The Vocabulary of Emergence

Gus performs an important service by trying to move us toward a more consistent lingua franca. There will always be an underlying tension between the Academie Francaise model of an artificially-managed vocabulary and the more organic, chaotic, and acquisitive, English language. The better explicated Gus can make his choices of preferred terms, the more likely his choices are to be appreciated and accepted by other researchers.

It seems reductionist and emergent phenomena are intended to be mutually exclusive, but after that the structure of the taxonomy seems muddled, because it remains unclear whether higher roman numeral headings are supposed to be subsets of the one preceding.

Saying that at least some chemistry may be emergent phenomena is tantalizingly vague. My problem is that reductionist phenomena can be fully described by physical theory, but according to the taxonomy, some chemical reactions cannot. Do you mean chemical reactions between molecules, or the emergence of complex molecules like DNA? Quantum-scale indeterminacy does not preclude reducing complex phenomena to a purely physical description. Schrodinger’s (1944) What is Life? argued that macroscopic physical laws result from small-scale chaotic indeterminacy. Embryological development is a deterministic process over the course of gestation unless you mean the evolution of new species, which is then brought in later in the taxonomy.

Then we come to complex adaptive systems. The distinction between simple and complex adaptive systems is unclear. Is cybernetics a reductionist phenomenon? Simple adaptive systems (I got the impression, but please clarify) work through either positive or negative feedback, but never both simultaneously? It might be helpful to define an additional category of simple adaptations, as opposed to simple adaptive systems. Often this would be an admittedly arbitrary distinction, which results from considering the purported adaptation in isolation, but defining this category would allow for a more precise delineation of what constitutes simple vs. complex adaptive systems.

You propose that all equilibrium systems are complex adaptive systems. Would you want this to apply to a chemical solution, where the equilibrium solution has been reached? Macroscopically deterministic, and certainly reducible to physical processes.

Then we come to Transformative Complex Adaptive Systems. Clearly one feature these have is that there is no final destination or equilibrium condition. Among the cultural examples you suggest is instrumental knowledge. Is this knowledge that is goal-directed in the manner institutions can be instrumental? Or is it just systematized or codified knowledge? Is instrumental knowledge captured by the expression “knowledge with an agenda?” This expression really flummoxed me!

The discussion of distinctions among different types of transformative complex adaptive systems is very well done and especially helpful, but I was left wondering if some of these categories might apply equally well to equilibrium-seeking complex adaptive systems (viola–new jargon!), simple adaptive systems, or even to simple adaptations considered in isolation. Everything in the taxonomy needs to be explicated to the same extent, or we run the risk of adopting a common vocabulary but continuing the process of talking past one another. Even I can see that Gus has made a terrific start here.

Last point—about law not being a pure spontaneous order because it is coercive. This suggests a further distinction between pure vs. impure or mixed spontaneous orders. Probably a very valuable idea, as which orders are ever going to be purely spontaneous in the cultural realm? The implication, however, is that if law represented a purely spontaneous order, people would observe it voluntarily and no coercion–and no law!–would be necessary. One observation is that most of us do voluntarily and automatically obey the law most of the time, even without being fully aware of its myriad requirements. Then again, spontaneous order accounts of the emergence of government and law–admittedly speculative–suggest they started as monopolies on coercion, and over time became progressively less predatory, but not necessarily less coercive. Finally, Hayek distinguishes between customary or common law, which has over time come to be accepted as the “rules of the game,” and positive legislation (positive law) which can be enacted without reference to accepted custom. Only customary or common law is a spontaneous order in Hayek’s view, though both kinds may involve coercion.

OK- I have had time to read Robert’s comments more carefully and have some additional comments. I thought my initial presentation was clear, but I will be more explicit. I believe I am not adding anything we did not agree on at our last conference, except perhaps some might disagree over specific examples. But I make absolutely no claim to originality here.

To me it obviously makes no attempt to be exhaustive nor was it presented as anything but a work in progress by those scholars who devised it. The schema has a strong bias towards the social sciences not because that is all that we consider emergent, but because that is where those of us who worked out this nomenclature felt reasonably competent. Stuart Kauffman, for example, has done exhaustive work on emergence in the natural sciences, but is much sketchier with regard to the social sciences, which seems reasonable as he is not one. We are. Speaking for myself, I think this schema can dovetail with Kauffman’s work pretty well, and anyone who did that competently would provide a real service to the study of emergent phenomena across disciplines. The attempt would be illuminating whether ultimately the two did dovetail, or they did not.

So there is first a division between phenomena that can be analyzed reductionistically and phenomena that have emergent qualities that cannot be comprehended reductionistically. We then turned to what can be analyzed as emergent phenomena.

That is Part. I.

Regarding Part II, within emergent phenomena there are at least three areas that appear to qualify, two of which are clearly not in the social sciences. I suggest Robert and others who are interested in areas like chemistry read Kauffman for starters. Many people will argue that water has emergent properties that cannot be discovered no matter how much we know about Hydrogen and Oxygen. Perhaps people more competent with the details of physical science can enter here and help flesh this category out or alternatively demonstrate why it is too big or has other flaws. I have no dog in that fight, as my work is in the social sciences and ecology. None of us were thinking of chemical systems here. Or of chemical equilibrium.

I would welcome what a chemist or physicist or other such person who takes emergent phenomena seriously might add to this description, or perhaps argue for modifications.

Robert suggests tweaking this. Fine by me, but first argue for why the tweaking leads to insights and distinctions we need to appreciate. (I have a very strong bias against the adequacy of abstract theoretical categories that cannot be linked to interesting empirical phenomena. Very very strong. One reason I was attracted to Hayek and away from Mises is because it could lead to a far wider range of useful insights that could be investigated to see whether they were useful.) So far I see no reason to tweak it and a good reason not to – to keep it simple. The study of emergence already suffers from too many similar concepts that ultimately cover different phenomena or different aspects of the same phenomena. This makes communication difficult.

When Timothy Sandefur used the example of students making pathways between new buildings as an example of spontaneous order, he was talking about something very different than what Hayek meant by that term. Consequently, his critique of Hayek in the Independent Review did not address Hayek’s argument or concepts. We stuck with Hayek’s mature concept of spontaneous order because both he and Michael Polanyi used it for systems like the market and science. It had a track record of providing interesting insights. Consequently for us Hayek’s term helped us focus on what was different about these complex adaptive systems from other kinds. All spontaneous orders are complex adaptive transformative systems, but not vice versa. All complex adaptive transformative systems are complex adaptive systems, but not vice versa. And so on. THAT is how to read this schema.

In Part. III we moved on to unpacking complex adaptive systems. They are distinguished as having BOTH kinds of feedback, unlike a cybernetic system like a thermostat, which responds to negative feedback only. Responding to both kinds of feedback enables complex adaptive systems to be far from equilibrium and still preserve adaptiveness to their environment in order to persist as systems. Some are transformative and some are not.

The difference between simple and complex adaptive systems appears to involve the concept of equilibrium. Think of the old model of a climax forest that may be approached in some instances, but not in others. A mature redwood forest or perhaps some Amazonian rainforests seem about as close to such a forest as we can get. Transformations must come from the outside.

A lake is quite different. Even without an inflow of silt from upstream a lake’s ecosystem will eventually transform it into a marsh, then a meadow, and perhaps then into a forest. It remains an ecosystem, but its particulars change not only as individuals, as in a climax forest model, but with the types of individuals, shifting from trout eventually to spruce over time.

Think also of Israel Kirzner’s concept of entrepreneurship as equilibrating or Thomas Kuhn’s concept of normal science, which does similar work in his thinking. They are analogous to the climax forest model. Now add Schumpeter’s insights of entrepreneurship as disequilibrating, and revolutionary science as disruptive to established paradigms. The market persists and so does science, but their similarities with earlier science and markets become increasingly abstract.

Instrumental knowledge is certainly a complex adaptive transformational system. Look at that huge dimension of scientific knowledge that is instrumental, as well as the continuing creation of inventions and adaptations. Upon reflection there is a possible ambiguity because not all instrumental knowledge is scientific, and for scientists not all scientific knowledge is instrumental. But scientific methods and procedures require knowledge that qualifies as scientific to be reliable and ideally subject to manipulation and prediction. If we cannot manipulate it (for example, a super nova) we seek at least to predict it.

Your comments on law suggest to me that you have a different concept of spontaneous order than has become clarified in our work. Spontaneous social orders often have severe sanctions if their rules are violated and this in no way at all makes them “impure.” What matters is that the rules enable independently arrived at projects to be pursued, these projects can be contradictory to one another, and their contradictoriness is not a flaw but a means by which the system discovers what plans fit best into the existing network of successful plans and enable those involved to acquire systemic resources.

The law issue was probably the most hotly discussed issue in our meetings, but law is inherently hierarchical in a way that a spontaneous order is not. Those who break the law suffer sanctions and some people have the power to order physical sanctions whereas others do not. This is radically different from science, markets, democracy as a system for discovering public values and deciding laws, and other spontaneous orders.