Maturana’s Autopoietic Biological System

(contrasted with the Allopoietic Enterprise)

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.


General system concepts. 1

Logical and physical boundaries. 2

Logical roles and physical actors. 3

Self-organising social groups. 3

The autopoietic biological entity. 4

The allopoietic enterprise. 5

Operational closure. 6

The operationally open enterprise. 6


General system concepts

“Enterprise architecture regards the enterprise as a system, or system of systems” (TOGAF 9.1)

A system is a bounded entity, separable from the rest of the universe.

It is divided into parts that are interrelated (directly or indirectly) else there would be two more systems.

But that is far short of defining the systems of interest in EA


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

In a social system, actors exchange information and perform other activities.

In a designed activity system, actors perform roles in processes that transform inputs into outputs, so as to meet the goals of sponsors/stakeholders.

EA is about designed activity systems, formalised social systems, where the principal actors are humans and information/communication technologies.


The term "autopoiesis" is a special (not general) system theory concept.

It refers to an activity system in which the outcome of system processes is to sustain or reproduce system components and so sustain the system itself.

Maturana insists autopoiesis is limited to – is a distinguishing feature of – natural, biological systems.


But “systems thinkers” often steal terms from one domain and use them in another.

Maturana sees the inputs and outputs of a biological system as matter and energy flows.

By contrast, many sociologists see the inputs and outputs of a social system as information flows.

Luhmann defined an autopoietic social system with reference to communication events about a concept.

And when others speak of an autopoietic social system, they may mean a self-organising group – a third idea.







Autpoietic life form



Autpoietic concept

Autpoietic social organisation


Sometimes, it is better to separate ideas that, though they may be analogous, are actually very different.

This is one idea

This is another

Logical system boundary

Physical boundary

Logical role

Physical actor

Self-organising social group

Self-sustaining (autopoietic) biological entity

Business actors hiring or making business resources

Biological entities generating their own elements

Evolution from one system generation to another

Adaptations made by a given system within a system generation

Logical and physical boundaries

The boundary of a biological entity (animal or plant) is physical – meaning it encloses a three-dimensional space.

When considering input/output flows, Maturana focused on matter and energy, rather than information.

An organism breaks down any complex materials it consumes into primitive chemicals before using them.

So it consumes and produces relatively simple chemicals (oxygen, water, proteins, lipids, and waste products).



Logical information flows

Physical matter and energy flows

Biologists focus on


Molecules. Chemical energy. Heat

Businesses focuses on

Complex data flows

Complex materials/goods


The boundary of a business system is logical; its components are distributed in space, connected now and then by flows.

When considering input/output flows, business architects focus on information, and sometimes material, but very rarely energy.

A business consumes and produces complex information flows: logical data structures, messages and contracts.

Some businesses consume and produce complex material flows: manufactured products, machines.

Logical roles and physical actors

A role defines a limited set of duties (activity types), performable by all actors in that role.

The role definition may include properties that actors need to play the role, but does not refer to any particular actor.




Logical type

Physical individual

Instantiated by actors

Embodies roles


An individual actor can be assigned to play several roles.
Generally speaking, an actor may use the same ability or resource in different roles.

(Much as objects of an OOP class use the same state data in different operations.)


Note that human actors are unconstrained, they often ignore given roles, and usually do more than is defined in any role.

Self-organising social groups

The roles in a social or business system may be divided between operational roles and meta system roles.

An operational system is defined by products and services delivered, and roles and processes that deliver them.

Operational system roles are performed by workers whose overarching aim is to ensure delivery of defined products and services.

Operational researchers observer operational systems, and consider changes to them.

(The term operational researcher is used here, where many might say enterprise or business architect.)



Role name

Actor  duties

Meta system

Operational researcher

Observes and envisages operational systems

Operational system

Operational worker

Works in an operational system


Operational systems are observed, envisaged and designed by the meta system.

Meta system roles are performed by operational researchers who redesign products and services, and roles and processes.

These roles may be played by actors (be they consultants or employees) who never work in operational system roles.

On the other hand, human actors can play roles in both operational system and meta system.


E.g. In an agile development project, there is a “stand up” meeting each morning.

In that meeting, the actors step up to play a role in the overarching meta system.

Then, having agreed an operational system change, they step down again to act in a changed role.

The meta system is stable, but the operational system changes every day (though not continually).


This switching between meta system and operational system may be iterative or appear seamless.

Nevertheless, actors are aware they are playing different roles; and play them at different times.

The autopoietic biological entity

"Autopoiesis" refers to a system in which the outcome of system processes is to sustain or reproduce system components and so sustain the system itself.

Three essential features of an autopoietic system are:

·         System components continually perform processes that sustain themselves

·         Self-sustaining processes occur within the same space occupied by the components that perform the processes.

·         The system’s components manufacture and sustain themselves from elementary inputs (proteins, lipids etc.)


“The theory of autopoiesis was developed by the two Chilean cognitive biologists Humberto Maturana and Francisco Varela in the sixties and early seventies.

They were trying to answer the question: What is life? Or: What distinguishes the living from the non-living?

Their answer was: A living system reproduces itself. This self-reproduction they referred to as autopoiesis.

They defined the autopoietic system as a system that recursively reproduces its elements through its own elements.” Seidl


Maturana defined a biological entity (e.g. a cell) as an autopoietic machine.

His original definition (in Autopoiesis and Cognition: the Realization of the Living - 1st edition 1973, 2nd 1980) is slightly simplified below.

An autopoietic machine is:


This definition can be seen as extending the basic concepts of general system theory.

First, it views a biological entity as a bounded system of components cooperating in processes.

Then, it adds that a living system is special because its processes and components are recursively self-sustaining.

In a circular way, living components perform processes that sustain the components that perform processes that.... and so on.


Today, Wikipedia says:

Autopoiesis was originally presented as a system description that was said to define and explain the nature of living systems.

A canonical example of an autopoietic system is the biological cell.

The eukaryotic cell, for example, is made of various biochemical components such as nucleic acids and proteins, and is organized into bounded structures such as the cell nucleus, various organelles, a cell membrane and cytoskeleton. These structures, based on an external flow of molecules and energy, produce the components which, in turn, continue to maintain the organized bounded structure that gives rise to these components.”


By the way

If you have a mechanical heart valve, it is neither made nor sustained by your body chemistry.

If it wears out, it has to be maintained or replaced by an external agent.

So your body (bounded by your skin, including the valve) can no longer be said to be fully autopoietic.

However, all system boundaries are a matter of choice, determined by system observers and describers.

So you can still say your biological system (excluding the heart valve) is autopoietic.

The allopoietic enterprise

The opposite of an autopoietic system is an allopoietic system.

“An allopoietic system uses raw materials to generate a car (an organized structure) which is something other than itself (the car factory).”


The following are allopoietic systems or processes.

Manufacturing a car is a production process that does not sustain the substance of the car manufacturing system itself.

Using a first tool (say, a lathe) to make a second tool (say, another lathe) does not sustain the substance of the first tool or its operator.

A first business actor may recruit a second business actor; this does not sustain the substance of the first business actor.

A business actor can hire or make business resources; this does not sustain the substance of the business actor.

Having a child is a kind of self-reproduction; it does not sustain the substance of the parents.


Assertion: no enterprise sustains its own components and processes entirely through the actions of its own components and processes.


Core business processes, as may be defined by operational researchers or enterprise architects.

·         are not designed to manufacture/sustain business actors

·         are distinct from the processes of any meta system (operational research or EA).

·         do not start from elementary molecules and energy


Some businesses manufacture some of their own resources, but very rarely from molecules and energy

And no business makes all its own resources.

Rather, it happily imports complex components from its environment.

The procurement department buys photocopiers, telephones and other complex but ready-to-go components.

The human resources department hires educated/skilled people (the most complex system components you can imagine).

Even a nuclear submarine, which may appear self-sustaining for a long time, must rest every so often for the import of replacement parts and submariners.


A business may be considered a self-organising entity in so far as it define and changes its own business roles and processes.

But self-organising does not mean operationally closed in the sense Maturana meant.

And the business may import not only resources but also process types from its environment.

Operational closure

Some authors devote many pages to explaining this concept, which suggests it is difficult to grasp.

Or else, that operational closure can be interpreted in various ways.

A simple interpretation is that system actors have no knowledge or understanding of what lies in the environment outside the system


This table is a generalisation of several explanations I have read.


In an “operationally closed” system

In an “operationally open” system


cannot cross the system boundary

can cross the system boundary

Complex resources from the environment

are not imported into the system

are imported into the system

Actors in the wider environment

cannot determine operations performed in the system

can determine operations performed in the system

Actors in the system

are not aware of actors or events outside the system

are aware of actors or events outside the system

The operationally open enterprise

Actors inside a business are aware of external processes. E.g. They help customers to place orders

Actors outside a business are aware of internal processes. E.g. They know that goods must be made or found and delivered by the business.


Think of an example enterprise and answer the following questions.

If you find yourself saying yes, then your enterprise system is operationally open.


Can processes cross the system boundary?

Can a process start outside, continue within the enterprise, and proceed outside again?

E.g. the process runs from order through invoice to delivery and payment?


Can complex resources enter or leave the system?

Are computers and other machines bought? Are people recruited?

Are complex parts assembled into larger products?


Can processes be imported into the system?

Does your enterprise ever copy a role or process from the wider world (another enterprise, a book or a consultant)?

Does legislation dictate any process in your enterprise, say, an accounting or redundancy process?

Does a legal contract with a customer or supplier oblige your enterprise to follow a pre-defined rule or process?


Can actors in environment determine operations within the system?

Can an external entity knowingly trigger a specific process performed within the system?

Can an external entity define or change a process type in the system description?


Are actors inside the system aware of actors or events outside the system?

Does any process within your enterprise depend on a process in the environment of the enterprise?

Does your enterprise outsource any "support function” or delegate any process to another enterprise?

Must your enterprise know the rules (inputs and outputs, pre-conditions and post conditions) of an outsourced process?



Published under the Creative Commons Attribution-No Derivative Works Licence 2.0       22/12/2016 12:09

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