The schism in systems thinking

Copyright 2017 Graham Berrisford.

One of about 300 papers at Last updated 29/05/2017 17:15


Some suggest modern social systems thinking (SST) derives from, or is an advanced application of, general system theory (GST).

Actually, SST approaches both preceded and depart from GST in one or more of the ways listed below.


General system theory (GST)

Can be contrasted with approaches that are




Specific to situations in which humans interact

About system roles and rules

About individual actors (purposeful people)

About systems (S)

About meta systems (M)

About describing testable systems

About solving a problem or changing a situation


This paper explores the second, third and fourth differences.

It draws sharp distinction between actor and role-centric viewpoints:

·         Actor-centric view: A social entity is a group of physical actors - who communicate with each other.

·         Role-centric view: A social system is a group of logical roles - which need actors to play them.


The schism.. 1

When did the two schools of thought diverge?. 2

The ambiguity in more recent social system discussion. 2

The difference between actor and role-centric views. 3

The difference between social entities and systems. 4

How to reconcile SST with GST?. 5

Ambiguous terms. 7

General conclusions and remarks. 8

Footnotes. 9



The schism

In first order (role-centric) cybernetics, a system is a conceptualization of some regular behaviors as a set of roles and rules (or entities and events, or stocks and flows).

If you change the roles and rules, then you change the system they define.


In second order (actor-centric) cybernetics, a system is a named entity that employs actors to meet some overarching goal(s).

Actors are expected to do what is necessary, which can include changing the roles and rules of the organisation they work in.


An “organization” (meaning, a named entity) that continually changes its “organization” (meaning, its roles and rules) is often called a “complex adaptive system”

Yet in role-centric cybernetics, this named entity is the very antithesis of a system; it is rather an ever-unfolding process.

All that can be described or tested are goals given to the entity by some external observer(s).


This paper discusses this schism, and shows how to reconcile the two kinds of cybernetics.

When did the two schools of thought diverge?

David Seidl (2001) said the question facing a social system theorist is what to treat as the basic elements of a social system.

“The sociological tradition suggests two alternatives: either persons [think actors] or actions [think roles].”

It might be argued this division in social systems thinking was detectable the 19 century.


Gabriel Tarde (1843 to 1904) was a French sociologist who conceived sociology as based on small interactions among individual actors.

It is commonly assumed that actors perform activities to sustain themselves meet their higher level goals.

It is sometimes assumed that systems are homogeneous, meaning all actors of a kind follow the same rules (like geese in a flight of geese).

Special interests of this school include how interactions between actors may generate so-called “complex” or “chaotic” behavior at a group level.

And the possibility that actors may define their own roles and rules.


Max Weber (1864 to 1920) was a German sociologist who profoundly influenced social theory and social research.

He set out three essential principles for bureaucratic organisations, be they public or private:

1.      Roles: labour is divided between roles – defined by the required activities.

2.      Assignment of actors to roles: roles are performed by actors qualified and hired to play those roles.

3.      Hierarchy: regulations describe chains of command, describe both roles and the capacity to coerce others to perform roles.

Weber’s other principles included the supremacy of abstract rules in a social system.

In short, activities are determined by defined regulations, roles and rules (not by actors themselves).


The ambiguity in more recent social system discussion

Boulding and Ackoff were management scientists who sought to import general system theory (GST) into social system thinking (SST).

Unfortunately, both blurred the distinction between actor and role-centric views of systems.


Kenneth Boulding (1956) regarded individual actors as the units of a social system.

In the same paper, he suggested that, instead, the units could be seen as roles.

A role is defined by the actions to be performed in the role.

What realises a defined role is not an actor; it is an actor’s performance of the actions expected in that role.

Social system

Abstract system description


Concrete system realization

Assignments of actors to roles


In other words, a system unit is that tiny part of an actor’s time and energy dedicated to performing activities needed in the system of interest.

What an actor does outside their defined system role is not a part of that system, but might be part of another defined system.

E.g. a person’s singing is irrelevant to their role in a tennis club or poker school, but vital to their role in a choir.


Russell Ackoff (1971) was ambiguous in “System of System Concepts” (1971).

Ackoff began by endorsing Ashby’s and Checkland’s point that a system is a conceptual model of reality.

“Different observers of the same phenomena may conceptualise them into different systems” Ackoff

In other words, observers of one named social entity may describe it as different systems (or no system at all).

Later, he contradicted himself about the nature of systems.

“A church, a corporation or a government agency is a system”. Ackoff

In other words, one named social entity is one system, regardless of any observer or conceptualization.


Today, discussion of systems still confuses the two viewpoints that Seidl distinguished in 2001.

And attempts to merge GST and SST tend to obscure what each has to offer.

Much of SST might more accurately be called “Social Entity Thinking”.

The difference between actor and role-centric views

This paper draws sharp distinction between actor and role-centric viewpoints:

·         Actor-centric view: A social entity is a group of physical actors - who communicate with each other.

·         Role-centric view: A social system is a group of logical roles - which need actors to play them.


Actor-centric social systems thinking (SST)

This is about how individuals relate in social groups.

It is evident in some “management science” or “organisation theory”.

It is actor-centric; its puts individual actors (usually people) first.

In SST, a system is a set of actors who each have their own purposes, motivations, goals, values.

It is usually a human “organisation”, be it a nation, society or business.


Role-centric general system theory (GST)

GST emerged after the second world war; it is generic, meaning cross-science.

It focuses on deterministic systems that repeat behaviors.

It is activity-centric; a core principle is the primacy of behaviour.

It emphasises the behavior of the whole (holism).

In GST, a system is a set of roles and rules for acting, exchanging and remember information

Orderly or deterministic behaviors are performed by structural actors/components playing those roles.


It is common to describe GST and SST as sitting at different ends of a spectrum.

But that is not satisfactory, partly because that implies there is an optimal middle position.

And partly because GST and SST are different tools, usually applied by different people in different contexts.


Some assume modern “systems thinking” is progression from, or advanced form of, general (cross-science) system theory.

This paper presents it instead as a shift away from science towards “scientism”.

And away from a general approach to a variety of specific approaches related to situations in which humans interact.

And away from thinking about systems towards thinking about meta systems.

The difference between social entities and systems

This paper draws the following distinction between a social system and a social entity.


A social system (say a tennis club, a choir or a poker school) is a set of logical roles, which need actors to play them.

All the actors who play its roles can be replaced.


A social entity is a collection of physical actors who communicate with each other.

The same actors can play unrelated roles in (say) a tennis club, a choir and a poker school.


Pointing to a social entity and calling it a system does not make it a system.

A group of people doing things is not a system just because people call it a “system” or an “organisation”.

The US economy, a church or IBM is not a system.

It is as many different systems as system describers can successfully describe and test.

Some of those systems may conflict with each other, or undermine each other.

With no system description, to assert that IBM is a system conveys no meaning (beyond saying it exists as a named social entity).


In casual conversation, a choir or poker school can be seen from either actor or role-centric perspectives.

When speaking of a named entity, people flip from one view or another within the same conversation.

However, social entities and social systems are different things, and they are related in a many-to-many association.

One social system (set of roles) can be realised by many social entities (e.g. many beehives, choirs and football teams).

One social entity (set of actors) can act in several social systems (e.g. its members can play unrelated roles a choir and a poker school).


(Whenever a social entity’s actors follow given roles and rules, then that group acts as a social system.

Honey bees do this when they follow rules (they inherit) to watch another bee’s dance, read the message and find the pollen.

If you could persuade the same bees to follow different rules, the same social entity would act as a different social system.)


How to reconcile SST with GST?

The “primacy of behaviour” is a fundamental principle (even the principal innovation) of GST.

However, it is important to distinguish three kinds of process.


The processes that sustain a system

A biological organism is sustained by the processes of life: sensation, respiration, digestion etc.

GST primarily addresses systems with such regular, repeatable processes.


The process that is the lifetime of a system

A biological organism is transient, it follows a process from birth to death.

Every system described according to GST has a finite life time.

It is a transient island of stability in the ever-unfolding processes of the universe.


The processes that replace one system generation by the next

Biological processes replace one generation of transient organisms by the next (slightly different) generation.

Note that the natural selection process of evolution applies outside of biology, to commercial enterprises for example.

Any system designed according to GST principles may be changed from one generation to the next.


Complexities arise when the three processes above are entangled in the description of a system.

The processes of life in an individual organism are very different from the cross-generational process of evolution.


Bridging the schism

It is possible to recast actor-centric cybernetics in a way that maintains the integrity of the system concept.

To put it another way, GST can be extended to reconcile role-centric and actor-centric cybernetics.


First, replace continual system change by incremental (generation-by-generation) system change.

Second, separate the system S from the meta system M.

·         S contains the operational processes that sustain S or meet its goals

·         M contains the evolutionary processes that changes the system S from version N to version N+1.

Third, allow an actor to switch between roles in S and M.

A basis for incremental system evolution

The roles and rules of S

<defines and changes>                  <are realised by>

Meta system M     <observes and envisages>      System S


When actors change the roles and rules of a business or other social entity they work in, then they create a new and different system.

The name of the social entity remains the same, but the system changes.

There is a transition from system version N to version N+1.


Meta systems as weak systems

An operational business system may be a strong system – meaning its actual behaviors are very regular, deterministic and repeated

A human meta system is usually a weak system – meaning its actual behaviors are not so regular, deterministic or repeated.

Consider for example the meta system that is SSM, or an EA framework like TOGAF, or an agile development method like SCRUM

Such a meta system is a social entity in which actors sometimes pay only light attention to given roles and roles.


How EA employs systems thinking

An EA framework is a meta system (M) to any business system (S) it monitors and steers.

An EA framework guides architects in how to design and plan incremental (generation-by-generation) changes to business systems.

Change control requires there to be an abstract system description that is agreed by stakeholders.

And a migration plan for the work to transform a system from version N to version N+1.

(Somewhat comparable to a year-long migration plan, is the daily stand up meeting agile methods like SCRUM.)


Architects certainly use principles of role-centric cybernetics to design processes and information flows that sustain S or meet its goals.

Methodologists may use actor-centric cybernetics in designing the weaker M that is the EA function.

Actor-centric cybernetics?                          Role-centric cybernetics?

EA framework                                      Architecture definitions

<create>          <realised by>                <create and use>              <realised by>

Methodologists <observe and envisage> System architects <observe and envisage> Business systems

Ambiguous terms

Scientists need a well-defined vocabulary both to communicate effectively and to produce testable system descriptions.

Scientific language helps shape thinking and provides the means for constructing new theories and test cases.


GST is supposed to be about terms and concepts that are common to different domains of knowledge.

However, specific domains of knowledge have their own specific terms and concepts.

And SST tends to uses words with meanings that are specific human social systems.


SST discussion often users scientific-sounding terms like “non-linear”, “chaos”, “complex”, “adaptive” and “entropy”.

But is it not clear that these words mean the same as in mathematics or relatively hard sciences.

Taking a word from one science to mean something else in another science can be misleading.


This table shows some systems thinking terms that have different meanings in different contexts.

System term

One meaning

Alternative meaning


A logical boundary chosen by observers

A physical, spatial boundary (e.g. of a farm, or human being).


A state change made by system processes

A rule change made to a system description by a designer


The self-sustaining nature of a biological entity

The self-organising nature of human groups


The biologists Maturana and Varela characterised living entities as being different from others by being autopoietic.

This means self-sustaining; an autopoietic organism manufactures its own body parts from primitive edible chemicals.

Some now speak of autopoietic social or business systems, meaning something different.

For example, some mean self-organising, mean that people can change the aims, variables or rules of the business they work in.


The table below shows some of the words that mean different things in different discussions of systems.


In describing social systems as per GST

In describing social entities as per SST


A named collection of roles and rules.

A named group of interacting actors.


A role which is constrained by definition

An individual who may act outside any defined role.


A structure connecting roles in processes

A structure connecting actors in a command hierarchy


Actors perform activities.

Actors change their roles and rules.


Unpredictable (meaning the opposite of linear)

Random (meaning the opposite of orderly)


The measurable complexity of a system description.

The un-measurable complexity of a system’s operation.


Changing state to achieve “homeostasis”

Changing its own roles and rules (“continual evolution”)


A social entity (a group of interacting actors) is often called a "complex adaptive system."

In GST terms however, if its roles and rules are little or not at all defined, then it is barely a system or not a system at all.

An ever-changing entity cannot be tested as conforming to a system description, and might better be called an ongoing ever-unfolding process.

General conclusions and remarks

The universe and human existence are ever-unfolding processes in which we perceive discrete entities.

We regard some of those entities as systems; but what makes an entity a system?


The terms “systems thinking” and “system theory” sound as though they refer to the same thing.

But not far down the road, one reaches a fork between what this paper calls SST (about social entities) and GST (about systems in general).


GST is about abstract system descriptions that can be realised or instantiated by concrete actors in a testable operational system.

SST is about social entities (groups of communicating actors) with less regard to roles and rules.


Much casual systems thinking discussion boils down to: "It is useful to think of some part of the world as a set of inter-connected things.”

GST says more; it asserts that different sciences share many more general ideas about “systems”.

Read system properties for a list of properties generally ascribed to systems, with links to explanations of them.


Contrasting social systems and social entities

It turns out that the two schools have different views of what named system properties mean.

Both are about whole-to-part thinking and part-to-whole thinking, connecting parts so they work together.

But after that, they have divergent views of what a system is, and what its elements and its properties are.

Giles Dalton suggests you might think of the two schools as wave theory versus particle theory.

The table below draws some contrasts between them.

Social system thinking

Social entity thinking

A system is a collection of roles and rules.

Actors are assigned to roles

Activities are ordered to given ends.

Discrete state changes.

Discrete generation changes.

A system has a name and generation number.

A system is a group of actors (usually human).

Actors choose their roles.

Activities may be ad hoc, to ad hoc ends.

Continuous state changes.

Continuous role and rule changes.

A system has a name only; there are no discrete versions


Using terms with various meanings undermines the notion of a general system theory.


The schism runs deep; consider for example “emergent properties”.

These are properties of whole system that cannot be found in a part of it.

E.g. bicycle transportation requires both bicycle and rider, neither part has the bicycle transportation property of the combination.


One social systems thinker told me he tried to explain emergent properties to some engineers.

The engineers couldn’t understand what he was banging on about; and he couldn’t understand why they couldn’t understand.

He spoke of emergent behaviour as being unpredictable, surprising or mysterious (perhaps also “non-linear” or “chaotic”).

The engineers were thinking that emergent properties are the very purpose of their system design efforts!


There is an anti-determinism streak in some social systems thinking.

Other papers argue that the critical property of human social systems is not whether humans behave deterministically or not.

It is that actors can choose to switch between social entities, and switch between playing different roles in social systems.

Further, they can step outside any system they play a role in, and step up to a meta system in which they change the roles of the first system.

(As software developers do in following an agile system development method like SCRUM.)


Later papers include analysis of thinking about systems by Boulding, Ackoff, Churchman, Beer, Ackoff, Luhmann and Snowden.

Their observations, opinions and models cannot be merged into a consistent body of knowledge.

It is often difficult-to-impossible to use their models as predictor of social system behaviour.

At the extreme, some systems thinkers (notably Luhmann) turn general system theory on its head.



All free-to-read materials at 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 in whichever social media you use.