David W. Bates
“Unity, Plasticity, Catastrophe: Order and
Pathology in the Cybernetic Era”
Abstract: Catastrophe is usually seen as something that befalls the organized,
adaptive system from the outside, threatening its future existence. While some
cyberneticians explicitly pathologized catastrophe, the French mathematician
René Thom in the 1970s redefined catastrophe as a sudden unexpected turn that
is generated from within the complex system. While “catastrophe theory” had
a limited impact, unlike the broader notions of chaos theory and complexity
theory that are now more familiar, I use this idea to turn back to the earlier twentieth century, to locate the ways in which catastrophic events were understood
to be essential to the functioning of a complex unity. I trace in Kurt Goldstein’s
The Organism (1934) the idea of “weak catastophe” and its relation to Georges
Canguilhem’s ideas of pathology and norm in order to demonstrate that in fact,
cybernetic-era theorists of the automatic machine were interested in developing
what we might call a “pathology of the machine” that was influenced by organismic ideas of internal catastrophe.
Cybernetics was centered on a fundamental analogy between organism and
machine. As W. Ross Ashby asserted: “I shall consider the organism … as a mechanism which faces a hostile and difficult world and has as its fundamental task
keeping itself alive.”1 Because cybernetics intentionally blurred the boundaries
between humans, animals, and sophisticated technological objects, it has often
been accused of reducing living beings to the mere interplay of mechanisms.
However, cybernetics always wanted to infuse machinic beings with the essence
of life—purpose, adaptive responsiveness, learning, and so on—while opening
up new insights by comparing organisms to some of the most innovative technologies of the era, namely servo-mechanisms, scanning instruments, electronic
communication systems, analog computers, and, perhaps most notably, the new
high-speed digital calculators that were emerging from secrecy in the postwar
era. Cybernetics drew together advanced automatic machines and organisms on
the basis of their shared capacity to respond flexibly to a changing environment—
1 W. Ross Ashby, “Homeostasis,” in Heinz von Foerster, ed., Cybernetics: Circular Causal and
Feedback Mechanisms in Biological and Social Systems (New York: Josiah Macy, Jr. Foundation,
1952), 73.
“Unity, Plasticity, Catastrophe: Order and Pathology in the Cybernetic Era”
33
whether the external world or that inner domain the physiologist Claude Bernard
called the “internal milieu.”
Incisive critics would question the analogy between machine and organism
not so much because it was inherently reductionist, but instead because the infusion of vital “purpose” into technological objects distorted the essential nature
of organismic life. As Georges Canguilhem argued in his 1947 lecture “Machine
and Organism,” machines of course have ends that govern their design, however
unlike an organism, the teleological orientation of a machine is always given to
it from the outside. This purpose is not intrinsic to the machine’s own organization. A machine governed by externally given ends is always a slave to its given
order: the machine must affirm “the rational norms of identity, consistency, and
predictability.”2 In contrast, the organism is organized by the goal of its own survival, and it was therefore capable, Canguilhem observed, of genuine improvisation in the event of crisis or even internal failures. Without a completely fixed
structure restricting its potentiality, the organism was able to find new paths to
success. The essential unity of the organism was maintained even in the event
of catastrophic error because its organs were fundamentally “polyvalent,” and
not limited to previously defined, specific deterministic functions, as the parts
of a machine were. The organism could be saved by what are called here pathological responses. There was, Canguilhem declared, no such machine pathology,
no machine “monsters,” because the machine could never create for itself new
norms, new forms of existence. If life, as Canguilhem said, is “an attempt from all
directions,” the machine is defined by a rigid singularity of purpose.
And yet, cyberneticists were intensely interested in pathological breakdowns. In his 1948 classic Cybernetics: Or, Communication and Control in the
Animal and Machine, Norbert Wiener claimed that certain psychological instabilities had rather precise technical analogues: “Pathological processes of a
somewhat similar nature are not unknown in the case of mechanical or electrical computing machines.”3 In a sense, cybernetics as a transdisciplinary science
had its very origin in the insight that pathological physiological performances
could be mapped, structurally, onto technological failures with mathematically
identical characteristics. While investigating the behavior of feedback systems in
steering mechanisms, Wiener and his colleague Julian Bigelow discovered that
excessive compensation could lead to increasing oscillations that eventually
became uncontrollable, leading to great disorder and a failure to find equilib2 George Canguilhem, “Machine and Organism,” in Knowledge of Life, tr. Stefanos Geroulanos
and Daniela Ginsburg (New York: Fordham University Press, 2008), 90.
3 Norbert Wiener, Cybernetics: Or, Communication and Control in the Animal and Machine (Cambridge, Mass.: MIT Press, 1948), 172.
34
David W. Bates
rium. When they asked the medical researcher Arturo Rosenblueth if there were
any similar pathologies known in human physiology, he immediately answered
that there was indeed an exact neurological parallel—voluntary motion could
degenerate into the same state of oscillation when the cerebellum was injured
in specific ways.4 Mathematical analysis of other forms of physiological disorder
(for example, cardiac arrhythmias) would soon reveal a number of these pathological parallels.
This early interest in pathology at the very foundation of cybernetics is hardly
surprising, given the fact that so many of its original practitioners were medical
professionals with interests in both physical and mental illnesses. Warren McCulloch was a neurologist who worked in psychiatric clinics, Ashby was trained as a
psychiatrist and practiced while developing his research projects in homeostatic
stability, and Rosenblueth was a cardiologist. In an important sense cybernetics
(especially its later incarnations) was a highly medicalized discipline, aimed at
identifying the origins of instability in large, complex systems, diagnosing the
sources of breakdown so as to eliminate them and recover unity and stability.
Cybernetics offers the hope of providing effective methods for the study, and control, of
systems that are intrinsically very complex…. In this way it offers the hope of providing the
essential methods by which to attack the ills—psychological, social, economic—which at
present are defeating us by their intrinsic complexity.5
Instability appears in this framework as a “self-generating catastrophe” that
leads to the collapse and death of the system if it is untreated.6
And yet, despite this extensive interest in failure, pathology, and catastrophic
breakdowns, cybernetics might still be accused of greatly simplifying the problem
of disorder and crisis, pathology and health. It was one thing to explain some
forms of adaptive response in terms of automatic technologies, but it was quite
another to explain the radical originality and inventiveness that one repeatedly
encountered in the living world. This capacity was not reducible to any specific
mechanism or logic: as the noted polymath Michael Polanyi observed, it flowed
from an “active center operating unspecifiably in all animals.”7 What thinkers
such as Polanyi (and Canguilhem) were trying to suggest was that organisms
4 Ibid., 15.
5 W. Ross Ashby, An Introduction to Cybernetics (London: Chapman and Hall, 1961), 5–6.
6 W. Ross Ashby and Mark R. Gardner, “Connectance of large dynamic (cybernetic) systems:
critical values for stability,” Nature 228 (1970): 784. On catastrophe, see Wiener, Cybernetics, 113
and Norbert Wiener, The Human Use of Human Beings (Boston: Houghton Mifflin, 1950), 49.
7 Michael Polanyi, Personal Knowledge: Towards a Post-Critical Philosophy (Chicago: University
of Chicago Press, 1958), 336.
“Unity, Plasticity, Catastrophe: Order and Pathology in the Cybernetic Era”
35
were not just able to respond to changing conditions; they were also able to enter
wholly new states of being, with new forms of order, and new potentials, when
they were confronted with extreme challenges, drastic injury, or internal failures
of communication. A pathological state was not simply the negative inverse of a
normal function, as cybernetics at times seemed to imply, but rather an opportunity for the invention of radically new behavior.
The historical question is whether cybernetics did in fact understand and
model this potentially productive relationship between genuinely pathological
conditions and the radical novelty that was the reorganization of the organism.
At stake is the legacy of cybernetics, which has reemerged as a question in the
twenty-first century with the rise of the global information society, the beginning
of new forms of human-computer symbiosis, and the acute conflicts over remote
killing technologies. The cybernetic moment, it has been said, initiated a revolutionary transformation in the world after the Second World War, one that accomplished a rationalization and systematization of the economic and the political
spheres, leaving the individual subjected to the teleological drive of these autonomous systems.8 But what did cybernetics really stand for?
Peter Galison, for one, has forcefully argued that cybernetics, born in the
midst of war and developed during the height of the Cold War, was at its heart a
“Manichean” science, obsessed that is with maintaining order against the constant active threat of disorder. Cybernetics could hardly embrace a link between
abnormality and creativity when it pathologized disorder as the ultimate enemy.9
Andrew Pickering’s recent sophisticated analysis of cybernetic science affirms
Galison’s claim. For Pickering, the pathological state was, for the cyberneticists,
just another way of looking at the normal, ordered state—“in medicine the normal
and the pathological are two sides of the same coin.” Cybernetic machines and
organisms were defined by their inherent structures and both were kept “alive”
by the processes that maintained homeostatic equilibrium. Breakdown might
reveal something of the normative order, however pathology was never conceptualized independently.10
In practice, however, it is somewhat difficult to maintain a clear boundary
between order and disorder in cybernetic discourse. The cybernetic analogy
between machine and organism was being forged at a moment when biological
thought had largely transcended the old opposition between mechanists and
8 “L’hypothèse cybernetique,” Tiqqun 2 (2001): 40–83.
9 Peter Galison, “Ontology of the Enemy: Norbert Wiener and the Cybernetic Vision,” Critical
Inquiry 21 (1994): 228–266.
10 Andrew Pickering, The Cybernetic Brain: Sketches of Another Future (Chicago: University of
Chicago Press, 2010), 132.
36
David W. Bates
vitalists. The key challenge in this period was to explicate organismic unity—first,
to ascertain how an organism developed into a coherent formal structure from
embryonic cells; and second, to identify how organisms could maintain those
forms despite ever-changing conditions and a dynamic metabolic process. Unity
of living beings was important because it linked transformation and destruction
with continuity and stability. Unity was predicated on the inherent plasticity of
organismic organizations, their capacity to take on new forms at key turning
points, even catastrophic ones. To understand the cybernetic effort to create a
science that bridged organismic and machine beings we must pay close attention to the way unity and flexibility were conceptualized in cybernetics. As we
will see, illness, pathology, breakdown, all were vitally important questions for
biological thinkers in this period. These states were not simply ascribed to “disorderly” Manichean forces attacking order, but instead were understood to be
potentially productive crisis conditions that revealed the essential transformative
potential that made the ongoing life of an organism possible in the first place.
This complex relationship between systematic, technologized order and the risky
opportunities of crisis seems particularly relevant to our own age. Tracing out this
occluded intellectual trajectory in cybernetics will offer a critical perspective on
the legacy of post-war technology and its conceptual underpinning.
From Cybernetics to Catastrophe Theory
One of the central claims of cybernetics was that a process of “negative feedback”
drove the homeostatic life of adaptive beings. The cybernetic entity (whether
living or machinic) first sensed its environment, alongside its own “state” of
being in that environment, then compared that information with its own “goal”
states embedded somewhere within this being, before effecting certain actions
that would bring the entity’s state in line with this ideal goal.11 To be sure, from
one perspective, this cybernetic concept of negative feedback could be read as
privileging a predetermined, precisely defined “order” that is imposed on these
active beings. Yet from another angle one might emphasize the importance of
deviation in the functioning of any cybernetic being—this being has an existential
relationship with error. It is not disorder or even error itself that threatens the
cybernetics being, but instead an inability to respond appropriately to an ongoing
state of disorder. Any dynamic adaptive being is in a condition of ceaselessly devi-
11 See the iconic essay: Arturo Rosenblueth, Norbert Wiener, and Julian Bigelow, “Behavior,
Purpose, and Teleology,” Philosophy of Science 10 (1943): 18–24.
“Unity, Plasticity, Catastrophe: Order and Pathology in the Cybernetic Era”
37
ating from its own formal ends. From this perspective, radical failure (even death)
is not so opposed to normal homeostatic operation because the catastrophic turn
is in fact continuous with normal efforts to maintain the health of the being in
its essential errancy. That is, the boundary between normality (health) and catastrophic failure (illness, death) is defined by the limits of the being’s own errancy,
and not at all by the mere presence of deviation. The great achievement of cybernetics was the demonstration that these limits to error were not at all arbitrary.
As Wiener remarked, “the conditions under which life, especially healthy life,
can continue in the higher animals, are quite narrow.”12 As Ashby explained, in
an essay on homeostasis, “if the organism is to stay alive, a comparatively small
number of essential variables must be kept within physiologic limits,” and this
applied to the “life” of technological entities as well. Their survival as unified
systems depended on keeping error within proper limits.13
The crucial arbiter of the limit of error was therefore the survival of the being
as a whole. As Ashby would make clear, the relationship between catastrophe and
normal error or deviation is governed by the threat to the fundamental unity of
the being in question. He gives the example of a mouse trying to evade a cat. The
mouse can be in various “states” or postures, and certain values may even change
drastically (it may lose an ear, for example), yet still, the mouse will survive. “On
the other hand, if the mouse changes to the state in which it is in four separated
pieces, or has lost its head, or has become a solution of amino-acids circulating
in the cat’s blood then we do not consider its arrival at one of these states as corresponding to ‘survival.’”14 The unity of the being is the ultimate mark of survival.
Unity amounts to the capacity to maintain some stable form even while experiencing drastic—perhaps even violent—transitional states.
There is no fundamental distinction between order and disorder here. The
definition of a formal unity determined the parameters of survival within the
unified system. Both in biology and cybernetics, the unity of the being is what
identifies the structural relations and variables that had to be maintained against
the threat of extinction. To rethink cybernetics from the perspective of the organism, we must zero in on the nature of this unity and its status. How could a breakdown, even a catastrophic failure, become the opportunity for an unprecedented
reorganization? How could a functionally determinate machine ever acquire this
organismic capacity?
In works such as Design for a Brain and Introduction to Cybernetics, Ashby
introduced a formal way of thinking about the behavior of complex, adaptive
12 Wiener, Cybernetics, 135.
13 Ashby, “Homeostasis,” 73.
14 Ashby, Introduction, 197.
38
David W. Bates
systems. He noted that if we take the basic states of a system as variables—
capable of change that is—then we could graphically plot them mathematically
as vectors. In turn, variables could be integrated in a multi-dimensional “phase
space” as a way of representing, with one single vector, what Ashby called a line
of behavior.15 Ashby’s innovation was to think of the cybernetic being as a set
of potential states, states that could be visually represented in phase space as
vectors whose functions were limited by boundary values. Ashby showed how
the interpretation of these lines of behavior as mathematical functions opened
up a new way of understanding complex systems. Mathematical analysis of
physical systems was normally limited to continuous systems of a linear nature.
A mathematical approach to biology, in contrast, would have to accommodate
non-linear and discontinuous features of organismic behavior.16 Ashby gave an
example: any sudden transformation of activity might be governed by a so-called
“step function,” where the value of the function changes abruptly when reaching
a certain point. This mathematical representation of behavior would suggest that
seemingly unpredictable events or even new, spontaneous behaviors in the life
of a being might well be governed by some hidden, dynamical process modeled
by a step function.17 The main insight was the idea that transformations of the
system’s behavior were strictly analogous to mathematical transformations of the
operands.18 For Ashby, this made it possible for cybernetics to study the innate
“determinateness” of a system formally, in its own mathematical terms that is,
and to ignore effectively the actual “material substance” of the system.19
Ashby’s graphical approach to cybernetic systems and self-governing stability had a precedent, in the work of one of the most prominent biologists of this
period, C. H. Waddington. In his 1940 book on the role of genes in embryological
development (morphogenesis), Waddington introduced the “epigenetic landscape,” a three dimensional virtual representation of the developmental pathways of an organism.20 Imagining a plateau cut through with a series of valleys
and ridges, Waddington explained that a ball rolled into this landscape would
first bounce between different valley formations before settling into a specific one,
where it would then roll up and down the sides, perhaps even exiting the valley
altogether if it was shallow enough. But as the ball continued, it would become
15 Ibid., 25, 31.
16 Ibid., 27.
17 W. Ross Ashby, Design for a Brain, 2nd. ed. (New York: John Wiley and Sons, 1960), 95.
18 Ashby, Introduction, 37.
19 Ibid., 24.
20 C. H. Waddington, Organisers and Genes (Cambridge: Cambridge University Press, 1940), 91.
The frontispiece depicts the epigenetic landscape.
“Unity, Plasticity, Catastrophe: Order and Pathology in the Cybernetic Era”
39
more and more confined in its path as the valley walls became steeper and the
path narrowed. The point of this virtual exercise was to account for the way early
embryonic cells, which lacked organizational order at the start, became increasingly specialized and distinct from one another as they divided. This representative landscape would also suggest why the path of the ball, despite perturbations
from outside, would continue to its destination. This was an important issue, as
Hans Driesch had famously shown in the last century that sea urchin embryos
would continue to develop normally even when they were partially destroyed,
split in half, or otherwise violently manipulated. The epigenetic landscape therefore was in part a description of what Waddington called the “canalization” of cell
development, but also an attempt to represent (at least qualitatively) the forces
that intervened to shape the paths of that development. The goal was to understand how the embryo itself was capable of developing this kind of internal organization, without invoking some mysterious vitalist force, as Driesch had done
with his concept of entelechy. Modern biologists were interested in the ways that
the “potential” of the organism, that is, its own teleological process of canalization, could be understood in strictly biological or physio-chemical terms.21
In his later, influential book The Strategy of the Genes, Waddington acknowledged Ashby’s innovative attempt to mathematically represent the behavior of
a complex system in geometrical form. In a chapter entitled “The Cybernetics
of Development,” Waddington approved Ashby’s effort to “map” the states of a
system, but disagreed with his claim that increasing the number of variables in
a system would lead to a greater probability of instability.22 Waddington argues
that this complexity in fact could have a positive function, in that it allowed for
a greater number of possible states that the organism could use to find order.
Like Ashby, Waddington emphasized the non-linear nature of biological systems,
which could suddenly shift into new states, both as they grew and differentiated
themselves in development and as they sought stability in their mature phases.
Waddington rejected Ashby’s use of step functions as too simplistic, returning
instead to his earlier notion of the epigenetic landscape, which was in essence a
qualitative (rather than strictly quantitative) representation of the behavior of an
organism as its cells found paths (his term for these paths was “chreods”) leading
from relative disorder to increasing order and organization. The landscape was a
neutral metaphor that eschewed both vitalist and mechanist assumptions. It was
a description of a process with its own internal dynamic.
21 See, for example, Alan Turing’s late work on the mathematics of chemical diffusions in morphogenesis.
22 C. H. Waddington, The Strategy of the Genes: A Discussion of Some Aspects of Theoretical
Biology (London: George Allen and Unwin, 1957), ch. 2.
40
David W. Bates
Waddington’s insight would become the foundation of a more precise science
of the complex. Later in his career, Waddington became very interested in the
work of the French mathematician René Thom, who in the 1960s was developing a theoretical and methodological technique that would become known as
“catastrophe theory.”23 Thom introduced his notion of “structural stability” in
the context of developmental biology and homeostatic constancy at an interdisciplinary gathering organized by Waddington to generate new theoretical models
in biology.24 Thom explicitly rejected cybernetic approaches to stability that drew
on the concept of feedback. “Like all notions taken from human technology, the
notion of feedback cannot properly be invoked to explain the stability of biological processes…. The only notion which is mathematically and mechanically
acceptable is that of ‘structural stability.’”25 His point was that stability had to be
inherent in the form itself, understood in terms of its formal organization that
is, and not the substrate of the formal order (whether biological or inorganic).
Central to Thom’s thinking was the idea that the response of the living being
to some “shock” should not be conceived as an actual feedback mechanism of
some kind but instead as a function of the mathematical organization inherent
in the formal order of the being itself.26 He redefined the organism as a geometric
object whose transformations could be mapped (as Ashby had argued earlier)
in a multi-dimensional phase space.27 Thom drew here on the groundbreaking
work of D’Arcy Thompson, whose On Growth and Form (1917) not only probed the
physical underpinnings of patterns in nature, but also suggested that the shapes
of organic forms may be related to one another geometrically.28 However, Thom’s
own topological approach was novel. He argued that changes of the organism (as
it developed as an embryo, or as it maintained itself in changing environmental
conditions) should be understood as topological transformations (that is, bending
and twisting) of a multi-dimensional geometrical object. It was this object that
23 For an overview of the theory, its genesis and reception see Alexander Woodcock and Monte
Davis, Catastrophe Theory (New York: Dutton, 1978).
24 The papers were published in C. H. Waddington, ed., Towards a Theoretical Biology (Edinburgh: Edinburgh University Press, 1968).
25 René Thom, comments on C. H. Waddington, “The Basic Ideas of Biology,” in ibid., 33.
26 René Thom, “Théorie dynamique de la morphogenèse,” ibid., 162–3.
27 René Thom, Structural Stability and Morphogenesis: An Outline of a General Theory of Models,
tr. D.H. Fowler (Reading, Mass.: W. A. Benjamin, 1975), 151–2.
28 D’Arcy Thompson, On Growth of Form (Cambridge: Cambridge University Press, 1917). Thom
explains the moment of insight that initiated catastrophe theory occurring while he was observing, in a natural history museum, a representation of the embryonic development of the frog and
saw it as a kind of continuous transformation of a single entity. See René Thom, Paraboles et
catastrophes: Entretiens sur les mathématiques, la science, et la philosophie (Paris: Flammarion,
1983), 45.
“Unity, Plasticity, Catastrophe: Order and Pathology in the Cybernetic Era”
41
would be represented graphically, not the numerical values of independent variables. What the technique may have lost in precise predictive power it gained
by enabling an understanding of the trajectory of a complex system (such as an
organism) as a unified set of values.
Thom’s major mathematical contribution was the insight that complex
systems with several controlling variables would inevitably undergo radical and
sudden transformations. These were called “catastrophes,” mathematical singularities. And Thom further demonstrated that if the number of control values were
limited, these systems would only exhibit seven kinds of catastrophe, which he
called the “elementary catastrophes”—they ranged from the simplest folds and
cusps to complex three-dimensional “slices” of higher dimensional objects. The
central claim was that the discontinuity marked by the catastrophe was nonetheless still a topological transformation of the original form, continuous with it
mathematically even as the system exhibited dramatic non-linear change when
perceived from within a particular phase space. Thom therefore believed he had
solved the vexing problem of morphogenesis, the origin of biological form. Discontinuities (for example, differentiations of the embryonic cells into specific
organs) were mathematical catastrophes and not radically new material phases.
Thus they did not have to be explained as the product of some specific mechanism or ethereal vitalistic force.29 The mathematical form was itself intrinsically
stable through the catastrophes, and moreover, this form could withstand a
certain amount of deformation imposed from the outside.30
Like Canguilhem, Thom resisted anthropomorphic concepts of biological
order derived from the realm of technology. However, Thom would admit:
This is not to say that comparisons of the life dynamic with some other manifestations of
human technology (automata, electronic computers, etc.) are pointless, but rather that
these comparisons have validity only for partial mechanisms, fully developed with their
complete functional activity, and they can never be applied to the global living structure or
to its epigenesist and physiological maturation.31
Thom rarely cited other thinkers in his work, but here he refered explicitly to
Jacob von Uexküll’s influential work in theoretical biology. Von Uexküll (who
died in 1944) emphasized the idea that living organisms were not systems made
up of various functional parts, but instead subjects capable of unified experience.
29 Thom, Structural Stability, 157–8.
30 Ibid., 209. Here we can see a connection between theoretical biology and the investigation of
self-organizing systems in the inorganic world, systems that Ilya Priogine would later describe
as “dissipative structures.”
31 Thom, Structural Stability, 200–201.
42
David W. Bates
While he echoed later cybernetic concepts of homeostasis and informational processing, von Uexküll insisted that the organs of biological beings come into existence to serve that unified subject as it navigates its environment, the Umwelt.32
Thom picked up on this idea when he claimed that machines must receive their
order all at once—the parts must be created contemporaneously then fitted
together in a broader organization that comes, as Canguilhem noted, from outside
its own organization. Organisms, in contrast, develop their internal organs; functions emerge as a product of the original unity of the being differentiating itself.
Life is therefore a process of individuation. For Thom, the specific configurations
of the organismic order make sense only within the encompassing unity of being.
The machine’s unity is artificially imposed on the parts, and for this reason there
is no active force of unity, as there is in the organism.
The importance of this contrast between machinic order and organismic
unity for Thom is evident when he came to address pathology and the breakdown
of order. Unlike a mechanical system, whose parts can only play a predetermined
role in the organization, the organism is capable of actively reorganizing its components in response to perturbations. However, there is a limit. As Thom explained,
if the “shock” to the life field is too great, it will not be compensated for and the
organism will “enter a zone of qualitative indeterminacy.” In this zone, greater
activity will take place and “larger variations in the global physiological state”
will be allowed. The organs that are not affected by the perturbation will become
extra-excited. Healing will occur if this extra activity stabilizes the affected (sick)
organ. Interestingly, Thom notes that the healed state will not be continuous with
the normal state prior to the injury or illness, since the system has moved into
a wholly new organization of activity due to the shock. The pathological state
here is understood to be a transitional phase, leading to a new form of the being
(a new topology) with its own stabilized order. The definition of life is tied, as it
was with Ashby, to this continuous articulation of a unified being, a stable formal
structure whose dissolution marks the end of that life. What is important about
Thom’s catastrophe theory is that it conceptualizes this life as a set of parameters
with an internal organizational relationship. It is entirely agnostic as to the metaphysical ground of this biological unity—even as it describes organismic development and dynamic homeostatic processes as a function of that formal unity. As
with cybernetics, then, a perturbation of the system is overcome in normal operation by a “correction” of the local deviation, a deviation defined by the formal
32 Ibid., 200. The epigraph to this chapter is a passage from von Uexküll’s Bedeutungslehre,
where he contrasts the centripetal (zentripetal) development of mechanisms with the centrifugal
formations in the organic world. Thom also cited Kurt Goldstein as another important influence
on his theory. The emphasis on holistic unity may be traced to his work as well, as we shall see.
“Unity, Plasticity, Catastrophe: Order and Pathology in the Cybernetic Era”
43
organization. However, as Thom revealed, a pathological state is one in which a
mere local correction gives way to a global reorganization of the system.33
This relationship between normal “catastrophes” and more profound transformations in crisis that potentially undermined the structural analogy between
organismic life and the engineered behavior of advanced machines because there
was no obvious technological counterpart to reorganization. Yet cybernetics was
in fact already closely tied to this catastrophic discourse of the organism.
Shock and the Plasticity of the Organism
We must remember that the theory of homeostasis, one of the central concepts
of cybernetics, was intimately linked to these notions of crisis and catastrophic
shock. The concept was first developed in the 1920s by the American physiologist Walter Cannon, who, as a collaborator of Rosenblueth later in his career,
directly introduced the idea to cybernetics. Cannon was intensely interested in
the dynamic quality of all forms of organismic stability. Although we can trace the
idea of bodily self-regulation to Claude Bernard’s influential study of the “interior milieu,” Bernard’s theory explicitly asserted the relative independence of the
internal regulatory mechanisms from the external environment and its effects
on the organism.34 In contrast, Cannon’s notion of homeostatic regulation was a
product of both his clinical experience of shock in the Great War and his interest
in what might be called “emergency states” of the body. In an early work from
1915, Cannon explored the endocrine system, which was capable in crisis situations of putting the body automatically into a heightened state of action, capable
of either “flight or fight,” as Cannon famous phrased it.35 Cannon would go on to
study the problem of physiological “shock” through examinations of numerous
soldier-patients suffering extreme injuries, and also through animal experiments
that induced shock artificially through massive bloodletting and the like.36 Can33 Of course, it is also possible that the pathological condition will lead to instability and eventual death, or what Thom calls a “generalized catastrophe,” distinct from the many specific catastrophic turns that the system will always undergo. As Thom puts it: “our everyday life, on the
physiological plane, may be a tissue of ordinary catastrophes, but our death is a generalized
catastrophe.” Thom, Structural Stability, 250–1.
34 See Steven J. Cooper, “From Claude Bernard to Walter Cannon: Emergence of the Concept of
Homeostasis,” Appetite 51 (2008): 419–27.
35 Walter B. Cannon, Bodily Changes in Pain, Hunger, Fear, and Rage: An Account of Recent Researches into the Function of Emotional Excitement (New York: Appleton, 1915).
36 Walter B. Cannon et al., The Nature and Treatment of Wound Shock and Allied Conditions
(Chicago: American Medical Association, 1918).
44
David W. Bates
non’s interest in these organismic reactions to extreme emergency conditions
led him to the idea that shock, like the “fight or flight” mechanism, was not just
a particular local “correction” of a physiological parameter, but instead a new
state of being that the body entered under stress: “every complex organization
must have more or less effective self-righting adjustments in order to prevent a
check on its functions or a rapid disintegration of its parts when it is subjected
to stress.”37 Cannon’s theory of homeostasis, further developed in the 1920s and
1930s, was intimately linked to these extreme shocks and the threat of peril.
Cannon’s approach to the dialectic of order and stress in physiology was consistent with other efforts in this period to understand how bodies reorganized
themselves in times of unexpected shock and injury. Of particular interest in this
context was the nervous system, whose plasticity (ability to change structure)
had been a topic of speculation and research since the discovery of the synapse
by Charles Sherrington early in the 20th century. And as Sherrington himself had
famously emphasized, the brain and nervous system functioned as an essential unity—every action and reaction had to be understood as a total response
involving the entire system.38 This theorization would revolutionize the approach
to brain-damaged patients (whose numbers dramatically increased in the Great
War). In contrast to an earlier interest in specifying the localization of various
functions in the brain by inferring what was lost when a part of the brain was
damaged (Broca’s area, responsible for language, was an early discovery), new
clinical research recognized how complex functional responses of the brain
involved the brain as a complex whole. In the Soviet Union, Alexander Luria,
studying the impact of brain injury on cognitive abilities, observed that the brain
had the startling ability to reorganize itself in order to compensate for the loss of
functions after stroke or accident. As he noted, numerous studies showed “the
high degree of plasticity shown by damaged functional systems, due to dynamic
reorganization and adaptation to new circumstances and not to regeneration
and restoration of their morphological integrity.”39 Karl Lashley’s research in
the 1920s, while initially aimed at finding the precise neural location of memory
traces, in fact ended up revealing the plasticity of the brain’s performance
grounded in a structural complexity that defied localization. After teaching
animal subjects certain tasks (maze running, for example), Lashley proceeded
to surgically destroy certain parts of the brain. Following injury, animals were
37 Walter B. Cannon, The Wisdom of the Body (New York: W. W. Norton, 1932), 25.
38 Charles Sherrington, The Integrative Action of the Nervous System (New York: C. Scribner
Sons, 1906).
39 A.R. Luria, Restoration of Function after Brain Injury, tr. Basil Aigh (New York: MacMillan,
1963), 33.
“Unity, Plasticity, Catastrophe: Order and Pathology in the Cybernetic Era”
45
still able to recover their earlier performances, revealing a significant reorganization of the brain’s activity. The unity of the brain’s systematic complexity was
the structural frame for these reorganizations, which were, according to Lashley,
themselves just one more example of the brain’s normal capacity to integrate its
activity across the areas of the brain. “The whole implication of the data is that
the ‘higher level’ integrations are not dependent upon localized structural differentiations but are a function of some general, dynamic organization of the entire
cerebral system.”40 Lashley called this capacity “equipotentiality,” alluding to
how the brain sought multiple paths for its activity, thereby giving it the capacity
to circumvent damage by relocating activity to some other part of the brain.
Extensive clinical experience with brain-damaged patients (many of them
soldiers with bullet wounds and other war injuries) furnished the data for Kurt
Goldstein’s innovative work on the nature of unity and plasticity in biological
systems. In his classic 1934 book, influenced by Gestalt theory, the theory of
brain “shock” championed by Constantin von Monakow, as well as broader holistic forms of thought in the interwar period, Goldstein defined the organism as a
unity, arguing that in its continual struggle with the world, within the essential
“milieu” of its activity, the organism maintained its stability by constantly reorganizing itself to accommodate new conditions.41 Goldstein’s focus on “pathological” data was, to be sure, aimed at illuminating the normal functions of a
dynamic organismic life. (50) This approach did not, however, involve an analysis
of the “mere defects”(48) of the organism as a way of capturing, in an inverse
image, the “normal” state. Instead, Goldstein saw that the pathological state had
its own particular characteristics, its own symptomology, its own way of being.
The difference between healthy and pathological states, for Goldstein, was the
difference between “ordered” and “disordered” behavior. In the first case, the
performances of the total organism were, he said, “constant, correct, adequate.”
The disordered state is defined by shock—Goldstein refers to the activity of the
organism in this state as a catastrophic reaction. This Katastrophenreaktion was
“disordered, inconstant, inconsistent” and had a “disturbing aftereffect.”(48–9)
In normal conditions, the organism is challenged by its milieu, and meets this
40 Karl Lashley, Brain Mechanisms and Intelligence: A Quantitative Study of Injuries to the Brain
(Chicago: University of Chicago Press, 1929), 176.
41 Kurt Goldstein, Der Aufbau des Organismus: Einführing in die Biologie unter besonderer
Berücksichtigung der Ehfahrungen am kranken Menschen (Hague: Martinus Nijhoff, 1934); in English as The Organism: A Holistic Approach to Biology derived from Pathological Data in Man
(New York: Zone, 1995). Further references (to the translation) will be in the text. An excellent
account of Goldstein’s work and the German contexts of theoretical biology can be found in Anne
Harrington, Reenchanted Science: Holism in German Culture from Wilhelm II to Hitler (Princeton:
Princeton University Press, 1996).
46
David W. Bates
challenge with a reaction that will bring the organism into equilibrium with its
environment. In the pathological state, the organism has no proper response
at hand. And yet, as Goldstein stressed, the organism is constantly seeking an
ordered condition—and injured, shocked, creatures do often return to some form
of health. His interest, then, was to show how the organism rediscovered stability
and normality after a catastrophic reaction.
As Goldstein argued, with numerous examples, organisms have the capacity
to modify themselves and their performances to reduce or minimize any defect that
had led to a catastrophe reaction. (52) New paths to successful performances are
found, or, alternatively, new milieus are sought out that do not require the same
kind of adaptation previously required. (105) This reorganization is explained, by
Goldstein, as the tendency of the unity of the organism to seek closure—what the
Gestalt theorists called the law of Prägnanz. (293) The catastrophic reaction, then,
is not a mere interlude between states of health, but instead an interruption that
demands a new foundation of order for the organism as a whole, because it must
overcome the persistence of a defect in its being. In this way the pathological
state can best reveal the essence of the organismic unity because it demonstrates
in sharp relief how a living being seeks out novel forms of order to overcome a
disordered state, whereas in the healthy being there is an occlusion of this capacity, due to the relative automaticity and predictability of the ordered responses.
The successful response to a catastrophic reaction is clearly not a return to
the previous state, a regaining of past performances—the organism seeks ordered
behavior in both normal and pathological states. Crucially, Goldstein observes
that every action of the organism is a response to a challenge, not just those of
a catastrophic interruption. “The normal person, in his conquest of the world,
undergoes, over and again, such states of shock. If in spite of this, he does not
always experience anxiety, this is because his nature enables him to bring forth
creatively situations which insure his existence.”(237) The catastrophic reaction
gives us insight into the creative action that is the organism’s essential nature,
according to Goldstein. The unity that marks the boundary is not a predetermined formal structure that is subsequently defended or repaired. The unity
is a tendency of the organism in its temporal existence to find order, although
this order is, at every moment, always put in question by the changing conditions of the milieu. “Therefore reactions scarcely ever occur that correspond to
a perfectly adequate configuration of the organism and the surroundings.”(227)
As Goldstein argued, the organism is never entirely “normal” because at each
moment it is being challenged by the environment and must continually seek
the proper adjustment: “normal as well as abnormal reactions (‘symptoms’) are
only expressions of the organism’s attempt to deal with certain demands of the
environment.”(35)
“Unity, Plasticity, Catastrophe: Order and Pathology in the Cybernetic Era”
47
And so Goldstein writes that the life of the organism can be considered a
series of what he calls “slight catastrophes” (leichter Katastrophenreaktionen),
where inadequacies are first confronted and then “new adjustments” or “new
adequate milieu” are sought to respond to this lack. (227) The serious catastrophe is in effect continuous with this normal, constantly repeated weaker form
of catastrophe; what is different is only the scale and intensity of the reaction.
The whole organism in its unity is always falling into states of shock and must,
over and over again, create new order to overcome these shocks. The key point is
that catastrophic shocks of some form are essential to the organismic being. And
not just in the physiological sense. Goldstein notes that the foundation of learning, for example, is an unpleasant confrontation with one’s inherent inadequacy
analogous to the experience of the brain-injured patient in recovery. (249) As he
later wrote: “we assume that ‘coming to terms’ with the world must proceed by
way of constantly recurring catastrophic situations, with concomitant emotions
of the character of anxiety.”42
In the end, Goldstein will admit that perhaps we should not even oppose
“Being-in-order” and “Being-in-disorder” because the catastrophic states of disordered behavior are foundational opportunities for achieving some degree of
order—and that order is always in question. “If the organism is ‘to be,’ it always
has to pass again from moments of catastrophe to states of ordered behavior.”
(388) The catastrophic reaction manifests the essence of life itself:
All the minor catastrophic reactions to which the organism is continually exposed thus
appear as inevitable way stations in the process of its actualization, so to speak, as the
expression of its inescapable participation in the general imperfections of the living world.
(392)
Thus for Goldstein, normative behavior is never really the norm. Normal ordered
existence is a product of the essential pathological disequilibrium of the organism and its surrounding condition, its milieu: “these shocks are essential to
human nature and … life must, of necessity, take its course via uncertainty and
shock.”43 The organism’s creative plasticity allows it to find a new form of being
that preserves unity and not any one particular privileged order.44
42 Kurt Goldstein, Human Nature in the Light of Psychopathology (Cambridge, Mass.: Harvard
University Press, 1940), 109.
43 Ibid., 112.
44 Goldstein makes the link between organismic plasticity and the catastrophic reactions of the
nervous system explicit in “Über die Plastizität des Organismus auf Grund von Erfahrungen am
nervenkranken Menschen,” in A. Bethe, ed., Handbuch der normalen und pathologischen Physiologie (Berlin: Springer, 1931), vol. 15: 1133–1174.
48
David W. Bates
Canguilhem’s theory of the normal and the pathological was greatly influenced by Goldstein’s work.45 The idea that catastrophic reactions are normal
informed Canguilhem’s approach to pathology, and lies behind his critique of the
analogy between machine and organism. While machines can be endowed with
a purpose, it was impossible to imagine a machine that would be capable of suffering something like a catastrophic reaction, let alone rising to this challenge by
reorganizing itself into an altogether new form of unity. Pointing to the great plasticity of the nervous system, Canguilhem noted that if a child suffers a stroke that
destroys an entire half of the brain, that child would not suffer aphasia (as is the
case with many brain injuries later in life) because the brain reroutes language
performance to other regions in order to compensate for the damage.46 Organisms
are stable as unities precisely because their organization is not fixed into any one
rigid structure—they are open, and thus equipped to surmount even a traumatic
loss of functions in some cases.
One major challenge for cybernetics, then, was to explain how this close
relationship between unity, plasticity and catastrophe, so characteristic of organismic life, could be engineered into the cybernetic machine. According to the
influential biologist and systems theorist Ludwig von Bertalanffy, cybernetics
could never account “for an essential characteristic of living systems,” namely
their ability to maintain stability despite the constant metabolic creation and
destruction of its own material essence.47 Like other critics of cybernetics such as
Polanyi, Bertalanffy believed that “a mechanized organism would be incapable
of regulation following disturbances,” since a machine could not radically transform itself—as the organism could—to accommodate shock and injury.48 Open
systems were plastic, and never fully determined; they possessed what Bertalanffy called “equifinality,” the capacity to follow multiple paths for the maintenance of life—an echo of Lashley’s depiction of the brain’s “equipotentiality.”49
While these sophisticated critiques cannot be easily dismissed, it is also the case
that cybernetics was entangled with these same questions. Tracking the problem
45 See George Canguilhem, The Normal and the Pathological, tr. Carolyn R. Fawcett and Robert
S. Cohen (New York: Zone, 1991), ch. 4, “Disease, Cure, Health,” where Goldstein’s work is discussed extensively, including his concept of the catastrophic reaction.
46 Canguilhem, “Machine and Organism,” 79.
47 Ludwig von Bertalanffy, Robots, Men, and Minds: Psychology in the Modern World (New York:
George Braziller, 1967), 68.
48 Ludwig von Bertalanffy, General System Theory: Foundation, Development, Applications (New
York: George Braziller, 1968), 213.
49 See Ludwig von Bertalanffy, “The Theory of Open Systems in Physics and Biology,” Science
13 (1950): 23–29.
“Unity, Plasticity, Catastrophe: Order and Pathology in the Cybernetic Era”
49
of plasticity in this new discipline can reveal some unexpected dimensions of the
cybernetic vision for the living machine.
Cybernetic Forms of the Plastic Being
Cybernetics put a great deal of pressure on the theory of the central nervous system
as an information processor that governed behavior. In his own speculations on
the nature of the nervous system, Wiener enumerated the fundamental analogy
between neural organization and the binary architecture of the computer, but
he raised many questions about this comparison. As early as 1948, he wrote that
the realization that the brain and the computing machine have much in common
“may suggest new and valid approaches to psychopathology, and even to psychiatrics.”50 He was equally interested in how the brain maintains its information
states without ever localizing them in particular spaces, the very problem that
prompted Lashley’s own neurological research. Memory, for Wiener, could not be
something merely physical. Instead, it had to be a constantly flowing circulation,
and therefore subject to deviation; this system, he said, could “hardly be stable
for long periods of time.”51
However, it was the nature of neural plasticity in particular that constituted
the main challenge to the cybernetic project of assimilating machine and organism into one comprehensive framework. From the very beginning plasticity was
on Wiener’s mind. A central question was how to explain the nervous system’s
agile flexibility; Wiener wondered “how the brain avoids gross blunders, gross
miscarriages of activity, due to the malfunction of individual components.”52
With instability and malfunctions in mind, Wiener would explore several different analogies between psychological pathologies and computer malfunctions. He
observed, for instance, that the drastic effort to “clear” the brain of its pathological activity with the use of electrical or chemical “shock treatment” (in lieu of the
more permanent surgical lobotomy) might well parallel the necessary purging
of the computer of all data when a pathological “configuration of the system”
disrupts its operations.53 Wiener also remarked more than once on the essential
plasticity of the brain. He gave the example of Louis Pasteur, who suffered a major
stroke early in his scientific career. After his death, it was discovered that he had
only “half a brain.” Yet Pasteur was, Wiener pointed out, only mildly affected by
50
51
52
53
Wiener, Cybernetics, 168.
Ibid., 171.
Ibid., 168.
Ibid., 172.
50
David W. Bates
some physical paralysis, and mentally he was not at all diminished, as his great
scientific achievements following the stroke proved. Wiener also used the very
same example Canguilhem had offered, noting that infants can suffer “an extensive injury” to one hemisphere of the brain, only to have the remaining half take
on all the functions of the destroyed one. “This is quite in accordance with the
general great flexibility shown by the nervous system in the early weeks of life,”
Wiener wrote.54
In this early, often speculative phase of cybernetics, Wiener could not offer
much in the way of a real explanation for this plastic quality. And yet he realized
that if the brain is basically a processor of information, and not a purely physical
system, we are likely to find that pathologies of the mind will not be traceable
to specific lesions of the mechanism (i.e., the level of the neurons) but instead
to a failure of the system as a whole to manage information flows between functional centers.55 Wiener claimed that because complex human behavior no doubt
involved a great deal of neuronal connectivity, it may well be the case that the
human brain is often running “very close to the edge of an overload,” that could
at any moment result in “a serious and catastrophic” failure. With increasing
neural flows, “a point will come—quite suddenly—when the normal traffic will
not have enough space allotted to it, and we shall have a form of mental breakdown, very possibly amounting to insanity.”56 Here Wiener hinted at a cybernetic form of the “catastrophic reaction,” but again, he was limited here by his
inability to say much about how the brain (or the computer network analogue)
could recognize the failure and reorganize itself in these moments of extreme
crisis and breakdown, as Goldstein and others had demonstrated. Still, Weiner
hoped one day that a “new branch of medicine” would be developed—he called
it “servo-medicine”—that would deal with these kinds of informational disorders
of control, when the “strains and alarms of a new situation” put demands on an
information system that was not equipped to deal with such situations.57
John von Neumann, Wiener’s cybernetic colleague and fellow mathematician, had more than a passing interest in the architecture of computing devices
and the challenge of malfunctions. Appointed to oversee the development of an
electronic computer at the Institute for Advanced Study at Princeton from 1945–
1951, von Neumann looked to neurology for inspiration. That the brain and the
nervous system exhibited an amazing robustness was, von Neumann observed,
in stark contrast to the immense fragility of the new high-speed computers then
54
55
56
57
Ibid., 178–9.
Wiener, “Problems of organization,” Bulletin of the Menninger Clinic 17 (1953): 132.
Wiener, Cybernetics, 178.
Wiener, “Problems,” 134.
“Unity, Plasticity, Catastrophe: Order and Pathology in the Cybernetic Era”
51
being constructed out of mechanical relays, telephone switches, or vacuum
tubes. Von Neumann was careful to draw attention to the critical differences
between digital machines and the nervous system.58 Yet he himself was drawn to
the nervous system as a model for the computer. One of the most important marks
of the natural communication and control system was its inherent flexibility.
It is never very simple to locate anything in the brain, because the brain has an enormous
ability to re-organize. Even when you have localized a function in a particular part of it, if
you remove that part, you may discover that the brain has reorganized itself, reassigned its
responsibilities, and the function is again being performed.59
Von Neumann was interested in finding out what he might be able to learn from
these natural, robust systems as he planned his own artificial thinking machine.
Speaking at the celebrated Hixon symposium on the brain, held at Cal Tech in
1948, von Neumann, alert to the holistic perspective of thinkers such as Goldstein and Lashley (the latter was also a Hixon participant), stressed that the elements of a living system were always organized into a unity, making it important
to understand how the “functioning of the whole is expressed in terms of those
elements.”60 This was a crucial point when it came to understanding how the
fragility of the system’s components could be related to the system’s plasticity,
for what was in question was precisely how the unitary order of the system could
be “reconfigured” among remaining component parts in the event of a failure or
defect within the system. For von Neumann, ideally one would design machines
that could imitate the organism’s ability to react to unforeseen errors; as someone
noted in the discussion of von Neumann’s paper, we must “not only be able to
account for the normal operation of the nervous system but also for its relative
stability under all kinds of abnormal situations.”61 While von Neumann never
succeeded at building a computer with such a flexible plasticity, he did insist in
his technical work in computer design that “error be viewed … not as an extraneous and misdirected or misdirecting accident, but as an essential part of the
58 John von Neumann, The Computer and the Brain (New Haven, Ct.: Yale University Press,
1958).
59 John von Neumann, Theory of Self-Reproducing Automata, ed. Arthur W. Burks (Urbana, Ill.:
University of Illinois Press, 1966), 49.
60 John von Neumann, “The General and Logical Theory of Automata,” in Collected Works, vol 5:
Design of Computers, Theory of Automata and Numerical Analysis (Oxford: Pergamon, 1963), 289.
61 Ibid., 323.
52
David W. Bates
process under consideration.”62 Unlike contemporary automatic computers, the
nervous system, he observed, is
sufficiently flexible and well organized that as soon as an error shows up in any part of
it, the system automatically senses whether this error matters or not. It is doesn’t matter,
the system continues to operate without paying any attention to it. If the error seems to
be important, the system blocks that region out, by-passes it, and proceeds along other
channels.63
There is an echo here of Goldstein’s claim that an organism in crisis will seek out
new paths when a normal performance is frustrated by obstacle or injury, and it is
perhaps an implicit allusion to Lashley’s equipotentiality, or Bertalanffy’s notion
of equifinality. In any case, von Neumann was always fascinated by how natural
organisms reorganize themselves in response to the challenge of what he called
“emergency” conditions.64
It was exactly this problem of reorganization in crisis that Ashby faced head
on in creating his cybernetic model of the adaptive organism, the Homeostat
machine. Ashby knew that by definition a machine was fixed, and its operations
wholly determined. A machine, in its ideal operating state, had only one form of
behavior and therefore could not change its fundamental design. Although it is
possible to have a machine that enters new states depending on changing environmental variables (a thermostat is a simple example), the design of the structure will still govern its operation at all times. In this respect Ashby would agree
with Canguilhem’s understanding of the machine. In a notebook fragment from
1943, Ashby cited the great psychologist William James, who compared the rigid
machine to the nervous system, an entity that paradoxically exhibits both fixity of
structure and open-ended adaptive plasticity.65 The notebooks show that Ashby
was also reading Sherrington’s revolutionary work on neural integration at this
time. Ashby wrote out several passages from Sherrington, where the neurologist
described the organism as a “moving structure, a dynamic equilibrium,” something constantly adjusting itself to ever-changing conditions. The living system
62 Ibid., 329. On von Neumann and the role of error in his thought, see my “Creating Insight: Gestalt Theory and the Early Computer,” in Jessica Riskin, ed., Genesis Redux: Essays in the History
and Philosophy of Artificial Life (Chicago: University of Chicago Press, 2006): 237–259, and Giora
Hon, “Living Extremely Flat: the Life of an Automaton; John von Neuman’s Conception of Error
of (In)Animate Systems,” in Giora Hon, Jutta Schickore, and Friedrich Steinle, eds., Going Amiss
in Experimental Research, Boston Studies in the Philosophy of Science (2009), no. 267: 55–71.
63 Von Neumann, Self-Reproducing Automata, 71.
64 Ibid., 73.
65 W. Ross Ashby, Journal, pp. 1523–4. Available at the W. Ross Ashby Digital Archive, rossashby.
info/index.html, accessed July 24, 2013.
“Unity, Plasticity, Catastrophe: Order and Pathology in the Cybernetic Era”
53
was “labile” and indeed, its greatest strength, for Sherrington, was its very fragility, because that fragility made it more sensitive to its surroundings.66 If his
own Homeostat was going to be an adequate representation of self-organization,
Ashby had to figure out how to model the behavior of a genuinely open system,
one that could assume a determinate structural organization, like any machine,
but at the same time was not eternally bound to any one order. Remember that the
influential systems theorist and biologist von Bertalanffy had criticized cybernetics precisely for their misguided use of closed systems to model the fundamentally open structures of natural organisms.
Ashby himself was hardly unaware of this issue, and he thought deeply about
how to create mechanical systems that could not only respond to environmental
changes, but in fact actually change its very organization as a way of finding new
paths to stability and equilibrium. Ashby’s insight was that if a machine, defined
by a specific form and purpose, was ever to reorganize, then logically it must in
fact be capable of becoming a wholly different and new machine. In 1941, he had
admitted that man-made machines that change their organizations were “rare”
(and he failed to give any concrete example, though he did point out elsewhere
that the inclusion of memory in a system would amount to such reorganization).67
Yet Ashby would push much further, seeking to conceptualize a machine that
had, like the nervous system, what James had called an “incalculable element”
that could interrupt, productively, the machine’s own fatalistic operation. Ashby
made a conceptual breakthrough by thinking about the potential value of failure,
something that was, after all, inevitable in any working machine. Ashby realized
that when a machine broke, it became in essence a brand new machine with a
new “design,” so to speak: “A break is a change in organization.”68 Ashby’s goal
was to engineer a machine that could take advantage of its own breaks as a way
of entering into a new state. Genuine breaks of this kind were exceedingly rare in
artificial machines. According to Ashby a break was: “1) a change of organization,
2) sudden, 3) due to some function of the variable passing a critical value.”69 If a
homeostatic machine were constructed in such a way as to “break” when pushed
to the limit of its ability to maintain equilibrium, this machine could acquire a
new organization, another possible path to equilibrium. “After a break, the organization is changed, and therefore so are the equilibria. This gives the machine
66 Ibid., 1906–09.
67 Ibid., 1054.
68 W. Ross Ashby, “The Nervous System as Physical Machine: With Special Reference to the
Origin of Adaptive Behaviour,” Mind 56 (1947): 50.
69 Ashby, Journal, 1054.
54
David W. Bates
fresh chances of moving to a new equilibrium, or, if not, of breaking again.”70 The
breakdown was in essence a temporary shock to a system that was not succeeding
in its quest to find equilibrium. Ashby invented a form of cybernetic plasticity by
taking advantage of the very weakness of all machines—their ultimate fragility.
As Ashby would point out, the brain was a special kind of machine in that
its many highly differentiated component parts—the neurons—were constantly
connecting and disconnecting with each other as the brain responded to perturbation and change. Built into its dynamic organization was an inherent tendency
to break down and thereby give way to new organizations, for the neurons have
a built-in latency period: after a certain amount of activity, neurons temporarily
“disappear” from the system, only to reappear fully active again once they have
recovered their potential. Ashby suggested that the cybernetic machine and the
organism could be linked by this shared capacity to self-organize in moments of
breakdown, a capacity that ultimately could be traced to the tendency to fail—at
least temporarily—on a repeated basis.71
Reading cybernetic concepts with and against contemporary theorizations
of open, living systems, it is possible to glimpse some important and provocative
intersections between the technological discourse of cybernetics and continental forms of thought that emphasized holistic and vitalist concepts of organismic
order. A crucial marker of the open living system was the important and productive role played by pathological and even catastrophic states in maintaining the
persistence of organismic unity. Having recognized this themselves, the cyberneticists had faith that it would be possible to create artificial machines that possessed genuine plasticity, so necessary for survival in times of extreme crisis. In
the cybernetic era, both organisms and advanced machines were sites of stability
and instability, truth and error, order and disorder, health and pathology; all were
intertwined, in sometimes strange ways, to produce fragile yet powerful beings,
endowed with astonishing and unpredictable creative powers. This is worth
remembering in our own age, where human thought and human identity is so
easily assimilated by the hyper-technologized visions of the brain and the body.
70 Ashby, “Nervous System,” 55.
71 Ibid., 57–8.