Introduction to Evolving Systems: Beyond Emergence to Agency

The Emergence of Form, Meaning, and Aesthetics

Evolving Systems:
Open Conversations

Introduction to Evolving Systems: Beyond Emergence to Agency

Paul Grobstein
8 October 2009

The issue:

Computer models show that quite complex systems can develop from quite simple interactions of quite simple things, in the absence of any architect, blueprint, or intention.  Biological evolution, as it is currently understood, seems to provide a real world example of such emergent phenomena.  But biological evolution has apparently also given rise to human brains and with those seems to have brought into existence architects and intentions.  How can we think about systems that are both emergent, lacking an architect, and also intentional?  How do the two interact?  What implications does such hybrid character have for thinking about a whole host of human phenomena, including the process of inquiry itself? 

Background:

Excerpted from "From Complexity to Emergence and Beyond" (Grobstein, Soundings 90(1/2), 2007, available as a Word file)

Emergence is increasingly “in fashion” (Resnick, 1994; Holland, 1999; Johnson, 2001; Keller, 2003), as was complexity fifteen years ago (Waldrop, 1992) and systems theory (von Bertalanffy, 1968), cybernetics (Wiener, 1948), and unified science (Reisch, 2005) among other things before that (cf. Clayton, 2006).  In the case of emergence, as with its predecessors, different people are drawn for different reasons to a something that isn’t exactly the same for each person but instead reflects to varying degrees varying senses of dissatisfaction with existing paradigms for asking and answering questions.  In this important sense, emergence exists because of emergence – because of inquiry as an emergent process" ...

The human brain consists of a very large number of relatively simple elements (neurons) interacting in relatively simple ways by exchange of information across synapses. Hence, the brain’s function would be expected to have an emergent character to it. What seems to give the brain its additional function as an architect is a bi-partite arrangement (see Figure 3; Grobstein, 2007b) in which a series of model building elements (circuits of neurons) that interact directly with the world also report their activities to a second set of neuronal circuits that in turn use them to develop goals and alternative behaviors for the system as a whole. The former corresponds, more or less, to what may be conveniently referred to as the “unconscious” and the latter to “consciousness”. Behavior reflects continuing interactions between the two, and is normally an expression of the blending of the emergent and the architect ....

what is significant about inquiry is not only the degree to which its stories account for the past in particular local situations but how well it recognizes additional patterns across wider realms in its own activities and how “generative” it is, how effectively it contributes to the ongoing explorations of what might be that characterize the larger emergent process of which inquiry itself is a part ...

the business of inquiry is as much about creation as it is about discovery, ... we ourselves are a result of and a continuing participant in a larger process of ongoing exploration/creation, and hence should not expect ever to achieve either a definitive description or properties and rules nor a definitive predictability.  We should instead be at least content, and perhaps even exhilarated by, being ourselves creative participants in the very phenomena that we inquire into.

Additional relevant materials:

• Emergent Systems: A Discussion
• Complexity and Emergence resources on Serendip

Discussion materials/notes

Variations of persistent pattern and explanations

• Some patterns don't depend on an architect/designer who had those patterns in mind before they came into existence, are "emergent"
• Other patterns do depend on an architect designer who had those patterns in mind before they came into existence, are products of "intention"
• Both sorts of patterns are products of evolution

Evolution: from randomness through emergent pattern to intentional pattern

• Intention is a product of evolution rather than a starting point for it
• Once brought into existence, intention neither replaces emergence nor is slave to it.  Both contribute to pattern and influence each other. 

Implications for intentional inquiry

• Patterns are not fixed but both perspective dependent and changing
• There is no possibility of establishing general and invariant first principles and good reason to doubt they exist/are relevant
• Intentional inquiry in general influences what is being inquired into, both directly and indirectly
• Intentional inquiries are inevitably interdependent
• Intentional inquiry should be as much about creation as it is about discovery/representation
• Intentional inquiry should use a multiplicity of perspectives to generate new ways of making sense of things

One (non-linear, multiply-voiced) summary of the meeting:

 "Turning the Tables" (by Anne Dalke)

(from Shepherd's "Turning the Tables")

Paul began the discussion by picking up on a problem that emerged repeatedly in the five years of discussion held by members of the Emergence Group. The emergence paradigm "leaves something out": it gives no clear place for wrestling with problems of agency. "We need a way to use the emergence paradigm to take account of agency," to study and understand, in particular, the intersections of phenomena that depend on agency, with those that don't.

As we looked @ a series of six non-random patterns ("ripples"), in a variety of quite different realms ("earth, water, air..."), both living on this earth and extraterrestrial, Paul asked us to acknowledge that all but one of them were "explainable without an architect or a planner." These are "emergent patterns": each one came into existence without a pre-defined conception, and developed as the result of "simple interactions of simple things," without the intervention of a conscious designer.

But beings who look @ such images are inclined to see patterns.
There are no ripples in the absence of human observers.
What is the difference between production and perception?
(Tabled.)

We can account for much patterning without evoking a pre-existing template that brought it into being. This is a generalization of Darwin's insights, and the core idea of both evolution and emergence: there is no architect, no planner, no designer of the world in which we study and live. This emergence perspective can also be usefully invoked to describe a wide range of cultural processes, including the "invisible hand" of economics, the "contingent development" of history, and ongoing changes in art and literature.

BUT IT IS NOT ENOUGH.

We cannot always tell whether a particular pattern was intentional or not. It is a mistake to deny the possibility that sometimes intention DOES PRODUCE ripples. Where do these architects and planners come from, and how do they intersect with the forces of emergence? Our shared project here will focus on the intersection between purely emergent phenomenon and deliberate creation, and include adaptations that emerge in a creative sphere.

In evoking the variable of adaptiveness, it was suggested that a third category might be useful in discussing emergence, one describing those patterns which arise in adaptive response to their environment: not guided by an invisible hand, but guided nonetheless. ("There is no template for the stripes on a zebra." But lions might still be described as "agents" in this scenario. Although they don't have zebra stripes "in mind" while hunting, their acts of predation exert direct pressure, which contributes to the development of zebra stripes.)

Are we seeking an explanation of how or what? If the former, we need to consider the terms of causation. If the latter, is there really a fundamental philosophical distinction between pressure that is exerted intentionally, and that which is not?
(Tabled.)

Evolving systems is a "broader category" than emergence, one that encompasses the created as well as the emergent products of evolution. It stretches both further back in time than biological evolution, and further forward into the future.

But is this "evolved" distinction necessary? Both nature and art are products of evolution: one just arose later in the process. Why can't we say that an artistic process is "just" emergent? (Is this question "opposite" to the one about lions being "agents"?)
(Tabled.)

It is precisely the emergence of this "latter" process that interests us here. Against a background of randomness, new non-random assemblies (stars, solar systems...) developed. The process still continues, from the development of inanimate objects, through that of "model builders" to that of "storytellers," with the "prototypical point of evolving systems" being the "substantial production of non-random, improbable patterns in the absence of a designer."

Is probablity a technical term?
Organization is probable.
Particular organizations are probable.
In what various senses is complex organization probable?
(Tabled.)

The early changes that took place in "proto-evolution" ("before biology") were both persistent and path-dependent. Over time, an important new characteristic and particular implementation of this process emerged, in the evolution of living beings, which we now call "differential reproductive success," or "fitness." Five billion years ago, increasingly complex assemblies discovered a "special trick for persistence": the ability to represent their surroundings within themselves. Paul called such an "improbable" entity a "model builder," and argued that its capacity to use such "representations" increased the likelihood of its persistence in the world. The representations of the model-builder, by allowing it (for example) to anticipate and avoid obstacles, store that knowledge within itself, and change its position in the world, enabled it to "stabilize its own existence." (Examples of characteristics within an entity that allow it to anticipate its surroundings include sensory systems that identify food, and so alter the direction an organism takes to find it; or circadian rhythms that tell a tree to anticipate periods of increased sunshine.)  Organisms acquire information, in other words, that allows them to incorporate feedback mechanisms that affect their activities.

There's always a way for the sophist to get out of this, but:
how might we best characterize this transition?
The language of "representation" is particularly confusing, because it implies consciousness.
Might we replace the term "biological evolution" with that of "adaptive evolution"?
Is the transition point really about the storing of information? About self replication? About self assembly?
(Tabled.)

Active inanimate objects remained in the world, of course, as the model builders came into existence, and were influenced by them in turn. Once masked by scale, model builders became increasingly powerful causal agents. There was a "bidirectional causal relation among things": beavers affected the inanimate, and vice versa.

But how far back did this reverse arrow of causation go? How deep the reciprocal interactions?
(Tabled.)

Watch out for confusing semantics. We are not looking for foundational ideas here, but just identifing themes, giving "a sense of problems to explore." Inanimate objects gave rise to model builders, which in turn gave rise to the next "theme," that of storytellers: sensate, conscious and able to construct "counterfactuals," that is, to use models to conceive of alternate possible futures. Storytellers in turn became a distinctive force in the world, because they are capable of imagining circumstances that do not (yet) exist. They can play with contingencies.

This concept of emergence-as-incorporating-agency leads to a series of interesting practical problems. For example, the economic crash might well be better described in terms of emergent interactions, rather than intentional agency; or better yet, in terms of the interplay between emergence, intention, and efforts to conceive of counterfactuals. Existing prior to life, evolving systems now continue with a peculiar human point: the need to understand how change is produced both with and without architects, as well as in the interactions among those systems.

The practical point at the core of all interesting intellectual problems is this: how much of what we see is caused by intention, as opposed to "actually being out there"? Storytellers, the makers of counterfactuals, have "lost touch"; we can never know if our perceptions are "in here" or "out there." This is the intellectual problem of the human as storyteller. Do the patterns we see become patterns only in the presence of us as pattern-seekers?  "All that's out there" is "mild statistical regularities in noise." We "put names on it," and call them stories.

We can push this question further: as the undertaking of understanding emergence becomes more and more capacious, what is left out? Can any temporal systems prior to life be said to not evolve? "Is anything non-evolving?" Is emergence generalizable to everything? Is there no fundamental difference between rabbits and rocks?
(Tabled.)

Might emergence be described, alternatively, as "wholes being more than the sum of their parts"?  The emergent process is a contingent one, not deterministic, and so accurately described as "not reducible to its parts." But such a definition does not highlight the focus of today's conversation: the role of agency and intentionality in evolving systems. It is actually difficult for intentional agents to tell stories about emergent processes that are not intentional. Intentionality has a strong bias towards seeing intentionality as cause. There's a curious recursiveness or feedback loop operating here. We explain what happened, post hoc, and in making up the story, make it deterministic.

If we think about evolution as "a mechanism for transferal of information between states," we may have been "confusing the mechanisms" here. Some are random, some selected, some transferable. Alternatively, this conversation may have added a "fifth mechanism" to the traditional four means of biological modification: natural selection, mutation, drift and gene flow. Because of contingency, our intentions can have unintentional outcomes. Whether intended or not, this conversation may have had the effect of making us be less arrogant about our intended outcomes: so much of what emerges is both unintended and unavoidable.

Unavoidable?
(Tabled.)

Supported by the Metanexus Global Network Initiative, by Bryn Mawr College Center for Science in Society, and by the Serendip web site.