On the philosophy of systems

Preface

This article is a postcript to five other articles you should read the first. This a collection of notes that didn't make the cut in the earlier articles!

• Systems, rules and self-organization
• On knowledge as description
• On systems as types and instances
• On description as classification
• On knowledge as a biological phenomenon

Philosophers today devote much time to rejecting or endorsing earlier philosophers, and positioning themselves in relation to them. They like to review, interpret and put a new spin each other's ideas, using language that is often impenetrable to outsiders.

To read the articles above you need next to no knowledge of philosophy, linguistics or mathematics. In fact. thinking in terms of traditional "isms" and "ologies" may hinder you from reading them right. To associate our philosophy of systems with one particular philosopher or school would bring a lot of baggage into the picture, endless room for debate, and cloud the story.

Having said, this article was written for readers interested in the relationship of philosopical positions to systems thinking. First, let me mention an approach deprecated here.

Constructivism - dismissed

Western governments and institutions are being captured by ideologies that are, in effect, substituting opinion for science.

Relativism is the doctrine that there is no absolute or verifiable truth, because knowledge emerges separately in each culture or society. Social constructionism is a radical form of relativism that denies the things studied by scientists are “out there”. Rather, it says things are entirely constructed through ways of observing and describing the world that are specific to each social context.

This philosophy leads some to conclude that opinions are as valid as evidence. Some sociologists have latched on to this philosophy as a rationale for not applying the scientific method, and found a way of talking ("my truth") that hides the underpinning philosophy.

Some interpret "my truth" as meaning that, since a receiver's interpretation of a message is true, then if a receiver is offended by the message, it is the fault of the sender. This is extraordinarily unjust. If a message receiver takes offence, the sender is blamed, however innocent they are or kindly they meant the message.

If someone misreads your message "Don't go near the cliff edge", should you be blamed when they fall down the cliff? If somebody on Tik Tok says "I read your post to be racist", does that prove you are a racist?

On description and reality

In a post-Newton, post-Darwin, post-Einstein world, scientists don't claim to understand reality directly or completely. They observe and envisage reality as being composed of interrelated things. They describe and typify those things in models that they judge to be true in so far as they help to predict what happens. They don't claim to understand reality beyond the models they make of it.

All models are partial; many are imperfect. We test and refine models by trial and error. Biological evolution has rewarded animals better able to model reality biochemically. And science has progressed by envisaging and constructing models, using them to make predictions, then observing and testing how well phenomena match those predictions.

Axiomatically: there is a reality out there (the alternative is useless sophistry), the descriptions we make of reality are also real, our descriptions may be verified empirically. All we can know of reality is what we can describe of it in models (in memories and in messages) that are true to the extent that they help us detect and predict what happens.

A reader has suggested this position close to critical realism. But classifying that wasy may obscure more than it clarifies; not least because there are many sub-schools, some overlapping.

Earlier articles advanced several more particular propositions about description and reality.

After A J Ayer: "To describe a thing is to classify it".

• To describe a thing using symbols is to conceptualize or typify it.
• Conversely, to conceptualize or typify a thing is to describe it.
• Every non-paradoxical description of one thing serves as a type to which other things might be observed or envisaged to conform.

After Humberto Maturana: "Knowledge is a biological phenomenon".

• The concepts or types we use in description were invented by us.
• A type exists in a location, whether encoded in a nervous system, in speech, in writing, or in another symbolic form.
• After a type has been invented, corresponding types are made by copying or translation from one location or symbolic form to another.
• If all members of the set of corresponding types are erased, a new member must be created before it can be used.

In short, we create and use types to describe things. A philosophy of systems must address how we create and use abstract system models to represent real systems.

Empiricism

This work takes a pragmatic empirical approach.

Science (and all reasoned argument of the kind that should inform government decisions) presumes not only that things do exist out there in the world, but we can observe features of those things, and describe them.

Social constructivists dismiss empiricism based on a false argument. They say empiricists cannot prove assertions. But empiricists don't presume that proof is possible or necessary.

I take a psycho-biological view of these things. Every animal with a brain lives its life as a pragmatic empiricist. How the brain works is irrelevant. All that matters is its evident ability to predict the future well enough to make decisions about actions, such as to stop before walking over a cliff edge.

On other "isms" and "ologies"

Empiricism is not enough, since not all knowledge is acquired by empirical testing of evidence against theories or models.

Logical positivism is not enough, since not all problems can be solved by logical analysis. (Having said that, testing and logical analysis are the primary ways to verify a model.)

Anthropomorphism is misleading, partly because it leads philosophers down a linguistics dead end.

Constructivist ontology, the idea that reality is socially constructed through language, is too anthropocentric. Evidently, non-social animals can remember things they have observed in reality, and interact with them. True, the human ability to make and share models was vastly increased by the evolution of verbal language. But to survive, every sentient being models things in its environment.

Moreover, the following psycho-biological view of thinking spans most if not all of the headings below

• Ontology: about reality as it is represented in our models
• Phenomenology: about models recorded by senses
• Epistemology: about how we create, test and use models
• Axiology: about the values to us of modelling and acting
• Praxeology: about connecting models we have or create and actions we take to motivations rather than stimuli.

Putting aside the "isms" and "ologies" above, the approach taken in these articles is to take a psycho-biological and Darwinian view of how thinking evolved, in accord with

• Maturana's "knowledge is a biological phenomena",
• Hilbert's "we think in models”,
• Wittgenstein's idea of "family resemblances" and
• Ashby's cybernetics.

Reductionism

Some speak of "system engineering" in a dismissive way by calling it systematic rather than systemic, reductionistic rather than holistic, analytic rather than synthetic.

Deprecating reductionism is naive. To begin with, the term is ambiguous. For some, it means reducing every explanation of things to the level of physics. Clearly, scientists do not do this. In higher level sciences, higher level phenomena are explained without reference to physics, and other lower level sciences.

• Sociology, and Social psychology
• Psychology, and Biopsychology?(neuroscience)
• Biology, and Biochemistry (organic chemistry)
• Chemistry (inorganic), and Chemical physics
• Physics

Others reject reductionism in favor of holism; yet they are inseparable sides of one coin. You cannot join up parts without first identifying the parts of interest, which is in effect to bound the whole composite of interest and decompose it as you see fit.

All scientific thinking involves both zooming in and zooming out as far as befits understanding or explaining the phenomena of interest. Zoom out far enough, to consider the cosmology of the universe, and you return to the domain of physics, in which the trials and tribulations of human kind are irrelevant.

Moreover, the mereologist's view that things are well explained in terms of wholes and parts is also naive, since one person's whole is another person's part, and systems thinking is more about the effects of interactions between things than the things themselves.

Further reading

This is one of several related articles that how show the ideas of famous systems thinkers (Ashby, Forrester, Ackoff, Meadows, Beer, Senge and others) can be embraced in a coherent system theory, provided you recognise that not everything is well-called a system, and not every approach to thinking about a social entity is well-called systems thinking.

Aside: Boulding on the limits of system theory

Boulding 1956 suggested a possible arrangement of system levels, rising (he said) in complexity.

1. Frameworks: the geography and anatomy of the universe. Structures and descriptive classification schemes.
2. Clockworks: the solar system, lever, pulley, steam engines and dynamos (examples include open activity systems).
3. Thermostats: controllers of systems that maintain their equilibrium.
4. Cells: open systems or self-maintaining structures, simple life forms.
5. Plants: differentiated and mutually dependent parts.
6. Animals: sense information, behave purposefully, self-aware.
7. Human beings: self consciousness as well as self-aware. (Research has since shown some other mammals are self conscious, what sets humans apart is symbolic language.)
8. Social organizations
9. Trascendental systems (a somewhat mysterious category).

Boulding suggested that higher-level systems are larger and/or more complex than lower-level systems. So social entities are larger and more complex than plants. However, the interactions between the many cells of a plant (level 5) may be more complex than the interactions between a few animals in a simple social organization (level 8).

In 1968, Bertalanffy wrote Boulding’s hierarchy was “impressionistic and intuitive with no claim for logical rigor”. However, some ideas below foreshadow the ones in this book.

• “The above scheme might serve as a mild word of warning even to Management Science. This new [system theory] represents an important breakaway from overly simple mechanical models in the theory of organization and control… Nevertheless… these advances do not carry us much beyond the 3rd and 4th levels, and in dealing with human personalities and organizations we are dealing with systems in the empirical world far beyond our ability to formulate.”

In other words, a real-world entity at levels 5 to 9 cannot be modeled as a system. We can only model particular ways of behaving as systems.

• "The unit of [social] systems is not perhaps the person but the "role" - that part of the person which is concerned with the organization or situation in question. Social organizations might be defined as a set of roles tied together with channels of communication."

In other words, actors' behavior in a real system in action is defined as roles in an abstract system.