General system theory
Copyright 2017 Graham Berrisford. One of several hundred papers at http://avancier.website. Last updated 15/10/2018 10:52
The first of the general system theorists was Ludwig von Bertalanffy.
This paper introduces some of the ideas in his book “General System theory: Foundations, Development, Applications” (1968).
Along with some ideas of Kenneth Boulding and Anatol Rapaport.
Bear in mind, system theorists distinguish abstract system descriptions from concrete entities that instantiate (realise) them.
A system description is a complex type; it symbolises or idealises both the structures and the behaviors of any entity that realises the system.
General system theory
Abstract / theoretical systems
<create and use> <idealise>
System theorists <observe & envisage> Concrete / empirical systems
The 1954 meeting of the American Association for the Advancement of Science in California was notable.
Four people at that meeting conceived a society for the development of General System Theory.
They included three thinkers whose ideas are introduced below.
· Ludwig von Bertalanffy (1901-1972) the cross-science notion of a system
· Kenneth Boulding (1910-1993) applying general system theory to “management science”.
· Anatol Rapoport (1911 to 2007) game theory and social network analysis.
Ludwig von Bertalanffy (1901-1972) was a biologist who promoted the idea of a general system theory.
His aim was to discover patterns and elucidate principles common to systems in every scientific discipline, at every level of nesting.
He looked for concepts and principles applicable to several disciplines or domains of knowledge rather than to one.
“There exist models, principles, and laws that apply to generalized systems or their subclasses, irrespective of their particular kind.”
This section outlines some of Bertalanffy’s ideas about systems in general.
It starts with information flow, because that relates system theory to the ideas already discussed.
Bertalanffy related system theory to communication of information between the parts of a system and across its boundary.
“Another development which is closely connected with system theory is that of… communication.
The general notion in communication theory is that of information.
A second central concept of the theory of communication and control is that of feedback.”
“Every living organism is essentially an open system. It maintains itself in a continuous inflow and outflow…”
So, general system theory incorporates cybernetic concepts such as:
· System environment: the world outside the system of interest.
· System boundary: a line (physical or logical) that separates a system from is environment.
· System interface: a description of inputs and outputs that cross the system boundary.
· System state: the current structure or variables of a system, which changes over time.
Read Introduction to Cybernetics for more.
In “General System theory: Foundations, Development, Applications” (1968), Bertalanffy wrote:
“General System Theory… is a general science of wholeness… systems [are] not understandable by investigation of their respective parts in isolation.”
He said the properties of a whole system “emerge” from interactions between its parts (say, a bicycle and its rider).
Bertalanffy encouraged people to take a holistic view of a system.
Holistic view: a description of how parts relate, interact or cooperate in a whole.
To a biologist, the boundary of the “whole” system (an organism) may seem obvious
To a sociologist, the boundary of the system is often debatable.
Some contrast holism with reductionism.
Reductionist view: identifying the parts of a whole, naming or describing parts without considering how the parts are related in the whole.
E.g. listing the organs and limbs of the human body without relating them.
There is a difficulty with discussing holism and reductionism.
The scope of the "whole" system is a matter of choice; and so too is the granularity of a “part”.
Bertalanffy may have promoted holism (studying how organs cooperate to the benefit of the body) and deprecated reductionism (studying organs in isolation).
However, when you study a subsystem in isolation, it is the whole system of interest to you.
And when you study how two systems (e.g. heart and lungs) are related, they become parts of a wider system.
In practice, people flip between holistic and reductionist views of things.
An emergent property is a behavior or structure of a whole that depends on interactions between its parts.
Bertalanffy said the properties of a whole system “emerge” from interactions between its parts.
Consider for example the forward motion of a cyclist on a bicycle, or the V shape of a flight of geese.
If you had never seen a bicycle ridden before, the forward motion of bicycle and rider would be a surprise.
But emergent does not mean unwanted or unexpected.
That forward motion was wanted and expected by the bicycle designer.
The requirements for any designed system must include emergent properties.
Note that a cross-boundary event that is external to a smaller system is internal to a larger system
And the emergent properties of a small system are ordinary properties of any larger system it is a part of.
Read Holism and emergent properties for more.
There are natural hierarchies in animal societies.
Animals toward of the top of a hierarchy are more competent in some way.
At the very least, they had whatever competence is needed to win the competition to get the top.
(Marxists tend to see hierarchies as power structures in which those higher oppress those lower.
This can lead to denying that hierarchies can be beneficial and to denying the concept of competence.)
We often design artificial hierarchies to help us describe and manage large and complex realities.
A hierarchy can be built from the top down by successively decomposing one node into several nodes.
E.g. The human body might be divided from the top down into organs, limbs and other structures.
A hierarchy can be built from the bottom up by successively composing several nodes under a higher node.
E.g. The human body can be seen collection of interacting cells; which are grouped into successively larger subsystems.
The designer of a human or computer system has to strike a balance between centralisation and distribution of process control.
Centralisation implies some kind of management hierarchy; and distribution implies its opposite – an anarchy, or a network.
For sure, there are many reasons why management hierarchies can become inefficient and inept bureaucracies.
Start with Parkinson's law and the Peter principle.
Add the difficulty of recruiting, motivating and retaining employees to do the most boring and/or difficult work
Add the impossibility of top-level managers knowing enough to do much better than random in big decision making
Add the “unintended consequences” that arise from setting targets and imposing them in a top-down manner.
But that doesn’t mean there is no advantage to, or need for, top-down management.
It has benefits as well as drawbacks; and there is a balance to be drawn between hierarchy and anarchy.
Read Hierarchical and network organisations for more.
Systems can be nested, and systems thinking can be recursive.
Given his background in biology, von Bertalanffy introduced the term organicism.
Organicism: the idea that systems are describable at multiple hierarchical levels.
He meant a system may be decomposable into subsystems, and/or composable (with others) into larger systems.
Another systems thinker, Russell Ackoff arranged aims, activities and actors in hierarchical structures.
He used different words for similar concepts at different levels.
It can be convenient to do this, as illustrated in this table (not one of Ackoff’s).
But pinning different words to different levels of decomposition is arbitrary.
And trying to do it can obscure the general nature of system theory.
The concepts are the same at whatever level of system composition you choose to model.
A process is process: a sequence of actions that may refer to system state, include choices and produce outcomes.
A choice is a choice: whether it is made by strict or fuzzy logic, deterministically or by free will (if you consider those to be incompatible).
Today, a general system theorist need not embrace all Bertalanffy wrote.
The follow ideas, which von Bertalanffy supposed to be general, don’t appear in all systems.
It is intuitively obvious that some systems are simple and some are complex.
But the term “complex system” is used loosely, variously and with questionable meanings.
The first difficulty with the concept of complexity is the difficulty of measuring it; there is no agreed measure.
Ashby considered complexity in terms of counting possible system states – which is well nigh infinite for all but the most trivial system.
Bertalanffy considered complexity in terms of counting system elements.
“In dealing with complexes of 'elements', three different kinds of distinction may be made: according to their number; their species; the relations of elements.” von Bertalanffy
OK, we can count the kinds
of system element and the relationships between them, and/or count the instances of system elements in a
How to turn that information into a measure of complexity?
Scores of complexity measures have been proposed; here are two I invented:
· Complexity = the number of relationships divided by the number of elements.
· Complexity = the number of event/state combinations * the average procedural complexity of the rules applied to one event/state combination.
Whatever complexity measure you choose, the important thing to note is this.
You can only measure the complexity of a concrete system by reference to a description of its structural and behavioral elements.
The second difficulty with the concept of complexity is the ambiguity of “complex system”.
Mathematicians speak of complex systems - though in some cases they mean a simple system that is unpredictable.
Some mean a system that behaves in a non-linear or stochastic way - with a complex state change trajectory.
But even a very simple deterministic system can behave in non-linear or stochastic way.
Sociologists speak of complex systems – though they often mean a complex entity rather than a system.
It is important in systems thinking discussion to distinguish description from reality.
A social system should be described in a way that enables us to test whether a real world social group instantiates it or not.
In a concrete system, actors perform activities to maintain or advance the state of the system or something in its environment.
In an abstract system, roles and rules describe the logic or laws that actors follow in performing activities.
The abstract system hides the infinite complexity of the actors and activities you may observe in the concrete system.
Any social group, though it is infinitely complex in reality, may realise one or more simple systems.
E.g. The recipe for a hamburger is a relatively simple – but every hamburger is infinitely complex.
E.g. The US constitution is relatively simple – but every real-world US government is infinitely complex.
Most of what exists and happens in a real world US government is not referred to the US constitution or predictable from it.
E.g. People speak of a social group or a business (like IBM) as a self-organising or complex adaptive system.
IBM surely does realise countless describable and testable systems.
But continual reorganisation undermines the notion of IBM being describable or testable as a system at all.
It is confuse the concept of a system actor with the concept of a system definer (an actor perhaps in a meta system).
Read Complexity for more.
Bertalanffy stretched his ideas into proposals about human psychology and the meaning of life.
“Life is not comfortable setting down in pre-ordained grooves of being; at its best, it is élan vital, inexorably driven towards higher forms of existence”.
It is possible Bertalanffy borrowed the idea of inexorable progress from Marxism.
The fact is, inexorable progress is not what one finds in nature.
Read Marxism, system theory and EA for a critique of this idea.
Read System Ideas for more on ideas traceable to von Bertlanffy’s writings.
Kenneth Boulding (1910-1993) was an economist who wrote in 1956 about applying general system theory to “management science”.
He described social systems at two levels, populations and individuals.
He said behaviors performed by an individual include:
· joining or leaving a population
· processing information and communicating meaning to others
· remembering and acting on mental images (internal state data)
· transcribing mental images into historical records and
· restoring system state to some kind of norm.
Boulding’s society might be described as group of humans fitting the definition in this table.
General system description
Boulding’s social system
A collection of active structures
that interact in regular behaviors
that maintain system state and/or
from/to the wider environment.
A population of individuals that interact by
processing information in the light of
mental images they remember, and
exchange messages to communicate meanings
to others and related populations.
Boulding’s essay prompts the question: What is a society?
How does an individual actor join or leave it?
How many societies can one individual be a member of?
(See earlier observations on Durkheim for more.)
Boulding on individuals as deterministic
Like many early systems thinkers, Boulding was concerned with how systems maintain their state.
Boulding presumed that a human actor, in a given state, will respond to a stimulus by acting in a predictable way.
He said the difficulty with applying classical system theory is that an actor’s state data (their “mental images”) is unknowable.
And since you cannot know the state of an actor, you cannot predict an actor’s response to an event.
Boulding didn’t mention that the state of a society is distributed between its individual members – which creates both theoretical and practical difficulties.
It makes it hard to maintain the integrity of a population, and difficult to collect management information about it.
Just as distributing a software system’s state data between objects makes it hard to maintain the integrity of the system, and collect management information.
Boulding on social systems as complex
Boulding classified systems into nine kinds, on a scale rising from “simple” to “complex”.
He placed social systems at level eight, below “transcendental” systems at level nine.
Boulding’s hierarchical classification is highly questionable.
It confuses entities with systems.
Entities in the real world are often more complex than systems they play roles in.
Consider these socio-technical systems: a game of chess, a game of tennis, an ATM or cash dispenser.
The human and computer entities who play roles in these systems are more complex than the systems.
It confuses composition with complication.
A system relates atomic structures/actors in larger structures, and atomic behaviors/actions in longer processes.
The complexity of the system lies in the complexity of those structures and processes.
We must, necessarily, ignore the internal structure of the “atomic elements”.
And by the way, there is no recognised way to measure the complexity of a thing.
“In dealing with complexes of 'elements', three different kinds of distinction may be made:
according to their number; their species; the relations of elements.” von Bertalanffy
Should we count the element and relationship types in the abstract system description?
Should we count the instances of those types in the concrete or operational system?
And then, how to combine those numbers in a measure of complexity? There is no agreed answer to that question.
Boulding on roles and actors
A human actor is not dedicated to one system; we all act in many roles in many systems.
So like Weber before him, Boulding suggested the essential parts of a social system might be roles rather than actors.
Boulding didn’t dwell on it, but role-centric and actor-centric views of a business are dramatically different.
And ever since Boulding, social systems thinking have tended to fudge the description/reality distinction.
Ashby told us it is meaningless to point to an entity and call it a system with no reference to a system description.
A system theory requires a philosophy of description and reality (as Chris Partridge proposed to me).
In a system description, the parts of the system are roles and rules.
In a social group that realises the description, the parts are performances of roles by actors.
Social groups are infinitely complex; but the systems they realise can be very simple.
Read Boulding’s ideas for more on his thoughts.
Anatol Rapoport (1911 to 2007) was a mathematical psychologist and biomathematician who made many contributions.
He pioneered the modeling of parasitism and symbiosis, researching cybernetic theory.
This gave a conceptual basis for his work on conflict and cooperation in social groups.
Game theory: cooperation and conflict resolution
In the 1980s, Rapoport won a computer tournament designed to further understanding of the ways in which cooperation could emerge through evolution.
He was recognized for his contribution to world peace through nuclear conflict restraint via his game theoretic models of psychological conflict resolution
Social network analysis
Rapoport showed that one can measure flows through large networks.
This enables learning about the speed of the distribution of resources, including information through a society, and what speeds or impedes these flows.
It is said that the most complex natural system is the human brain.
Today, unforeseen by Weiner, Ashby, Bertalanffy and Boulding, the most complex designed systems are software systems.
Generic system description
An object-oriented software system
A collection of active structures
that interact in regular behaviors
that maintain system state and/or
from/to the wider environment.
A population of objects that interact by
processing information in the light of
state data they remember, and
exchange messages to communicate
with other objects and system users.
Usually, a concrete activity system matches an abstract system description only well enough.
But a concrete software system matches its abstract system description perfectly.
At run time, it can only do what is described in its code – no more, no less.
Bertalanffy eventually committed to a book on General System Theory (1968) in which he said GST brings us “nearer the goal of the unity of science”.
The quotes below are drawn from selected passages, which you can find on this web page.
“There exist models, principles, and laws that apply to generalized systems or their subclasses, irrespective of their particular kind, the nature of their component elements.”
“conceptions appear in contemporary science that are concerned with what is somewhat vaguely termed 'wholeness'.
I.e. problems of organization, phenomena not resolvable into local events, dynamic interactions manifest in difference of behavior of parts when isolated or in a higher configuration, etc.
In short, 'systems' of various order not understandable by investigation of their respective parts in isolation.”
“General System Theory… is a general science of 'wholeness'.”
“Closed and Open Systems: Every living organism is essentially an open system. It maintains itself in a continuous inflow and outflow…”
“Information and Feedback: Another development which is closely connected with system theory is that of… communication.
The general notion in communication theory is that of information. A second central concept of the theory of communication and control is that of feedback.”
“Causality and Teleology: You cannot conceive of a living organism, without taking into account what variously and rather loosely is called adaptiveness, purposiveness, goal-seeking and the like.”
“The System Concept: In dealing with complexes of 'elements', three different kinds of distinction may be made: (1) according to their number; (2) according to their species; (3) according to the relations of elements.”
Other sources say:
“Systems theory is the interdisciplinary study of systems in general, with the goal of elucidating principles that can be applied to all types of systems at all nesting levels in all fields of research.
The term does not yet have a well-established, precise meaning.” Wikipedia
Larry L. Constantine
“Some scholars consider general system theory to be broader than a theory, but rather an alternative Weltanschauung—a unique worldview” (Ruben & Kim, 1975).”
Not everybody accepted that GST is valuable.
The schools of systems thinking have different roots and perspectives; different schools hold sway in different regions.
“General system theory, like other innovative frameworks of thought, passed through phases of ridicule and neglect. (Laszlo and Krippner)
A respectable summary of GST is quoted below:
"von Bertalanffy.emphasized that real systems are open to, and interact with, their environments, and that they can acquire qualitatively new properties through emergence, resulting in continual evolution.
than reducing an entity (e.g. the human body) to the properties of its parts or
elements (e.g. organs or cells), systems theory focuses on the arrangement of
and relations between the parts which connect them into a whole (cf. holism).
This particular organization determines a system, which is independent of the concrete substance of the elements (e.g. particles, cells, transistors, people, etc).
Systems concepts include: system-environment boundary, input, output, process, state, hierarchy, goal-directedness, and information." Principia Cybernetica Web
“General System Theory” (1956) von Bertalanffy.
“General System theory: Foundations, Development, Applications” (1968) von Bertalanffy.
Read System Ideas for more ideas associated with general system theory.
All free-to-read materials at http://avancier.website are paid for out of income from Avancier’s training courses and methods licences.
If you find the web site helpful, please spread the word and link to avancier.website in whichever social media you use.