Ackoff’s ideas – for applying system theory to management science

Copyright 2017 Graham Berrisford. One of many articles on the “System Theory” page at . Last updated 07/05/2021 10:06


Reading on line? If your screen is wide, shrink the display width for easier reading.


This article focuses on Ackoff’s fundamental ideas about the nature of systems. It reviews how Ackoff defined and classified systems in three documents that you can probably find on the internet. It was written during research for the book at


Ackoff’s viewpoint 1

Ackoff’s basic ideas. 5

Ackoff’s hierarchy of activities. 8

Ackoff’s hierarchy of aims. 10

Ackoff’s hierarchy of systems. 11

Ackoff’s three system classifications (1971, 1999 and 2003) 13

Ackoff’s further ideas. 16

Ackoff’s five conditions (1999) 16

Remarks related to other work. 17


Ackoff’s viewpoint

Russell L Ackoff (1919-2009) was an American organizational theorist, operations researcher, systems thinker and management scientist. He is respected for a large body of work focused on managed human social entities, and perhaps best known the following works.


·       Towards a system of system concepts(1971)

·       “Re-Creating the Corporation - A Design of Organizations for the 21st Century” (1999

·       “On The Mismatch Between Systems And Their Models”. (2003)


In the above, and in this article, Ackoff says a few things disputed below


Ackoff was concerned with the management of organized social entities – like universities and government agencies. His work on systems concepts draws on two distinct “system” traditions. First, a sociological tradition that can be traced back to the 19 century. Second, activity systems thinking, which emerged after the second world war in form of cybernetics and general system theory, and later embraced system dynamics.


Whereas Ackoff strove to combine these two traditions, it is argued below he might have done better to distinguish the actor-centric and activity-centric viewpoints.

Ackoff’s first view of systems

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


In “Towards a System of Systems Concepts” (1971) Ackoff started some distance from human society. He started with eleven terms and concepts that appear to derive from cybernetics.

1.     System: two or more interrelated objects (parts or elements).

2.     Abstract system: a system in which the elements are concepts. 

3.     Concrete system: a system that has two or more objects.

4.     System state: the values of system properties (state variables) at a particular time. 

5.     System environment: those elements and their properties (outside the system) that can change the state of the system, or be changed by the system. 

6.     System environment state: the values of environment properties at a particular time. 

7.     A closed system: one that has no environment. 

8.     System/environment eventa change to the system's state variable values.

9.     Static (one state) system: a system to which no events occur (it does not change).

10.  Dynamic (multi state) system: a system to which events occur (its state changes).

11.  Homeostatic system: a static system whose elements and environment are dynamic.


In EA, a concrete business system (3) realizes an abstract system (2). The system is open (7) dynamic (10) and interacts with its environment (5). It recognizes input events that represent changes to the state of entities (people, processes, materials and machines) in the environment (8). It maintains system state variables (4). It informs, directs or enables consumers or customers to change the system environment state (6) and meet their aims. The communication between the business and the external actors it monitors and serves or directs may be seen as a regulator-to-target feedback loop.


Business activity system

Feedback loop

Business environment


ßstate information



Consumes inputs

Produces outputs

Maintains system state

ß Inputs

à Outputs

External actors

Environment state


Ackoff on abstraction

Above, Ackoff’s abstract system is a model of how a concrete entity behaves, or should behave. He wrote that: “Different observers of the same phenomena may conceptualize them into different systems and environments.” Moreover, one concrete entity can realize several abstract systems. This table illustrates the difference between an abstract system and a concrete system instantiated by a concrete entity.


How an abstract activity system is realized

Orchestral music

Abstract system: a description of roles for actors, rules for processes and variable types

A musical score

Concrete system: a performance of defined activities, which gives values to variables

A performance of the score

Concrete entity: two or more objects able to perform the activities

An orchestra


Ackoff wrote wisely about the management of human institutions, but lost the plot when speaking of them as systems. There is much to enjoy in his view of motivating people to learn, the self-serving nature of the leaders of human organizations and other sociological phenomena. And there are things to dispute in his use of the term "system".


In cybernetics, system dynamics and soft systems methodology there is

A.   one observable entity or phenomenon

B.   any number of systems abstracted from it


Although Ackoff began by saying “Different observers of the same phenomena may conceptualize them into different systems”, later, he started using the term system with reference to the observed phenomenon: "Every human organization is a system."


Ackoff’s 2003 collaboration with Gharajedagh further confused matters by saying "Most contemporary social systems are failing" and "There is a … mismatch between most social systems and the models of them that are in use". What Ackoff usually meant was

A.   some government institution or university is failing

B.   some model we have of it is inadequate to solve its problems.


On the other hand, is it not obvious that:

A.   most such institutions also succeed to some extent

B.   none of the infinite systems abstractable from one could solve all its problems.


Worse, this use of “system” to mean the observed phenomenon, rather than a model of it, increased the confusion in wider systems thinking discussion between a social entity and any activity system it may realize. The main conclusion one can extract from the joint 2003 paper is that human creativity and delinquency make any attempt to equate 1 social entity to 1 system not only contrary to system theory, but daft.

Ackoff on aims

In an effort to build a unifying system theory that bridges the schism between activity systems and social entities, Ackoff built elaborate hierarchies of aims, activities and four system types. The graphic below is my attempt to stitch Ackoff’s system ideas together in a coherent whole.


Ackoff’s system classes

Actors (parts)

Activities (parts)

Aims (purposes)

State maintaining system

Play roles in

an activity system

No optional activities

Fixed aims

Goal-seeking system

Some optional activities

Purposive system

Are members of

a social entity

Define their own activities

Fixed aims

Purposeful system

Define their own aims


The first two kinds of system are activity systems with a fixed range of possible activities. Ackoff referred to them deterministic; I don't know if he recognized they could instead be probablistic or possibilistic (as defined elsewhere in this work).


Ackoff's primary interest was in social entities of the third and fourth kinds. He assumed actors can define their own activities, and perhaps aims. He characterized these social entities as being "purposive" or "purposeful".


The aims of Ackoff’s purposive social entity lie in the desire of external actors (sponsors, stakeholders, managers, designers and others) for the entity to produce particular state changes in its environment. The internal actors may be seen as slaves to that end.


The aims of his purposeful social entity are found in the desire of its internal actors to produce internal or external state changes that benefit themselves. Which implies the entity is “self-organizing.


Surely well-nigh every human institution is both? It may be called purposive in so far as it is shaped by the aims of external actors; and called purposeful where it is shaped by the aims of actors who play roles in the system (who may change activities and aims in ad hoc ways).


If there is no pattern regularity or repetition, then what is systematic or systemic about a social entity? Ackoff said his purposeful system "can change its goals in constant environmental conditions; it selects goals as well as the means by which to pursue them. It displays will.” However, he said it does have one stable feature - an "ideal" - which is unobtainable. Who defines it? Who knows what it is? Why can it not be changed?

Ackoff on changes to parts and wholes

Ackoff characterized a system as a whole in which two or more parts interact, and parts as elements that interact to produce the behavior of the whole. He wrote


"A part is never modified unless it makes the whole better, that is a systemic principle, you don't change the part because it makes the part better without considering its impact on the whole, that is systemic thinking" Ackoff


A difficulty with this is that the granularity of a whole and its parts are whatever the observer decides, and that systems can be nested, one a part of another, several times over. So, if you took Ackoff’s advice to its extreme, you might never improve a system, for fear its impact on one or more wider systems (of which it is a part) cannot be predicted. (Holism is not wholeism; you can never consider everything.)


Ackoff has been quoted as saying "Improving a part does not necessarily improve the whole". Nevertheless, improving a part on its own is a reasonable way to improve the performance of the whole. Attending to the highest cost part, and removing the largest bottleneck (one at a time) are recommended practices for improving a system. Such incremental development is a feature of biological evolution and agile system development.


Moreover, contrary to some interpretations of Ackoff, you can indeed remove parts from a whole without affecting its ability to meet its main aim. E.g. You can remove the spell checker from a word processor. You can remove the glove compartment, airbag, safety belt, carpet, arm rests and radio from a motor car, and still drive from A to B. You can remove the spleen, gall bladder and appendix from a human body with no significant effect on the functioning of the body.


Ackoff did a great job of highlighting the failings of organized social entities. Especially "public" ones such as government agencies and universities. On the other hand, he is partly responsible for spreading confusion between social entities and activity systems.


EA is about the design and planning of activity system changes. In Ackoff’ terms, you might say a human organization is one purposeful social entity, whose actors may play roles in any number of activity systems, and may change them from time to time. By the way, Ackoff’s “socio-systemic view of organizations” bears some similarity to today’s Open Group Architecture Framework (TOGAF).


For analysis of Ackoff’s “system of system concepts” read on

Ackoff’s basic ideas

In “Towards a System of Systems Concepts” Ackoff declared more than thirty ideas about systems.


Ackoff idea 1- System: a set of interrelated elements.

All the elements must be related directly or indirectly, else there would be two or more systems.


Note: This definition embraces both passive structures (e.g. tables) and activity systems. However, Ackoff’s main interest was multi-state systems in which actors perform activities.

·       In Forrester’s system dynamics, a system’s elements include both stocks and flows.

·       In Ashby’s cybernetics, a system’s elements include both state variables and transformations

·       In Ackoff’s management science, a system’s elements include properties and functions.


Ackoff idea 2- Abstract system: a system in which the elements are concepts.


Note: Like Ashby in discussing cybernetics, Ackoff distinguished abstract systems from concrete systems. An abstract system may describe an observed or envisaged reality. Abstract systems take concrete forms, in mental, documented and material models. What matters here is not the form the description takes but the correlation of an abstract description to a concrete reality observed or envisaged as instantiating that description.


Ackoff idea 3- Concrete system: a system that has two or more objects.

A concrete system is a realization (in physical matter and/or energy) of an abstract system description.


Abstract system

Concrete system

The Dewey Decimal System

A set of books sorted on library shelves

“Solar system”

A set of planets orbiting a sun

The laws of tennis

A tennis match

The score of a symphony

A performance of a symphony

A computer program

A performance of a computer program


Note: One abstract system (or type) can be realized (or instantiated) by many concrete system realities.  One concrete system (or type) realize (or instantiate) many abstract systems.


Which comes first? A natural concrete system (like the solar system) runs in reality before it is observed to be a system. A designed concrete system (like a motor car) cannot run in reality until after it has been envisaged by a designer.


People find this hard to understand and accept, but here goes…  In cybernetics, it is meaningless to say a named entity is an activity system without reference to a particular system description. With no system description there is just stuff happening. The universe is an ever-changing entity in which stuff happens. A concrete activity system is an island of regular or repeatable behavior carved out of that universe. It is a set of describable activities performed by describable actors, which we can test an entity as realizing.


Ackoff idea 4: System state: the values of the system’s properties at a particular time.

The current structure of a system is represented in state variable values, which changes over time.


Note: Ashby’s cybernetics defines systems in terms of the state variable changes that result from events detected. Ackoff used the term “property” to mean a system state variable.


Abstract state: typically a complex type, a composite of variables.

Cyclist Name, Direction, Speed, Gear Number

Concrete state: an instantiation of the type, a composite of variable values.

Guy Onnerbeich, South, 20 mph, 7


The state variables of a concrete entity may be part of one system but not part of another. E.g. the temperature of the earth’s atmosphere is vital to its role in the biosphere, but irrelevant to the earth’s role in the solar system.


Ackoff idea 5: System environment: those elements and their properties (outside the system) that can change the state of the system, or be changed by the system.

“The elements that form the environment… may become conceptualised as systems when they become the focus of attention.” Ackoff


Note: the environment is the world outside the system of interest. The system’s boundary is a line (physical or logical) that separates a system from its environment. Encapsulation is the enclosure of a system as an input-process-output “black box”.


Every brain and business can be described as encapsulated in a wider environment.

·       It receives information (in messages) about the state of actors and activities in its environment.

·       It records information (in a memory)

·       It sends information (in messages) to inform and direct actors and activities in its environment.


Thus, every brain and business can be seen as a control system connected in a feedback loop with external actors and activities in its environment. And note that a systems thinker may conceptualize those external actors and activities as a target system.


Ackoff idea 6: System environment state: the values of the environment’s properties at a particular time.

The current state of an environment is defined by giving values to the variable types in a description of that environment. The remainder of the real-world does not count as part of that environment (though it might count as part of another system’s environment).


Note: “Different observers of the same phenomena [actors and activities] may conceptualise them into different systems and environments.” Ackoff

Yes, observers of the same business may see different systems that are nested, overlapping, discrete, coupled by output/input flows and cooperating, competing or conflicting.


Ackoff idea 7: A closed system: one that has no environment.

“Such conceptualizations are of relatively restricted use”. Ackoff

Whereas an open system interacts with entities and events in a wider environment, a closed system does not interact with its environment.


Note: It is normal in using Forrester’s System Dynamics to model a closed system of stocks that interact by flows. Ashby’s more general cybernetics models open systems that respond to inputs from their environment. A conventional business system design process proceeds along these lines. 1) Define customers, outputs, inputs, suppliers and processes. 2) Define roles in the processes 3) Hire, buy or build actors to play the roles. 4) Organise, deploy, motivate and manage the actors – to perform the processes.


Ackoff idea 8: System/environment event: a change to the system variable values.


Note: Ackoff was concerned with how state changes inside a system are related to state changes in its environment. He did not distinguish events (input messages received) from state changes caused by those events. One may think of external events are crossing the boundary from the environment into the system.


In response to an event, a system refers to its current state, then “chooses” which actions to perform, to change its current state and/or produce outputs. A state change is the modification of one or more state variable values. One event can cause different, optional, state changes, depending on the current state of the system.


(See the “Causality and choice” section in the concluding remarks.)


Ackoff idea 9: Static (one state) system: a system to which no events occur.


Note: his example of a static system, a table, does experience events; it can be laid for dinner, be painted, stripped and polished, and have its wonky leg replaced.


Ackoff idea 10: Dynamic (multi state) system: a system to which events occur.


Note: the system of interest to most systems thinkers is not only dynamic (changes state in response to events) it is also an activity system in which actors perform activities.


Ackoff idea 11: Homeostatic system: a static system whose elements and environment are dynamic.

“A house that maintains a constant temperature… is homeostatic”. Ackoff


Note: Ackoff considered a homeostatic system to be a static system. But the state of such as system does fluctuate, and it does experience events. A truly one state system could be a passive structure (as in ArchiMate), which is inanimate; it can experience events, can be acted on or in; but cannot act.


Ackoff’s example here is also questionable. Surely the house is merely a container for the air whose temperature is controlled? What maintains that air at a constant temperature is a heating/cooling system, controlled by a thermostat.


Aside: Several early social entity thinkers viewed sociology as analogous to biology; they looked at social entities as akin to homeostatic organisms. Ackoff and other systems thinkers have deprecated using this analogy.


About the ideas to follow

From idea 9 onwards Ackoff’s ideas grow more questionable. He built elaborate hierarchies of aim, activity and system types. He used the term “function” sometimes mean an activity and sometimes an aim. The table below is an attempt to stitch many of his ideas into a coherent whole. I believe it gives a good impression of how a cybernetician like Ashby would see Ackoff ideas.


Ackoff’s system classes

Actors (parts)

Activities (parts)

Aims (purposes)

State maintaining system

Play roles in

an activity system

No optional activities

Fixed aims

Goal-seeking system

Some optional activities

Purposive system

Are members of

a social entity

Define their own activities

Fixed aims

Purposeful system

Define their own aims


When Ackoff strove to apply the terms and concepts of general system theory and cybernetics to human society, he contradicted himself. He said (re. points 2 and 3 above) “Different observers of the same phenomena may conceptualize them into different systems”. But later said that every human organization is a system, regardless of any particular observer's conceptualization. Thus, he confused social entities with activity systems.

Ackoff’s hierarchy of activities

Most system theorists are concerned with activity systems. In System Dynamics, a system may be defined thus:

 “A set of elements or parts that is coherently organized and interconnected in a pattern or structure that produces a characteristic set of behaviors." Meadows


And in the context of cybernetics:

 “Cybernetics deals with all forms of behavior in so far as they are regular, or determinate, or reproducible.” “[It] treats, not things but ways of behaving. It does not ask "what is this thing?" but ''what does it do?" It is thus essentially functional and behavioristic.” (Ashby 1956)


Ackoff built an elaborate hierarchy of behavior concepts. He distinguished acts, reactions and responses with reference to state changes inside the system and its environment.


Ackoff idea 12: System reaction: a system event that may be triggered directly by another event.

E.g. “A light going on when the switch is turned.” Ackoff


Note: A system reaction is simplistic: there is no reference to memory and no other option. If the reaction depends on the state of the world, then Ackoff would call this a response.


Ackoff idea 13: System response: a response that goes beyond the naive reaction at 12.

E.g. “A person’s turning on a light when it gets dark is a response to darkness”. Ackoff


Note: in Ackoff’s example, a person compares two variables: an internal state variable (visual acuity level needed for their current activity, say reading) and an external environment variable (current light level).


If the required acuity level is higher than the current light level allows, then the person switches on the light. Presumably, people apply what might be called fuzzy logic to such internal variables rather rigid rules.


Ackoff idea 14: System act: a self-determined, autonomous, behavior.


Ackoff said a system act may be triggered by an internal event (rather than an external event).  An internal event happens when the value of an internal state variable crosses a significant threshold. E.g. when a person’s required acuity level is higher than the current light level allows, this may trigger them to switch on a light.


Ackoff’s example of a system response also fits his definition of a system act. Any activity system can “choose” its response to an event, and produce different outcomes, by applying rules that refer to the state of its memory. Any observer may see the response to an event as unpredictable if a) the rules include applying fuzzy logic and b) the internal state of the system is unknown. (See the “Causality and choice” section in the concluding remarks.)


However, surely Ackoff was intending to distinguish determined behaviors (responses) from self-determined behaviors (acts). Meaning that the latter are made by self-aware entities with free will.


Ackoff idea 15: System behavior: system events which… initiate other events.


Ackoff says behaviors are state changes whose consequences are of interest. It is arguable that all behaviors (acts, reactions and responses) have both consequences of interest at some level of investigation, and antecedents of interest at some level of investigation.


I can’t find a definition of “outcome” in Ackoff’s 1971 work. Are they consequential changes in the state of the system? or its environment?


On the hierarchy of behavior types

Ackoff arranged behaviors (ditto aims and systems below) in a hierarchical structure using different words at different levels. I am not confident that his classification is wholly coherent; however, this table represents my attempt to summarize it.




Make choices

wrt system state

Outcome (event/state change)





12 Reaction






13 Response






14 Act


Not defined?


Not defined?

Not defined?

Ackoff’s hierarchy of aims

Ackoff’s main interest was in managed human organizations. So, although other things (the solar system, a virus, a tree, an ant, an ant colony, and a person) might be viewed as a system, it is best to consider his hierarchy of aim concepts in the context of a human institution.


Ackoff idea 22: The relative value of an outcome: a value (between 0 and 1) compared with other outcomes in a set. The highest value outcome is the preferred outcome.

Ackoff idea 23: The goal of a purposeful system: a preferred outcome within a time period.

Ackoff idea 24: The objective of a purposeful system: a preferred outcome that cannot be obtained in a time period.

Ackoff idea 25: An ideal: an objective that cannot be obtained in a time period, but can be approached.

Ackoff idea 26: An ideal-seeking system: a purposeful system that seeks goals leading towards its ideal.


From top to bottom, the hierarchy of aims seems to be ideals < objectives < goals < outcomes.


25 Ideals

are persistent aims that appear

“unobtainable in principle or in practice”.


24 Objectives

can be ordered with respect to


23 Goals

can be ordered with respect to


22 An outcome

can be valued and preferred as a



Again, outcome is undefined in Ackoff’s 1971 work, but appears to be a consequential change in the state of a system or its environment.


How Ackoff’s purposes and functions relate to each other and to ideals, objectives, goals and outcomes not entirely clear, but purposes appear to sit at a higher level than goals.


How is the aim hierarchy affected by system composition?

Systems may be composed or decomposed in space, time or logic. So, an event that is external to a smaller system is internal to a larger system, and vice-versa. 


What happens to aims if you expand the time period ? Presumably, an ideal becomes an objective and an objective becomes a goal?


What happens to aims if you expand the system boundary (say from department to division to enterprise)? Presumably, a goal of a small subsystem is seen as one of many outcomes in a wider system?

Ackoff’s hierarchy of systems

A system is composed from interacting parts or subsystems. Systems are recursively composable and decomposable. Like Boulding before him, Ackoff blurred the idea of system decomposition with the idea of different system types. His aim and behavior hierarchies (above) may have been designed to fit his system type hierarchy (this section), in which he classified systems by dividing behaviors and outcomes between “determined” and “chosen”.


Ackoff idea 16: State maintaining system: the most naïve of reactive systems (cf. 12 reactions).


When an event of a given type occurs, Ackoff’s state maintaining system reacts to produce a fixed outcome. “Fixed outcome” implies there is no choice as to the outcome of the event. Ackoff’s discussion of this system type is dominated by self-regulating or homeostatic entities. In this category, is maintaining a system state variable (within defined bounds) definable as a goal or objective?


What if maintaining one system state variable involves making choices related to the values of other state variable? I guess that is a goal-seeking system - the next idea below.


Ackoff idea 17: Goal-seeking system: a system that chooses its response to events (cf. 13 responses).


This category of system seems to be defined by its ability to retain and use a memory to make choices. Every business remembers facts in the form of state variables (in data stores) and uses that memory to inform its actions. It refers to its memory when choosing between actions that produce different outcomes, with an end goal “in mind”.


(See the “Causality and choice” section in the concluding remarks.)


Ackoff refers at this point to a system that does more – it learns – it adapts its behavior by conditioning according to experience. E.g. a rat in a maze increases its efficiency in meeting a goal by choosing not repeat a behavior that failed earlier. That primitive kind of learning can be mimicked by a computer program.


It is unclear whether Ackoff intended to embrace all kinds of learning, for example:

·       Learning from simple physical sensation (e.g. that your lips may stick to a cube ice)

·       Learning facts (e.g. the colours of the rainbow)

·       Learning a physical process (e.g. to swim or play music)

·       Learning a logical process (e.g. multiplication, algebra, a business process)

·       Learning a cultural norm (e.g. to say please and thank you)

·       Machine learning (detecting patterns in data).


Ackoff idea 18: Process: a sequence of behavior that constitutes a system and has a goal-producing function.


You might call this process a one-time system. It runs from trigger event to an outcome that meets its goal. Ackoff says each step in the process (be it an act, reaction or response) brings the actor closer to the goal. It isn’t clear whether his process can have control logic, or lead from one event to a variety of different outcomes/goals. Also, it isn’t clear whether the goals of a system = the sum of all goals of processes performed by the system.


Ackoff idea 19: Multi-goal seeking system: a system that seeks different goals in different states.


So, the goal(s) of a system depend on the state it is in. Presumably, Ackoff means the state-goal relationship predetermined. Here, Ackoff appears to presume system actors behave in a deterministic manner.


Ackoff idea 20: Purposive system: a multi-goal seeking system where goals have in common the achievement of a common purpose.


The common purpose could be to survive, win the game or make a profit. Purposive systems and sub-animate actors may be given goals but can’t change them. Again, it seems purposes sit at a higher level than goals, but the relationship of purposes to objectives and ideals is not clear. Here, Ackoff appears to presume system actors can choose between acts, but cannot determine goals.


Ackoff idea 21: Purposeful system: can produce the same outcome in different ways and different outcomes in the same and different states.


Ackoff believed that free will enabled purposefulness, which in turn demonstrated free will. A book Ackoff wrote with Emery about purposeful systems (1972) put it more clearly.

“a purposeful system is one that can change its goals in constant environmental conditions; it selects goals as well as the means by which to pursue them. It displays will.” Ackoff

Animate actors (being self-aware) might change the aims of any system they participate in.

"members are also purposeful individuals who intentionally and collectively formulate objectives and are parts of larger purposeful systems.” Ackoff


(How far the goals of a social entity relate to or derive from individual actors’ goals is a question beyond the discussion here.)


On the hierarchy of system types

The table below is a sketchy attempt to stitch ten of Ackoff’s ideas together into a coherent whole. I believe it gives a good impression of what he was striving towards.


System types

Behavior types

Aim types

16 A state maintaining system

12 Reacts as predetermined to produce

22 one predetermined Outcome

17 A goal-seeking system chooses between

13 Responses as predetermined to meet

23 one predetermined Goal

20 A purposive system chooses between

14 Acts by self-determination to meet

23 predetermined Goals with a shared 24 Objective

21 A purposeful system chooses between

14 Acts by self-determination to meet

25 Goals and Objectives it chooses, leading to an Ideal


If (20) actors can define the acts they perform, there isn’t a system in the normal sense.


Later, he defined “organizations” as purposeful systems (21). If their only stable feature is an ideal (which Ackoff described as unobtainable), there is nothing systematic about the organization.


What might a cybernetician like Ashby make of this? Find an answer in the next section.

Ackoff’s three system classifications (1971, 1999 and 2003)

Remember Ackoff’s primary system of interest was managed human organizations. This section reviews how his classification of system types evolved over twenty years.

A hierarchy of system type before Ackoff

Ackoff’s hierarchy of systems appears to have been an attempt to turn hierarchical system decomposition into a science. In this regard, he seems to have followed in the footsteps of Kenneth Boulding. However, Boulding's (1956) system classification was highly questionable.


Boulding proposed systems be classified in hierarchy with seven levels that rise in two ways by aggregation from part to whole, and by complexification from simple to complex. The hierarchy is topped by societies at level 6 societies and ecologies at level 7. This has encouraged people, ever since, to apply the term "complex system" to social networks.


One trouble is that levels of aggregation and levels of complexity are different ideas. Every system is a perspective, or partial representation, of a reality. To take a holistic view of an aggregate is to deliberately ignore the complexity of each part. E.g. a card game is a very simple activity system; the biology of a human’s cardio-vascular system is very complex. The staggering complexity of a card players’ biology is irrelevant to the card game (and to other social interactions).

1971: Classification by behavior and outcome

At first, Ackoff differentiated his systems of interest by whether behaviors and outcomes are “determined” and “chosen”. Remember this table (copied from above) is my interpretation of what he was saying in 1971.


System types

Behavior types

Aim types

16 A state maintaining system

12 Reacts as predetermined to produce

22 one predetermined Outcome

17 A goal-seeking system chooses between

13 Responses as predetermined to meet

23 one predetermined Goal

20 A purposive system chooses between

14 Acts by self-determination to meet

23 predetermined Goals with a shared 24 Objective

21 A purposeful system chooses between

14 Acts by self-determination to meet

25 Goals and Objectives it chooses, leading to an Ideal


Later, Ackoff defined “organizations” as purposeful systems (21).


A cybernetician’s view

This table shows what Ashby might make of Ackoff’s classification.


Ackoff’s system classes

Actors (parts)

Activities (parts)

Aims (purposes)

State maintaining system

Play roles in

an activity system

No optional activities

Fixed aims

Goal-seeking system

Some optional activities

Purposive system

Are members of

a social entity

Define their own activities

Fixed aims

Purposeful system

Define their own aims


Note that Ackoff’s purposive and purposeful systems are only systems if the definitions of responses and aims are changed under change control, else they are continually mutating entities in which no regular activity system can be found.

1999: Classification by purposefulness

Later, Ackoff differentiated his systems of interest by the “purposefulness” of the parts and the whole. The table is edited from table 2.1 in “Re-Creating the Corporation - A Design of Organizations for the 21st Century” .


Systems type





Not purposeful

Not purposeful

Motor cars, fans, clocks,

 computers, plants


Not purposeful






Corporations, universities, societies



Not purposeful

An environment that serves its parts.

E.g. ocean, atmosphere


By 1999, Ackoff had defined “purposeful” more neatly than in 1971.

"An Entity is purposeful if it can select both means and ends in two or more environments."

He then contrasted goal-seeking entities…

“Although the ability to make choices is necessary for purposefulness, it is not sufficient. An entity that can behave differently (select different means) but produce only one outcome in any one of a set of different environments is goal seeking, not purposeful. For example, automatic pilots on ships and airplanes and thermostats for heating systems have a desired outcome programmed into them; they do not choose them.”

… with purposeful people and social networks

“people can pursue different outcomes in the same and in different environments and therefore are obviously purposeful; so are certain types of social networks.”

2003: Classification by choice

Ackoff subtly revised his system classification scheme as shown below. Now, the key differentiator is the ability of the parts and the whole to exercise “choice”.


System Model




Deterministic / mechanistic

Clock, Tree, Ant, Fish

No choice

No choice

Not just man-made machines!

includes plants and “lower” animals.


Human, Gibbon

No choice


Not all animals!

excludes “lower” animals, includes only “higher” animals Ackoff considered to exercise free will


Church, Corporation



Not all social networks!

excludes “lower” animal groups and informal groups, is primarily if not only formal organizations




No Choice

Not an environment served

an environment that serves its parts.


Clearly, the meaning Ackoff attached to “choice” is vital to his system classification. He had had previously allowed that machines can make choices.


See the “Causality and choice” section in the concluding remarks. And perhaps read this article for a more detailed critique of this 2003 system classification.


Ackoff was a doomsayer about the management of organised human social networks. His joint article On The Mismatch Between Systems And Their Modelsstarted: “Most contemporary social systems are failing.” However, Ackoff’s examples of failing organizations were questionable.


Ackoff’s example

Alternative possibilities

“Head Start is said to be a failure.”

Impossible targets? Or the impossibility of garnering enough resources to meet the huge challenge?

“The US has a higher percentage of its population in prison than any other developed country, but has the highest crime rate.”

The prison population is consequence of the crime rate? Or shows the success of criminal investigation organizations?

“Most corporations formed each year fail before the year is up.”

A flourishing entrepreneur base?

“Half the corporations on the Fortune 500 list twenty-five years ago no longer exist.”

A healthy free market adapting to changing times? (Biology requires that older individuals die to make way for new ones.)


“One could go on citing deficiencies in the management of our principal social systems.”

Surely most of organizations also succeed in some way to some extent, at least they employ some people for a while?

Is it reasonable to expect every organization to succeed? The difficulties of managing large organizations are multitudinous. What if managers have impossible targets? Or they cannot garner enough resources? Or market forces change the game?


In short

Ackoff’s system classifications (1971, 1999 and 2003) appear plausible on the surface, yet are often misunderstood. And there is a problem with classifying real-world entities into system types. That is, it undermines the distinction Ackoff originally drew between an entity, such as a human organization, and the many different systems that might be abstracted from observation of that entity.

Ackoff’s further ideas

This section lists Ackoff’s further ideas with only a couple of comments.


Ackoff idea 27: The functions of a system: production of the outcomes that defines its goals and objectives.

Whether and how purposes and functions relate to each other and to the hierarchy of ideals < objectives < goals < outcomes is not clear.


Ackoff idea 28: The efficiency of a system: defined in mathematical terms beyond the analysis here.


Ackoff idea 29: An adaptive system: reacts or responds to reduced efficiency by changing system or environment state.

Note that adaptation here appears to mean system state change rather than system mutation.


Ackoff idea 30: “To learn is to increase one’s efficiency in the pursuit of a goal under unchanging conditions.” Ackoff


Ackoff idea 31: Control: “An element or system controls another element or system if its behavior is either necessary or sufficient for subsequent behavior of the element or the system itself and the subsequent behavior is necessary or sufficient for the attainment of one of more of its goals.” Ackoff


Ackoff idea 32: Organization: “An organization is a purposeful system that contains at least two purposeful elements which have a common purpose relative to which the system has a functional division of labor; its functionality distinct subsets can respond to each other’s behavior through observation of communication; and at least one subset has a system-control function”. Ackoff


Since Ackoff later deprecated the mechanical and biological views of systems taken by cyberneticians, it is a little surprising that (in 1971) he described inter-system relationships in terms of controls.

Ackoff’s five conditions (1999)

Remember: Ackoff’s primary system of interest was a managed human organization. In 1999, Ackoff defined his system of interest by five conditions.


Condition 1- The whole has one or more defining properties or functions.

Ashby (cybernetics) defined a system as having both state variables and behaviors. Forrester (system dynamics) defined a system as having both stocks and flows. Ackoff (management science) defined a system as having both properties and functions. He used the term “property” to mean a system state variable.


Condition 2- Each part in the set can affect the behavior or properties of the whole.

At first sight, this seems to go without saying, but what is meant by a “part”? Is it a type in an abstract organization – a role, rule or variable? You cannot remove a type without having an impact on the behavior or variables of the whole. Or is it an instance in a concrete organization – an actor or activity instance? You can remove an instance (remove a fish from a shoal) with no effect on the behavior or properties of the whole.


Condition 3- There is a subset of parts that is sufficient in one or more environments for carrying out the defining function of the whole; each… necessary but insufficient for… this defining function.

The division of a system into parts, and the granularity of those parts, is entirely in the gift of the describer. This third condition is not true of systems described at the level of coarse-grained parts that are tightly coupled. E.g. there is no functioning subset of a rider and bicycle system. Or a marriage between two people. Or a commercial business divided into sales, delivery and accounting parts.


Condition 4- The way that each essential part of a system affects its behavior or variables depends on (the behavior or variables of) at least one other essential part of the system.

Again, the division of a system into parts, and the granularity of those parts, is entirely in the gift of the describer. This fourth condition is not true of systems in which one subsystem (part) encapsulates all essential parts.


Condition 5- The effect of any subset of essential parts on the system as a whole depends on the behavior of at least one other such subset.”

The fifth condition seems an elaboration of the fourth.


Some system descriptions fit all Ackoff's 5 conditions. Some do not, where all the described parts are essential and there is no functioning subset of parts.

Remarks related to other work

This article was written during research for the book at


References below are to other chapters in that book, which discusses ways that management scientists (Boulding, Ackoff and Beer) tried to meld management science with cybernetics or general system theory. Unfortunately, attempts to merge different approaches can obscure what each has to offer. And when management scientists borrow terms from cybernetics or general system theory, they often use them with different meanings.


On abstract and concrete systems

In 1971, one of Ackoff’s insights was this: “Different observers of the same phenomena may conceptualise them into different systems”.


Ackoff distinguished the conceptualization of a system from its realization. Abstract systems are conceptual - as are the roles and rules in a system description. Concrete systems are physical - as are the actors and activities in a working system. Systems thinkers often conflate the two ideas, yet there is a many-to-many association between them. One abstract system can be realized by several concrete social networks. One concrete social network can realize several abstract systems.


In 1999, Ackoff’s first condition was that a system’s variables or functions must be defined. So, a group of people or social network doing things is not a system just because people call it an “organization”. It becomes an activity system when, where and in so far as it performs functions defined in a system description.


So, systems thinking involves separating abstract systems (e.g. laws like apartheid) from concrete things (e.g. societies) that realize them.  Just as a thing is infinitely more than any system it realizes, so, a social entity is infinitely more than any law that constrains its behaviors.  A social entity that realizes a system will continue to evolve. Sometimes a social entity evolves in a way that reinforces the system (cf. a social cell); sometimes a social entity evolves in a way that forces the system to be changed or abandoned.


The social entity and the system are two different things. The trouble is - Ackoff later went on to contradict himself by equating a social network with a social system. He referred to a church, a corporation or a government agency as being a social system. Whereas IBM, for example, is as many different activity systems as system describers can describe and test. Some of those activity systems may conflict with each other, or undermine each other.


Moreover, much of what happens in a business like IBM is not at all systematic or systemic. It relies on humans knowing what to do, choosing what to do, and inventing what to do when the need arises. In equating an organization that depends on human knowledge and abilities (rather than system description) to a system, Ackoff undermined the basic ideas he set out at the beginning.

On system composition/decomposition

Ackoff arranged aims, behaviors and systems in hierarchical structures, using different words at different levels. It can be convenient to use different words for different levels of system concept. The table present an example of using different words at different levels of composition.


Division in time




Division in space






Long term


Value stream


Large part

Short term




Small part





Atomic part


However, systems are infinitely nestable; the level of decomposition is arbitrary – a choice made in a particular situation. Pinning different words to different levels of a decomposition hierarchy can obscure the general nature of system theory. At every level, a process is an event-triggered 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.

On system change or adaptation

Ackoff said all animated systems are organisms, which he said are defined as being autopoietic. In discussion of social entities, he used the term self-organising, which is a different concept.


First, we need to distinguish self-sustaining from self-organising. Then, if the latter means actors continually change the variables and functions of the organization they work in, this conflicts with Ackoff’s first condition for a system. For a more coherent discussion, we need to distinguish several possible meanings of “self-organising”.




It may seem Ackoff’s attempts to merge social entity thinking with activity system thinking are doomed. However, it is possible to reconcile the two traditions, as indicated in the chapter above.

Causality and choice

Ackoff took an anthropomorphic view of systems. Even so, he did not see all human social entities as organizations. He showed little interest in those with minimal bureaucracy (a tennis club or pick-pocket gang).  His primary interest was in organizations of the kind addressed by “management science”.


Management science addresses human institutions that – in more or less bureaucratic ways - organize most or some of what human actors do. Typically, these social entities employ many discrete activity systems. And when observing actors’ behavior, we can classify their responses to stimuli into four kinds.



In theory, when an event happens, we can predict


exactly which action an actor will perform in response.


how likely an actor will perform activity type A or activity type B.


the actor will choose from the range of activity types in our model.


nothing – because actors can invent activities outside any model made.


In so far as a social entity realizes a known activity system, its actors always act in the first three ways. We may not be able to predict which action they choose to perform, but we can say they will choose one of the actions available to them in the activity system. (A designer has to make allowances for “exception paths” actors may choose to follow.)


In so far as actors act in the fourth way, they act outside any activity system we know of. Even if they are in acting in an activity system we don’t know about, we must treat them as having free will and the ability to do what they choose.



Ackoff’s view of choice

Ackoff defined animate actors (humans and other higher animals) as ones that choose freely between possible actions. He said they make choices in the light of their “ideals”.


"The capability of seeking ideals may well be a characteristic that distinguishes man from anything he can make, including computers".

Really?  Doesn’t a chess-playing computer choose between possible chess moves in the light of its ideal – to win the game? Doesn’t a humanoid robot direct its limb movements in the light of its ideal – to stay on its feet?


Surely the main issue is not how actors make choices - by free will or not? Rather, it is that animate actors may have different ideals from any activity system they play roles in. And so, the actors have to choose between conflicting ideals when choosing their next action.


“All organizations are social systems”

Ackoff said the reverse is not true; not all social systems are organizations. And for sure, not all social entities are organized to the extent we can define an activity system they realize


“Most contemporary social systems are failing”

Did he mean to say most organizations are failing? Sometimes he seems to be presenting something akin to a political manifesto. Does he really mean government agencies rather than commercial businesses? Most UK government agencies surely succeed a bit and fail a bit. Is his concern US government in particular?


“An organization is a purposeful system that contains at least two purposeful elements which have a common purpose”.

Can purposes be ascribed to a purposeful organization by entities inside or outside the organization?


“An aggregation of purposeful entities does not constitute an organization unless they have at least one common purpose”

Does it count if employees and suppliers have the common purpose of maximising their income from the organization? Or all employees are motivated by self-respect from being employed and social interactions with their colleagues?


“Organizations display choice.”

Geoff Elliot told me that in the sociological perspective, only people can be “purposeful”.  But Ackoff said social systems (meaning social organizations) are also purposeful.


What does it mean to say a social organization make choices or exercises free will? Not all the purposes of an organization are abstracted from the purposes of the actors in it. What about purposes ascribed to an organization by entities outside the organization – like the government?


Would you say an investment bank’s trading division chooses which stocks to buy and sell? Those choices, formerly made by human actors, are now by made mechanistic computer actors. So, in what sense does the organization display choice? And how does it make a choice independently of choices made by its animate or mechanistic parts?

a deterministic process? If so, does that undermine the meaning of purposeful?

On people not being “parts”

The word “part” suggests containment within a boundary, be it physical or logical. A man can contain a machine (say, an artificial heart) both physically and logically. The heart has no aim or behavior outside of the man. A machine (say, motor car) can contain a man physically but not logically. E.g. a taxi driver has aims and behaviors outside of their role as driver.


Suppose you employ human actors, each with their own aims and behaviors, in your business. Your business needs only some of the time and talents of each employee. You ask an employee (Jane) to perform the activities expected of her given role. Within her work hours, Jane will sometimes act outside her defined role, and sometimes contrary to it. Outside of work hours, she may play a role in other (perhaps even competing) businesses. In what sense is Jane a part of your business?

On organising, deploying, motivating and managing actors perform processes

“A group of unwilling slaves can be organised to do something that they do not want to do, but they do not constitute and an organization, even though they may form a system” Ackoff


There is a difference between not being keen to do some work and not being willing and able to do it well when asked.  We may baulk at starting a job, but get some enjoyment from it nevertheless. And being appreciated by colleagues and managers has a lot to do with that.


Your new employee brings their own aims (ideals, purposes, objectives, goals) to your business. Their aims may be in conflict with each other, in flux, unconscious or unrecognised. Some of their aims may be contrary to the aims of you and your social organization.


Your employee may act to make your organization inefficient, or even to sabotage it (e.g. sharing information with a competitor). OK, you can design security and exception handling processes to address the conflicting aims and contrary behaviors of employees.


And yes, human actors are special, they need special attention. This is the territory of social psychology, neuro-linguistic programming and the like – rather than system theory.


Business managers and project managers are responsible for managing and motivating their employees. It isn't the responsibility of enterprise architects. Attending to the "machinery" of human and computer activity systems is difficult enough. Where people's roles are affected, the EA function work with business managers, HR and sometimes a "business change function" to ensure people are well managed and motivated.