Systems Research Vol. 13 No. 1, pp.

13-23 1996

Research Paper

Reflections on Systems and their Models

Russell L. Ackoff and Jamshid Gharajedaghi
INTERACT, 401 City Avenue, Suite 525, Bala Cynwyd, PA 19004, USA

Distinctions are drawn between deterministic (e.g., mechanistic), animate (e.g.,

organismic), social, and ecological systems and models. Historical examples of applying
models of one type to systems of another type are provided and the current consequences
of such mismatches affecting social systems are discussed.
Keywords systems; models; deterministic (e.g., mechanistic); animate (e.g., organismic); social systems;
ecological systems; mismatches; consequences

INTRODUCTION
There are different types of system and different
ways of representing (modeling) them. Our
concern here is with the consequences of applying a model of one type to a system of a different
type. This is a common practice with what we
believe to be serious consequences.
A system is a whole defined by one or more
functions, which consists of two or more essential parts. (1) Each of these parts can affect the
behavior or properties of the whole. (2) None of
these parts has an independent effect on the
whole; the effect an essential part has on the
whole depends on what other parts are doing. (3)
Every possible subset of the essential parts can
affect the behavior or properties of the whole but
none can do so independently of the others.
Therefore, a system is a functioning whole that
cannot be divided into independent parts.

An essential part of a system is one without

which the system cannot perform its defining
function(s). For example, a motor is an essential
part of an automobile, and a heart is for a person.
On the other hand, the cigarette lighter is a nonessential part of an automobile, and a person's
body hair is also non-essential.
CCC 0731-7239/96/010013-11
(& 1996 bv Tohn Wilev fc finns T.tH

Whatever one considers a system to beand

there is a considerable agreement as to what a
system isthere are obviously different ways
of classifying them. For example, they can be*
classified by size, by discipline (physical,
chemical, biological, psychological, and so on),
by location, by function, and many other ways
as well. The choice of a classification scheme
normally depends on its intended use. For our
purposesexamining the consequences of mismatching systems and their modelsthe critical classifying variable is purpose and purpose is
a matter of choice.
An entity is purposeful if it can produce (1) the
same functionally defined outcome in different
ways in the same environment, and (2) functionally different outcomes in the same and different
environments. Although the ability to make
choices is necessary for purposefulness, it is not
sufficient. An entity that can behave differently
but produce only one outcome in any one of a set
of different environments is goal-seeking, not
purposeful. Servo-mechanisms are goal-seeking.
In contrast, people are obviously purposeful
systems, and so are certain types of social
groups. An entity can be multi-goal-seeking it if
is goal-seeking in each of two or more different
Received 6 September 1995
Accepted 16 September 1995

RESEARCH PAPER

Systems Research

controllers, or users. Provision of that service is

theii function. Although the parts of a mechanistic
Systems and
system do not have purposes of their own, they do
Whole
models
Parts
have functions serving the function of the whole.
Not purposeful
Not purposeful
Deterministic
Therefore, all the subsystems of a deterministic
Purposeful
Not purposeful
Animated
system are also deterministic systems.
Purposeful
Social
Purposeful
The behavior and properties of a deterministic
Not purposeful
Ecological
Purposeful
system are determined by its structure, the
causal laws, and its environment if it is an open
system, but not by its environment if it is closed.
environments, and it seeks different goals in at
Even closed deterministic systems have a funcleast two different environments.
tion: to serve the purposes of an external entity.
For example, Descartes and Newton conceptualized the universe as a mechanistic system, as a
TYPES OF SYSTEMS AND MODELS
hermetically sealed clock, and they believed that
this system had been created by God and served
There are three basic types of systems and
His purposes, doing His work. Even such
models of them, and a meta-system: one that
conunonplace open deterministic systems as
contains all three types as parts of it (see Table 1).
automobiles, generators, and computers have
(1) Deterministic: systems and models in which no purposes of their own but serve the purposes
of their producers and users.
neither the parts nor the whole are purposeMechanisms are not the only deterministic
ful.
systems; plants are also even though they are
(2) Animated: systems and models in which the
alive. Neither they nor their parts can display
whole is purposeful but the parts are not.
(3) Social: systems and models in which both the choice; neither they nor their parts have purposes of their own.
parts and the whole are purposeful.
Since deterministic systems and their parts
These three types of system form a hierarchy in
cannot display choice, there is only pne thing
the following sense: animated systems have
they can do in any particular environment. Their
deterministic systems as their parts. In addition,
behavior and properties are determined by their
some of them can create and use deterministic
internal structure, their environment (if there is
systems, but not vice versa. Social systems have
any), and the causal laws of nature. Open
animated systems as their parts. All three types
deterministic systems have an environment
of system are contained in ecological systems, some (external variables that affect their behavior and
of whose parts are purposeful but not the whole.
properties); closed systems don't.
For example. Earth is an ecological system that
Despite the fact that computers are mechanhas no purpose of its own but contains social and
isms, they appear to make choices. But not so.
animate systems that do, and deterministic
Their behavior is completely determined by the
systems that don't.
information and program put into the computer
Consider each of these types of system in more
by external sources. If we know these, then, in
detail.
principle if not in practice, we would be able to
predict with certainty what the computer would
do in any situation. Programmed instructions in
Deterministic Systems
a computer are its causal laws. These together
with
its internal structure and externally proSystems that have no purpose, and whose parts
vided inputs completely determine its behavior.
do not either, are systems whose behavior is
Deterministic systems can be differentiated by
determined. They are exemplified by mechanthe
number of functions they have. An ordinary
isms. Although deterministic systems, including
clock has one function: to tell time. On the other
mechanisms, have no purposes of their own,
hand,
an alarm clock is multifunctional, since it
they normally serve the purpose(s) of one or
also has a wake-up function. Some clocks have
more entities external to them, their creators.
Table 1

14

Types of systems and models

Russell L. Ackoff and Jamshid Gharajedaghi

Systems Research

RESEARCH PAPER

many additional functions; for example, they can

measure lapsed time, and reveal the temperature.

Included in the long list of eminent biologists

who also had a mechanistic point of view were
Reil, Lamarck (the evolutionist), Rudolphi, Berzelius, Verworn, and Loeb.
Animated Systems
Opposition to a mechanistic conception of
organisms surfaced in the mechanist-vitalist
Animated systems have purposes of their own,
controversy which arose out of this conception's
but their parts don't. The most familiar examples
inability to account for the nature of life
are, of course, animals, including human beings.
adequately. Today, when life tends to be defined
All animated systems are orgarusms but not all
in terms of self-organization and self-renewal
organisms (e.g. plants) are animated systems.
(autopoiesis), it is apparent that essential aspects
(Unless otherwise indicated, we use 'organism'
of organisms are not included in mechanistic
to mean an animated organism.) Animated
models of them.
systems are alive. Life is currently defined in
On the other hand, mechanistic entities have
terms of autopoiesis: 'the maintenance of units and seldom been conceptualized as organisms. The
wholeness, while components themselves are
only exceptions we can think of occur among
being continuously or periodically disassembled
primitive people whose beliefs are said to have
and rebuilt, created and decimated, produced and
been 'animatistic', which (according to the Encycloconsumed' (Zeleny, 1981, p. 5). As will be apparent,
paedia Britannica, 11th edition. Vol. 2, p. 53) is 'the
it follows from this defirution of Hfe that sodal and
doctrine that a great part, if not the whole, of the
ecological systems are also living systems.
inanimate kingdom, as well as aU animated beings,
Plants do not have purposes but like all living
are endowed with reason, intelligence, and volithings have a goal: survival. Plants react to
tion, identical with that of man'.
changing external conditions in such ways as to
make their survival possible, but their reactions
are determined, not matters of choice, and choice
Social Systems
is necessary for purposefulness.
For animals, survival is also a^if not the^most
Social systems^^for example, corporations, uniimportant purpose. They are purposeful organisms
versifies, and societieshave purposes of their
whose parts (some of which are called 'organs')
own, contain parts (other sodal systems or
have functions but no purposes of their own. The
animated organisms) that have purposes of their
behavior of an organism's parts is determined by its
own, and are usually parts of larger social systems
state and acfivity. For example, a person's heart,
that contain other social systems (for example,
lungs, brain, etc. have no purposes of their own, but
corporations and nations). (Some primitive sodetheir functions are necessary for the survival and
ties lived in complete isolation, hence were not
pursuit of the purposes of the whole.
part of a larger sodal system.) We are not aware of
anyone trying to model organisms or mechanical
Historically, animated systems have often
systems as social systems, but clearly, social
been treated as though they were nothing but
systems have often been modeled organismically
complicated mechanisms. Mechanistic biology
(e.g., Stafford Beer, 1972) and even mechanically
dominated biology, the study of living things, for
(e.g., social physicists and Jay Forrester, 1961,
centuries. For example, the biomechanist Roux is
1971). For example, the sodologist, P. Sorokin, in
said to have taken the following position:
his book. Contemporary Sociological Theories (1928),

According to Roux, biology admits of exact

formulation because matter alone exists; there
is no ground for a fundamental distinction
between the living and non-living. The animate,
appearing as cells with nuclei, developed from
the inanimate by the operation of mechanical
laws, and is governed by them. (E.F. Flower,
1942, p. 72)
Reflections on Systems and their Models

summarized the mechanistic interpretations of

two prominent social physicists, Haret and
Barcelo, as follows:
In their works the translation of the nonmechanistic language of social science into that of
mechanics goes on in the following way: The
individual is transformed into a material point.

15

RESEARCH PAPER

and his social environment into a 'field of

forces',... As soon as this is done, there is no
difficulty in applying the formulas of mecharucs
to social phenomena; all that is necessary is to
copy these formulas, inserting the word individual instead of material point, and the term
social group instead of physical system or a field
of forces. 'An increase in the kinetic energy of an
individual is equivalent to a decrease in his
potential energy.' 'The total energy of an individual in his field of forces remains constant
throughout all its modifications ... and so on.'
(pp. 17-18)
In addition, Sorokin wrote that 'H.C. Care/s
Principles of Social Science is one of the most

conspicuous attempts in the second half of the

nineteenth century at a physical interpretation of
social phenomena' (p. 13). Carey applied such
laws as those of gravitation to social phenomena.
If an individual is taken as a molecule and the
social group as a body, then the attraction
between any two bodies is in direct proportion
to their masses (the number of individuals per
unit volume) and inversely proportionate to the
square of the distance between them. In addition,
Carey took centralization and decentralization of
populations to be the same as centripetal and
centrifugal forces.
Herbert Spencer, the nineteenth-century evolutionary philosopher, provides an excellent
example of biological modeling of social systems.
His position was summarized by A.M. Hussong
(1931) as follows:
Spencer himself groups together under four
heads these comparisons of life and society
which result in showing three phenomena well
known to characterize life, to be no less characteristic of anything to be called a society. They
are: (1) growth; with which is associated (2)
increasing differentiation of structure; and (3)
increasing differentiation of function, (p. 23)
Consider the first of Spencer's points to help
clarify his position:
In both biological and social organisms, growth is
evidenced by the same phenomena. In both, there
is increases in massin the biological individual,
and expansion from germ to adult form; in the

16

Systems Research

social, and expansion from small wandering

hordes to great nations. In both, aggregates of
different classes reach various sizesamong
biological organisms, the Protozoa rarely increase
beyond a microscopic size; among social organisms the primitive Tasmanians seldom form large
groups, while the empires of civilization include
millions of people. In both, increases by simple
multiplication of units is followed by union of
groups and unions of groups of groups. In both,
finally, a multipUcation of individuals goes on
within each group of units, (p. 23)
Stafford Beer, in Brain of the Firm (1972),
introduced his celebrated model of a viable
system which is derived from a model of the
brain. He takes the requirements for effective
individual and organizational behavior to be the
same. The difficulties in implementing his model
derive from the purposeful nature of the parts of
the organizations involved. Organismic models
do not take the purposes of the parts of an
organism into account. However, these models
are useful in social systems in those rare cases in
which the purposes of the parts are very limited
or not relevant; for example, in organizations
that are managed or ruled autocratically. The
more autocratic an organization, the more
appropriate is the use of an organismic model.
A problem arises with increasing education of
the members of a social system, the increasing
technology that they must master to do the tasks
assigned to them, and the greater the variety of
demands made on them. When those managed,
governed, or ruled know how to function better
than those who manage, govern, or rule them,
the less effective autocratic management or rule
is. A democratic organizationthat is, one in
which the members have considerable freedom
and opportunity to make choicescan't be
adequately modeled organismically precisely
because such modeling misses this most important characteristic of such a social system: the
ability of its parts to make choices. This inadequacy is particularly apparent where problem
solving is involved.
Consider the use of case studies in management
education. We recently asked a group of managers in an executive development program that
had just completed working on a case what
would happen if they presented their solution to
Russell L. Ackoff and Jamshid Gharajedaghi

Systems Research

the relevant corporation's senior management.

The class members said that these managers
would probably find a number of reasons for not
accepting it, and if they accepted it, it would
probably not be implemented as intended
because of opposition to it by those who would
be responsible for its implementation. We then
pointed out that the managers and the implementers
were part of the problem, not external to it. In the

orgarusmic model of the corporation which the

class had unconsciously used, the purposes of
those who had to approve of any proposed
action, let alone those who had to carry it out,
were not taken into account. Had the class used a
social systemic model they would have treated
acceptance and implementation of its solution to
the problem in the case as part of the problem,
not as separate from it.
In the political arena finding what is normally
thought of as a solution to a problem, and getting
it accepted and implemented, are usually treated
separately rather than as necessary aspects of the
problem. For example, many laws are simply not
obeyed or enforced and therefore solve nothing.
President Clinton's recent proposed solution of
the national health care problem was rejected by
Congress whose approval was required before it
could be enacted. Furthermore, many problem
solutions that are implemented are sabotaged by
those who implement them. This is the case
when alleged solutions, when implemented,
encourage and facilitate corruption.

Ecological Systems
Ecological systems contain interacting mechanistic, organismic, and social systems, but unlike
social systems have no purpose of their own.
However, they serve the purposes of the organisms and social systems that are their parts, and
provide necessary inputs to the survival of the
non-animate biological systems (plants) that it
contains. Such service and support is their
function.
An ecological system is affected by some of the
behavior of its component organismic and social
systems, but their effects are determined, as are
the behavior and properties of a mechanistic
system. For example, the purposeful use of
fluorocarbons as a propellant affects the ozone
Reflections on Systems and their Models

RESEARCH PAPER

layer in a way that is determined, not as a matter

of choice.
Like animated and social systems, ecological
systems also live because they are capable of
'maintaining their unity and wholeness, while
components themselves are being continuously
or periodically disassembled and rebuilt, created
and decimated, produced and consumed'
(Zeleny, 1981, p. 5).
Although the function of an ecological system
is to serve its parts, many people assume the
existence of a deity whose purposes the universal
ecological system is believed to serve. The deity
is also assumed to have created this system.
Variations within Types of Models
When we talk of deterministic, animate, and
social systemic models, we refer to classes of
models within which there are many variations.
However, these variations derive from different
treatments of non-essential variables.

THE EVOLUTION AND CONSEQUENCES OF

MISMATCHING MODELS AND SOCIAL
SYSTEMS
One can model a part or an aspect of a social
system as a mechanism and by so doing improve
its performance, but this can reduce the performance of the whole. Optimization of parts can
suboptimize the system as a whole. Recall that
every essential part of a system can affect the
performance of the whole but cannot do so
independently of all the other essential parts.
Therefore, in changing performance of any of
these parts its effect on the system as a whole
should be taken into account. Nevertheless, some
(non-essential) properties of even a system's
essential parts may have no effect on the
system as a whole: for example, the color of an
automobile's motor may have no effect on its
performance and therefore can be modified
without affecting performance of the automobile.
Similarly, workers may be essential in a corporation, but the color or type of clothing they wear
may not be.
The effectiveness of any model used to
describe and understand behavior of a particular

17

RESEARCH PAPER

system as a whole ultimately depends on the

degree to which that model accurately represents
that system. Nevertheless, there have been and
are situations in which application of determirustic or animate models to social systems have
produced useful results for a short period of time.

However, in a longer run, such mismatches

usually result in less than desirable results
because critical aspects of the social system
were omitted in the less complex model that
was used. This is the point that we argue and
illustrate here.

Deterministic Models Applied to Organizations

In the early stages of industrialization, machines
replaced thousands of agricultural workers.This
resulted in a very large number of unemployed
unskilled agricultural workers and had a destabilizing effect on Western societies. It was then
that a new concept of manufacturing 'came to the
rescue'. Production processes were designed
much as a complicated tractor is, as consisting
of an assembly of parts each of which involves a
very simple and repetitive task. Then unskilled
workers could be assigned to these elementary
tasks and be treated as replaceable machine
parts. In time this mechanical model of producfion converted the army of unskilled agricultural
workers to semi-skilled industrial workers. The
impact of this mechanistic model of organizations on productivity was so great that in one
generation it provided an amount of goods and
services that surpassed all previous expectations.
Henry Ford's phenomenal success in the
creation of a mechanistic mass production
system marked the beginning of the production
era but contained the seeds of its demise. He
failed to appreciate the potentiality of the process
he initiated when he said, in effect, 'they can
have any color [of automobile] they want as long
as it is black'. This gave Alfred Sloan of General
Motors the opportunity to gain domination of the
market. He took mass production for granted
and focused on the question: how to sell. The
marketing era emerged. It gave rise to a new set
of challenging questions, the most important of
which were: (1) how to respond to increasing
demand for variety and diversity; and (2) how to
organize and manage the increases in size and
18

Systems Research

complexity that resulted from increasing variety

and diversity.
As the size and complexity of organizations
increased, the effectiveness of managing them as
though they were machines decreased. Decentralization of control became necessary and this
was incompatible with a mechanistic conception
of organization. A machine requires centralised
control and invariance of output. No driver in his
right mind would drive a car with decentralized
front wheels.
However, in an organization that requires
invariant functiorung of its parts (as bureaucracies do), decentralization leads to disorganization, if not chaos, and less than the best
possible performance of the whole. This is so
because improving the performance of one part
of an organization taken separately, as occurs in
decentralization, often reduces the effectiveness
of other parts. For example, the best solution
to a production problem taken in isolation
(such as mirumizing inventory) may well be in
conflict with the best solution from a marketing point of view (which consists of holding
an adequate stock of all items offered for sale).
It is for this reason that organizations continually oscillate between centralization and
decentrahzation.
An example of a social-systemic treatment of
an inventory problem is provided by an abrasive-manufacturing company that offers a discount on purchases of its products that are
proportional to the amount of lead time given
by the customer for delivery of the goods
ordered. This solution converted 'demand' into
a partially controllable variable by taking into
account the purposefulness of the customers. It
reduces inventory several times as much as
could be obtained by use of (mechanistic)
economic lot sizes, and it had a very positive
effect on customer attitudes toward the company.
Summarizing, deterministic and animate
models may work very well when applied to
parts or aspects of a social system when these
parts or aspects are considered in isolation.
However, it is possible to improve the performance of each part or aspect of a system taken
separately and simultaneously reduce the performance of the whole. This follows from the fact
that no part or aspect of a system has an
independent effect on the performance of the
Russell L. Ackoff and Jamshid Gharajedaghi

Systems Research

system. A system's performance is the product of

the interactions of its parts or aspects.
The extent of the harm done by using a model
that is not of the same t5^e as the system
modeled depends on the level of development
of the system involved. For example, when it was
initiated, Henry Ford's production system
approximated a mechanical system for the
following reasons. First, the workers were
poorly educated and relatively unskilled, but
adequate for the simple repetitive tasks assigned
to them. These tasks required behavior that was
more machine-Hke than human. Second, since
there was virtually no social security available,
unemployment implied financial destitution for
many. This resulted in workers who were wOling
to tolerate working conditions suitable for
machines, but not people. Third, there was a
large pool of people looking for work, hence
workers could easily be replaced, like machine
parts; and the workers knew this. Fourth,
business organizations were normally managed
by their owners, who had virtually unlimited
control over their organizations; they were not
yet subject to significant intervention and constraints coming from government and unions.
Animate (Organismic) Models Applied to
Social Systems
There were a number of developments between
the two world wars that made deterministic
modeling of social systems less and less appropriate. Business and government organizations
took growth to be necessary in order to respond
effectively to demand for an increasing amount
and diversity of products and services. In
addition, technological developments required
workers with increasingly greater skills. The
educational level of workers increased but so
did union and government intervention. However, the threat of financial destitution associated
with unemployment decreased as social security
emerged. Finally, as businesses went public in
order to raise the capital required to fuel growth
and technological improvements, management
and ownership were separated. The publicly
owned company became a corporation (derived
from 'corpus', meaning 'body') and the chief
executive became the 'head' of the firm. Because
Reflections on Systems and their Modeis

RESEARCH PAPER

of these developments organismic models were

increasingly used when dealing with social
systems.
Sloan's concept of an organization was essentially organismic, that of a single-minded biological entity. This provided a relatively effective
way to manage organizational growth and
increase the diversity of organizational outputs.
In his (implicit) model, corporations, like the
human bodies, were divided into two distinct
parts: (1) management, the brain (Beer, 1972); and
(2) the operating unit, the body.
The operating unit, the body, was considered
to have no choice, no consciousness. It was
restricted to reacting determiaistically to instructions coming from management, the brain, and/
or events in its environment. Ideally, an operating unit would be a robot programmed to carry
out without deviation a set of procedures defined
by headquarters. Military organizations, governmental bureaucracies, and autocratic corporations closely approximated the behavior of robots.
Social-Systemic Models Applied to Social
Systems
As a result of World War II a large portion of the
workforce in Western nations was drawn into the
military. Replacements included Rosie the Riveter and Tillie the Toiler who were motivated at
least as much by patriotism as by additional
income. Such workers could not be treated as
replaceable machine parts or organs of a body
that only required consideration of the effects of
their work on their health and safety. They had to
be treated as human beings with purposes of
their own. In addition, unions discovered that
corporations working under cost-plus contracts
or demands for increased production could
easily be induced to make concessions to workers' interests. Work rules went through a
profound transformation. Furthermore, because
of the great increase in technological developments, the skills required of the workforce
increased dramatically. The more skilled the
workers, the harder they were to replace.
Technological developments required a significant investment in specialized training of the
workforce. Management had to obtain an
adequate return on this investment. All this
19

RESEARCH PAPER

produced a need to treat employees as human

beings with purposes of their own.
Children of the post-World War II workforce
made up the permissive generation whose
members were not about to be treated as other
than purposeful entities. They expected their
interests to be taken into account by their
employers and, where they weren't, they were
alienated from work and their productivity
decreased (Work in America, 1973). The quaUtyof-work-life movement was an effort to correct
this. In addition, protest groups formed outside
orgaiuzations insisting that their interests be
better served by the organization that affected
them, for example, consumerists and environmentalists. Management's social responsibility
and its work-related ethics emerged as major
concerns.
By the end of the 1960s it was apparent that the
West was experiencing both an accelerating rate
of change largely due to technological developments, and increasing complexity produced by
an explosion of interconnections resulting from
continuously improving transportation and communications. The socio-economic environment
became turbulent: one in which predictability of
the future diminished significantly and the only
equilibrium that could be obtained was dynamic,
like that or an airplane fiying through a storm.
These changes undermined whatever effectiveness had been obtained by applying organismic
models to social systems; centralized control and
the treatment of subordinates as mindless parts
were no longer good practices.
Increasingly, employees could do their jobs
better than their bosses, but only if they were
given the freedom to do so. Therefore, the
mechanistic and organismic concept of management as command and control, or even softer
supervision, became less and less appropriate.
The functions of management became one of
enabling and motivating subordinates to do as
well as they know how, to develop them so they
can do better in the future than the best they can
do now, to manage their interactions, not actions,
and manage the interactions of the unit managed
with other internal and external organizations.
This can only be done with a social systemic
model in mind.
Furthermore, the unprecedented generation
and distribution of both wealth and knowledge
20

Systems Research

resulted in increasing choice and greater interdependency. This changed the nature of social
settings and individual behavior. The greater the
interactions and interdependencies, the more
vulnerable social systems became to the actions
of a few. The more knowledge available, the
greater the value of communication and information. However, advances in information technology and communication did not yield the
quality or quantity of control managers hoped
for. Since it was assumed that members of
organismically conceptualized organizations
would behave like organs in a human body by
reacting mechanistically to information provided
by the brain, it appeared reasonable to conclude
that malfunctioning of organizations was due
either to the lack of information or noise in the
communication channels. Therefore, more and
more information and better and better communications were provided. Unfortunately, this
mode of thinking is ineffective in deaUng with
the complexities of increasing social interactions
and interdependencies. It fails to recognize that
members of an organization, unlike the parts of
an organism, have a choice and do not react
passively to the information they receive. Imagine a thermostat that developed a mind of its
own. When it received information about the
temperature in the room that it did not like, it
would not react to it. This would result in a
chaotic air-conditioning system. The effectiveness of a servo-mechanism is based on the fact
that it does not have a choice and can only react
in a predefined manner to the events in its
environment. Our organsheart, lungs, and so
oncannot dedde on their own not to work for
us. Even when they are defective, we do not
conclude that they 'are out to get us'.
Furthermore, increases in information eventually produces a condition Meier (1963) called
'information overload'. As the amount of information received increases beyond the amount its
receivers can handle effectively, they use less and
less of it. Not only do receivers become saturated
with informationand therefore cannot receive
any morebut they can and do become supersaturateddiscard some of the information they
already have.
An organization with purposeful parts almost
inevitably generates internal conflict. Wherever
there is choice, conflict is likely; without choice.
Russell L. Ackoff and Jamshid Gharajedaghi

Systems Research

there can be no conflict. In conflict situations,

organismic thinking is ineffective because it tries
to resolve conflict by increasing the flow of
information between the conflicting parties.
Unfortunately, when conflict is based on differing values or scarcity of resources, an increased
flow of information, contrary to conventional
wisdom, does not improve but aggravates the
conflict. For example, the more information
enemies at war have about each other, the more
harm they can inflict on each other.
However, the biological mode of organization
can be successful: in the short term, in the
particular context of paternalistic cultures;
where loyalty, conformity, and commitment are
considered to be the core virtues. These virtues
are reinforced by the security of belonging to a
group which in turn protects and provides for its
members. For example, Japan, an industrialized
society, with a relatively strong paternalistic
culture, closely approximates an organismic
system. Therefore, it has been able to capitalize
more effectively on the strength of the biological
model of organization. In the context of a strong
paternalistic culture, conflict can be resolved by
the intervention of a strong father figure, whose
command, 'Give the apple to your sister', would
be respected without much hard feelings. To
appreciate the power of this type of leader, recall
that such American corporate giants as Ford,
DuPont, General Motors, and IBM owe much to
their paternalistic founding fathers.
The nature of highly developed social systems
is fundamentally different from that of a paternalistic culture. Members of societies that have
matured past the secure and unifying umbrella
of a paternalistic culture, insist on the right to
make choices. But there is a price to be paid for
this right; it can induce insecurity and conflict.
Purposeful actors, individually or in groups,
who pursue incompatible ends and/or employ
conflicting means, generate conflict. Consequently, because of its organismic orientation,
corporate America is ill equipped to deal
effectively with internal and external conflict.
Furthermore, it finds it almost impossible to
make the changes required to flourish in its
rapidly changing and increasingly complex
environment. A significant part of its energy is
wasted on futile efforts to deal with such conflict.
The frustration that results reinforces its inability
Reflections on Systems and their iVlodeis

RESEARCH PAPER

to change. This in turn creates a feeling of

impotency and hopelessness that immobilizes
Western governments, institutions and organizations.
A conflict-free organization can be created by
reducing choice, reducing members to robots.
Fascist societies and autocratic organizations
have attempted to approximate such a state.
Such systems are dehumanizing and, over time,
result in reduced productivity of the workforce
and reduced quality of its outputs. This in turn
produces a precipitous decline of an economy as
is occurring in many Western nations. On the
other hand, relying exclusivelyas organismic
modelers of organizations doon an increasing
flow of information and compromise to reduce
confUct does not produce encouraging results.
Witness the situation in the UN, which has
dramatically increased the flow of information
between nations and compromise among them.
Therefore, the challenge before us is to create
a type of organization that is capable of
continuously dissolving conflict while increasing
choice. This requires an organizational concept
that is not compatible with either a deterministic
or animate model of organization. It requires
application of social-systemic models to social
systems.
THE SOCIAL SYSTEMIC ORGANIZATION
In other places we (Gharajedaghi, 1985, 1986;
Ackoff, 1981, 1994) have proposed an organizational design that is based on a social-systemic
model. It has the following features, none of
which are compatible with other than a socialsystemic model:
(1) It is a democratic organization, one in which
every individual who is affected by what that
organization does has a voice in deciding
what it does, and in which anyone who has
authority over others taken individually, is
subject to their collective authority.
(2) It has an internal market economy, one in which
every part of the organization can purchase
the goods and/or services it requires from
any internal or external source it chooses,
and can sell its output to any buyer it wants.
Both these types of decision, buying and
selling, are subject to overrides by higher
21

RESEARCH PAPER

authorities who must, nevertheless, compensate the part of the organization affected for
its loss of income or increased costs due to
the higher-level intervention.
(3) It has a multidimensional organizational structure,

one in which units of three different types are

located at each level of the organization: units
defined by (a) their function (i.e., units whose
output is primarily consumed internally), (b)
their output (product or services primarily
consumed externally), and (c) their users
(markets, defined by type or location of
customers). This type of organization eliminates the need for continual restructuring.
Restructuring is replaced by reaUocation of
resources.
(4) It uses interactive planning which involves
idealized redesign of the organization, and
determination of the closest approximation
to that design that can be realized. Such
planning then involves selection of the
means by which the approximation is to be
pursued; provision of the resources required
by the pursuit; specifying the implementation steps to be taken, when and by whom;
and finally, design of monitoring and controls of both the implementation and the
effects of the plan.
(5) It contains a decision support system that
facilitates learning and adaptation by (a)
recording the expectations associated with
each decision of significance, (b) the assumptions and information on which they are
based, and (c) the process by which the
decision was reached, and by whom. It then
monitors the implementation, assumptions,
and effects of every decision, corrects them
where assumptions turn out to be wrong or
expectations are not met, and retains in an
easily accessible memory what has been
learned. Finally, it carries out continuous
surveillance of the environment to detect
changes that have occurred or are about to
occur that require adaptation by the organization.
Any one or subset of these changes can significantly improve organizational performance.
However, when all are made together, there is a
powerful multiplicative effect, one that is much
greater than the sum of its parts.
22

Systems Research

CONCLUSION
We have argued that it is useful to cast systems
and their models into one of three tj^es:
deterministic, animate, and social-systemic. The
difference between them is a matter of 'choice'.
Deterministic systems and their parts display no
choice. Animate systems can display choice but
their parts can't. Social-systemic systems display
choice, their parts do as well, and they are part of
larger systems that also display choice and
contain other systems that do so as well.
Our point has been that when models of one
type are applied to systems of a different type,
at least as much harm is done as good. The
amount of harm (hence good) that is done
depends on the level of maturity that social
systems have reached.
Our society and the principal private and
public organizations and institutions that it
contains have reached a level of maturity that
eliminates whatever effectiveness applying
deterministic and animalistic models to social
systems may once have had. Finally, we showed
five characteristics that we believe social systems
designed as social systems should have in order
to function as effectively as possible.

REFERENCES
Ackoff, R.L. (1981). Creating the Corporate Future, Wiley,
New York.
Ackoff, R.L. (1994). The Democratic Corporation, Oxford
University Press, New York.
Beer, S. (1972). The Brain of the Firm, Allen Lane/
Penguin Press, London.
Flower, E.F. (1942). Two Applications of Logic to
Biology. In Clarke, F.P., and Nahm, M. (eds).
Philosophical Essays in Honor of Edgar Arthur Singer,
Jr., University of Pennsylvania Press, Philadelphia.
Forrester, J.W. (1961). Industrial Dynamics, WrightAllen Press, Cambridge.
Forrester, J.W. (1971). WorM Dynamics, Wright-Allen
Press, Cambridge.
Gharajedaghi, J. (1985). Toward a Systems Theory of
Organization, Intersystems Publications, Seaside, CA.
Charajedaghi, J. (1986). A Prologue to National Development Planning, Greenwood Press, New York.
Hussong, A.M. (1931). An Analysis of the Group Mind,
Doctoral Dissertation, University of Pennsylvania.
Meier, R.L. (1963). Communication overload: proposals from the study of a university library. Administrative Science Quarterly 7, 521-544.
Russell L. Ackoff and Jamshid Gharajedaghi

Systems Research
Sorokin, P. (1928). Contemporary Sociological Theories,
Harper, New York.
Work in America: Report of a Special Task Force to the
Secretary of Health, Education, and Welfare, MIT Press,
Cambridge, MA.

Reflections on Systems and their Models

RESEARCH PAPER
Zeleny, M. (ed.) (1981). Autopoiesis: A theory of Living
Organization, Elsevier, New York.

23