So equilibrium is the big picture in a very real sense.

[apeiron]

So I see your emphasis on dichotomy as appreciating the deep “two-sidedness” of the world. But I don’t see dichotomy per se as a principal, a starting-point... rather as the fundamental expression of any particular stage in the evolution of relationships.
.

Dichotomisation is certainly not the starting point, just the transition, and hierarchies are the destination.

So it seems sensible that all causal stories would follow a three step pattern: a before, a during, an after. There was (1) a state of some kind, there was (2) a change for some reason, then there was (3) a new state of some kind.

So my argument has been that (1) is a state of vagueness, (2) is a process of dichotomisation, and (3) is a hierarchical outcome when what has been divided has also mixed. Two things in interaction make three things altogether.

This is the basic insight of Anaximander and Peirce. Not so much Hegel of course.

an ability to think from a standpoint utterly removed from daily life and personal relationships. (A viewpoint vectorcube has been expressing forcefully in several recent threads.)

Perhaps that can be said of some of the people Vectorcube quotes.

But the basic question for both of us seems to be – where does possibility come from, in our world? In that possibility isn’t "just given" any more than actuality is – something is going on here that makes new things possible, all the time. And we think we can learn to make it understandable somehow.

And vagueness is not this kind of ocean of pure possibility?

I think it is safe to apply a precursor argument to establish what must have come "before". We can reason that whatever we find now must have once been somehow in that original state.

So what do we find now? Well we find something for a start, rather than nothing, or everything.

And we find that that something is also highly dichotomised, highly asymmetric. As you say, philosophy is all based on complementary dualisms, opposites which arise out of the negation of each other. There are many dozens - one~many, local~global, substance~form, stasis~flux, discrete~continuous, atom~void...on and on...

So to recover the origins of our world, we should attempt to reverse what we find, go backwards in a way that steadily erases it. If our world evolved, then to find its initial conditions we must devolve it.

So from asymmetry we would go to symmetry. What was broken gets reunited. This is exactly the thought path taken by Anaximander.

Thus possibility is the unbrokenness of a symmetry. That is the basis of our model. The idea of a state of infinite symmetry.

Of course, people will want to say if the start was a symmetry, well who cooked that up? That too would seem to require a prime mover.

But vagueness, as far as it is possible to imagine such a thing, does seem the kind of thing that can "just be" because it isn't really there. It is the everything and nothing.

Imagine you are rowing a boat on a mist shrouded lake. You paddle hard. But you might have traveled a long way, or no where at all. As far as you can visibly tell. If your motion could have been anything, then what has actually happened seems vague. From an internalist perspective - which is part of what we assume for this way of looking at reality anyway.

So when action looks the same as inaction, you do have everything and nothing. But when scale arises, when a dichotomy of event and context emerges, the fog lifts and a boat's motion can be judged against a reference frame. It become crisply a something.

In a reference frame, we can count the multiple possibilities - the total ensemble of microstates. In the boat example, with no fog, we can see that the boat underwent motion x, and also count every other not-x possibility - all the other possible actions that did not take place because we went in that direction, at that rate, and in no other.

But in vagueness, the other microstates cannot be counted. They are symmetric. They are indistinguishable. In the fog, all actions look the same so we can no longer stack the not-x's up against the putative x. There is no possibilities (in the usual sense of countable microstates, an ensemble) in vagueness, only pure unbroken potential.

It is not easy to convey what vagueness means it seems. But I find when you get it, it clicks right into place. It is just a very different concept to any we would normally encounter in modern anglo-saxon education.

Louis Kauffman wrote this very helpful paper on x/not-x.

http://www2.math.uic.edu/~kauffman/Peirce.pdf

 

[ConradDJ]

Dichotomisation is certainly not the starting point, just the transition, and hierarchies are the destination.

So it seems sensible that all causal stories would follow a three step pattern: a before, a during, an after. There was (1) a state of some kind, there was (2) a change for some reason, then there was (3) a new state of some kind.

I think it is safe to apply a precursor argument to establish what must have come before . We can reason that whatever we find now must have once been somehow in that original state.

So what do we find now? Well we find something for a start, rather than nothing, or everything. And we find that that something is also highly dichotomised, highly asymmetric.

So to recover the origins of our world, we should attempt to reverse what we find, go backwards in a way that steadily erases it. If our world evolved, then to find its initial conditions we must devolve it.

Whether or not there is a “safe” or logical argument here, I do think we are both trying to take what seems most fundamental to us in the world we experience now, and project back to where that came from. And of course what ultimately counts is – to what extent does this projection help us appreciate and understand the empirical world?

So let me back up and ask about something that came up earlier –

I wrote that –

The fundamental constraints in biology are given by the fact that organisms have to reproduce themselves. This is the basic functionality that opens up the possibility of life. Once you have self-replicating systems, then you get internal constraints requiring that these systems maintain themselves long enough to reproduce, and also external constraints from the environment to which these systems have to adapt.

You responded –

That would be the definition of complex systems (rather than simple ones) - the ability to create their own boundary conditions. To be systems unto themselves.

So the simple system in this case would be the second law of thermodynamics. Life has no control over that, and is in fact completely entrained by it. However life exists by its ability to constrain metabolism - a cell is a membrane that constrains chemistry, and further constrains those equilibrium processes through enzymes and physical channels.

That makes sense to me – but it makes me think that hierarchy theory misses the key point, in an attempt to establish principles that apply to all systems. Organisms are able to maintain boundaries and constrain physical processes not because they’re complex, but because they reproduce themselves and so can evolve. The radical difference here is not one of scale or complexity, but functionality. In a complex environment such as must have existed on this planet long ago, there may have been many kinds of more and less complex homeostatic systems (like those Prigogine studied), channeling energy gradients and perhaps creating boundaries. The only one that was relevant to the emergence of life was the one that somehow by accident was split into multiple copies, some of which were capable of being split into multiple copies, again... At the physical level there was probably nothing special about these systems – what was special, to begin with, was just that they were lucky enough to be replicated, and managed to sustain their homeostatic processes in enough copies, that selection could make them better and better replicators.

So my question is – Salthe’s work surely recognizes the role of evolution – reproduction and selection – and I understand that he and Oyama and others want to emphasize that there are other kinds of processes also at work in biology. But it almost sounds as though the concept of generative hierarchy overlooks the radical discontinuity that occurred with the origins of life – when a different kind of relationship-context was accidentally established, without necessarily (at first) involving any major difference of scale or complexity.

To me this kind of “creation of possibility” is basic. I’m trying to imagine the origins of human communication in a similar way – and the origins of the physical world – through the accidental beginning of a new kind of relationship-context, involving a new functionality that sustains and is sustained by it.

 

[apeiron]

So my question is – Salthe’s work surely recognizes the role of evolution – reproduction and selection – and I understand that he and Oyama and others want to emphasize that there are other kinds of processes also at work in biology. But it almost sounds as though the concept of generative hierarchy overlooks the radical discontinuity that occurred with the origins of life – when a different kind of relationship-context was accidentally established, without necessarily (at first) involving any major difference of scale or complexity.

First, I do argue that the notion of hierarchy can be generalised to cover all natural outcomes. Boiled down to its essence, a hierarchy needs only two things in interaction. And these things must have opposing scale in some definite sense as the interactions are dichotomised. There must be a difference between the top-down actions and the bottom-up actions. Even though together they must have a creative mutuality, a synergy, that makes the whole greater than its parts.

Further, a hierarchy is stable because it is an equilibrium state. The two sources of action must develop to have a steady-state balance so the hierarchical state can exist (or rather persist).

So cannonically, the fundamental unit or cell or a hierarchical relationship involves an uppper, lower and middle level. But then, these atomistic triads can become stacked up to form the kind of multilevel structures we think of as "true hierarchies".

Then on your second question, what theoretical biologists have been trying to do is make a clear distinction between development and evolution. The two concepts have been muddled together too long.

Development is the open, self-organising, side of things. Put together a bag of organic molecules and an equilibrium state will self-organise. An outcome freely develops. Evolution is then the imposition of constraints on such freedoms. So if a system can throw different enzymes into the mix at chosen moments, it can regulate that development. It can switch it down paths. It could evolve the state of the mixture.

So development is the naked action that generates creative potential, evolution is about the application of constraints that forces development down forking paths - crisp choices that mean the system went this way, and not that way.

Salthe and others then distinguish between bios and abios on these criteria. All material systems are based on development - ruled by the second law, dissipative structure. So a dust devil or whorl in a stream both develop in a self-organising fashion into organised energy-transacting structures. They seem a bit alive in their order and the way they dance about their landscapes creatively. But they are not alive (or mindful) because they have no way of constraining the "choices" or paths they find themselves taking. They are not evolving structures.

However, if we take a longer term view of a structure like a river network, we do see the beginnings of evolution because a river is a fast changing thing (the current) that carves out a persisting thing (a riverbed). So there is a kind of system memory. The river ends up taking choices about cutting across this part of a landscape and not that part.

But still this is not really evolution in a strong sense because a river's branching is fractal. It is ruled by powerlaw statistics - the hallmark of simply constrained dynamics, dynamics with the same constraint acting over all scales. It is a generalised action rather than a more powerfully focused one. Really strong systems constraint has not the open-ended nature of powerlaws but the closed nature of gaussian statistics.

So we move to life and mind, which are about harnessing developmental potential with strong evolutionary constraints.

Take genetic replication. It is about controlling development so as to produce a gaussian variety - instead of a powerlaw range of possibility, possibility is restricted to a mean and a normal curve. So the height of a population has a controlled statistics. There is a longterm memory for how high it is sensible for a population to be, and then enough variety around that mean for a very focused selective competition - a fine-tuning of a parameter within bounds.

This should give you the flavour of the difference. The discontinuity that makes the boundary between bios and abios was the jump to gaussian control over boundary conditions. The restriction of possibility from something too open to be meaningful (powerlaw variation like fractal branching) to something more focused like gaussian variation that allows selection within a single scale, selection within a longterm context.

But the dissipative structure view also preserves what is continuous here - the fact that both dust devils and lifeforms are founded on developmental potentials.

Development is what happens when things self-organise within a set of constraints. Evolution is what happens when constraint becomes hierarchical so that there is a nesting of levels of constraint. Simple constraints produce powerlaw statistics and more complex constraints produce gaussian statistics. We go from systems exploring spaces of broad possibility to systems exploring spaces of much more narrowly defined possibility. Or from vaguer to crisper paths.

The machinery of life and mind - the means by which constraints are exerted - does come in many grades and forms. We have membranes, pores, channels, vessels and other stuff closer to the riverbed or powerlaw end of the spectrum. And then neurons, enzymes, genes, words, which are towards the gaussian end of the spectrum.

Genes and words - both serial codes that can preserve system information, memories for constraints, over many generations of individual examples of a system - are of course the most power level of constraint. They carry over the paths taken, the paths not taken, a whole weight of evolutionary history, from instance to instance of some general dissipative structure. They are constraints freely transported across time and space.

 

[ConradDJ]

I have to say, based on the above – it seems that hierarchy theory is developing general descriptive categories that may well become widely useful, but I'm not sure it's working out real explanatory principles. It seems to gloss over basic differences in the structure of what’s going on and what’s possible at different levels.

Concepts like equilibrium and dissipative structure are indeed relevant to most categories of physical order. But they are fundamental only at a certain level. Specifically – the laws of physics make it natural for many kinds of systems to arise that recreate their own structure again and again over time. A planetary orbit is the simplest example, though here the concept of equilibrium doesn’t yet apply – if an orbit is perturbed there is no tendency for it to return to a previous state. A star is an example of a system that does maintain an equilibrium-state through opposing forces of gravitation and nuclear fusion. In this kind of dissipative structure, many perturbations can easily be absorbed – but if this kind of system is pushed beyond a certain limit, it fails and will never recover equilibrium.

Once a self-replicating system is established, we have a completely new kind of possibility-generating dynamic. Each organism by itself is a complex of self-equilibrating chemical systems, that can maintain itself over time or else fail and die. But the failure of individual organisms to survive and reproduce is the basis of “selection”, which plays a necessary constructive role in the evolutionary process. This is completely new.

As you say, biological evolution evolves constraints that channel homeostatic processes. But it’s not as though some system invented the possibility of imposing constraints on itself! The constraints come from the new "requirement" of self-reproduction.

And equilibration is now playing a secondary role. The primary biological dynamic is a one-way process that improves the homeostatic efficiency of organisms as a means of improving their reproductive success, and adapts the species to a given physical environment for the same reason.

Once life gets going, this species-level evolutionary process itself begins to evolve. Since the physical environment changes over time, there is a selective advantage for species that evolve more effectively than others. As species become more and more interdependent, this co-evolutionary dynamic becomes more important, and the evolutionary process itself starts to evolve more and more rapidly – in the emergence of sexually reproducing species, the emergence of flowering plants, etc.

So there are stages in which one level of order emerges “naturally” from a previous level – as homeostatic systems arise naturally in any complex physical environment, or as ecological structures of inter-species dependency emerge naturally once life begins to proliferate.

But there are other stages – notably the origin of life itself – that are unique, one-time events that may be extremely improbable... but eventually create an entirely new playing-field — a new game with entirely new rules. And I think that human evolution is based on a similarly unique and improbable emergence of a new kind of relationship – communication between individuals – that opens an entirely new evolutionary dynamic.

Now each stage takes up and incorporates the fundamental structures of previous stages. At every stage of evolution there are new kinds of self-equilibrating processes, e.g. in economics... new kinds of “gradients” and “dissipative structures”. That means that hierarchy theory can indeed find instances of its basic principles almost everywhere. But the fact that homeostatic equilibrium is a more generally applicable concept than reproduction doesn’t make it more fundamental as an explanatory principle, when it comes to biology.

Philosophically, my personal proclivity is to emphasize the unique, the breakthroughs to a new level of possibility. I think there is an overarching “logic” to the way new possibility-structures have emerged in the world, but I don’t think concepts like equilibrium are adequate, even as metaphors. The “logic” needs to include both the “natural” emergence of new kinds of structure over time, within given constraints, and unique events (like the origin of the universe itself?) that in certain special contexts can create a whole new ball game, making accidents meaningful in new kinds of relationships.

 

[apeiron]

Concepts like equilibrium and dissipative structure are indeed relevant to most categories of physical order. But they are fundamental only at a certain level. Specifically – the laws of physics make it natural for many kinds of systems to arise that recreate their own structure again and again over time. A planetary orbit is the simplest example, though here the concept of equilibrium doesn’t yet apply – if an orbit is perturbed there is no tendency for it to return to a previous state. A star is an example of a system that does maintain an equilibrium-state through opposing forces of gravitation and nuclear fusion. In this kind of dissipative structure, many perturbations can easily be absorbed – but if this kind of system is pushed beyond a certain limit, it fails and will never recover equilibrium.
.

Quite true. But I am trying to go beyond these everyday imperfect examples to extract the general truths. So we actually want to forget all the local particulars which hide the deeper principles.

This has been one of the problems in the development of hierarchy theory I have argued. The urge to include unnecessary details.

But there is something that is deeply counter-intuitive in my approach, agreed. The normal idea of dissipative structure is as a closed and stable system transacting energy. So you have a cell that takes in stuff at one end and pushes it out the other. The cell itself stays the same size. Likewise a star. A balance of gravity and fusion, it stays the same size by varying its burn rate.

But I instead am thinking about systems like scalefree networks and universes which instead are dissipatively expanding, not statically transacting. A big difference, though thermodynamically the same second law applies.

So a universe, in effect, is making its own heat sink. The expansion of spacetime is paid for by the cooling of spacetime. This is more like a star than a cell because it is more freely self-organising. A cell has to construct its walls, a star just collapses to a balance point. And a universe just freely expands and cools.

It is all about removing the constraints until you have a hierarchy as naked as possible.

Once a self-replicating system is established, we have a completely new kind of possibility-generating dynamic. Each organism by itself is a complex of self-equilibrating chemical systems, that can maintain itself over time or else fail and die. But the failure of individual organisms to survive and reproduce is the basis of “selection”, which plays a necessary constructive role in the evolutionary process. This is completely new.
.

True, true, true. But again, this is the very familiar story of complexity, and I am talking about a hierarchy theory approach to simplicity as well - that is where the novelty lies.

I was for many years very focused on issues of complexity. And it was a big surprise to find that the same systems logic - stripped to its fundamentals - would seem to give a view of the origins of simplicity as well. I shifted track to this new area of study. But I don't deny my original one.

As you say, biological evolution evolves constraints that channel homeostatic processes. But it’s not as though some system invented the possibility of imposing constraints on itself! The constraints come from the new "requirement" of self-reproduction.
.

The evolvability of species has in fact itself evolved. Bacteria for example will exchange RNA at any time with any species (OK, within limits, but in handwaving way...). So this is quite unconstrained. Then higher levels of biotic complexity gradually tightened the moment of genetic experimentation and the choice of partner. We have fruit flies doing little mating dances and copulating/laying eggs. Genetic variation in each new generation is a highly constrained matter, compared to the very random bacterium.

Interesting to think how many replicational constraints the human species is now removing to create its new freedoms.

And equilibration is now playing a secondary role. The primary biological dynamic is a one-way process that improves the homeostatic efficiency of organisms as a means of improving their reproductive success, and adapts the species to a given physical environment for the same reason.
.

In populations, genes form equilbriums. There is coming and going and a resulting gaussian distribution - the definition of equilbrium. Same in ecosystems. Species come and go, and there is no great change in entropy throughput overall.

There may be some progressive trend in the history of life towards richness, complexity, entropy throughput, but it is perhaps surprisingly weak. Major steps forward would have to depend on new breakthroughs, like photosynthetic pigments early on, mitosis, somewhat later, grammatical language very recently. Then we see punctuations in the evolutionary equilbrium!

Once life gets going, this species-level evolutionary process itself begins to evolve. Since the physical environment changes over time, there is a selective advantage for species that evolve more effectively than others. As species become more and more interdependent, this co-evolutionary dynamic becomes more important, and the evolutionary process itself starts to evolve more and more rapidly – in the emergence of sexually reproducing species, the emergence of flowering plants, etc.

So there are stages in which one level of order emerges “naturally” from a previous level – as homeostatic systems arise naturally in any complex physical environment, or as ecological structures of inter-species dependency emerge naturally once life begins to proliferate.
.

So we agree generally. But I think I am making the point that major game-changing advances are more random and unpredictable than natural and automatic - steps just waiting to be found.

Another problem with life is that it exists in a dynamic context - shifting tectonic plates, variable sun brightness, wobbly Earth orbit, asteroid smashes - so this would obscure any natural path of innovation that did exist.

However, I myself would argue that there is a deep natural path to be seen if we drill down a few levels to that of a succession of increasing dimensional constraints. So as I mentioned, I think, the innovation of serial codes - OD symbols on a 1D line - is this kind of deep structure, the explanation of the causal power and dramatic shifts that occurred with genes, then words.

So there is something for sure in this search for a progressive tendency. But that is why you need a hierarchy theory approach stripped right down to raw geometry really.

Now each stage takes up and incorporates the fundamental structures of previous stages. At every stage of evolution there are new kinds of self-equilibrating processes, e.g. in economics... new kinds of “gradients” and “dissipative structures”. That means that hierarchy theory can indeed find instances of its basic principles almost everywhere. But the fact that homeostatic equilibrium is a more generally applicable concept than reproduction doesn’t make it more fundamental as an explanatory principle, when it comes to biology.
.

And this is the value of dichotomistic logic. It does not force us to chose one or the other as more fundamental. Instead it forces us to chose both as equally fundamental.

So homeostasis = innovation. Stasis = flux.

This is explicit in the MR (metabolism and repair) systems approach of Robert Rosen - the doyen of recent mathematical biology. See...

http://www.people.vcu.edu/~mikuleck/PPRISS3.html

Philosophically, my personal proclivity is to emphasize the unique, the breakthroughs to a new level of possibility. I think there is an overarching “logic” to the way new possibility-structures have emerged in the world, but I don’t think concepts like equilibrium are adequate, even as metaphors. The “logic” needs to include both the “natural” emergence of new kinds of structure over time, within given constraints, and unique events (like the origin of the universe itself?) that in certain special contexts can create a whole new ball game, making accidents meaningful in new kinds of relationships.

Perhaps I have persuaded you that it is one-sided to have such a preference - either for what changes/progresses, or what stays the same. You need both these things to have anything - simple or complex.

The unique can only exist with the context of the general. But a holistic model would model both and not expect to construct the world merely by reference to the unique.

And a last comment on placing progressive tendencies and complexity on a pedestal. In the long run, the second law wins. Destiny is heat death. So simplicity is there at the start and there again at the end. And globally, slice the universe at any moment along the way and it has to be at equilibrium, regardless of whether it is all the simplest kind of equilibrium, or whether as now, around about the middle of its history, that simplicity is seasoned with a small dash of complexity.

Remember that the cosmic background radiation constitutes most what exist so far as the universe is concerned. The rest is just a few percent at best.

http://www.mso.anu.edu.au/~charley/papers/LineweaverEgan2008v2.pdf