[Work in Progress – Draft will be edited without notification.]
[Feedback appreciated on the “See / Refs” – where more are needed? Meantime all those indicated will be elaborated and worked into the text. And obviously on the intelligibility of the text so far. Drafting arose out of the “Three Essays” post, particularly “Algorithms for Humans” and the comment thread on “Balancing the Future”. It represents one (of many) draft of my “thesis”.]
[At this stage I don’t intend any revision-control within the page / document, but if anyone is up for it we could turn it into a collaborative document? As in fact is being done with “coda” in the Active Inference Lab.]
[See also comments on the original post that inspired this page.]
Information & Systems Thinking
I’m going to talk and think in terms of systems, information and processes and some of that will use language often associated with machines and computers. The fact that until recently computers were actual humans doesn’t seem to lighten the mechanistic baggage that we tend to see computers and computing systems as electro-mechanical machines – and indeed much of their research and development uses (eg Turing) machine language.
Using the terms above, almost anyone with a humanist / humanities bent will baulk if I start using them to talk about how humans actually work biologically, psychologically or socially. Any use of them being at best metaphorical in underlying biological processes. Crude talk of living things as machines and brains or minds as computers generates a pretty negative knee jerk response. Part of mechanistic scientism over-reaching into the humanities, and missing – devaluing – its true richness.
Believe me, I share that reaction. My main quest here has for 22 years been precisely to point out the shortcomings of logical, objective, scientific thinking when it comes to human enterprises of all kinds. So, I’m not going to use the language of physical machines or computers if I don’t mean to. What I am going to use is systems thinking, and I will be talking about how it, they, we process information.
So, firstly I need to reassure you that when I say the words – systems, information and processing – and the shorthand of computation – (and god forbid, algorithms) – for what systems that process information do, I’m not, and you mustn’t therefore, be thinking of these as machines or computers, mechanical or electronic.
System is simply the word for any collection of stuff, thought of in terms of the arrangement / architecture of its component systems. Systems made of systems in any arrangement, any level of nesting within each other. Notice I could have said “parts” for the sub-systems, provided you didn’t assume I meant material-physical things. We really are talking any stuff as the subject of any thought or dialogue.
- Two (or more) dots on this screen. Two (or more) points separated in any frame of reference in fact.
- A pair of quarks, an electron-proton pair, a DNA molecule …
- Two hemispheres in a brain, a fore-brain and a mid-brain, a whole brain, a whole brains-trust of many brains.
- Me writing and you reading this.
- Two sides of a debate, the set of concepts in a dialogue.
- The idea of logical positivism, the entire works of Dostoevsky.
- A trolley problem, a termite mound, the Sagrada Familia.
- The thought experiments of philosophers everywhere.
The reason to think of them as systems – to use systems thinking – is to focus on their architecture: how they’re arranged, how they relate and communicate with each other and with their parts as more sub-systems.
A system is literally anything conceivable, but thought of in terms of its architecture. Systems Thinking. Literally.
[Communication is a fundamental information exchange process. In fact information, communication and computation are as fundamental as – if not more than – physics itself. See xxx / later.]
[Cybernetics – the root topic in this Psybertron blog – has itself suffered the same skewing of perceptions as the systems and computation talk above. Very much from Plato to the 1946 Macy conference it was about kybernetes, the self-governance of human affairs, but it was taken over in implementation by computer geeks for the past 80 years, to the point that the original human cybernetics became known ironically after 1963 as “the second cybernetics”. In further irony, the term governor had already been borrowed metaphorically from human affairs to name the mechanical devices regulating the speed of early machines like 19th C steam-engines, long before we reinvented Roman thermostats. See yyy / later.]
[For more on Systems Thinking itself and positively practical application to everyday use by real humans in business and in government, see Anatoly Levenchuk (Ref). Unfortunately, in this piece I’m going to climb & dive into layers further removed from the everyday – using the abstractions of systems thinking to address some fundamental issues with science and humanity.]
Systems & Sub-Systems
When we talk of wholes and parts in everyday life – or of systems and sub-systems in systems thinking – there are common throwaway opinions:
“Some things – significant real-world things – are more than the sum of their parts and anyway, context is everything.”
To which orthodox scientific “rational” responses would be:
“Well no, once you’ve taken proper causal account of “all” component contributions, there can be no more. And, no, context is just more stuff to be taken into account.”
As ever the initial problem is at least partly linguistic – “taking into account” is more than a “sum”. We have apples and pears involved. 2 apples “plus” 2 pears/nuts/shells/seals gives us neither 4 apples nor 4 pears/nuts/shells/seals – but 4 fruit at best, 2 pairs of different kinds. And the process itself introduces or creates kinds we might not have started with. Pairs are not the only fruit. The best solution depends on what was the question.
[Classification / Naming / Definitions / Taxonomies / Good Fences / Tabletop / Rules & Exceptions / Identity Politics / Levels of Abstraction / “Small & Large” Facts]
The combination of component “parts” is more like systems architecture (above) and more like integration than addition. What’s more, an integration / assembly over multiple axes and levels of “kinds” of stuff and “types” of things, not to mention integration of processes over multiple timescales and histories in most real world contexts.
At human, individual and social scales, and at both micro and macro physical levels, and indeed at any number of biological, geological and cosmological scales, the outcomes of so many real world processes are statistical, dependant on probability distributions. Different kinds of distribution over different axes and timescales on top of the distributions on the kinds of stuff and things. Accounting as simple addition doesn’t get us very far. Being charitable, every discipline knows which particular set of mathematical / computational operations helps with their day job.
[History matters – Ergodicity / Uncertainty / Bayes / Taleb]
Complicated, or complex, or both. I don’t intend to make a big thing of the difference here, but different they are. Which is another complication.
Leaving aside whether in any real world case it could ever be practical, tractable in a computational sense, to take literally everything into account in a useful finite time, is moot.
In reality we draw control volumes – boundaries – around our problems. Boundaries within which and across which we are either in control or have workable levels of predictability and uncertainty. In a scientific, logical or mathematical context those control boundaries may simply be credible working assumptions, which can themselves be varied and validated in due course. That validation may often be the point of the scientific exercise. In most human endeavours it’s about holding the alligators at bay whilst we drain the swamp, or whatever it was we were trying to achieve.
This process / strategy looks like good practice learned from shared human experience, but it turns out to be much more fundamentally natural than that.
Faced with complexity, systems thinking helps. Remember, that means thinking in terms of stuff being systems as described above, not simply systematic / systemic / disciplined / formalised ways of thinking / planning / acting. Systems Thinking says treat the thing in front of you as a system whether or not it’s called a system in an everyday usage. A computer application or a collection of work processes or protocols or any electro-mechanical network assembly – we readily see as a system. Systems Thinking says treat every collection of stuff whose existence or behaviour depends on interactions with other stuff as a system. Systems Thinking addresses complexity.
[Park the argument for now between the strict orthodox scientist that says everything is ultimately physical, including our thoughts – the kind of thing material / energy / physicalists say – and the humanist who would reject that. Remember we already hinted earlier that information and computation are more fundamental than either physics or thought. More later.]
Thinking of the complex stuff / complicated thing in front of us as a system we will identify component sub-systems – the parts – and we will also identify the context – the super-system outside our control boundary – the world. We can further subdivide our thinking as necessary. Our system its internal sub-systems and that external world super-system can be further divided into as many more sub-sub-systems as suits our task.
Thing, with boundary, with internal view and with external view.
Now the clever – and surprisingly fundamental bit – that describes and makes an ontological commitment to the reality of both physics and thought evolving from this basic view of systems and information processes within and between them.
At this point you need to believe the whole of the world and everything in it, can at least be thought in these “systems thinking” terms above, without any losses? No point elaborating if there’s already any objection up to this point.
So given seemingly universal applicability of Systems Thinking, without any other presumptions about what exists in the world, the entire world, micro and macro, past and future, how it all works and fits together – no other presumptions – the world is my oyster. No? Just a way of thinking about stuff, including thinking itself.
I’ve already noted asides in the text so far, and below we have dozens more angles and corollaries to come. And this is the history of the entire world, so far and forever – a very long story – so clearly I’m going to need to be choosy about which bits I try to articulate and why, what for? I need to apply systems thinking to my systems thinking – which details matter where?
Remember I may not mention every detail, or give examples of everything, that has happened or may happen in the entire (previous and future) history of the world, but this story really is about all of that. Seen through systems thinking.
In many ways, apart from my aim to synthesise the story around systems thinking, there is nothing here I’m creating that’s actually new. All my sources and ideas are already published – some of them long ago, some of the syntheses more recently. ‘Twas ever thus. Nothing new under the sun.
And obviously lots has been written scientifically and philosophically, and mythologically or creatively metaphorically in the humanities – in literature generally. And a lot of that reflects prehistorical oral traditions before there was even literature. In some ways, from the humanities side, all I’m doing is clearing a path through ancient wisdom and intuitions that have become overgrown with scientific progress. From the ever-contingent science side, I’m mostly pointing out missed opportunities in choices made to best fit the data at the time, according to the methods of science. Stuff taken to be right – best current model – because it worked for some recognisable STEM purposes. And which touch-points / differences / controversies are invisible in the undergrowth may be obscure archaeology unless they’re part of your specialism.
This story is necessarily polymathic and necessarily of unequal depth and rigour in any given specialism – an occupational hazard of multi-discipline working. Indeed as we shall see – having the courage (*) to know which details do and don’t matter in any given context are a fundamental part of systems thinking. So this is mostly a story of rhetoric and persuasion, bringing together science and humanities where each may have missed a trick in reconciling with the other. Creativity and metaphor are not confined to either. No truth without reconciliation.
[Sun Tzu “Art of War” – Even if he could know them, the general does not need the position of every blade of grass on the battlefield, not even every individual warrior. (*) And “courage” – see risk aversion in “science” generally.]
And, I need to start somewhere. One thing I skipped so far, was that in claiming universality for “systems thinking” – I also mentioned information, processes and computation, and I subsumed them into that systems thinking umbrella without saying much more about them.
[Placeholder – I could come back and branch differently from this point? Why don’t you start with … ]
I can’t quite believe myself, but in order to start talking about information and processes in a systems thinking way, I’m going to start with one of the most abstruse technical areas of philosophy, metaphysics itself as a basis for all existence (ontology) and meaning (epistemology) in the world.
[No prizes for spotting I will be relying on Whitehead when it comes to process metaphysics.]
Information however …
[Counterfactuals and conceivability. Deutsch & Marletto.]
Systems thinking about thinking? Let’s do a thought experiment. Not one of those that asks you to suspend disbelief about whether it could possibly ever be arranged to happen – and please god, not a trolley or a zombie – just literally a thought experiment.
Imagine nothing and imagine the least thing that could conceivably exist in or in addition to that nothing, to make it something other than nothing.
Firstly, whatever you managed to conceive of as nothing (you’ll have done better than Larry Krauss anyway) – whatever you conceived of as that smallest addition would, in some pretty absolute sense, be indivisible – atomic in Democritus original conception, a thing with no parts, a Euclidian point. If it were in any way divisible, you could have conceived of one of those parts (sub-systems) existing, no?
[Avoid the confusion of the pretty complex systems that science first labelled “atoms” in order to make distinctions between chemical elements. Elemental elements in a chemical sense only. These are a long way from the Democritan / Euclidian conception. That’s just naming and language – already noted above.]
It’s not possible to conceive of anything less than the least conceivable thing. No cheating now. (And this is true whatever you had to settle on for that absolute void of no-thingness before you added that thing to it – not surprising that many may have to posit some ill-defined cosmic unity thing for that nothingness, just to be able to have the thought, but no matter.)
We have in mind that no-thing and a thing. If you can think of any things more primitive than that thing and that thing<>no-thing relation, I’d like to hear it. (We’ve not said anything about things like space or time, where or when yet, nor anything else we might call physical or material? This least thing is conceivably the dimensionless thing without any properties a Democritan / Euclidian point.)
If we chose binary notation, this no-thing and point-thing looks very like most primitive 0 & 1 but we don’t have to jump to “bits” quite yet, and anyway as I said above we don’t need this line of thought to be dependent on what we thought of as that 0 (no-thing) in the first place. So patience.
Now imagine a second thing. Hard to imagine it being much different to the first thing – after all we have placed some severe constraints on our thinking so far – no time or space, no materials or physics. Whatever kinds of things these things are they are simply both the same kind of indivisible atomic point-things.
So what makes them different things? What is their difference? What could conceivably be the difference between two point-things? Their difference is their separation, what separates them?
[We might think all that separates them so far is that they arose from two thoughts – but have no fear, we’re not jumping to pan-psychism – we can keep distinct the thought we’re having (here and now in the real world) and the conceivable thing thought about – the map and the terrain, the finger and the moon.]
[Insert magic – aka – Sense and Experience, Free Energy Principle / Markov Blankets / Active Inference / McGilchrist, Solms & Friston / Emperors’s New Clothes / Finger / Moon – Map / Terrain – Model / Reality. Information & Quality / The Subjective Perspective.]
[All models are wrong, or are they? Active Inference makes the ontological commitment as well as providing the model.]
[Corollaries / Negative-Corollaries / un-Adages: It’s the thought that counts. Things that can be counted are not the things that count.
Nothing new under the sun / ‘Twas ever thus / This s all already “out there”. Free Won’t, Vive la Differance, Devil in the Details / Examples / Pennies & Pounds / Butterflies & Cow-farts / One subject & one audience at a time?]
Meantime, any Psybertron content searches on Markov, Solms, Active Inference, and the Information Metaphysics (2019 version) will all be recast in this Work-in-Progress essay. It’s all there 🙂
Pass me the machete?