Luhmann’s Autopoietic Social System (after Seidl): a critique

This page is published under the terms of the licence summarized in the footnote.


Biological organisms and human societies clearly have special properties of their own.

Does biology or sociology provide a general model or metaphor for enterprise architecture?

This is one of a pair papers looking at the notion of autopoiesis in biology and sociology, in relation to enterprise business systems.


Maturana’s Autopoietic Biological System

This first paper suggests enterprises are not well described in Maturana’s terms as autopoietic or operationally closed.


Luhmann’s Autopoietic Social System (as described by Seidl)

The paper shows that Luhmann gives radically different meanings to “autopoietic”, “social” and “system” from other sources.

It looks like only two of Luhmann’s many concepts can reasonably be applied to enterprise architecture.


This paper does not reject Luhmann's theory per se; it is an interesting construction, impressive in some ways.
But is it useful? It is incompatible with the more general system theory that underpins enterprise architecture (EA).


Preface. 2

Luhmann’s concepts. 2

Luhmann’s re-interpretation of autopoiesis. 3

Luhmann’s re-interpretation of society. 3

Luhmann’s re-interpretation of system.. 4

Luhmann’s concepts. 4

Luhmann’s view of organisations. 7

How Luhmann departs from other organisation theory. 8

How Luhmann departs from enterprise architecture. 10

Summary mapping of 12 concepts to EA.. 13

Further discussion of incompatible views of a system.. 14

Conclusions and remarks. 15

More quotes from Seidl 16

b) Interpenetration: the relation between social and psychic systems. 16

c. Communication and action. 18

Seidl’s references. 18



“Systems concepts include: system-environment boundary, input, output, process, state, hierarchy, goal-directedness, and information.”     Principia Cybernetica

“Though it grew out of organismic biology, general system theory soon branched into most of the humanities.” Laszlo and Krippner.

Kenneth Boulding’s essay in the journal of Management Science 1956 explicitly applied general system theory to social entities.

Since then, various gurus have looked at social organisations as systems of actors and actions.


The various schools of sociology address questions about society and the individual.

There is a history running from Weber, Mead, Durkheim, Pareto and Tarde, through Parsons to Luhmann and Habermas.

They ask: how does society relate to the individual? What are the sociological origins of modern law, states, organisations and economics?

David Seidl (reference below) says 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 [actors] or actions.”


Luhmann proposed the components of a social system are communication events rather than actors or actions.

The aim here is to look at Luhmann's work from outside its context in sociology.

To look at his ideas from the stand point of general systems theory, enterprise architecture and business systems.


Most see a social system as a group of actors who interact.

Actors communicate by creating/sending and receiving/interpreting messages.

The normal presumption is that message receivers find the meanings created by message senders.

But (as I understand it) few if any normal presumptions hold in Luhmann’s social system.

A physical message only becomes a communication event when a meaning is extracted from that message by a receiving actor.


Luhmann’s social system is composed of all events related to one "code" such as justice/injustice.

If one message is read has having different meanings, then each of these different communication events belongs in a different system.

Physical messages may be retained in some kind of memory, but there is no persistent record of the communication events in one system.


Luhmann’s social system is the purest of pure abstractions, it exists only in moments when receiving actors find a meaning in a message and act on it.

How could anybody demonstrate the existence of such an autopoietic social system?

I cannot conceive of a test that could confirm or disprove its existence in the world. Can you?


Luhmann’s concepts

We ought to be clear about what Luhmann said before we can discuss his theory’s applicability.

The paper follows and summarises David Seidl’s explanation of Luhmann’s concepts.

You may well find it helpful to print out Seidl’s paper and read it alongside this one.

If I have misunderstood Seidl’s summary of Luhmann’s concepts, or if my analysis of those concepts is illogical, I will welcome redirection.

Luhmann’s re-interpretation of autopoiesis

For an introduction to autopoiesis and operational closure read Autopoietic (Biological) Systems.

Luhmann was a radical; Luhmann changed the meaning of autopoiesis.

“Luhmann did not apply the original concept directly to the social domain but (in line with the general systems tradition) tried to abstract from the originally biological concept of autopoiesis.” Seidl

In his more abstract view, autopoiesis means only this: the elements of a system are reproduced by the elements of the system.

Luhmann’s re-interpretation of society

Luhmann’s rejected the idea that social systems contain people.

“The first decision Luhmann as a theoretician had to make for constructing his general sociological theory of autopoiesis is what to treat as the basic elements of the social system.

The sociological tradition suggests two alternatives: either persons or actions.

Luhmann rejects both as incompatible with the concept of autopoietic social systems.

Instead, he chooses a completely different element… the communicative event; promoting a 'conceptual revolution' (Luhmann 1986: 178).”


Many consider how a persistent system components responds to transient events.

What makes Luhmann’s view different is that his social systems contain nothing but events.

Luhmann’s social system not only depends on communication events, it is a set of communication events.

The social system elements are communication events about a specific "code" such as justice/injustice.

All the system elements are related to that code, and therefore related to each other.


Luhmann deconstructed society by separating people from communications.
His social systems contain only communicative events, not the people who utter or receive the communications, or their thoughts

“The relation between social system and 'human being' is a very controversial aspect of Luhmann's theory; it is also the most misunderstood aspect.

For an adequate appreciation it is thus necessary to carefully outline this relation.

In Luhmann's theory the 'human being' has to be understood as a conglomerate of organic and psychic systems.”

“Although psychic systems trigger communication processes and vice versa - we repeat this point, since it is very important – the processes of the psychic system and the social system do not overlap in any way.”

“Although psychic systems are necessarily involved in bringing about communication, the communication (as this unit) cannot be understood as the product of any particular psychic system.”

“What the ‘involved’ psychic systems think during the communication processes, i.e. how the psychic systems understand the communication, is (at first) completely irrelevant for the communication.”


To Luhmann, communication senders and receivers (such as people) are considered to be psychic systems

They intersect with a social system, but are not part of it.

Luhmann’s re-interpretation of system

Traditionally, a system features both entities and events, both actors and activities, both people and processes

But in Luhmann’s social system, there are no physical or persistent elements.

Luhmann shifted the focus from a system of persistent entities to a system of transient events.


But Luhmann said: Communication is the only genuinely social operation”  

His kind of social system is composed of only of communication events.

A communication event involves two or more psychic systems (think people).

A communication event:

·        can be accepted or rejected,

·        can lead to further communication,

·        may convey or lead to decisions.


In Luhmann’s view, only communication can produce communication.

A social structure creates expectations that lead to communications, and steer communications, which in turn tend to re-reinforce the expectations of that society.

Luhmann’s concepts

Below are 12 concepts in Luhmann’s view of an autopoietic social system.

They are presented, for convenience of comparison, in much the same order as Seidl presents them.

Note that there is some overlap between concepts as Seidl presents them.


1. An autopoietic system sustains itself at an elementary level

To Luhmann, an autopoietic system reproduces its elements through its own elements.

Luhmann distinguishes three kinds of autopoietic system by their different kinds of element.

“Living systems - reproduce themselves on the basis of life

Social systems - reproduce themselves on the basis of communication

Psychic systems - reproduce themselves on the basis of consciousness or thoughts.”


2 An autopoietic system is operatively closed

Operational closure means a system works alone; it has no understanding of the way other systems work or perceive their environment.

Resources from their environment do not become part of the systems' operation.

No operations can enter or leave the system (though matter, information and energy can).

Maturana and Varela argue that operative closure is a precondition for interactional openness (see trigger causality).


3. A system cannot know its environment

Meaning or knowledge is always a human construction.

Every cognition (idea) is a construct of a cognitive system.

It does not in any way reflect any kind of external reality.


4. Trigger causality

The autopoietic system does not receive any directions or operations from the environment.

It receives only perturbations (or irritations), which can trigger internal operations in the system.

In other words, external events may trigger internal processes but they cannot determine those processes.


5. Self-organisation

An autopoietic system is self-organising and “self-organisation refers to the determination of structures”.

Actually, the term self-sustaining would be more accurate than self-organising.

This means the system’s structures are determined by the system, are a product of the system.


6. Structural coupling

The common view is that a system is structurally coupled to its environment.

This means the structures of the system are in some way or other ‘adjusted’ to the structures of the environment (or to selected entities and activities in that the environment).

This coupling enables the structures of the system to react to ‘important’ events in the environment.


7. The system has no persistent components

Luhmann radicalises the temporal aspect of autopoiesis.

Maturana and Varela considered the elements of biological systems as relatively stable chemical molecules.

By contrast, Luhmann’s system contains no stable elements – no psychic systems (people).

He considers the system elements to be communication events that are

fixed as points in time… occur only once and only in the briefest period necessary for their appearance.”


8. The system has no persistent memory structure

“Because the elements of the system have no duration the system is urged to constant production of new elements. If the autopoiesis stops the system disappears immediately.”


At first reading, Luhmann appears follow the biologists understanding that when an organism’s processes stop, the organism dies

Whereas a photocopier, for example, merely rests between processes.

However, I understand Luhmann to say the system disappears between communication events rather than dies.


The justice/injustice system does not die when people are not communicating about justice/injustice.

In any gap between communication utterance and response, the system has no substance, no memory.

But at any moment, an individual psychic system may come to the understanding that a past communication event is related to the justice/injustice code, then respond to that event by uttering a new communication - thus resuming that particular system.


9. Communication events have no meaning on their own

Luhmann deontologises the concept of system element.

The elements are communication events that become meaningful only through their integration with other elements in the system.

A communication event only acquires meaning after it has been received and used by an event receiver.


10. Communications are the elements [of the autopoietic social system]

The only elements of a social system are communications, which are recursively produced and reproduced by a network of communications, and which cannot exist outside of such a network. (Luhmann 1986: 174).


11. The meaning of a communication follows its use

It is the receiver who gives meaning to a communication event by using it.

Luhmann (1995a: 143) writes: “Communication is made possible, so to speak, from behind, contrary to the temporal course of the process.”  

This is also called “the  principle of hermeneutics: that not the speaker but the listener decides on the meaning of a message, since it is the latter whose understanding of the set of possibilities constrains the possible meaning of the message, no matter what the speaker may have had in mind.” (Baecker 2001: 66).


Re points 9, 10 and 11

There nothing "going on" inside Luhmann's social (communication) system.

It contains only logical communications about a theme (say justice). These are

·        not physical messages sent by living systems

·        not messages in the course of transmission from senders to receivers.

·        not thoughts in the psychic systems or memories of message senders or receivers.

A communication comes into existence only *after* a message transmission, when a message receiver has attributed that message to a theme.

So, one physical message may turn into two logical communications, in different systems, or never be given any meaning.

There is no record of a communication; there is no system state.

There is no point at which anybody knows what is inside the system.

The system boundary is undetectable.


12. Interpenetration: the relation between social and psychic systems

The psychic and social systems intersect, but their contents and boundaries are different.

Communications events take place between psychic systems (think people).

Luhmann defines ‘person’ as the ’social identification of a complex of expectations directed toward an individual human being’ (Luhmann 1995a: 210)

Psychic systems reproduce themselves on the basis of consciousness: only thoughts can produce thoughts.


Seidl says: "Although psychic systems trigger communication processes and vice versa - we repeat this point, since it is very important - the processes of the psychic system and the social system do not overlap in any way."  


How to make sense of this? If a social system contains only communications, then how can it contain processes as well? What components execute those processes? The utterance of a communication event is a process of the sending psychic system.

The receipt and understanding of the event is a process of a receiving psychic system.

What other processes are there in Luhmann’s social system?  I don’t know.

Luhmann’s view of organisations

In Luhmann’s view, organisations are social systems that reproduce themselves on the basis of decisions.

A decision is a communication.

It communicates (explicitly or implicitly) that other alternatives have been not been selected.


Organisations are systems that consist of decisions.

Organisations produce the decisions of which they consist through the decisions of which they consist.

Every decision is the product of earlier decisions and gives rise to ensuing decisions.


Luhmann reinterpreted not only the original idea of autopoiesis, but also the idea of what a system contains.

His kind of social system has no persistent state.

With no state at its centre, how to draw a boundary around what this kind of system includes? 

What does sit at the centre? Luhmann suggests a system is centred on the “code” of a functional subsystem.


In Luhmann’s view of history, social systems have evolved through four forms.

Model of society



division into equal subsystems, tribes, clans or families

Centre and periphery

division between city and country.


in the late medieval times a hierarchy of different social strata or classes

Functional subsystems

specialisms in modern society, e.g. law, science, economy, art,


The Domesday book suggests stratification may have occurred before cities.

And even the most primitive tribes had functional subsystems (hunters, gatherers, food preparers, carers, weapon makers, etc).

But Luhmann allows that each primary form can be combined with the other forms of differentiation on a secondary level.

And the historical sequence is not important here.


What matters here is Luhmann’s  focus on functional subsystems and the communications required to sustain them.

In Luhmann’s view, in the self-reproduction of a particular functional subsystem, all communications taking part ‘carry’ a specific code.

For example:


Communication code

legal system


economic system


system of science


political system



In Luhmann’s view, functional systems cannot ‘exchange’ their communications.

An economic communication cannot take part in the scientific system.

A political communication cannot take part in the legal system etc.


Hmm…  If two message receivers give different meanings to one communication event, then that one event can appear in different “functional subsystems”.

Is that compatible with Luhmann’s proposition that functional subsystems cannot ‘exchange’ communications?

How Luhmann departs from other organisation theory

General systems theory concepts are woven through all software and enterprise architecture methodology/framework developments over the last half century.

An enterprise is also a social organisation in the people/process sense introduced by Boulding.

An enterprise can be seen as a collection of individuals (animate components or roles) cooperating in behaviour (processes), to meet some goals of the system (of its individuals, owners, customers or other observers).

However, this definition of an enterprise as a social organisation does not fit the characteristics of an autopoietic social system as Luhmann defines it.


In management science, a social system is usually expected to contain some or all of

- an organisation of individual components who send and receive communication events

- processes for sending, receiving and understanding communication events

- a record of past communication events, or a summary thereof.


Luhmann's kind of social system contains only momentary communication events related to a specific meaning – a code.

The code (e.g. justice/injustice) is relevant to a "functional subsystem" (e.g. judicial system).

So, when Luhmann "radicalises the temporal aspect of autopoiesis", he distances his theory from more conventional organisation theory.



Ton van Reehnen: Luhmann says that communication is what creates social systems.


Graham: Luhmann takes that idea to a logical extreme by saying that:

·        A functional subsystem of human society comprises nothing else but communication events about a specific code (point 10 above). 

·        “If the autopoiesis stops the system disappears immediately.” (point 7 above).


Ton: Every utterance must be followed by another utterance as a reply in order to maintain the social system.

This second utterance shows how the former utterance is understood/interpreted/ giving meaning.


Graham: Yes (see point 11 above); given a communication event, Luhmann discards the sender’s meaning in favour of the receiver’s meaning.

The challenge then is to understand what kind of system is created out of nothing but communication events that have no meaning until they are used.

An utterance is not enough, there must be both utterer and utteree – a sender and receiver.

If an utterance has not yet been used by any receiver, then the system is in abeyance.

The system does not resume until a communication receiver replies in some way (and further, that reply is understood to be a reply on the same topic).


Ton: I think Luhmann’s theory of society starts from assumptions that differ fundamentally from other theories of society and organisations.

A basic difference in my view is the one between:

·        actor-based /person based theories (Talcott Parsons may be the most prominent) and

·        communication-based theories (Luhmann may be the most prominent representative).

In the end, those two families of social theory are incompatible.

Present day management science belongs to the actor-based family, generally speaking.


Graham: If Luhmann's social system contains only communication events, then it cannot contain any actors or people.

Perhaps there are three views of a human social system?


In the view of

a human social system contains elements that are

which are related by

Common man

human actors

family, authority or other structural relationships


human roles

cooperation in recognised processes.


inter-human communications

reference to the same code.


Luhmann's kind of system seems to sit outside of general system theory.

It has no persistent components.

Further, it is impossible to say which communication events are inside it and which are not, since the meaning of the events lies only in the minds of psychic systems that receive the events.

This means the system has no definable boundary; it has only a theme, the code of the "functional subsystem" in which psychic system receivers use the event.

How Luhmann departs from enterprise architecture

The analysis below suggests that an enterprise, as it is viewed in enterprise architecture, fits only 2 of Luhmann’s 12 points.


Consider the communication events in a process that runs from sale through delivery and invoice to payment.

This is a good example of how enterprise systems run contrary to Luhmann's ideas about a functional subsystem.

Enterprise systems exchange communication events with external entities.

A business communication exchange, such as invoice and payment, depends on sender and receiver having a shared understanding

They must both understand terms like order item description, delivery date, payment due amount and payment due date.

The goal is that invoice receiver pays the amount that was itemised on the invoice.

So, the success of the enterprise depends on invoice sender and invoice receiver finding the same meaning in the message. 


Even the social subsystem of an enterprise contains more than just communication events.

The meaning of a message is usually agreed before the message is sent.

Enterprise architects might formalise this shared meaning in a persistent data structure - a data dictionary or canonical data model.

(Which is to say there is a some structural coupling between sender and receiver systems.)


1. An autopoietic system sustains itself at an elementary level?

Businesses consume inputs (energy, information and materials) from their environment.

Furthermore, businesses import resources (component instances, component types and process types) from their environment.

Some businesses do manufacture some of their own resources, but no business makes all its own resources.

For more on why an enterprise is allopoietic, rather than autopoietic, read Autopoietic (Biological) Systems.


2 An autopoietic system is operatively closed?

A business is “operationally open” in the way outlined above.

Also, its actors are aware of components and processes in the wider environment.

For more on why a business is operationally open, read Autopoietic (Biological) Systems.


3. A system cannot know its environment?

In EA, by contrast: The enterprise will contain a model of its environment.

The enterprise does have to interpret the events it detects.

However, it must know its environment well enough to interact with it successfully.

It will contain a model of those entities and activities in the environment that the enterprise system is supposed to monitor, inform or direct.

And if the system works, then that internal model must reflect the reality of the external environment.


4. Trigger causality?

In EA, similarly: the environment will trigger or cause processes in our enterprise.

The enterprise will react to processes in its environment, in a way that can be modelled as a “discrete event dynamic system”.

The enterprise’s internal structures and processes will be designed to respond to events detected at the system boundary.

However, the internal processes are not entirely mysterious to an external observer.

E.g. an invoice, payment or change-of-address event leads to internal processes an external observer would expect.


5. Self-organisation?

This is tricky ground because people do not agree on the meaning of “self-organising” or its application to systems.

Maturana himself stated he would: "never use the notion of self-organization, because it cannot be the case... it is impossible. That is, if the organization of a thing changes, the thing changes"

Maturana, H. (1987). Everything is said by an observer. In Gaia, a Way of Knowing, edited by W. Thompson, Lindisfarne Press, Great Barrington, MA, pp. 65-82, p. 71. Quoted in Wikipedia


Luhmann might equate self-organising with self-sustaining.

But while biological entities do maintain their internal structures, they do not “determine” their structures in the sense that an architect does.

Nor can a biological entity determine to change the nature of its own structures and processes.


Another source suggests self-organising means a process that moves from disorder to order.

A process where order or coordination arises out of the local interactions between the components of an initially disordered system.

Spontaneous, not directed or controlled by any agent or subsystem inside or outside of the system.

However, the laws and its initial conditions of the process may have been chosen or caused by an agent.

It is often triggered by random fluctuations that are amplified by positive feedback.

The resulting organization is wholly decentralized or distributed over all the components of the system.

As such it is typically very robust and able to survive and self-repair substantial damage or perturbations.


Hmm… The enterprise may employ a meta system to organise itself.

An enterprise is not like a biological entity whose structures and processes are predetermined in DNA.

The extent to which people employed in everyday operational business systems determine the organisation of those systems is limited.

An enterprise likely employs specialists (perhaps called enterprise architects) to determine its business structures and processes.

These specialists and their processes may be viewed as a meta system that sits over and above the operational system.


6. Structural coupling?

In EA, similarly: The enterprise will be structurally coupled to its environment.

Structural coupling applies to all systems that involve information processing ("higher" systems in Boulding’s classification) and maintain internal state variables (after Ashby).

Structural coupling is at the heart of much systems thinking, be it in machine design, software design, biology, cybernetics or enterprise architecture.

Every business enterprise monitors, informs and directs entities and activities in its environment. 

The structures of the enterprise are in some way or other ‘adjusted’ to related entities and activities in the environment.

Events detected at the system boundary will trigger internal processes designed to keep the external and internal structures in step.


7. The system has no persistent components?

In EA, by contrast: The enterprise's components and processes are stateful. The enterprise will exist in operation, in state and in description.

If the event stream stops, then system processes stop, the system rests.

Machines will be idle; employees may be laid off.

However, the system persists between events.

It rests in its last updated state, along with any recorded memory, and any roles and processes (be they documented or remembered).

So, the enterprise can be restarted from its resting state.


8. The system has no persistent memory structure?

In EA, by contrast: The enterprise will remember past events and the current state of entities within the system and beyond, in its environment.

An event is a transient happening, detected at the system boundary.

The system will process transient events so as to maintain the system state in which data entities represent external entities and activities.

The system’s state persists and is continually updated in response events.

Any event may trigger an “atomic” state change inside the system. (It may also be logged, replayed or reversed.)


9. Communication events have no meaning on their own?

In EA, by contrast: The enterprise will draw the intended meaning from events.

Events are assumed to be meaningful per se.

The meaning of an invoice, payment or change-of-address event must be shared between the system and entities in its environment, if they are to cooperate effectively.

(See the discussion at the start of this section.)


10. Communications are the elements [of the autopoietic social system]?

In EA, by contrast: The enterprise’s architectural entities include also roles and processes.


11. The meaning of a communication follows its use?

In EA, by contrast: Enterprise communications are made meaningful through shared data schemas. (See the discussion at the start of this section.)

Message receivers do read, interpret and process messages in their own way.

However, communications are made meaningful through shared data definitions and translators. 

Purely social communications - interpreted by the receiver on a case by case basis - are outside of conventional EA.


12. Interpenetration: the relation between social and psychic systems?

In EA, by contrast: The boundary of the enterprise/system that contains human roles is the same boundary as that of the enterprise/system that sends and receives communication events.

The enterprise employs human actors to play human roles.

Human role types persist in the form of role and procedure descriptions.

Human actors persist in playing those roles – which is recognised by the use of instance identifiers (e.g. employee numbers) used by the actors.

Summary mapping of 12 concepts to EA

We don’t have to reject Luhmann's theory per se; it is an interesting construction, impressive in some ways.
But is it useful? It is incompatible with general system theory
It stands the original GST concept of a system on its head, and illustrates how far systems thinkers have departed from GST.


Only two of the concepts Luhmann uses to define an autopoietic social system naturally fit conventional enterprise architecture assumptions.

Luhmann’s concepts

Similar conventional EA assumptions

Trigger causality

The enterprise will react to processes in its environment

Structural coupling

The enterprise will be structurally coupled to its environment


The other ten concepts do not naturally fit enterprise architecture description.

Luhmann’s concepts

Contrasting conventional EA assumptions

An autopoietic system sustains itself at an elementary level.

The enterprise will import some components

An autopoietic system is operatively closed

The enterprise will import some processes

A system cannot know its environment

The enterprise will contain a model of its environment


The enterprise may employ a meta system to organise itself

The system has no persistent components

The enterprise has stateful components

The system has no persistent memory structure

The enterprise will exist in operation, in state and in description

Communication events have no meaning on their own.

The enterprise will draw the intended meaning from events

Communications as the elements

The enterprise’s architectural entities include roles and processes

The meaning of a communication follows its use

The enterprise’s communications are made meaningful through shared data schemas

Interpenetration: the relation between social and psychic systems

The enterprise identifies people as actors in roles


Some EA concepts are analogous to but actually different from Luhmann’s.

Further discussion of incompatible views of a system

You may think of the legal system as per general system theory or as Luhmann would have it, but not both at once, because they are very different kinds or systems.


Way of thinking


Legal system example

Allopoietic enterprise

People, processes and state

employs law makers and lawyers,

processes legal cases

maintains statutes

records cases and judgements

Autopoietic social system

Communication events

the binary code of justice/injustice.


If you  take the second view, then where are the components and processes that send and receive communication events? Again, you can think in two ways about this:


a)      The system is continually importing and exporting those components and processes (so the system is operationally open, which is to undermine Luhmann’s point on that).

b)      The system is stateless: it includes only transient communication events ; it excludes the components and processes the send and receive the events.


Luhmann advances the second interpretation.  So, does the system boundary momentarily contain one or more communication events? 


Unfortunately, there can be no record of events related to a functional subsystem, because, in Luhmann view,  a communication event does not identify the system it belongs to.

The meaning is not in the message per se. Meaning only emerges after a message has been interpreted and used in a network.


So, you cannot recognise and encapsulate communication events when they are sent.

You cannot even recognise them in retrospect, since that would require monitoring all message receivers to see how they interpret and use a message, and that (surely) implies making your own interpretation of the use.


In short, it appears the kind of system Luhmann describes has no persistent parts; it is stateless; and it is never at any moment in time separable from its environment or enclosable behind a boundary. It is hard to imagine a kind of system that is further removed from the core concepts of general system theory we started with.

Moreover, a general systems theory does not (cannot) assume that the components of a system are self-aware psychic entities - with all that implies.



Reader: Organisations exist in time. So they have a history.


Graham: Yes. They have a life, continuity, history. Ordinary organisations have a persistent state.

Their components can rest between processes.

They maintain a central record of essential variables, a shared memory, some kind of ledger or transaction history.

However, it appears Luhmann’s system has no persistent components or memory. Its history not recorded or remembered, so it cannot be used.


Reader: Organisations need to continually recreate themselves by acts of communication.


Graham: Yes. Ordinary human enterprises process a flow of information events, with a variety of meanings.

But they can rest. A choir may do nothing between rehearsals and performances.

Systems can rest and resume activity when the next input or time event triggers them into action, because they are able to pick up from the last recorded state. 

Similarly, in Luhmann’s social system, the flow of communication events can stop.

A communication receiver may reflect on the meaning of an event before replying. And during this period of reflection, the system has no substance.


In the moments of silence between communication events there is almost nothing.

All that persists is the code to the "functional subsystem" that communication events are related to.

There are no individuals or components waiting to process the next event. There is no record of past events. 

There are no current state variables awaiting update. 

In short, there is no "organisation" as it is more generally understood in organisation theory, management science or systems thinking.

Conclusions and remarks

General system ideas are applied every day, all over the world, in the modelling of systems of many kinds.

They are employed in defining and designing business systems.

Today, Ross Ashby would surely recognise general system theory concepts in

·        Discrete Event Dynamics Systems (DEDS)

·        Models built using “System Dynamics” (after Forrester)

·        Maturana’s autopoietic biological machine.

·        Software design

·        EA frameworks like TOGAF


This paper does not reject Luhmann's theory per se; it is an interesting construction, impressive in some ways.

His view of a social system might lead to some useful outcomes.

However, Luhmann's theory does seem to be academic kite flying.


Luhmann’sautopoietic” is unlike that of Maturana, only analogous with the original idea.

His “social system” is unlike social systems as seen in conventional management science.

Moreover, his “system” differs from that in general systems theory.

It stands the original GST concept of a system on its head, and illustrates how far systems thinkers have departed from GST.


More broadly speaking, there is no presumption in enterprise architecture that biology or sociology provides a model or metaphor for enterprise architecture

Biological organisms and human societies have such special properties of their own.


“The “paradigm” mode has its dangers.

Attempts to understand (pieces of) the world often start with metaphor:

one studies a … system which “looks like” the real object of interest,

but one does not insist on fundamental connections or derivations “from first principles”. Philip HOLMES

More quotes from Seidl

Below are more quotes from “Luhmann’s theory of autopoietic social systems” by David Seidl (Munich Business Research).

His research looks thorough, and convincing about what Luhmann meant.

b) Interpenetration: the relation between social and psychic systems

The relation between social system and 'human being' is a very controversial aspect of Luhmann's theory; it is also the most misunderstood.

The 'human being' is not conceptualised as forming a systemic unity.

Instead it has to be understood as a conglomerate of organic and psychic systems.

The former consists of biochemical elements, the latter of thoughts.


Both systems are operatively closed against each other:

neither can contribute elements to the other.

The systems are however structurally coupled; i.e. their respective structures are adjusted to each other in such a way as to allow mutual irritations.


Although the ‘human being’ does not constitute a systemic unity, the social system treats it as such: it constructs it as a person.

In other words, ‘persons’ do not exist as such – they are no systems – but they are a construct of the social system with which it refers to the conglomerate of organic and psychic systems.


Particularly important for the social system is the psychic system.

[Both] systems are meaning-constituted systems.

However, in contrast to social systems, the meaning events do not materialise as communications but as thoughts.

Psychic systems are not only closed with regard to other types of systems but also with regard to each other.

No psychic system has direct access to another psychic system; my thoughts can never enter your psychic system.

As operatively closed systems psychic and social systems constitute environment for each other: thoughts cannot become communications and communications cannot become thoughts.


Mutual influences are restricted to the structural level.

Both types of systems are structurally adapted to each other in a way which allows for mutual irritation.

Luhmann calls the specific structural coupling - interpenetration.


Without psychic systems social systems are impossible – and probably vice versa.

Every communicative event presupposes 'parallel' events in the psychic systems.

Because of their structural coupling social systems can expect their communications to cause irritations in the psychic systems and to receive irritations from the psychic systems.

They can, for example, count on psychic systems to trigger further communications after every communication.

Although psychic systems trigger communication processes and vice versa - we repeat this point, since it is very important - the processes of the psychic system and the social system do not overlap in any way.


The most important evolutionary achievement for the coupling of social and psychic systems is language [which] ensures that psychic systems are irritated through the communication processes.

Language is a purely social phenomenon (psychic systems do not think in language) but thought processes can be structured in a complementary way to language; particularly during communication processes: thoughts are broken down into equivalent chunks to those of sentences and words.


In other words, psychic processes are synchronised with communication processes and,

in this way, they 'know' when to contribute irritations to the communication process in order to make the reproduction of the social system possible.

Although Luhmann's strict distinction between social and psychic systems runs counter to our everyday beliefs and almost all social and psychological theories, it has one important theoretical advantage.

It allows for a concept of the social which is clearly distinguished from the psychological.

Consequently, social and psychic phenomena can be analysed in their own right.


The treatment of human beings as environment of the social system (and not as part of it), as Luhmann writes, does not mean that the human being is estimated as less important than traditionally.

Anyone who thinks so has not understood the paradigm change in systems theory.


I prefer and promote the universal usefulness of General System Theory.

In this context, I can't imagine a system with processes, but no components or state.

For processes in Luhmann's theory, I refer you to the Seidl paper below.

As I understand it, processes are constructs a social system attributes to actors.

An actor is a living system + a psychic system.

But I do find the explanation below rather baffling.

c. Communication and action

While Luhmann suggests treating communications - and not actions - as the elements of social systems, the concept of action does not become completely irrelevant. .

Often communication is treated as some kind of action; in this sense Habermas (1987), for example, speaks of 'communicative action'.

But Luhmann’s communication - and this is very important - is not a kind of action.

As explained above, communication is constituted as a synthesis of a threefold selection of utterance, information and understanding.

The concept of action cannot account for all three selections.

It might capture the first two selections but certainly not the third: understanding.

[T]he perfection of communication implies understanding and understanding is not part of the activity of the communicator and cannot be attributed to him. (Luhmann 1986: 178)

Thus, a central element of Luhmann’s concept of communication would be missing if interpreted as action.

Apart from that, the original intention of an action is not important for the communication.

For example, looking at one's watch might be understood as communicating one's boredom, although one only wants to know what time it is.

Luhmann suggests treating action as a (fictive) construct of social systems for observing, and communicating about, their communications: social systems observe their communications not as communications but as actions, which they causally attribute to ‘persons’ (‘actors’).

Seidl’s references

Most of this paper depends on my reading of Seidl’s translation and explanation of the following of Luhmann’s works, mostly published in German.


(2000) Organisation und Entscheidung. Opladen: Westdeutscher Verlag.

(1986) "The Autopoiesis of Social Systems." Pp. 172-92 in Sociocybernetic Paradoxes: Observation, Control and Evolution of Self-Steering Systems, eds. F. Geyer and J. Van d. Zeuwen. London: Sage.

(1992b). "Organisation." Pp. 165-85 in Rationalität, Macht Und Spiele in Organisationen, eds. W. Küpper and G. Ortmann. Opladen: Westdeutscher Verlag.

(1993a). Soziologische Aufklärung 3: Soziales System, Gesellschaft, Organisation. Opladen: Westdeutscher Verlag.

(1993b). Soziologische Aufklärung 5: Konstruktivistische Perspektiven. Opladen: Westdeutscher Verlag.

(1993c). "Die Paradoxie des Entscheidens." Verwaltungs-Archiv: Zeitschrift für Verwaltungslehre, Verwaltungsrecht und Verwaltungspolitik 84:287-310.

(1995a) Social Systems. Stanford: Stanford University Press.

(1995b). Soziologische Aufklärung 6: Die Soziologie Und Der Mensch. Opladen: Westdeutscher Verlag.

(1997). Die Gesellschaft der Gesellschaft. Frankfurt a.M.: Suhrkamp.

(2003) “Organization”. In Bakken, T./Hernes, T. (eds.) Autopoietic Organization Theory. Drawing on Niklas Luhmann’s Social Systems Perspective. Copenhagen et al.: Copenhagen Business School Press, pp. 31-52.


Published under the Creative Commons Attribution-No Derivative Works Licence 2.0                                          25/12/2016 14:53

Attribution: You may copy, distribute and display this copyrighted work only if you clearly credit “Avancier Limited:” at the start and include this footnote on each page . No Derivative Works: You may copy, distribute, display only complete and verbatim copies of this paper, not derivative works based upon it.. For more information about the licence, see