilibriumand trigger a Hopf bifurcation, transforming a steady-state
equilibriuminto a stable oscillatory motion.
Modeling markets as self-organizing structures has nothingto do
with the famous invisible hand that, according to some economists,
guides a market toward optimal levels of efficiency when left to its own
devices. On the contrary, the presence of nonlinear effects throughout
the economy from multiplieracceleratoreffects in investment and

finance to the multiplicity of delays, bottlenecks, surpluses and shortages

stemming from the limited rationality of economic agents means that

real markets must be able to cope with life far from equilibrium, and
indeed, to create special buffering structures (inventories, retail stores,
banks) for this purpose. In short, to the extent that markets emerge
and operate spontaneously,they are incapable of achievingoptimal
equilibria on their own. Other kinds of human organizations, on the
other hand the State, sedentary armies, large corporations are

more capable of goal-directed optimization; these latter can be treated

as crystallizations in the flows of matter and energy.

And yet, not even these are perfect crystalline structures, for just as
the dynamics permitting actual formation of crystals is complicated by
the Earths real conditions by gravity, notably so are the dynamics

of human institutions. If one wants to grow pure crystals, one must

do so (for now) aboard a space shuttle, where the single point of equi-
librium guiding their formation is artificially isolated from other influ-
ences. Should one attempt to do so on Earth, crystals require special

Z0118 157
 Nonorganic Life

nonequilibriumstructures, dislocations, which trap energy locally to

allow a crystal to overcome limitations to its growth. Iberall has com-
pared the role of dislocations to thatof human elites in the development
of State institutions: these elites operate, for example, by monopolizing a
surplus of grain thatallows the State to extend its regular structure over
the more amorphous agriculturalvillages around it.53 And these organi-
zations, precisely because of theirhierarchical nature permitting the con-
trolled application of reason and power, are more capable than pure
markets of reaching more or less optimal solutions to social problems.
If hierarchies have this advantage over markets, though,they pre-
sent other dangers of their own. Robert Crosby has described these in
terms appropriate to this analysis: the goals of the elite often tend to Great Wall of China.
diverge from those of lower hierarchical levels, and when this occurs
ever-greater amounts of energy must be diverted to the task of making
the lower levels comply with official goals rules multiply, enforce-

ment units proliferate, punitive means increase.59 According to Crosby,

we may have to search for the right mixture of fluid market structures
and more stratified hierarchical organizations: a hybrid system combin-
ing the resilienceand adaptabilityof markets withthe goal-seeking be-
havior of hierarchies.7 Ideally, such a hybrid would be able to avoid the
dangers of markets (like the formation of monopolies due to economies
of scale, or the disregard of external effects and costs, like pollution),
as well as the need to exercise thateternal temptation of command
structures: too much control.
Is it possible to design stratometers capable of performing the neces-
sary measurements to give this hybrid the right consistency and viscosity?
In a sense, yes. For example, among the Considerations involved in such
a design could be thatof creating a new conception of control. Crosby

suggests that if we could manage to map the attractors and bifurcations

governing the dynamics of a given human Community, we Could then
articulate a new philosophyof control in a few policy rules: if a local
sector of a hybrid system were trapped in a nonoptimal attractor, then
moving this sector toward a more efficient attractor might be prohibi-
tively costly, so the best policy would be to exercise no control at all; if,
on the other hand, this sector were near a bifurcation (or near the bor-
der of a basin of attraction),then the proper role of control would be to
guide it toward the most favorable outcome (the most favorable available
attractor).7 Such a philosophy,of course, would imply accepting modest
gains and accepting suboptimality as a fact of life; but in return, it could
allow society to trackthe machinic phylum, to follow the more natural
paths of development, those allowing for a greater consistency or ade-
quacy in our relationships with each other and the environment.
There are, in fact, many proposals as to how to utilize knowledge of
selforganizing processes for the purpose of formulating policies, or more

158
 generally, strategies of life. Ecologist C. S. I-Iolling, for example, distin-
guishes betweenthe notions of stability(defined as being locked into
an attractor) and resiliency (defined as existing at the border of a basin
of attraction). In his view, ecosystems (whether natural or humanmade)
are not so much stable as resilient. In other words, the abilityto main-
tain a steady or cyclic state is not considered by him to be as important
as the abilityto switch rapidly between stable states (attractors),thus

allowing more responsiveness to environmental changes. He claims that

if mechanisms existed for maintaining a position at the border of sev-
eral attractors, that is, for achievingresiliency,very little effort would be
needed to choose betweenalternative stable states. If Hol1ings hypothe-
sis is Correct, it means that government policies that attempt to drive,
say, the economy toward stabilitymay be approaching the problem in
the wrong way and that economic policies should be designed to achieve
resiliency instead.
A similarbut more radical approach is suggested by the work of sci-
entists like Stuart Kauffman and Christopher Langton. These authors
recommend that systems should be located not at the border of a basin
of attraction, but rather in the vicinity of certain bifurcations. Through
careful exploration of the dynamics of a wide variety of systems, they
have found a magic region of phase space where self-organizing effects
are maximized, that is, where spontaneous self-organization reaches a

peak in complexity. This region is reached, they say, by avoiding attrac-

tors. In their view, point and limit-cycle attractors are too rigid and
predictable to generate anythingnovel, while chaotic attractors are too
wild to produce anythingviable. At the frontier between the two (at a
bifurcationfrom periodic to chaotic, for instance), magic happens, and
so Kauffman and Langton claim thatnatural evolution may have in fact
selected for mechanisms designed to keep organisms poised at the edge
of chaos.73
Another way of thinkingabout such poised systems is in terms of
phase transitions in matter. The solid state would represent the existence
of too much order, and the gaseous state that of too much chaos. It is
well known thatthe liquid state has more interesting dynamics than the
other two states since it can, for example, give rise to vortices and soli-
tons (and other coherent structures). The state at the edge of chaos
exists in the vicinity of the solidliquid bifurcation. It is a state that is
not too liquid, but not too solid either, that is, a state with the right vis-
cosity or consistency. Matter in this magic state, according to Langton,
acquires spontaneous computational abilities,almost as if complex
computational devices could self-assemble at the frontier with chaos.
If this is indeed true, it will have serious consequences for economic
policy. In market economies, prices serve as transmitters of information
(and incentive), and for this reason economic systems can be seen as

Z0118 159
 Nonorganic Life

complex computers, the sum of the decision-making abilitiesof many

micro-economic agents (consumers and producers). Mainstream eco-
nomics regards these agents as perfectly rational, so the computational
abilitiesof the economy as a whole are seen simply as a reflectionof this
underlying rationality. But if nonlinear economics is right, if agents are
essentially limited in their rationality as well as in their access to informa-
tion, then the economy as a whole may have to acquire its power of com-
putation dynamically,that is, by operating poised at the edge of chaos.
Because the problems raised by hybrid systems involve both dealing
with flows of matter and energy and studying the stratificationsthat
form in those flows, we might do well to learn from the experiences of
those who were the first to trackthe machinicphylum artisans and

metallurgists. Before the advent of modern methods, they did this by

using their instincts and the empirical knowhow accumulated through
the ages. They had to track everything from stratified ore deposits to a
metals melting and crystallization points, and then to experiment with
different ways of crossing those bifurcations throughforging techniques;
they had to look for their own hybrids, that is, synergistic combinations
of metals in which the whole spontaneously becomes more than the sum
of its parts (alloys), and to allow the materials to have their own say in
the final form produced. All of this involved following a given materials
local accidents and imperfections, ratherthan imposing a rigid, pre-
planned form on it. .

According to metallurgist Cyril Stanley Smith, this know-how had

been developed well before the Greeks began to apply formal reasoning
to these problems, and it was therefore mostly of a sensual nature.
These artisans, in a sense, developed a special abilityto follow the phy-
lum, to trackthe machiniceffects created by nonlinear phenomena in
nature, and it is this abilitythat we must acquire again if we are to help
social hybrid structures evolve. (As sensual knowledge, this know-how
would constitute yet another stratometer, one built into our own bodies.)
Artisans and metallurgists, though, are not the only source of insight
for such a project: the Earth itself has been dealing with flows of matter
and energy for millennia, and herein lies the wisdom of the rocks from
which we might derive our inspiration. In geological, hydrological and
organic strata (as well as in the two other strata thatbind us, intention-
ality and language), we can always find some elementary components
that are not only less stratified, less bound to their attractor, but that
are also capable of destratifying their own and other systems compo-

nents by pushing them away from their attractors, toward the border of
csryastlatl.
Polyhedra
theirbasin of attraction, where small fluctuations can then tip the sys-
tem into acompletely new regime. In the lithosphere, for example, this
role is played by metallic catalysts: by interactingwith various other ele-
ments and therebyallowing them to transform each other chemically,

16o
 they enable inert matter to explore the space of possible chemical com-
binations, in a nonconscious search for new machinelikesolutions to
problems of matter and energy flow. It is as though catalysts were, to use
Deleuze and Guattaris term, the Earth's own probe heads, its own
builtin device for exploration; and indeed, to the extent that autocat-
alytic loops and hypercycles were part of the machinery involved in the
discovery of life, these probe heads allowed physicochemical strata to
transform themselves and their milieus into completely new worlds.
We ourselves must become this kind of probe head in our own strata,
and allow society to explore and experiment with the possible machinic
solutions to its own problems. We must create stratometers of every
kind mathematicaland nonmathematical and get to work mapping

the attractors thatdefine our local destinies and the bifurcationsthat

could allow us to modify these destinies. And though this is undoubt-
edly an enterprise fraught with dangers, we can derive some comfort
from the hints of the machinicphylum thathave recently becomevisible
to us and seem to indicate there may be ways of evading our currently
doomed environmental destiny.

NOTES
1. Thomas Kuhn, The Structure of$cienzific Revolutions (Chicago: University of
Chicago Press, 1970), p. 116.
2. Ian Stewart, Does God Play Dice? The Mathematics(J Chaos (Oxford: Basil
Blackwell, 1989), p. 186.
3. Hans Degn, Lars Olsen and John Perram, Bistability,Oscillationand Chaos
in an Enzyme Reaction, in Okan Gurel and Otto Rossler, eds., Bifurcation Theory
and Applications in Scientific Disciplines (New York: New York Academy of Sciences,
1979), p. 623.
4. Ilya Prigogine and Isabelle Stengers, Order out of Chaos (New York: Bantam,
1934), p. 14.
5. See, for example, Philip W Anderson and Daniel L. Stein, Broken
Symmetry, Emergent Properties, Dissipative Structures, Life: Are They Related?
in Eugene Yates, ed., SefOrganizin(g Systems (New York: Plenum, 1987), p. 454.
6. Peter Decker, Spatial, Chiral and Temporal Self-Organizationthrough
Bifurcation in Bioids: Open Systems Capable of Generalized Darwinian Evolution,
in Gurel and Rossler, eds., Bifurcation Theory, p. 236.
7. Gregoire Nicolis and Ilya Prigogine, Exploring Complexity (New York:
Freeman, 1989), p. 187.
8. David Campbell, Nonlinear Science: From Paradigms to Practicalities,in
Necia Grant Cooper, ed., From Cardinals to Chaos (Cambridge, Eng: Cambridge
University Press, 1989), p. 225.
9. john Briggs and F. David Peat, Turbulent Mirror (New York: Harper and Row,
I989), p. 121.
10. Ibid., p. 123.

Z0116 161
 Nonorganic Life

11. Prigogine and Stengers, Order out ofChaos, p. 155.

12. Ary L. Goldberger, David R. Rigney and Bruce]. West, Chaos and Fractals
in HumanPhysiology,Scientyric American 262.2 (Feb. 1990), p. 42.
13. Alwyn C. Scott, Solitons in Biological Molecules," in David Pines, ed.,
Emerging Syntheses in Science, Santa Fe Institute Studies in the Sciences of Complexity
(Reading, Mass.: Addison-Wesley, 1988).
14. Briggs and Peat, Turbulent Mirror, p. 123.
15. Alwyn C. Scott, Introduction, in Peter Christiansen and R. D. Parmentiert,
eds., Structure, Coherence and Chaos in Dynamical Systems (Manchester, Eng:
Manchester University Press, 1989), p. 2.
16. David Campbell, Introduction to Nonlinear Dynamics, in Daniel Stein,
ed., Lectures in the Sciences r_;fCompIexity, Santa Fe Institute Studies in the Sciences
of Complexity (Reading, Mass: Addison-Wesley, 1989), p. 90.
17. Stewart, Does God Play Dice? p. 83.
,

18. Hermann Haken, Synergetics and Bifurcation Theory, in Gurel and

Rossler, eds., Bjurcation Theory, p. 365.
19. Scott, Introduction, in Christiansen and Parmentiert, eds., Structure,
Coherence and Chaos, p. 1.
20. We always get back to this definition: the machinicphylum is materiality,
natural or artificial, and both simultaneously; it is matter in movement, in flux,
in variation, matter as conveyor of singularities and traits of expression (Gilles
Deleuze and Flix Guattari, A Thousand Plateaus: Capitalism and Schizophrenia,
trans. Brian Massumi [Minneapolis: University of Minnesota Press, 1987], p. 409).

Here, the term singularities" refers to phase transitions and other bifurcations;
traits of expression are the emergent properties that arise in material systems
when their global behavior is guided by attractors; and emergentproperties are
those synergistic properties that distinguish a whole as being more than the sum of
its parts (properties that cannot be deduced from the components properties, but
only from their interactions).
21. Many may well have seen nonorganic life long before the advent of com-
puters: Deleuze, for example, credits several philosophers (from the ancient Stoics
to Spinoza, Nietzsche and Bergson) with having tracked the machinic phylum in
various ways.
Ralph Abraham divides the historical study of bifurcations into three periods,
according to the instrument of study used: The period of direct observation may
be much older than we think, but let us say that it begins with the musician Chladni,
contemporary of Beethoven, who observed bifurcationsof thin plate vibrations. . ..

Analogous phenomena discovered in fluids by Faraday are still actively studied.

These experiments, so valuable becausethe medium is real, suffer from inflexibility

especially in choosing initial conditions. The next wave of bifurcation experi-

. ..

ments, which I shall call the analog period, begins with the triode oscillator. The
pioneering work of van der Pol (in the l920s) produced a flexible analog com-
. . .

puter, and institutionalizedthe subharmonic bifurcations. The development of

. ..

the early computing machinesushered in the digital period. Well-known numerical

162
 methods were implemented from the start, and graphical output began to appear in
the literature. The pioneering papers of Lorenz, and Stein and Ulam, are still stud-
ied (Dynasim: Exploratory Research in Bifurcations Using Interactive Computer
Graphics, in Gurel and Rossler, eds., Bifurcation Theory, p. 676).
22. Useful introductions to phase space can be found in: Stewart, Does God Play
Dice?, ch. 5; James Gleick, Chaos: Making a New Science (New York: Viking, 1987),
pp. 49-52; and Ralph Abrahamand Christopher Shaw, Dynamics: The Geometgz of
Behavior, The Visual MathematicsLibrary (Santa Cruz, Cal.: Aerial Press, 1985),
vol. 1.
23. Prigogine and Stengers, Order out of Chaos, p. 206.
24. Sec Campbell, Introduction to Nonlinear Dynamics, in Stein, ed., Lectures
in the Sciences gfComplexity, p. 26.
25. A more detailed account of phase transitions from plasma to crystal can be
found in Hans Gutbrod and Horst Stocker, The Nuclear Equation of the State,
Scientific American 265.5 (Nov. 1991), p. 61.
26. Although the transition from plasma to gas is not in fact sharp in the same
way that those from gas to liquid or from liquid to solid are, this does not substan-
tially affect this example.
27. According to quantum physics, the virtual orbits around naked nuclei are
not attractors but minima of free energy that is, the equations of quantum physics
are linear, and hence do not give rise to self-organization as nonlinear equations

do. Many of the pioneers of quantum physics, though Enrico Fermi, Werner

Heisenberg, Louis de Broglie, John von Neumann envisioned a future in which

their discipline would become nonlinear. Various methodshave since been proposed
thatwould permit elementary particles to be described in terms of attractors. In
Toward a General Science of Viable Systems (New York: McGraw Hill, 1972), p. 33,
Arthur Iberall offers a nonlinear theory of matter and energy (ranging from atoms
to human societies) relyingexclusively on limit cycle attractors (which tends to
make it mechanistic at points). Iberall challenges received quantum theory, arguing
that it provides only a statistical algorithm that undoubtedlygives the right results
but no conceptual structure with which to understand them (an ad hoc numer-
ology); he proposes that it be replaced with a theory thatcharacterizesthe eigen-
values (fundamental parameters) of a field in terms of nonlinear limit cycles.
28. Richard Palmer, Broken Ergodicity, in Stein, cd., Lectures in the Sciences
ofComplexity,p. 275.
29. The clothes you wear, be they wool, cotton, or silk, are animal or plant
gcls. They are dyed with colors, which in many instances. are colloid in type.
. . . ..

The leather in your shoes is an animal gel, closely related to the prototype of the
colloids, gelatin. The wood of the chairs in which you rest is made of cellulose,
. ..

which in all its various forms is colloid in nature. [They] are held together by
. ..

glue or by steel nails and steel is a colloid solid solution. The paper upon which
. ..

you write. .[and] the ink in your fountain pens [are] probably also colloid.
. . ..

[C]olloid, too, is the hard rubber of your pen holders, prepared from that notori-
ously colloid mother substance, soft rubber (Wolfgang Ostwald quoted in Milton

Z0118 163
 Nonorganic Life

Gottlieb, Max Garbuni and Werner Emmerich, Seven States tfMatter [New York:
Walker, 1966], p. 92).
30. Harvey Blatt, Gerard Middleton and Raymond Murray, The Origin of
Sedimentary Roclz (Englewood Cliffs, N.].: Prentice Hall, 1972), p. 20.
31. ibid., p. 300.
32. Michael Bisacre, Richard Carlisle, Deborah Robertson and John Ruck, eds.,
The Illustrated Encyclopedia of Natural Resources (New York: Simon and Schuster,
1984), p. 77.
33. Deleuze and Guattari, A Thousand Plateaus, p. 40. Hardening, whether soft
or rigid, is in no way limited to the form of the example here: for example, soft
structures may result from forms of statistical accumulation quite different from
sedimentary sorting. These two modes of stratification may also occur simultane-
ously ratherthan sequentially.
34. Bisacre et al., eds., The Illustrated Encyclopedia of Natural Resources, p. 79.
35. My essay may be seen as expanding on the philosophicalinsights of Deleuze
and Guattari found in the following quote: The Earth. is a body without organs.
. .

This body without organs is permeated by unformed, unstable matters, by flows in

all directions, by free intensities or nomadic singularities [i.e., bifurcations], by mad
or transitory particles [e.g., solitons].
. [But] there simultaneously occurs upon the
..

earth a very important, inevitable phenomenon that is beneficialin many respects

and unfortunate in many others: stratification. Strata are Layers, Belts. They con-
sist of giving form to matters, of imprisoning intensities or locking singularities into
systems of resonance and redundancy, of producing upon the body of the earth
molecules large and small and organizing them into molar aggregates. The strata. ..

are judgments of God; stratification in general is the entire

system of judgments of
God (but the earth, or the body without organs, constantly eludes thatjudgment,
flees and becomes destratified, decoded, deterritorialized) (A Thousand Plateaus,
p. 40). Deleuze and Guattari call the ethics we might derive from this process the
geology of morals.
36. Arthur Iberall and Harry Soobak, A Physics for Complex Systems, in
Yates, ed., Say-Organizing Systems, p. 509.
37. Deleuze and Guattari, A Thousand Plateaus, p. 411.
38. Peter Schuster and Karl Sigmund, Self-Organization of Macromolecules,
Self-Organizing Systems, p. 75.
in Yates, ed.,
39. Howard H. Patee, instabilitiesand Information in Biological Self-
Organization, in Yates, ed., Sey10rgani2ingSystems, p. 334.
40. Vladimir Glisin, Ana Savic, Radomir Crykvenjakov, Sabera Ruzdijic and
Neveenka BajkovicMoskov, Molecular Biology in Embryology: The SeaUrchin
Embryo," in Yates, ed., Self-Organizing Systems, p. 163.
41. Rupert Sheldrake, A New Science oflife: The Hypothesis ofFormative Causation
(Los Angeles: ]. P. Tarcher, 1981), p. 19.
42. V. French, Pattern Formation in Animal Development, in C. F. Graham
and P. F. Wareing, eds., Developmental Control in Animals and Plants (Oxford: Basil
Blackwell, 1984), p. 242.

164
 43. HansMeinhardt, The Random Characterof Bifurcations and the
Reproducible Processes in Iimbryonic Development, in Gurel and Rossler, eds.,
Bifurcation Theory, p. 188.
44. C. H. Waddington, New Patterns in Genetic Development (New York:
Columbia University Press, 1964), pp. 44-45.
45. Sheldrake, A New Science qflye, p. 71.
46. Stuart A. Kauffman, Principles of Adaptation in Complex Systems, in
Stein, ed., Lectures in the Sciences c_Jf'Complexity, pp. 621, 651.
47. P. T. Sounders, An lntroduction to Catastrophe Theory (New York: Cambridge
University Press, 1986), pp. 118-19.
48. Brian C. Goodwin, The Evolution of Generic Form, in Maynard Smith
and G. Vida, eds., Organizational Constraints on the Dynamics of Evolution (Manchester,
Eng: Manchester University Press, 1990), pp. 114-15.
49. I. G. Simmons, Biogeography: Natural and Cultural (London: Edmund
Arnold, 1979), pp. 58-62. Further, The flows of energy and mineral nutrients
through an ecosystem manifest themselves as actual animals and plants of a partic-
ular species (p. 79).
50. Many aspects of the reproductive budget (includingparental investment
and battles of the sexes and the generations) are treated in Richard Dawkins, The
Selfish Gene (Oxford: Oxford university Press, 1989), chs. 7, 8, 9.
S1. A good discussion of catalytic loops in evolutionary arms races and male
decorations can be found in Richard Dawlcins, The Blind Watchmalzer (New York:
W. W. Norton, 1987), chs. 7, 8.
52. john Maynard Smith, Evolution and the Theory of Games," in Did Darwin
Get It Right? Essays on Games, Sex and Evolution (New York: Chapman Hall, 1989),
pp. 206-207.
Evolutionary stable strategies, according to Maynard Smith, are those that are
optimal for a given context and cannot thereforebe invaded by a mutant strategy.
To this extent, they are not stabilizednonlinearlyby attractors, but rather linearly,
like a crystal. The idea thatevolution optimizes is becoming more and more fre-
quently the target of nonlinear science. The arguments are similar to those I use
below for the case of human economics (basically,the existence of mutually incom-
patible constraints). For a criticism and a version of these concepts in terms of
attractors, see Kauffman, Principles of Adaptation, pp. 687-95.
53. Niles Eldrige, MacroEVoluti0na2yDynamics: Species, Niches and Adaptive Peaks
(New York: McGraw Hill, 1989), p. 151.
54. Leon Glass and Michael C. Mackey,From Clocks to Chaos (Princeton:
Princeton University Press, 1990), pp. 114-15.
55. Ibid., ch. 9.
56. Alan Garfinkel, The Slime Mold Dictyostelium as a Model of Self-
Organizing Systems, in Yates, ed., Self-Organizing Systems, p. 200.
57. See Deleuze and Guattari, A Thousand Plateaus, ch. 6: How Do You Make
Yourself a Body without Organs?
58. Gleick, Chaos, pp. 59-77; R. M. May, Chaos and Dynamics of Biological

ZUIIE I65
 Nonorganic Life

Populations, in Dynamical Chaos (London: Royal Society, 1987); and Manfred

Peshel and Werner Mende, The Preclator-PreyModel: Do We Live in Volterra World?
a

(Vienna: Springer, 1986).

59. Arthur Iberall, A Physics for the Study of Civilizations,"in Yates, ed.,
.

Self-Organizing Systems, pp. 531-33.

60. Ibid., p. 533.
61. See Pierre Clastres, Society against the State, trans. Robert Ilurley (New
York: Zone Books, 1987), p. l89ff. Societies against the State" have also been
termed anticivilizationsto stress their effective resistance to crystallization into

a stratified society: see Gordon W. Hewes, Agriculture and Civilization,in

Matthew Melko and Leighton R. Scott, eds., The Boundaries cyfCivilizazion in Space
and Time (Landham, Md.: University Press ofAmerica, 1987), p. 199.
62. See, for example, Gilles Deleuze, Politics, $emiotex1(e) 8 (1978),
pp. 15463.
63. Deleuze and Guattari, A Thousand Plateaus, ch. 11: The Refrain."
64. In some mathematicalmodels, periodic and even strange attractors can
result if we maintain these two assumptions: in other words, even if the model
allows markets to clear perfectly (supply always exactlysatisfying demand), and the
agents are given perfect foresight, cycles and chaos can result if we include non-
linear financial constraints (for example, limited access to borrowing). See Michael
Woodford, Finance, Instabilityand Cycles, in Willi Semmler, ed., Financial
Dynamics and Business Cycles: New Perspectives (Armonk, N.Y.: M. E. Sharpe, 1989).
65. On the phenomenon oflocl<ing in suboptimal equilibria, see W. Brian
Arthur, Self-Reinforcing Mechanisms in Economics," in Philip W. Anderson,
Kenneth].Arrow and David Pines, eds., The Economy as an Evolving Complex System
(Reading, Mass: Addison-Wesley, 198 8).
66. A recent nonlinear model of Kondratieffcycles can be found in: D.
Sterman, Nonlinear Dynamics in the World Economy: The Long Wave," in
Christiansen and Parmentiert, eds., Structure, Coherence and Chaos. Also of note
is the pioneering work of Richard Goodwin, who has been using nonlinear models
since the 19505: see his Essays in Nonlinear Economic Dynamics (Frankfurt: Peter
Lang, 1989). There are many models of Kondratieffcycles, both linear and non-
linear; for a survey, see Christopher Freeman, ed., Long Waves in the World Economy
(Stoneham, Mass.: Butterworth, I983).
Day, Adaptive Economics, in Robert Crosby, ed., Cities and
67. Richard H.
Regions Non-LinearDecision
as Systems (Washington, DC: AAAS, 1983), p. 10.
68. On the role of dislocation structures in real crystals, see Allan Bennet,
The Importance of Imperfections, in Sharon Banigan, ed., Cgrstals, Perfect and
Imperfect (New York: Walker, 1965), p. 85. On human elites as dislocations, see
Iberall, Toward a General Science, p. 208.
69. Robert Crosby, Asking Better Questions, in Crosby, ed., Cities and Regions,
p. 10.
70. Ibid., p. 12.
71. Ibid., p. 21.

166
 72. C. S. Holling, Resilienceand Stabilityin Ecosystems, in Erich jantsch
and Conrad H. Waddington, eds., Evolution and Consciousness (Reading, Mass.:
Addison-Wesley, 1976), p. 87.
73. Stuart A. Kauffman, Antichaos and Adaptation, Scientific American 265.2
(Aug. 1991), p. 82; and Christopher Langton, Life at the Edge of Chaos, in
Langton, Charles Taylor,]. Doyne Farmer and Steen Rasmussen, eds., Artgicial Life
(Reading, Mass.: Addison-Wesley, I992), p. 85.
74. Langton, Life at the Edge, p. 82. If as Langton and Kauffman think,
poised systems are where complexity and variety peak (i.e., are crossed in the
most intense form by the machinicphylum),then we may imagine a different set
of policies (or even philosophiesof everyday life). We would have to trackthese
special zones, for example, by destratifying ourselves (that is, by pushing our solid
components a little, but only a little, toward the liquid state). just exactlyhow
much to liquify should be established through careful experimentation, since
there are dangers if one goes too far (one may be swallowed up by a chaotic attrac-
tor). This is how Deleuze and Guattari put it: You dont reach the [machinicphy-
lum] by wildlydestratifying. If you free it with too violent an action, if you blow
. ..

apart the strata without taking precautions, then instead of [tapping into the phy-
lum] you will be killed, plunged into a black hole, or even dragged toward cata-
strophe. Staying stratified organized, signified, subjected is not the worst that

can happen; the worst that can happen is if you throw the strata into demented or

suicidal collapse, which brings them back down on us heavier than ever. This is
how it should be done: Lodge yourselfon a stratum, experiment with opportuni-
ties it offers, find an advantageous place on it, find potential movements of deterri-
torialization, possible lines of flight, experience them, produce flow conjunctions
here and there, try out continuums of intensities segment by segment, have a small
plot of new land at all times (A Thousand Plateaus, pp. 160-61).
75. Cyril Stanley Smith,A Search for Structure (Cambridge, Mass.: MIT Press,
1982), p. 113.
76. Intentional strata correspond to the world of beliefs and desires as they
exist in the animal kingdom, past a certain level of complexity of informational
flow. The sedimentations and hardenings that make up these strata grow on top
of the organic stratum (theyuse the brain as their substratum). They, in turn, pro-
vide the raw materials for the emergence in humans of linguistic strata. Deleuze
and Guattari offer theories on these two strata, which they call subjcctification
and signiliance. I shall produce a more detailed exposition of these topics in my
Forthcoming book Chaos and the Millennium.

Z0118 167