The philosophy of system theory and system architecture

Copyright 2016 Graham Berrisford. One of about 300 papers at http://avancier.website. Last updated 17/10/2017 23:03

 

Enterprise architecture is about business system planning.

It can be seen as applying the principles of general system theory – which we’ll get to later.

First, to understand system theory is to make a paradigm shift as radical as is needed to understand Charles Darwin’s evolution theory.

The aim here is not to present a new philosophy or theory (the paradigm shift is evident in Ross Ashby’s system theory).

It is to present a coherent and consistent philosophy, compatible with modern science.

Side effects of the discussion include challenging an ISO standard and rearranging Karl Popper’s “three worlds” view of the universe.

 

Contents

The necessary paradigm shift 1

Description and reality. 2

What is the nature of description?. 3

How accurately do we describe things?. 6

What does it mean to exist?. 7

An aside: on the value of pi and the instance/token distinction (help needed here) 9

Mental models. 10

Externalisation of mental models. 11

Types in natural verbal languages. 12

Science as a kind of idealism.. 13

Abstraction of system description from concrete system realization. 14

Conclusions and remarks. 16

Footnote 1: The questionable triangle in the ISO 42010 standard. 18

Footnote 2: The questionable triangle of Popper’s three worlds (a tentative analysis) 20

 

The necessary paradigm shift

You may find what follows is straightforward.

You may find it is at odds with some instinct or philosophy you already adhere to, and therefore find difficult to put aside.

Many people’s instinct is to divide the universe into mental and physical worlds.

However, the two-way mental/physical dichotomy of Cartesian Dualism has long been rejected by philosophers and scientists.

And since Descartes, people have proposed various three-cornered views of description and reality.

Some are drawn to "the semiotic triangle" and/or Karl Popper's "three worlds view" of the universe.

This work proposes these are not the best way to model the describer/description/reality trichotomy.

 

First, one has to shake off:

·         the mental/physical dichotomy presumed in Cartesian Dualism (really shake it off, further than Popper did)

·         a language-centric view of philosophy (really shake it off, unlike Wittgenstein)

·         a human-centric view of the universe.

 

Then acknowledge that:

·         describers are actors who have some intelligence about their environment

·         the ability of actors to describe the world is a side effect of biological evolution

·         in natural intelligence, the mental world is a physical world - though the bio-chemistry of that is deeply mysterious

·         in artificial intelligence, the most notable ability is the ability to abstract descriptive types from observations of similar things

·         describers and descriptions are themselves part of reality - and can themselves be described (if need be).

 

And finally, acknowledge that:

·         descriptions include models of every kind, both private and public

·         mental, spoken, documented and physical models are all descriptions

·         humans and their machines can translate every description of one kind into a description of another kind.

 

The philosophy here can be expressed in a triangle.

Description theory

Descriptions (private & public)

<create and use>                   <idealise>

Describers     <observe & envisage>     Realities

 

The distinction between private and public description is important, especially in any collaborative design exercise.

But it is not as fundamental as the description/reality distinction, and the need for describers to be included in the picture. 

The premise here is that description started in the domain of biology, and the ability to share descriptions is the basis of sociology.

Many kinds of animal share their private mental models through social communication.

Human social systems are more elaborate, and some have been formalised into business systems.

Description and reality

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

Countless philosophers have proposed countless philosophies – both overlapping and contrary.

Today, there is little debate about the existence of material things.

Most presume that there is physical matter/energy out there – the questions are rather:

1.      What is the nature of description?

2.      How do we test things match descriptions?

3.      How accurately do we describe or measure things?

4.      What does it mean to exist?

 

Thinking about these questions usually starts by distinguishing universal descriptive types from particular things that instantiate them.

E.g. A circus ring instantiates (embodies, exemplifies or realizes) the circle type (how accurately it does this is to be discussed later).

 

Universal type

“Circle”

described using other types such as “radius” and “circumference”.

Particular thing

A circus ring

described by giving particular values to the types “radius” and “circumference”.

 

One type can be realised by many things: e.g. the “arena” type is realised by very many circus rings and theatre stages.

One thing can realise many types: e.g. a circus ring realises the two types “circle” and “arena”.

 

Universal type

“Circle”

“Arena”

“Arena”

Particular thing

A circus ring

A theatre stage

 

Any material structure can instantiate very many unrelated types.

So to describe all the properties of a structure is impossible, and the attempt is never made.

We must focus on properties relevant to some type(s) of interest to us.

 

Our main concern is describing activity systems rather than passive structures.

This paper extends the normal type/instance distinction by dividing instantiations into behaviors and structures.

 

Behavior type

“Orbit”

“Rotation on axis”

Behavior instance

An orbit

A rotation

Entity that performs behavior

A planet

 

Again, one behavior type can be realised by many things: e.g. the “orbit” type can be realised in many orbits by many planets and asteroids.

One thing can realise many behaviour types: e.g. a planet can realise the types “orbit”, “rotation on axis”, “rocky” and “smaller than earth”.

 

People often point at an identifiable entity and call it “a system”.

However, a social or business entity can realise very many unrelated and uncoordinated behaviors.

E.g. one social group can realise the two behaviour types “card game” and “dinner party”.

 

Behavior types

 “Card game”

“Dinner party”

Behavior instances

A game of bridge

A game of whist

A dinner

Entity that performs behaviors

The social group: John, Joe, Joan and Jo

 

A business can be seen as a social entity in which actors play roles in many systems

Those business systems may be inconsistent, unrelated, or even in competition with each other.

To study all the behaviors an entity realises is impossible, and the attempt is never made.

To apply system theory is to select and study those behaviors of an entity that are relevant to some given aims(s) and/or concern(s).

What is the nature of description?

 “In describing a situation, one is not merely registering a [perception], one is classifying it in some way, and this means going beyond what is immediately given.”

Chapter 5 of “Language, truth and logic” A J Ayer.

 

To describe particular things is to relate them to general property/quality types.

We describe two-dimensional shapes in terms of universals/qualities such as “circular”, “rectangular”.

We describe particulars/things in terms of universals/qualities such as “beautiful”, “yellow”, “planet-like”, “dangerous” and “friendly”.

Such universal qualities are also known as types, concepts or properties.

 

“The problem of universals”

The existence of universals has been debated by realist and idealist philosophers since Plato.

One popular source distils the debate thus.

“Taking "beauty" as example, three positions are:

·         Platonic realism: beauty is a property that exists in an ideal form independently of any mind or description.

·         Aristotelian realism: beauty is a property that exists only when beautiful things exist.

·         Idealism: beauty is a property constructed in the mind, so exists only in descriptions of things.” (“The Problem of Universals” Wikipedia 2017.)

 

Philosophers have proposed variants of realism and idealism, and alternatives.

However, the classification above is enough for us here.

 

Carving colours out of light

Consider the observable reality commonly called “colours”.

The sun emits radiation across a wide range of the electro-magnetic spectrum..

A tiny part of that spectrum is visible to humans in the form of a rainbow.

There were no colours in the rainbow until organisms could perceive and distinguish them.

Rainbow colours

7 colours of the rainbow

<defined>            <abstracted from>

Newton     <observed>  The visible spectrum

 

Different people have divided the rainbow into differently-named colours.

And other animals see a different rainbow.

“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

 

Carving roles out of realities

Consider the observable reality commonly called “the solar system”.

Before life on earth, it did not have a description, awaiting discovery.

Rather, astronomers created a description to help them understand and predict the reality.

 

Our solar system

“Planet”

<define & refer to>                <idealises>

Astronomers <observe and predict> Orbital motion

of bodies in space

 

Our “solar system” has changed: astronomers have added/subtracted bodies to/from the set of planets.

Meanwhile, those bodies have carried on as they did before – uninfluenced by any typification of them as planets.

 

It is impossible to describe every potentially describable feature of a reality.

E.g. to define a planet’s role in the solar system is to say little about any individual planet, and nothing about life on earth.

Similarly, business role definitions say very little about the human actors employed to play the roles.

A role can define only very limited aspects of a person’s behavior.

And regardless of their role descriptions; employees carry on breathing, eating and drinking.

 

Business systems

Roles

<create and use>             <idealise>

Employers     <observe and envisage>   Activities

of actors

 

All definitions of the solar system and of business roles are created and used by mankind.

They exist in two forms:

·         An abstract description

·         One or more realisation(s) of that description in concrete matter and motion.

 

Abstracting (idealising) a description from realities

Most believe that reality is out there – we can observe and envisage at least some of it.

However, most scientists believe the universe is not fully understandable by us.

We can only know reality to the extent we can perceive and describe it.

To help us navigate the world and predict its behavior, we name and describe things we perceive and envisage.

And to describe a particular thing, we refer to types (like “rose”) previously learnt or created.

 

You may reasonably say pink roses existed before man saw them.

But they could not be typified as either "roses" or "pink” before man created those types to describe them.

On their own, the words “rose” and “pink” are not types; they are only the names of types – of plant and colour varieties.

A type name becomes meaningful to us only when associated in our minds with one or more mental models (discussed below).

 

Instantiating (realising) a description in realities

Given a type, we can envisage the existence of several things matching that type.

Given one description, several realities can match it (cosmologists now talk of parallel universes).

Given one abstract system description, several concrete systems can realise it.

 

This triangle expresses some essential relationships between describers, descriptions and realities

Description theory

Descriptions

<create and use>         <idealise>

Describers <observe and envisage> Realities

 

How accurately do we describe things?

We test the truth of a description by either:

a)      pure logic – starting from some presumed axioms – as in mathematics

b)      measuring property/quality instances – in most practical situations.

 

The circumference and diameter properties of a circle are always related by the same ratio (pi), so one can be derived from the other.

But even that purely logical calculation is only as accurate as the value used for pi (3.1? 3.142? etc.)

And in measuring real world things, there is always a margin of error.

 

Strict types in pure mathematics

In pure mathematics, types are perfect, incontrovertible, immutable forms.

Consider these types: number, prime number, ratio, triangle, quadrilateral, rhombus, parallelogram, rectangle, square.

Mathematicians often treat types as strict - or monothetic.

They presume a thing conforms to a type if and only if it instantiates (gives a value to) every property type of that type.

 

Unfortunately, discussion of types tends to be dominated by mathematicians and philosophers who live in this world of abstract description.

They devise logical rules for manipulating abstract types and things: e.g. circle circumference = circle diameter * pi.

Some look at nature as a realisation of their rules, rather than look at their rules as idealistic and questionable models of nature.

 

Fuzzy matching of things to types in applied mathematics

There are no perfect circles in the natural matter and energy of universe.

There are only countless shapes that are measurable as approximating to the "circle" type.

 

As soon as the "circle" type was first conceived, many things were measurable as approximating to it.

What is the value of the C, the total number of circles in the universe today?

The answer depends on the degree of accuracy you demand in measuring the radius of a shape from its central point.

And the accuracy of the measurements that are made.

And perhaps worse, the number will change during the measurement process.

So, the value of C could only ever be an approximation – a fuzzy truth.

 

Fuzzy types in the natural biological world

Things were typified long before mathematicians, or even humans, came along.

How would Charles Darwin have looked at “the problem of universals"?

He would see the process of typifying things (e.g. into food, friends and enemies) as a natural biological process.

The survival of biological entities is enhanced if they can remember types and recognize things that conform to them.

 

Types are descriptive tools that evolved because they help organisms to survive.

However, no animals naturally classify things into types in a strict or perfect way.

Rather, the types important to them are fuzzy (or polythetic) and/or their pattern matching process is fuzzy.

 

For example, birds peck at things that look to match the food type, and sometimes mistakenly swallow inedible items

Although animals do make such errors, it seems likely that fuzziness in their descriptions of the world is essential to life.

They cannot adapt to their experience unless their mental models are flexible, and can evolve.

 

So, fuzzy types emerged in biology long before humans and their verbal languages.

These types must somehow be encoded in the biology of an organism’s body (perhaps but not necessarily in its brain).

And they existed in the universe before the written word, before the spoken word, and before mankind

 

However, there is no good reason to suppose types existed before life.

And no reason to suppose types exist in an eternal, ethereal or Platonic form outside of physical matter and energy.

What does it mean to exist?

Remember the distinction between realist and idealist philosophies?

·         Do you side with Plato: The “ratio” type exists in an ideal or ethereal form independently of any mind or description?

·         Do you side with Aristotle: The “ratio” type exists only when perceived by a person?

·         Do you side with Idealists: The “ratio” type is constructed in the mind, and so exists only in description?

·         Or adopt some other position?

                              

Many philosophical debates come down to arguments about what words mean.

The questions above (especially the word “exist”) can be interpreted and answered in various ways.

So, how to choose the best answer – one that fits system theory?

 

There are several objections to the notion of an ethereal or Platonic ideal.

First, people disagree what a named type (say “beauty”) means.

They disagree which things embody or give a value to the type “beauty”, and how to measure it.

 

Second, after people agree a definition of a named type, they can change it.

Pluto existed in space before man detected it, though unnamed at that time.

It was then described as a "planet", because the "planet" type had already been created by man.

Later, having typified Pluto as a planet, astronomers changed their mind, and declassified it.

 

That probably limits Platonic ideals to the seemingly eternal types of pure mathematics.

In pure mathematics, not only types but also things are abstract descriptions.

E.g. The values of pi and C (the total number of circles in the universe today) are abstract descriptions.

 

So, let us consider the mathematical type “ratio” and the value of pi – both of which are useful to mankind.

Our hypothesis is this: either the type “ratio” or the value of pi has existed for eternity in an ethereal or Platonic form.

Occam and Popper have given us two useful tools to examine this hypothesis.

·         Occam's razor: is this hypothesis necessary or practically useful? No, it is redundant.

·         Popper's principle: could this hypothesis ever be disproved by measure or experiment? No, so it is not a scientific concept.

 

We may not need the Platonic ideal, but many (especially mathematicians like Roger Penrose) want it; they cannot resist it

And yes, there are types so commonly created, so widely understood and so useful that they are unavoidable in practice.

We can reasonably conclude these types do accurately represent or encode something that is inevitably repeated in reality.

But however accurately it models reality, a type is an abstraction from some structure or behaviour that is repeated.

The type itself has no concrete existence – until it is described.

 

There were no types before life

The idea advanced here is that the roots of typification are in biological evolution rather that mathematics.

And the only useful meaning of “to exist” is - to be found in physical matter/energy, including physical biology.

 

Of course, roses were pink before anybody looked at them; but that does not mean the type "pink" existed before life.

If you say it did, then you have to presume an infinity of conceivable types (including imaginary ones like “unicorn”) have always existed.

 

Of course, one thing was twice as heavy as another before anybody weighed them; but that does not mean the type "ratio" existed before life.

To propose it did is a) not necessary or useful in a practical way and b) not disprovable.

Occam and Popper lead us to the view that there was no "ratio" type before life; it did not exist.

 

If there was no “ratio”, “diameter” or “circumference” type before life, then there was also no value for pi either.

That does not mean the value we ascribe to pi in inaccurate or untrue; it only means it is a man-made description rather than a reality.

(Having said, it is presumed the complete/accurate value of pi cannot be computed as number of finite length).

 

Surely, no concept can exist without a conceiver to conceive it; there can be no description without a describer to make it.

The "circle" type has been conceived and described by many.

It exists now in countless mental, documented and other models made of things in the universe.

If all those models were deleted today, the concept would disappear.

But “circle” would be reinvented tomorrow, because it is such a readily conceived and practically useful model.

The next person to look at the sun and the moon would surely re-create the “circle” type.

 

What is said of types can be said of descriptions

An unexpected outcome of systems thinking is the conclusion that there were no types before life

Furthermore, what is said of types can be said of descriptions.

 

The colour “yellow” and “the solar system” are man-made descriptions of realities.

Descriptions are created and maintained in physical forms (in the mind, on paper, however) by living entities.

Descriptions exist only in those physical forms – not in any ethereal or eternal sense.

Delete all physical instances of a description, and it no longer exists.

 

Suppose a description is deleted from all memories - mental and documented.

Surely, a complex description (say, a Beethoven symphony score) would be lost forever.

But a simple description (say, of a triangle, or a unicorn) might soon be reinvented and so exist again.

And any description that seems universally true (say, Newton’s second law of motion) will be reinvented - even though is not universally true!

An aside: on the value of pi and the instance/token distinction (help needed here)

I am far from sure the last paragraph in this section is correct, or that I understand the instance/token distinction at all.

 

It doesn’t take a genus to recognise there is a relationship between the types "circumference" and "diameter".

It is clear that if you increase one you increase the other.

That relationship is a particular instance of the universal "ratio" type.

And since that ratio relates 1 universal type to 1 other universal type, it is a relationship with only 1 instance.

 

As you know, the particular value of that particular ratio is called pi – it has both a name and a value.

The first calculation of pi’s value was made by Archimedes of Syracuse (287–212 BC).

The ancient Egyptians appear to have used the value 3.1605, but now we use an abbreviation of the irrational number that starts 3.1427.

 

Mathematicians have found several algorithms that will accurately calculate pi to many decimal places.

One of the simplest is the Gregory-Leibniz series, which, though inefficient, gets closer to pi with every iteration

After 500,000 iterations, it produces a value for pi that is accurate to 5 decimal places.

 

Like "diameter", pi is the name of an abstract concept, but unlike "diameter", it has one and only one true value.

Let us focus on the one commonly-used approximation to the value that is 3.14.

Confusingly, that 3.14 value appears many times in many mental and documented models.

Does this mean that the value is both a particular value and a type that has many instances?

You might say that, but the multiple appearances of the same value are simply copies of it.

 

I have the impression philosophers distinguish copies from instances thus:

·         the multiple uses of the ratio type (e.g. in circle and triangle shapes) are “instances” – each giving a different value to the type named "ratio".

·         the multiple numbers that are approximate values of pi are “instances” – each giving a different value to the type named "pi".

·         but the multiple copies of one pi value in different places are called “tokens” – since each is the same.

 

Is that last right?

Mental models

Answers have emerged to the four questions at the start.

1.      A thing is described with reference to general property/quality types.

2.      We test things match descriptions by measurement of property/quality instances.

3.      We can describe or measure things only imperfectly, with a degree of truth.

4.      To exist means to be found in physical matter/energy.

 

That last is counter intuitive to many people.

They believe their thoughts exist, but are immaterial.

Historically, theologians drew a division between what is mental and what is physical.

And Cartesian dualism was the position that the universe divides into mind and matter.

Descartes declared that the mental can exist outside of the body, and that the body cannot think.

And that the mental and material worlds interact via the body.

 

Cartesian dualism

Minds   <perceive and manipulate>  Matter

 (via bodily sensors and motors)

 

The view taken today in cognitive science and psychology is that the mind has a physical biological basis.

Nobody knows exactly how brains work or how memories are stored.

Or understands the processes by which animals idealise, or abstract concepts from, realities.

One may speculate the brain has a communication stack from chemistry to consciousness, with different processes at every level.

But none of those things matter here.

 

All that matters is that we can prove the existence of a “mental model”.

Can an animal maintain a mental model that is accurate enough, for long enough, to recognise things it perceived before?

Can it find food, avoid danger, find a mate and reproduce?

Can it recognise similar events and respond to them by acting appropriately - to achieve a desired effect?

Tests show yes - it can.

 

Presumably, none of us inherits a mental model of pink roses, since they are not critical to human existence.

We have two other ways to form a mental model of pink roses, which may be used separately or together:

·         By perception: remembering perceptions of pink things and rose-like things.

·         By description: hearing or reading descriptions that associate “pink” and “rose with other type names we already have mental models of.

 

Descriptions appear in physical forms, in mental, oral and written models, and can be translated between those models.

Physically, descriptions need not even faintly resemble the realities they describe – they are instead encoded representations of them.

Externalisation of mental models

Biological evolution led first to animals with the ability to form and retain internal mental models.

The more intelligent the animal, the greater its ability to abstract universal types from particular things.

And this extends its ability to recognise and manipulate things of those types.

 

Evolution led further to organisms able to communicate mental models to their relatives.

The communication can take many physical forms; facial expressions, gestures, noises, etc.

Famously, honey bees share mental models of their world by wiggling/dancing.

Naturally, members of the same species tend to conceptualise realities in the same ways

Because they share a common bio-chemistry, have similar experiences and similar needs.

 

Humans evolved the ability to translate mental models into and out of oral speech – a huge step forward in communication.

We further align our mental models under social pressure to conform to the norm.

And we have a second huge advantage when it comes to communication of mental models.

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

 

It may seem natural to draw a division between internal mental models and external communications and documented models.

However, we continually, well nigh automatically, translate between mental models, spoken words and written words.

Translating mental models into/out of speech is no different in principle from translating speech into/out of writing.

And no different in principle from translating up and down the mysterious communication stack from chemistry to consciousness

(For more, read Information and communication)

 

Remember that human describers translate descriptions between mental and documented models.

A mental model is unconsciously encoded in a physical biochemical form – which is fuzzy, fluid and easily forgotten.

A documented model is more or less consciously encoded in a different physical form – which is more stable and persistent.

(For more, read Description.)

Types in natural verbal languages

Verbal languages vastly increase the ability of human actors to create and use types.

This table represents the abstraction of types from particular situations, by describers.

 

Type theory

Universal Types

<create and use>               <idealise>

Describers <observe and envisage> Situations

 

Consider the peculiarly human situation that is the telling of a particular joke.

 

A particular/real joke

John

Joan

Knock knock

Who’s there?

Ivan

Ivan who?

No, Ivanhoe

Groan

 

You surely recognise this as a knock knock joke, and would describe it as one.

That is to go beyond what is immediately given, and classify the joke.

Thus, you abstract from the particular thing to a universal type.

This table documents that abstraction, putting the variable names in italics.

 

The universal/abstract joke

Actor in initiator role

Actor in responder role

Knock knock

Who’s there?

Name

Name who?

Name elaboration

Groan

 

In short, realities do not have descriptions, awaiting discovery.

Rather, describers create descriptions to help them understand and predict realities.

They abstract a general or universal type from properties attributable to similar things.

They do this first in the mind, before they communicate it or record it.

(For more, read Information and communication, Language, Knowledge and Description.)

Science as a kind of idealism

Remember the distinction between realist and idealist philosophies?

Take Newton’s second law of motion (f = ma) as an example.

·         Do you side with Plato: The law exists in an ideal or ethereal form independently of any mind or description?

·         Do you side with Aristotle: The law exists only when measured in the motion of a thing?

·         Do you side with Idealists: The law is constructed in the mind, and so exists only in description?

·         Or adopt some other position?

 

A description arises from the interaction between the describer and the thing observed

In quantum mechanics, Heisenberg’s uncertainty principle says we cannot accurately measure both the position and speed of a particle, since any measurement affects its properties.

In this connection, the famous double slit experiment <https://en.wikipedia.org/wiki/Double-slit_experiment> produces counter intuitive results.

If the position of one particle (at one or other slit) is measured, it goes through only one slit, and does not produce an interference pattern.

If the position of one particle is not measured, it must have gone through both slits, since an interference pattern appears on the screen

In the latter case, the particle’s path is not so much uncertain as certain – it has travelled two paths and been in two places at once.

The largest entities for which the double-slit experiment has been performed were molecules that each comprised 810 atoms.

 

According to the relational interpretation of quantum mechanics, first proposed by Carlo Rovelli,

“Observations such as those in the double-slit experiment result specifically from the interaction

between the observer (measuring device) and the object being observed (physically interacted with), not any absolute property possessed by the object.”

 

In other words, a description arises from the interaction between the describer and the thing observed.

A description is not any absolute property possessed by the thing itself.

 

Scientific hypotheses

The word science is rooted in an old word for knowledge.

Science might be simplified to three propositions.

1.      Scientists <observe and predict> the universe – meaning the behaviour of entities observable in reality

2.      Scientists <create and use> hypotheses - which are regarded as knowledge if they successfully predict real-world behaviour that is measured.

3.      Hypotheses <conceptualise> the universe – meaning the behaviour of entities observable in reality

 

The three propositions are arranged in a triangle below.

Science

Hypotheses (& Knowledge)

<create and use>               <conceptualise>

Scientists      <observe and predict>    The universe

 

This view of science is “idealist” rather than either Platonic or Aristotlean realist.

Curiously, you might find others call the philosophy of science “realism” or “materialism”.

So why call it idealism here? Partly, because it matches the distinctions quoted above.

And partly because it matches a common sense reading of that term - as used in enterprise architecture frameworks.

 

Again, remember that human describers translate descriptions between mental and documented models.

A mental model is unconsciously encoded in a physical biochemical form – which is fuzzy, fluid and easily forgotten.

A documented model is more or less consciously encoded in a different physical form – which is more stable and persistent.

Abstraction of system description from concrete system realization

Finally, to understand system theory is to make a paradigm shift as radical as is needed to understand Charles Darwin’s evolution theory.

I say that because I find most people find it difficult to understand the paradigm shift expressed in 1956 by Ross Ashby.

 

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

Our first impulse is to point at [a concrete entity repeating a behavior] 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.” Ashby 1956

 

Remember: to apply system theory is to select and study those behaviors of an entity that are relevant to some given aims(s) and/or concern(s).

The describer must abstract from the infinite describable facts that could be found in observing the activities of the concrete entity.

 

"Cybernetics does not ask "what is this thing?" but ''what does it do?" It is thus essentially functional and behavioristic.”

“[It] deals with all forms of behavior in so far as they are regular, or determinate, or reproducible.” Ashby 1956

 

In other words, our concern is activity systems rather than passive structures.

This table identifies three levels of thinking about the realisation of an activity system.

 

Concept

System 1

System 2

Abstract system description

(mental, spoken, documented, mechanical, whatever)

A symphony score

The design-time code of a computer program

Concrete behavior

(a behavior that matches the above well enough)

A performance of the above

A run-time execution of the above

Concrete entity (which performs the above)

The orchestra members in a concert hall.

A computer in a data centre.

 

Again one entity can realise many systems; e.g. one computer can perform many unrelated programs at once

And you may realise the three systems in the table below, at least two of them at the same time (three at once is not advisable).

 

Concept

System 3

System 4

System 5

Abstract system description

“Oxygen to carbon dioxide converter

“Gene reproduction”

“System modelling”

Concrete behavior

Person breathing

Person making love

Drawing ArchiMate diagrams

Concrete entity

A person – you for example

 

As Ashby said: our first impulse is to point at a concrete entity repeating a behavior and to say "the system is that thing there".

However, an entity is rightly called a system only when and where it conforms to a system description.

No entity is rightly called a system without reference to a specific system description or model – one to which the entity demonstrably conforms.

 

In practice, it is normal to regard the concrete entity as part of the system realisation.

Because the remainder of that entity, and whatever it does outside the system of interest, is out of scope.

So, the next table wraps up the concrete entity and behavior into concrete system description, adds the describer into the picture.

 

Describer

Astronomers

LTA

Composer

Playwright

Abstract system description

“Solar system”

The laws of tennis

The score of a symphony

The roles in a radio play

Concrete system realisation

Planets in orbits

A tennis match

A performance of that symphony

Actors playing those roles

 

The premise is that all of these are all material in form, but the abstract system description is a passive structure rather than an activity system.

 

Concretion

The interaction between describers and real things is especially intimate in human and/or computer activity systems.

The describer’s premise is that a concrete system will realise (gives values to) variables in an abstract system description.

For example, the members of an orchestra will give values to the notes in a musical score.

 

In operation, the concrete system must realise (embody, instantiate) the property types invented by the designer.

And we test this by observing and measuring the values the system gives to those property types.

To run at all, the system must employ actors (human and computer) who can read the types in a description and assign values to them. 

 

Abstraction

By necessity, a description abstracts from most aspects of what it describes.

It is impossible to describe the infinite potentially describable feature of a real-world entity.

But it is possible to describe every feature of a system – because the system is whatever it is described it to be.

 

·         a definition of the solar system says nothing about life on earth.

·         the score of a symphony says nothing about the personalities of orchestra members.

·         role definitions usually say almost nothing about the human actors that play the roles.

 

If you consider friendliness to be an important property of your run-time system, then you should put it in your role definitions.

The only features of the real-world entity that count as part of the system are the features included in your system description.

 

Architecture

The level of abstraction in any system architecture is a matter of choice.

The architecture of a computer program can be described at several levels, from program code up to very abstract architectural models/diagrams.

The “system of interest” to a system architect extends so far as the architect chooses to describe the structures and behaviors of the system.

Conclusions and remarks

The universe includes planets and plants, memories and messages, tractors and tennis matches, symphony scores and performances, brains and computers.

We know these things as we describe them - in mental, documented and other kinds of model.

We cannot know them in any other way.

Descriptions are physical: whether organic or inorganic, internal or external, mental or documented.

Descriptions of one kind are translatable into descriptions of another kind.

The distinction between description and objective knowledge is fuzzy and/or ephemeral.

Other papers explore: Information and communication,  Language,  Knowledge and Description.

 

Realising abstract descriptions as concrete systems

This table draws correspondences between different views of description and reality.

Viewpoint

Description

Realities

Philosophy

A universal concept

helps people to understand and describe

particular things

Science

A scientific theory

helps people to understand and predict

real world events

System architecture

An abstract system description

helps people to understand and build

concrete systems

Software architecture

A software program

enables computers to make

run-time systems

 

When system theory 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 idea that there is a cross-science general system theory.

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

Today, artificial intelligence software is marked by its ability to abstract types from things to help in dealing with future things.

 

The philosophy of system theory

In short: every system description abstract facts important to the describer, from the infinite describable facts that could be found in the reality of the system that runs.

The view of science above can be generalised beyond science as follows.

1.      Describers observe and envisage realities (which they perceive as composed of discrete entities and events).

2.      Describers create and use descriptions (stored in memories and conveyed in messages using brains, speech, writing, pictures and other forms).

3.      Descriptions idealise realities (either mimic selected features of things, or represent them in some encoded form).

 

This is both our description theory and our philosophy of system theory.

It can be expressed more graphically in our triangle.

Scientific Idealism

Descriptions

<create and use>                <idealise>

Describers     <observe & envisage>     Realities

 

Philosophers may well reject using the term "idealism".

But I am wary of giving this philosophy a name already used by others, and don't have a better term for it.

And whatever it is called - I believe the paper presents a rational argument for this triangular philosophy as a platform for system theory and EA.

 

FOOTNOTES ON OTHER TRIANGULAR VIEWS

Some are drawn to "the semiotic triangle" and/or Karl Popper's "three worlds view" of the universe.

This paper proposes these are not the best way to model the describer/description/reality trichotomy.

One has to acknowledge that:

·         describers are actors who have some intelligence about their environment

·         the ability to describe the world is a side effect of biological evolution

·         in natural intelligence, the mental world is a physical world - though the bio-chemistry of that is deeply mysterious

·         in artificial intelligence, the most notable ability is the ability to abstract descriptive types from observations of similar things

·         describers and descriptions are themselves part of reality - and can themselves be described (if need be).

·         descriptions include models of every kind, both private and public

·         mental, spoken, documented and physical models are all descriptions

·         humans and their machines can translate every description of one kind into a description of another kind.

 

The philosophy here can be expressed in a triangle.

Description theory

Descriptions (private & public)

<create and use>                   <idealise>

Describers     <observe & envisage>     Realities

 

The distinction between private and public description is important, especially in any collaborative design exercise.

But it is not as fundamental as the description/reality distinction, and the need for describers to be included in the picture. 

The premise here is that description started in the domain of biology, and the ability to share descriptions is the basis of sociology.

Many kinds of animal share their private mental models through social communication.

Human social systems are more elaborate, and have been formalised into business systems.

Footnote 1: The questionable triangle in the ISO 42010 standard

ISO 42010 expresses system architecture in this curious three-way relation.

 

ISO 402010 standard triangle

 1Architecture Description (public)

<is expressed in>          <identifies>

1 Architecture     <is exhibited in>     System

 

Premises for analysis of the ISO 42010 triangle

A type is more than a type name, it is a type definition composed of property types associated with that type name.

E.g. A “square” is a “rectangle” <with> “sides” <of> “equal length”.

 

Axiomatically:

·         1 Type definition <can describe> N Things - which each instantiate that 1 type.

·         1 Thing <can exhibit the property types in> N Type definitions - which each abstract different properties from that 1 described thing.

 

And in the philosophy of system theory:

·         1 System Description <can describe> N Concrete Systems - which each instantiate that 1 description.

·         1 Concrete System <can exhibit the property types in> N System Descriptions - which each abstract different properties from that 1 described system.

 

Applying system theory, a “system architecture” must be one or both of these two things.

·         A collection of inter-related property types recorded in a System Description.

·         A collection of inter-related properties instances observable in a Concrete System.

 

It seems reasonable to say an Architecture is a high-level System Description and express the relations thus.

·         1 System (Architecture) Description <can describe> N Concrete Systems

·         1 Concrete System <can exhibit the property types in> N System (Architecture) Descriptions.

 

(Like most people, I use the term System ambiguously for the description, the reality or both.

But when the difference matters, I say “abstract system description” or “concrete system realisation”, or something along those lines.)

 

Analysis of the ISO 42010 triangle

ISO 42010 expresses system architecture in this curious three-way relation.

·         1 Architecture Description <identifies> 1 System, which < exhibits the property types in> 1 Architecture, which <is expressed in> 1 Architecture Description.

 

Consider this first two-way relation.

·         1 Architecture Description <identifies> 1 System.

 

Why only 1 System? Is that the System in somebody’s head? Whose head?

Is it the Platonic idea of the System, or the Architecture Description? Which of them?

 

Now consider the other relations.

·         1 System < exhibits the property types in> 1 Architecture which <is expressed in> 1 Architecture Description.

 

Where else could the property types be, other than in an Architecture Description - be it mental, documented or other?

What is the ISO 42010 Architecture? Is it the Architecture Description in somebody’s head? Whose head?

Is it the Platonic idea of the System, or the Architecture Description? Which of them?

 

Discussion with the ISO 42010 editor of the 1-1-1 triangle led me to conclude the standard's "system" and "architecture" are both Platonic ideals, which seems to duplicate a redundant concept.

 

ISO 42010 triangle revised

Suppose we agree the System is a reality out there, but take the ISO 42010 view that all is defined from the perspective of 1 and only 1 Architecture Description

Then the relations in the standard could correctly be expressed as:

·         1 Architecture Description <can describe> N Similar Systems

·         1 System <can exhibit the property types in> 1 Architecture Description.

 

Or expressed in our triangle thus.

ISO 402010 standard triangle revised

1 Architecture Description

<create and use>              <idealises>

Architects     <observe and envisage>  N Similar Systems

 

Is the standard about systems or architecture at all?

The ISO standard’s web site <http://www.iso-architecture.org/ieee-1471/cm/> says.

"The Standard takes no position on the question, What is a system?"
"The premise of the Standard is, For a system of interest to you, the Standard provides guidance for documenting an architecture for that system." ISO 42010

 

In other words, the standard has nothing particular to do with a "system" as defined by anybody - even by other ISO standards. 

And it has nothing particular to do with "architecture".

You can change all to "design" (I have done that) without changing its intent or potential application.

 

In effect, the standard is about documenting different views of any thing that different stakeholders have concerns about.

And ensuring the documented description meets certain quality criteria.

Footnote 2: The questionable triangle of Popper’s three worlds (a tentative analysis)

Karl Popper split the universe into three worlds.

I am struggling to make sense of his three-cornered model of description and reality.

Here is a simplified version of it

Popper’s three worlds

3: Objective knowledge

<produces>         <describes/predicts>

2: Subjective knowledge <monitors and controls> 1: Reality

 

The division is not self evident, but I believe it to be as follows.

 

World 1: some parts of reality: physical objects and events

I believe this includes the solar system, hurricanes, brains and all kinds of organism.

But does not include human dreams, memories, mental models or documented models

Or even the physical products of mankind?

 

World 2: subjective knowledge: internal private mental models

I believe this includes internal descriptions, memories and mental models of worlds 1 and 3.

But does not include their translation into external messages or models.

It includes thoughts of imaginary things, not found in the universe – like unicorns.

 

World 3: objective knowledge: external public products

I believe this includes all products created by humans.

Descriptive products such as scientific theories, stories and myths.

Social institutions such as a choir, a church, IBM, Google and the United States.

And even physical products such as works of art, bridges and machines?

 

A tentative analysis of Popper’s three worlds is included in another paper.

I might not have got Popper's three worlds' view right.

But whatever that right is - this paper presents a rational argument that his three-way division is human-centric and/or a little odd - and better rearranged.

This is an attempt to place Popper’s three worlds in our triangle.

 

The philosophy here modifies Popper’s three worlds view thus.

·         World 1 The universe: all physical matter and energy (including worlds 2 and 3 when they are described).

·         World 2 Descriptions: abstract models of the universe as it is perceived and described in terms of discrete things.

·         World 3 Objective knowledge: descriptions that, when tested, match the universe well enough.

·         Describers: biological entities and AI machines capable of forming and using descriptions.

 

This is an attempt to place the revised worlds in our triangle.

Popper’s three worlds

Worlds 2 and 3

<create and use>                   <idealise>

Humans       <observe and envisage>      World 1

 

I believe this refinement of Popper’s triangular view stands up to scrutiny.

And, being more distant from Cartesian dualism than Popper, it is a more productive way of looking at description and reality.

 

 

All free-to-read materials on the http://avancier,web site are paid for out of income from Avancier’s training courses and methods licences.

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