“Open” Systems, Closed Minds
John A. Seeger
Bentley College, Department of Management
175 Forest Street
Waltham, Mass., 02154-4705
U.S.A.
ABSTRACT
System Dynamics has not achieved widespread recognition as a paradigm of substance in the business-
related disciplines of Strategic Management, Organization Behavior, Organization Theory, or Opera-
tions Management. One reason for its slow acceptance by academicians in these fields and related
social sciences may lie in the specialized meanings and usages attached to common words by the
System Dynamics lexicon. Words such as “open,” “closed,” “feedback,” and “structure” -- used differ-
ently than established scientists might expect -- may create perceptions that System Dynamicists
simply don’t understand systems theory. Writers in the field need pay special attention to the semantic
implications of their presentation.
INTRODUCTION
The human legacy of feedback thinking, as traced by George Richardson (1991),
covers some two millennia of practical inventions, two centuries of technological
application, 100 years of analytic exploration, and 50 years of modern development.
Richardson documents the growth of two quite separate threads of modern feedback
thought, originating independently in cybernetics and in servomechanisms theory.
These competing threads often use the same words with different definitions, without
recognizing either that they represent different schools of thought or that they use
different languages.
In the cybernetics thread, Richardson finds (pp 332-341), systems thinking begins with
an equilibrium model, subject to exogenous disturbances and relying on homeostatic
mechanisms to restore balance. Discrete stimuli from the environment are taken as
causing the system to respond. The system’s internal causal mechanisms are believed
unobservable or beyond comprehension, so attention is concentrated on measurement
of discrete inputs and resulting behavior. Systems are described in verbal terms, and
complex social systems are seen to be capable of self-change: they can evolve, or “re-
write” their internal structure. Dynamic behavior stems from the randomness of events,
and management's problem is to anticipate and compensate for these disturbances.
In the servomechanisms thread, in contrast, systems thinking begins with a dynamic
model, subject to exogenous disturbance but capable of departing from equilibrium and
generating its own behavior through the interaction of internal forces. Environmental
stimuli may stress the system, but they are not necessary to cause behavior. Internal
causal mechanism are explicitly represented. Systems are described as quantitative,
continuous relationships between variables arranged in positive or negative feedback
- 665 -
loops, where nonlinearities can produce differing modes of system behavior without
changing the basic model structure. Dynamic behavior stems from the feedback
structure, and management’s problem is to understand how its policies influence
internally- generated behavior.
System Dynamics as an organized body of knowledge and practice clearly falls within
the servomechanisms thread as described by Richardson. Yet the field has not
achieved widespread respect in other disciplines. “The influence of this [servomech-
anisms] thread on feedback thinking in the social sciences appears to be slight com-
pared to the influence of the cybernetics thread,” says Richardson (1991, p 313). “The
tendencies of thought that characterize this thread have not spread as widely...” (p
271). This paper examines one possible reason for the failure of the servomechanisms
school of thought -- more specifically, of the System Dynamics school of thought
growing from the writings of Forrester -- to penetrate the main stream of academic
thinking.
THE PROBLEM
From the standpoint of the field’s own professionals, the desired state is one where
System Dynamics is widely recognized, understood, and accepted as an appropriate
approach for analysis of complex nonlinear systems. In the ideal scenario, universities
would vie for outstanding faculty capable of teaching the field; doctoral candidates
could choose between prestigious teaching posts and lucrative practitioner positions;
masters and bachelors graduates could look forward to influential positions in policy
analysis and planning. ‘
The current state of the system is not measured on any rigorous basis, but most System
Dynamics people will agree the desired state is far removed from the present. Although
the approach is appreciated by many in industry (as witnessed by the number of small
consulting firms blossoming in the Boston area), it is far from achieving recognition in
academic circles, especially in the United States. However, the field is growing, even
if from a very small base. Given patience and a publicized base of successful applica-
tions, recognition should come. In other words, there is no real problem.
Some in the field point to Thomas Kuhn’s The Structure of Scientific Revolutions
(1970) to explain the slow progress of System Dynamics. Kuhn wrote that new
paradigms ascend to prominence in an intellectual field only when existing science
Tecognizes anomalies that cannot be explained by old theories. However, in a recent
paper in Science, Lightman and Gingerich (1992) suggest that anomalies may exist for
a very long time before people take them seriously. Only when a new paradigm is
shown to explain a widely-recognized anomaly does the established field begin to
tespond, they say.
I suggest there is a problem for System Dynamics, and its professionals can influence
the field’s rate of acceptance. If they are to explain anomalies unresolved by existing
science, they must do so in terms the establishment scientists can readily understand.
They must recognize the vital role of communications across intellectual boundaries,
and take pains to write in the language of the target audience.
The hypothesis of this paper is that faculty members in all the social sciences and their
related business disciplines already have what they believe is an adequate grounding in
systems theory and see little need to alter what they “know.” Their operative paradigm
is anchored in the cybernetic thread of feedback theory identified by Richardson. They
studied systems theory in their own doctoral programs and seminars, and they have
trained thousands of doctoral candidates in those same systems beliefs since receiving
their own degrees. They share a common vocabulary and common beliefs about such
elementary concepts as cause-and-effect. They are the essence of what Kuhn called
“established science.” Further, they are the majority viewpoint, and it is naive to
expect them to take the initiative in giving System Dynamics “equal time.”
When System Dynamics writers attach different meanings to common words -- words
already “owned” by established systems thinkers -- it should not be surprising that the
establishment misunderstands. We might expect many faculty to simply ignore the
System Dynamics paper, assuming its author didn’t know enough about systems to
merit serious attention.
Like Paul Newman in the film, Cool Hand Luke, what we have here is a classic
“failure to communicate.” Luke paid with his life for his refusal to acknowledge the
dominant norms and definitions of those in power.
DEFINITIONS IN THE CYBERNETICS THREAD
The pertinent set of definitions for concepts in the cybernetics thread of feedback
thinking must be drawn from old literature -- from the books studied by the people
most influential now in the scientific fields we would like to influence. Those faculty
members are the full professors, the textbook authors, the dissertation chairs, the hiring
committees, the editors, the referees -- all the senior faculty who lead opinion in their
disciplines. Those are people whose operating definitions of systems theory were
formed in the 1960s and 70s; it is the literature they studied that we must look to.
This paper takes the business-related fields of study -- Strategic Management, Organiza-
tion Behavior, Organization Theory, Production/Operations Management, etc. -- as the
domain most pertinent to System Dynamics. In these fields, an extremely influential
book at the time of interest was The Social Psychology of Organizations, by Katz and
Kahn (1966 and 1978). A full generation of management scholars learned its basic
system concepts and definitions from this book. Since those are the scholars whose
models System Dynamics seeks to change or replace, it is necessary for us to under-
stand the viewpoint they bring to conversations or debates about systems.
Ten Characteristics of “Open Systems”
Katz and Kahn fall squarely within the cybemetics thread, crediting their major source
(1978, p 22): “This model of an energic input-output system is taken from the open
system theory as promulgated by von Bertalanffy (1956).” Katz and Kahn defined
their “energic input-output system” as marked by ten distinct characteristics (1978, pp
-30):
1) Importation of energy from the system’s environment;
- 667 -
2) Throughput, or work, by the transformation of energy or the reorganization of
inputs;
3) Output: the export of something to the environment;
4) Cycles of events: for example, from production to the product’s marketing, to
the collection of revenues, to the purchase of additional raw materials, and back
to production;
5) Negative entropy, importing more energy than expended, permitting storage of
energy and growth;
6) Information input, negative feedback, and the coding process;
i) The steady state and dynamic homeostasis: continuous energy flow and stable
characteristics of the system’s equilibrium;
8) Differentiation: movement toward greater specialization;
9) Integration and coordination: development of “fixed control arrangements.”
(The other nine points were identical in the 1966 edition (pp 19-26); this point
was added in 1978.)
10) Equifinality: a system can reach the same final equilibrium state from different
initial conditions and via different paths.
Toa System Dynamicist, this may seem a strange list of the defining characteristics of
a system. It is drawn directly from von Bertalanffy, whose work in biological and
chemical systems was aimed at resolving a pressing philosophical issue of the 1930s
and 40s: how could systems develop, organize, and grow in the face of the second law
of thermodynamics? Von Bertalanffy himself cautioned against applying open systems
theory to sociocultural systems (1968, p 28), but that caution and the other conditions
he placed on his original concepts were disregarded by the social scientists Katz and
Kahn. It is useful to elaborate on the above list of characteristics, to demonstrate how
far afield from System Dynamics usage are the assumptions and definitions of the
cybernetics thread.
System Boundaries
The first three characteristics of Katz and Kahn’s open systems imply boundary
definitions, which the authors amplified ina number of passages. “Our basic
model...imports energy, transforms it, and exports a product to the environment that is
the source for re-energizing of the cycle” (1978, p 55).
“the social system is more open than physical systems; it must constantly import
both production and maintenance materials [both raw materials and human resources]”
(1966, p 31)
“System boundaries refer to the types of barrier conditions between the system and its
environment which make for degrees of system openness....Boundaries are the
demarcation lines...for admission of members into the system and for other imports
into the system. The boundary constitutes a barrier for many types of interaction
between people on the inside and people on the outside...” (1966, p 60).
“Social systems will move...toward incorporating within their boundaries the external
resources essential to survival” (1978, p 28).
System Structure
The fourth defining characteristic implies the cybernetics concept of structure as
embodied in cycles of discrete events: “Social systems have a structure, but it is a
structure of events rather than physical parts.” (1966, p 69)
“Structure is to be found in an interrelated set of events that retum upon themselves to
complete and renew a cycle of activities.” (1978, p 24; Katz and Kahn here credit F.
H. Allport and James Miller).
“The basic method for the identification of social structures is to follow the energic
chain of events from the import of energy through its transformation to the point of
closure of the cycle.” (1978, p 25).
Negative Entropy
Katz and Kahn’s fifth characteristic derived directly from von Bertalanffy’s efforts to
settle the debate over how systems could self-organize and grow. The importation of
energy, information and materials across system boundaries was seen as counteracting
the natural process of entropic decay. To von Bertalanffy, it was important to discredit
the unscientific idea of ”vitalism,” which endowed all enduring or growing systems
with some life force of their own.
Feedback
Richardson has documented the cybemetics thread’s use of “feedback” as oriented
toward communications and message handling, as opposed to the servomechanisms
usage of interrelated causal linkages. Katz and Kahn’s sixth characteristic links feed-
back with information inputs and coding processes.
“The simplest kind of informational input found in all systems is negative feedback.
Information feedback of a negative kind enables the system to correct its deviations
from course....to keep the system on target.” (1978, p 26).
The reference to “information feedback of a negative kind” suggests the authors’
readiness to equate negative feedback with “constructive criticism,” as is frequently
found in the literature of interpersonal behavior and organization development. As
Richardson noted, the cybernetics thread is not conscious of positive feedback or self-
reinforcing “vicious circles.”
~- 669 -
The Steady State and Dynamic Homeostasis
Open systems theory was originally based on tightly constrained, rigorously defined,
continuous-flow chemical systems, where a steady state could be mathematically
described. In its application here by Katz and Kahn, the concept has lost much of its
precision. “Steady state” seems to take on some aspects of “consistent behavior.” The
authors were aware of the internal inconsistency between an equilibrium model and a
universe of growing systems. They drew on Kurt Lewin’s field theory (1951) to
reconcile the differences.
“...in counteracting entropy, these [homeostatic] systems move toward growth and
expansion. This apparent contradiction can be resolved, however, if we recognize the
complexity of the subsystems and their interaction in anticipating changes necessary for
the maintenance of an overall steady state.” (1978, p 26).
“Tn terms of Lewin’s quasi-stationary equilibrium, the ups and downs of the adjustive
process do not always result in a return to the old level. Under certain circumstances a
solidification or freezing occurs during one of the adjustive cycles. A new base line is
thus established and successive movements fluctuate around this level, which may be
either above or below the previous plateau of operation.” (1978, p 28).
Differentiation and Integration
The eighth and ninth characteristics of open systems appear to be an effort to embrace
the contingency theory model of Lawrence and Lorsch (1967), published concurrently
with Katz and Kahn’s early work.
Equifinality
Von Bertalanffy’s original concept of equifinality was a mathematical statement of
general transport equations, stating that the rate of change of a variable in the open
system is equal to the velocity of transport through the system plus the rate of
production of the variable by the system. Provided that both the velocity and the
production rates are linear with respect to the variable, and provided both are indepen-
dent of time, von Bertalanffy said, then a steady state attained by the system has a value
“equifinal” or independent of the initial conditions.
Equifinality in its original conception was a severely constrained system condition -- a
necessary condition for the system to qualify as “open.” The constraints, however, do
not survive in the modem social science interpretation.
System Dynamics
One further definition from the Katz and Kahn treatment is essential to this paper. The
authors define “system dynamic” as the maintenance of equilibrium-in social organi-
zations by constant adjustment and anticipation.
- 670 -
“The operation of system dynamics can be seen in the acquisition and extractive mechan-
isms employed by systems and subsystems. Much of the organizational literature...is
strangely silent...about how a social system acquires the input necessary to maintain
itself and carry on its functions” (1966, p 67). The authors go on to suggest that
competition, in many dimensions, is the answer.
IMPLICATIONS OF THE DOMINANT PARADIGM
It is tempting for System Dynamicists to evaluate the Katz and Kahn characteristics, to
refute or deny their applicability, or to ignore them entirely as unworthy of attention. I
suggest these are the same reactions cybernetics-trained scientists might experience on
reading one of our manuscripts. They would see we used words incorrectly, failed to
understand the most basic concepts, and failed even to see that we could not see. Some
of them might conclude our material was not worth reading.
It is vital to recognize that the Katz and Kahn definitions represent the dominant
paradigm in the social sciences, or at least in the business-related descendants of the
social sciences, as it was studied by the generation of senior academics who now stand
in judgment of our work. To the extent the definitions here linger in the perceptions,
the writing, or the teaching of today’s opinion leaders, System Dynamics suffers an
immense handicap in presenting its message. The words we use -- even the name of
our field -- may evoke negative responses wholly unrelated to the content of our work.
For example, our use of “closed” loops is often taken as indicating our work is
restricted to closed systems -- systems of utterly no interest to scholars in the cyber-
netics thread (or to us). This confusion is seen in the early work of Stafford Beer
(1959) as reported by Richardson (p 184). It has also arisen frequently in conversa-
tions between disciplines; a major figure in international behavioral science reported to
me in 1975 he had once asked a graduate student to look into System Dynamics, but
had lost interest on hearing the method dealt only with closed systems.
Richardson see less conflict and more potential for convergence between the cyber-
netics and the servomechanisms threads of feedback thinking than do I. “Over time,”
he says, “...as feedback ideas become more persuasive in the social sciences, we
should expect a blending of the two threads. Some natural selection process ought to
take place, guided by whatever governs the ecology of ideas” (1991, p 166). All those
with an interest in System Dynamics will concur: of course the threads ought to
converge. But we must recognize that scientists of the established point of view will
see no reason to “blend” with us. ‘Richardson hints at no mechanism for bringing about
the hoped-for convergence.
He does recognize the short-term disadvantage of System Dynamics. “:In the long run,
the feedback ideas that persist are likely to be those with the most promise. But there
seems to be no guarantee that in the short run less promising aspects of a new view will
not be selected. The prevailing wisdom will strengthen certain aspects of a new view,
those that are most easily assimilated, even if they will prove in the long run to be the
less promising aspects.” (1991, p 286).
- 671 -
For members of the prevailing school of thought, non-linear, multi-loop continuous
system representations are not easily assimilated. They are a foreign language.
Management practitioners, on the other hand, are not handicapped by training in the
existing science. They recognize the applicability of the System Dynamics approach.
This is both a blessing and a curse. Practitioners provide employment for our gradu-
ates, but they also lure promising students away from academic careers where they
could, by teaching, increase the creation rate of dynamic modelers.
TOWARD RESOLUTION OF THE PROBLEM
This paper is based on opinion and assumption as much as on analysis. I assume that
nearly all management scientists, regardless of discipline, are familiar with at least a
superficial appreciation of “open systems theory” and accept it at face value, without
conscious thought (although most do not attempt to use it in their own work). I
assume the established scholars see little or no need to invest time and effort in re-
examining their systems concepts. I assume they form quick judgments about new
ideas, based on their apparent ease of assimilation into the existing body of knowledge
and their apparent utility to the scientist.
If these assumptions hold, then arguments posed as confrontational debates between
paradigms are unlikely to attract opponents of stature. Explanations detailing closed
loops of cause and effect relationships will only close the minds of the audience. Pro-
testations about positive feedback will produce little response, because the concept has
no corresponding point in the cybernetics thread. Similarly, nonlinearities are not
necessary to explain system morphogenesis, if the audience believes systems are
capable of self-reorganization. All the fundamental aspects of System Dynamics that
make the field attractive to its proponents, make it irrelevant or trivial to holders of the
dominant paradigm.
How, then, can we influence the acceptance rate of the servomechanism-based
approach? I suggest the first imperative is to recognize the mind-set of the cybernetics-
oriented scholars, and to phrase our communications with them in language that links to
their own frame of reference. It should not be necessary (and it is surely difficult) to
teach them the details of dynamic modeling. It is unrealistic to try to “convert” them to
our enlightened point of view.
Rather, we can avoid direct confrontation and recognize we are the minority. As such,
it is in our interest to consider the community of scholars as a customer, to whom we
need to sell ideas. We could position System Dynamics in the intellectual marketplace
as a Vital tool for addressing a class of systems not easily handled with existing theory.
(We ourselves may be convinced that class is all-encompassing in scope, but there is no
need to convince the majority of that point.) We could seek out and use more descrip-
tive terms for our work -- using words linked to positive images in the management
disciplines. And we could do this, in my opinion, without compromising the integrity
of System Dynamics in any way. The key lies in recognizing that many (perhaps most)
of the established scholars are non-quantitative thinkers, unequipped to judge the merits
of a technique at the detail level. They need explanations at their own level of compre-
hension and analysis. As customers, they deserve no less.
- 672 -
Systems of Bounded Rationality
For example, we might tell others we are concerned with a subset of open systems,
called perhaps “Systems of Bounded Rationality.” Such systems, we could say, fall
outside the strict domain of economic analysis because they incorporate decision-
makers’ perceptions and their nonlinear responses to stimuli originating both inside and
outside the system. Since the problems we address are poorly defined, our methods
are aimed first at theory-building, and then at testing the theories for internal consisten-
cy and potential utility by the client. We are conscious of the
Such a description does no injustice to either System Dynamics or to Herbert Simon,
whose phrase (1945) so aptly describes the kind of systems we are good at. This
description might help others recognize the unique niche of System Dynamics, without
implying that all other techniques are now obsolete and without creating defensive
reactions. Indeed, any set of tools proclaiming success at treating intractable problems
might be widely accepted.
Many other approaches might also be devised, to gain credibility for the field. The
important first step is to recognize within System Dynamics that our language is not
common, even though other scientists use the same words. We must recognize that
the paradigmatic debate is not restricted to proponents of the cybernetic and the servo-
mechanisms traditions; it extends to the whole community of established academic
professionals. The vast majority of those scholars would be intensely disinterested in
the details of debate between competing systems points of view. They know enough
about open systems theory to recognize it when they hear it, but not enough to judge
technical merit. On technical issues, they will follow the leaders of conventional
wisdom.
I suggest that System Dynamics might increase its rate of acceptance by adopting a
niche strategy, in which it complements existing techniques by demonstrating success
at theory building and testing in specialized problem applications. Scores of such
applications are documented in the literature, but they are typically presented on a
technical level or as confrontational invitations to combat. If recognition can be
achieved by presenting these cases as evidence for competence in a small niche, then
the future question becomes one of enlarging the niche.
REFERENCES
Beer, S. 1959. Cybernetics and Management. London: English Universities Press.
Bertalanffy, L. von. 1968. General Systems Theory. New York: George Braziller.
Forrester, J.W. 1961. Industrial Dynamics. Cambridge, MA: M.I.T. Press.
Katz, D. and R. L. Kahn. 1966 and 1978. The Social Psychology of Organizations.
New York: John Wiley & Sons.
Kuhn, T.S. 1970. The Structure of Scientific Revolutions. Chicago: The University
of Chicago Press.
Lawrence, P.R. and J.W. Lorsch. 1967. Organization and Environment.
Homewood, IL: Richard D. Irwin.
Lewin, K. 1951. Field Theory in Social Science. New York: Harper & Brothers.
Lightman, A. and O. Gingerich. 1992. When Do Anomalies Begin? Science 255:
690-694.
Richardson, G.P. 1991. Feedback Thought in Social Science and Systems Theory.
Philadelphia: University of Pennsylvania Press.
Simon, H.A. 1945. Administrative Behavior. New York: The Free Press.
- 674 -
