General System Theory Principles

Copyright 2017 Graham Berrisford. One of about 300 papers at http://avancier.website. Last updated 23/07/2017 22:30

 

This paper follows others that introduce general system theory.

Read “System theory terms and concepts” for a brief introduction to core ideas

Read Introducing General System Theory and Cybernetics” for a longer explanation of them.

 

This paper is also a companion to another that distils and reviews Ackoff’s ideas about systems.

Key references: Introduction to Cybernetics” (1956) W. Ross Ashby.

And “Towards a System of Systems Concepts” (1971) Russell Ackoff.

Contents

General system theory. 1

Principles about encapsulation. 1

Principles about description and reality. 4

Principles about state, behavior and change. 8

FOOTNOTES on philosophy, science and GST. 11

 

General system theory

Some of Bertalanffy’s views of life, evolution, and the human condition are questionable.

But his core ideas are widely accepted and applied in business systems analysis and design.

Building on these core ideas, this paper defines the following GST principles.

·         Principles about encapsulation

o   Principle: an open system interacts with its environment

o   Principle: a closed system is sealed from its environment

o   Principle: systems can be composed and decomposed

·         Principles about description and reality

o   Principle: descriptions idealise observed or envisaged realities

o   Principle: concrete systems realise abstract ones

o   Principle: an entity is a system whenever and wherever it realises an abstract system description

o   Principle: realisation differs from translation

·         Principles about change

o   Principle: continuous behavior can be modelled as driven by discrete events

o   Principle: system change differs from system state change.

Principles about encapsulation

“Every living organism is essentially an open system. It maintains itself in a continuous inflow and outflow…” Bertalanffy

 

System environment: the world outside the system of interest.

System boundary: a line (physical or logical) that separates a system from is environment

System interface: a description of inputs and outputs that cross the system boundary.

Organicism: the idea that systems are describable at multiple hierarchical levels (as von Bertalanffy named it).

Principle: an open system interacts with its environment

The environment outside a system contains structures/entities and behaviors/events that can change the state of the system, or be changed by the system.

 

System encapsulation (IPO) as system scoping

There are usually feedback loops between a system and its environment.

To encapsulate a system means defining its input-process-output (IPO) boundary.

The inputs and outputs can be flows of information, material or energy.

 

“Any suggestion that we should study "all" the facts is unrealistic, and actually the attempt is never made.

What is necessary is that we should pick out and study the facts that are relevant to some main interest that is already given.” Ashby

A system describer defines the inputs and outputs with reference to an already-given interest or aim.

 

Remember, the system-environment boundary is a choice made by system describer(s).

It meaningless to point to an entity and call it a system if its boundary and properties are obscure.

First, those in the discussion must agree its boundary and which I/O flows are relevant to some shared interest.

The I/O flows of interest in business system models are rarely energy flows, sometimes material flows and often information flows.

A business each application/technology component it uses can be defined by the I/O services it offers.

 

System design as a process

SIPOC is an acronym that captures GST ideas used in business systems analysis, design, process and quality improvement.

The conventional design process defines the system-environment boundary thus:

1 Define Customers - entities in the environment that need outputs to meet their goals

2 Design Outputs - that customers need from the system

3 Design Inputs – that the system needs to produce the outputs

4 Define Suppliers - entities in the environment that will supply the inputs

5 Design Processes - to produce the outputs from the inputs.

 

To which one might reasonably add:

6 Define Roles - in which actors can perform the process steps

7 Hire and/or make Actors - to play the roles

8 Organise, deploy, motivate and manage the actors – to perform the processes.

 

Many sociological systems thinkers speak to the concerns at step 8.

 

Of course, the process is iterative in practice.

Systems and processes can be, and are, composed and decomposed.

Decomposition continues until individual actors can be hired or made to perform the required behaviors.

 

At every level of system composition, there are cross-boundary input/output flows.

If you expand the boundary then external events become internal events that pass between subsystems.

If you contract the boundary then internal events become external events, crossing that boundary.

 

Aside on archetypes or design patterns

Archetypes are used in designing the structures and behaviors in which subsystems interact.

General, cross-science, archetypes appear in contrasting design patterns listed in this table.

Centralisation of control

Distribution of control

in one place or component. 

between places or components.

Hierarchy

Anarchy or Network

Hub and Spoke

Point-to-Point or Mesh

Client-Server

Peer-to-Peer

Fork or Orchestration

Chain or Choreography

 

The last of these is a behavioral rather than structural pattern.

System architects choose between alternative patterns by trading off their pros and cons on the light of given requirements.

Principle: a closed system is sealed from its environment

An open system interacts with entities and events in a wider environment.

A closed system does not interact with its environment.

 

Aside: Every “system dynamics” model is a closed system.

It is a model of populations (stocks) that grow and shrink in response to continuous inter-stock event streams (flows).

The whole system is closed, so all events are internal events.

However, each stock can be seen as a subsystem, to which every event is an external event.

 

System dynamics models represent continuous inter-stock flows as events-per-time-unit (e.g. total births per month).

When running the model to simulate a reality, the time unit is the discrete event that drives the system to change state incrementally.

The real-world events being modelled appear as units (1s) aggregated in the quantitative variables (e.g. total deaths per month).

M A Jackson (c1975) credited this insight to Mike Woodger of the National Physical Laboratory in Teddington UK c4 miles from where I sit.

 

A System Dynamics model is an abstract system description (a theory).

Running the System Dynamics model is sometimes presumed to be a test of the theory, but it isn't; it is only an animation of the theory.

System testing requires that the results of running the model are compared with the result of whatever reality is modelled.

Principle: systems can be composed and decomposed

Systems (along with aims, behaviors and structures) can be described at different levels.

Systems can be hierarchically nested: one system can be a part or subsystem of another.

Systems can be recursively composed and decomposed not only in space but also in time or logic.

 

So, how to describe different levels of 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, e.g.

 

Time-frame

Aims

Behaviors

Active structures

Granularity

Persistent

Business mission

 

Enterprise

Whole

Long term

Goal

Value stream

Division

Composite

Short term

Objective

Process

Team

Part

Immediate

Requirement

Action

Actor

Atom

 

However, the level of composition or decomposition is arbitrary – a choice made in a particular situation.

It is impossible to be scientific about pinning different words to different levels of a three, four or five level decomposition.

And trying to do so can obscure the general nature of system theory.

 

The concepts are the same at whatever level of system composition you choose to model.

A process is 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 (if you consider those to be incompatible).

Principles about description and reality

An entity is a system whenever and wherever it matches an abstract system description.

 

Description: a memory, message, model or view that captures/encodes knowledge of a thing’s properties.

Name: an identifier or label for a thing whose properties can be described.

Abstract system description: a description or model of a concrete system.

Concrete system (aka System): a system that runs in reality.

Principle: descriptions idealise observed or envisaged realities

This Scientific idealism triangle is a generalization of how the world is described.

Scientific idealism

Abstract descriptions

<create and use>              <idealise>

Describers  <observe and envisage>  Concrete realities

 

Aside: describers and descriptions are realities that can be described.

 

The triangle is specialised below to relate role definers, roles and actors who play the roles.

Bear in mind, a person may double as role definer and actor in the same business; we’ll return to that point later.

Social organization

Defined roles

<define and change>              <idealise>

Role definers <observe and envisage> Real-world actors

Principle: concrete systems realise abstract ones

A system is a set of elements that relate or interact in a structured or orderly way.

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

This definition embraces both passive structures (e.g. tables) and activity systems.

The concern of GST is activity systems, in which structural elements interact in orderly behaviors.

 

A system takes two forms: a concrete system realises (or instantiates) an abstract system description (or type).

Abstract system description

Theoretical system

System description

Concrete system realisation

An empirical system

A system in operation

 

Abstract system description: a description or model of a concrete system.

It may be purely conceptual, or describe an imagined or envisaged reality, or describe an observed reality.

Abstract system description

The Dewey Decimal System

“Solar system”

Laws of tennis

Defined roles (e.g. Orchestral parts)

The score of a symphony

 

Abstract descriptions do take concrete forms; they are found in mental, documented and physical models.

What matters here is not the form but the relationship of the description (model, conceptualisation, idealisation) to a reality that is observed or envisaged.

 

Concrete system (aka System): a system that runs in reality.

It is realization in physical matter and/or energy of an abstract system description.

Abstract system description

The Dewey Decimal System

“Solar system”

Laws of tennis

Defined roles (e.g. Orchestral parts)

The score of a symphony

Concrete system realisation

Books sorted on library shelves

Planets in orbits

A tennis match

Actors (e.g. Orchestra members)

A performance of that symphony

 

Which comes first? Abstract system description or concrete system realization?

A designed concrete system (like a motor car) cannot run in reality until after it has been described, however abstractly.

A natural concrete entity (like the solar system) runs in reality before it is recognised and described as a system.

Principle: an entity is a system whenever and wherever it realises an abstract system description

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

 

“At this point we must be clear about how a "system" is to be defined.

Our first impulse is to point at [an entity displaying behaviors] and to say "the system is that thing there".

This method, however, has a fundamental disadvantage: every material object contains no less than an infinity of variables and therefore of possible systems.

Any suggestion that we should study "all" the facts is unrealistic, and actually the attempt is never made.

What is necessary is that we should pick out and study the facts that are relevant to some main interest that is already given.” Introduction to Cybernetics (1956) W. Ross Ashby

 

In other words: a substantial entity has well-nigh infinite detectable and measurable properties.

Any system description that entity conforms to can include only a tiny fraction of that entity’s describable properties.

We abstract relatively simple abstract system descriptions from infinitely complex realities.

A system is a perspective, a selective idealization of an observed or envisaged reality.

It is an island of stability imposed by a system describer in the ever-unfolding processes of the universe.

 

As Ackoff and Ashby said in their different ways.

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

The US economy, a church or IBM is a not a system; it is infinite possible systems.

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

 

To be called a system, an entity must exhibit (manifest, instantiate, realise) the properties of a system.

Until those properties have been described and observed, the entity is an unbounded, amorphous part of the universe.

An entity is a system whenever and wherever it matches an abstract system description.

 

An entity:

·         is a thing that has continuity of identity over time

·         is a system where and in so far as it realises an abstract system description.

·         can be zero, one or many systems at once.

·         can be different systems over time (e.g. caterpillar and butterfly).

 

People find this hard to understand and accept, but here goes.

It is meaningless to say a named entity is a system until you are sure it realises a system description.

Until there is an abstract system description an entity cannot fairly be presumed to be a system.

(An exception might be made for a life form, which may be presumed to realise the system described in its DNA.)

 

A philosopher is reputed to have said: “I know there is stuff out there; I am just not quite sure what it is.”

In what sense does the stuff called the “solar system” exist?

Its boundary, the structures it contains and how they behave, were not only defined by astronomers, but have also been changed by astronomers.

Today, real-world phenomena (planets in orbits) can be observed and measured as conforming – well enough - to the system description

Gradually, those phenomena will change until they are longer recognisable as matching that description.

 

GST note: The universe, or IBM, is an ever changing entity in which stuff happens.

A concrete system is

·         an island of repeatable behaviors carved out of that universe.

·         a set of describable entities that interact in describable behaviors.

·         an entity we can test as doing what an abstract system description says.

With no system description, there is no testable system, just stuff happening.

Principle: realisation differs from translation

Here, idealisation or conceptualisation means abstracting a description from a reality that is observed or envisaged.

Whereas translation means transforming one description of a reality into another description of the same reality.

 

Realisation = the operation of a concrete system that is testable against a system description
The US constitution is a document that conceptualises the structures and behaviors of the system that is a US government.

This public document was agreed by its authors and is understandable by all who share the authors' understanding of the words in it.

It has been realised in concrete / operational / run-time systems throughout the history of the US by successive government bodies.

Translation = transformation of one system description into another (more or less refined) description

The US constitution authors translated and collated their mental models into a documented model.
Readers of the published document translate it into their private mental models.

All documented and mental models are abstract system descriptions.

Obviously, documented models are more stable and shareable, which is why we create them.

We document models also so that we can demonstrably test the behavior of real-world systems against the models.

Principles about state, behavior and change

State: the current structure of a thing, as described in values of its variable properties.

Event: a discrete input that triggers a process that changes a system’s state, depending on the current state.

Process: one or more state changes over time, or the logic that determines which state changes lead to which other state changes.

System change: a change to the state or nature of a system.

System adaptation: a change to the state of a system, which changes its variable values.

System evolution: a change to the nature of a system, which changes its variable types.

Meta system: a system that defines a system or changes it from one generation to the next.

Principle: continuous behavior can be modelled as driven by discrete events

A concrete system’s property values realise property types or variables in its abstract system description.

 

System properties

Abstract description of system state

Property types (air temperature, displayed colour)

Concrete realization of system state

Property values (air temperature is 80 degrees, displayed colour is red)

 

The current state of a concrete system realises (gives particular values to) property types or variables in a system description.

Other qualities of that entity are not a part of that system, but might count as part of another system.

E.g. the temperature of the earth’s atmosphere is irrelevant to its role in the solar system, but vital to its role in the biosphere.

 

A system’s state may change continually, or in discrete steps

In either case, the system may be modelled in terms of discrete state changes that result from discrete events detected.

 

GST note: On discrete event-driven behavior.

External events cross the boundary from the environment into the system.

Within a system, internal events pass between subsystems.

In response to an event, a system refers to current system state.

It then “chooses” what actions to take, including actions that change its own state.

The choice depends on the values of input event variables and internal state variables.

Principle: system change differs from system state change

It is important to distinguish:

·         changing state within a system generation - system adaptation

·         changing nature between system generations - system evolution.

 

Change in biological systems

The biologists Maturana and Varela characterised living entities as autopoietic.

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

 

This table suggests there is no one definitive list of what characterises a life form.

According to this source

living entities share 8 characteristics

General System Theory

concepts

According to this source most

living entities have 7 characteristics

Heredity

system description

Adaptation through evolution

system change

Reproduction

system instantiation

Reproduction

Growth and development

system state change?

Growth

Cellular organization

system structure/composition

material input

Nutrition

Metabolism

material processing

Respiration

material output

Excretion

information input

Sensitivity

Response to stimuli

information processing

Movement

Homeostasis

information processing

 

From a GST perspective, these characteristics are interesting:

·         Sensitivity to stimuli - the entity can detect changes in the state of its environment.

·         Response to stimuli – the entity responds to events, typically by triggering some motor actions.

·         Homeostasis – the entity responds to internal and external state changes so as to maintain system state.

·         Reproduction – the entity creates a new generation, a new version, of the system.

 

Many kinds of system change can be distinguished in animal entities.

It is curious that nobody mentions decay and death (or taxes) as characteristics that define a living entity.

 

System adaptation: a change to the state of a system.

·         Update – the processes by which a system’s information state is changed to reflect its concrete state or environment.

·         Homeostatic adaptation – the processes by which a system responds to internal and external state changes so as to maintain its state variables in acceptable ranges.

·         Autopoiesis – the processes by which a biological cell, given simple chemical inputs, sustains/replicates its own complex chemistry.

·         Decay – the gradual loss of an entity’s ability to act according to a system description (if we didn't age and die, our species could not evolve).

 

System evolution: a change to the nature of a system.

·         Growth by maturation – the processes by which we develop from egg to adult.

·         Evolution by reproduction - the processes by which we creat a new, different, member of our species (system generation N + 1).

·         Learning – the processes by which a intelligent man or machine responds to differently to a new input as a result of recognising some “family resemblance” to past inputs (implies fuzzy logic).

·         Change by design – the processes by which a designed system is changed - the purpose of all enterprise and solution architecture efforts.

 

Evolution in biological systems

In the special case of a biology entity, there are many possible abstract system descriptions.

It may be argued that an organism’s DNA is the most fundamental description of an organism.

Incrementally, generation by generation, biological evolution changes the DNA of an organic system.

 

Evolution in designed systems

In the special case of a designed system, there must be at least one abstract system description (be it a mental or document model).

Incrementally, generation by generation, changes to a system description can be realised in designed system.

With no system description, there is no design, and no way to plan design and build effort.

 

EA languages like ArchiMate are used to describe and design business systems.

Enterprise architects analyse a baseline system and design a target system.

They then plan and govern the transition from the baseline system (version N) to target system (version N+1).

They presume changes to concrete systems are governed (under change control) against documented system descriptions.

FOOTNOTES on philosophy, science and GST

I don't mean to say anything below about how people idealise, or abstract concepts from, realities.

(Inheritance, learning, induction, deduction, trial and error, logical reasoning or guesswork; it makes no difference.)

I only mean to say that people’s concepts of realities are abstractions of those realities.

 

On the philosophy and science of description and reality

In reality, the sun emits radiation across a wide range of the electro-magnetic spectrum.

Only a small part of that spectrum is visible to us, and the colours we describe in that light are artefacts of our human senses.

Other animals see light differently:

“It was natural for scientists to assume that bird vision is like human vision… after all, birds and humans are both active by day, we use bright colors as cues.”

But… systematic testing of bird vision revealed something unexpected: Many bird species can see UV light.”

http://www.nwf.org/news-and-magazines/national-wildlife/birds/archives/2012/bird-vision.aspx

 

The relationship of description to reality is curious, and has been debated by philosophers for thousands of years.

At least some idealist philosophers believe that reality is out there – we observe it and envisage it – but can never know it directly.

Our minds idealise realities in the form of concepts or mental models.

For most day to day purposes, our concepts must model realities well enough, else we would not be here.

Let me call this “Scientific idealism”.

 

Scientific idealists presume universals (concepts) are abstractions of particulars (things) that help us recognise, predict and deal with future things.

Similarly, scientists’ theories abstract from realities.

(I don't mean to say scientists arrive at their theories by a process of abstraction from realities, or say anything about that process at all.)

To test an abstract theory, a scientist envisages how a future reality will match the theory, and then conduct a test.

In so far as reality matches prediction, they consider their theory adequate – for now.

 

On mental models

Scientific idealism is a philosophy that not only fits how science works, but also suggests how biological evolution led to intelligence.

Evolutionary biologists presume animals form mental models to help them deal with realities.

In response to events, and in accord with their mental models, animals act to achieve desired effects.

To act effectively, animals’ mental models must describe their environment accurately enough.

 

Nobody knows how a brain works, but that does not matter here.

What matters is only that the presence of mental models is empirically demonstrable.

The mark of an intelligent animal is its ability to abstract types from things, to help it recognise, predict and deal with future things.

For example, you see a thing approaching, you recognise it as an instance of a train type, you predict it will run you over, you step off the track.

This demonstrates you have a mental model of a train’s behavior on a track that describes reality well enough.

More generally, if animals’ mental models didn't conform well enough to reality, they would be unable to find food and avoid danger.

Thus, the ability of animals to form mental models is an outcome of biological evolution that enhances survival prospects.

 

Sharing of mental models through communication

Further, evolution has favoured organisms able to communicate mental models of their environment to their relatives.

Honey bees share experiences through communication of their mental models.

One bee tells other honey bees where it found the pollen - and those other bees fly off to share that experience.

Clearly, sharing mental models of experiences and visions gives honey bees an evolutionary advantage.

 

Naturally, members of a species tend to conceptualise realities in similar ways.

We share inheritance of common bio-chemistry; we learn the same thing from similar experiences (e.g. grasping nettles).

We further align our models by intra-species communication; and by accepting social pressures to conform to the norm.

 

Humans have special advantages when it comes to communication of mental models.

Eons ago, cave men were able to translate mental models into and out of sophisticated speech – a big step forward in communication.

Modern humans have the huge additional advantage of using written words and mathematical symbols.

We translate mental models into documented models for posterity, for agreement and for testing.

 

Realising abstract descriptions as concrete systems

As Scientific idealists presume universals (concepts) are abstractions of particulars (things) that help us recognise, predict and deal with future things.

So scientists’ theories are abstractions of realities that help them predict and deal with future realities.

And system architects system descriptions are abstractions of envisioned concrete systems that help us to build them.

An entity is a system whenever and wherever it matches an abstract system description.

 

In the special case of a software system, the run-time system is the concrete system, and the software is the abstract system description.

Perhaps the pinnacle of our descriptive ability is the ability to write a description so precise (in software) that a computer can realise it in an operational system.

When GST ideas were aired by von Bertalanffy, Boulding and Ashby in the 1940s and 50s, computers were not in the picture.

The subsequent birth of computer science can be seen as a vindication of the notion that there is a cross-science general system theory.

And now, artificial intelligence software is marked by its ability to abstract types from things to help it recognise, predict and deal with future things.

 

Software systems can be seen as a special application of GST.

However, they don’t exhibit some features von Bertalanffy’s speculated might be “general” to all systems.

That is why some of von Bertalanffy’s notions don’t appear in this update of GST.

And why homeostasis, a focus of early system theorists, is not presented here as a general system property.

 

Activity systems are temporal

An entity is a system whenever and wherever it matches an abstract system description.

In the special case of an activity system, the system description defines behaviors.

When an activity system stops exhibiting behaviors, the system no longer exists.

When a living tiger’s organs stop working, the tiger dies.

When the planets orbiting our sun fall out of orbit, there will be no solar system.

When the players in a tennis match go home, the tennis match is finished.

 

What remains when an activity system stops running?

An abstract system description may persist long after the concrete system has disappeared.

But the question here is rather: what remains of the concrete system?

In some cases, the system rests in the form of a structure containing parts that can be started up again.

E.g. An airplane’s mechanical parts are dedicated to their role in flying the plane

When an airplane rests in a hangar the parts stop playing their roles.

But they remain contained within the airplane’s carcass, and may resume their roles the next day.

 

When a bank closes for the night, its employees stop playing roles in it.

Overnight, those employees play other roles, in families, in bars, in part time jobs as football coaches.

Those people are not “parts” of the bank system in the sense they are dedicated to the bank, or the bank “contained” them.

Rather, they are actors hired to play roles in the bank system.

The “part” they play in the system is limited to the activities expected of those roles.

Outside of that, they are participants in the bank viewed as a social entity rather than viewed as a system.

An entity is a system whenever and wherever it matches an abstract system description.

 

 

All free-to-read materials at http://avancier.website are paid for out of income from Avancier’s training courses and methods licences.

If you find the web site helpful, please spread the word and link to avancier.website in whichever social media you use.