Chapter 3: Constraint Closure and Evolution

Chapter 3: Constraint Closure and Evolution

It’s not only the parts of a bigger system that influence the system, but it’s also the system that influences the parts. In this case, the level of organization can constrain and thereby guide the evolution of the entities at that level, and at the same time, the level is made up from those entities themselves. In this sense, levels define coevolving niches for their composing entities.

Johannes Jaeger, excerpt from his lecture Rainforest Ontology – Beyond Networks: The Evolution of Living Systems https://m.youtube.com/watch?v=3Jw8N3cUZYA&list=PL8vh-kVsYPqOKJOboONJIQBd8ds0ueM_W&index=5

This epigraph applies to the project of ontogenology (a typology of ontically precedent generative processes) and a rainforest ontology of processes per Jaeger (2020) and Wimsatt (2007). This also applies to the present chapter’s topics: evolvability, and the constraint closure of evolution. The flow of evolutionary constraint, the evolutionarily constrained, and evolutionary flow itself are in coevolution. This constraint closure yields evolvability of evolutionary systems over time. 

Constraint Closure

Organisms actively self-organize and individuate via operational closure; three types of closure are catalytic task closure, work task closure, and constraint closure. This chapter emphasizes the role of constraint closure.

We miss in our science the idea of a system that constructs itself. I will introduce the requisite concept due to Maël Montévil and Mateo Mossio (2015) called “Constraint Closure”. [...] [W]e can think of constraint closure like this: it is a set of both constraints on the release of energy in non-equilibrium processes, and those processes, such that the system constructs its own constraints. (Kauffman, 2019 p. x)

By combining an operational definition of work, boundary conditions and constraint work cycles, Kauffman explains how constraint closure contributes to developing an organism’s holistic operational closure.

Work is the constrained release of energy into a few degrees of freedom [...] without the boundary conditions that serve as constraints on the release of energy in a non-equilibrium process, no work is done. [...] [L]iving cells, as we will see, really do work to construct constraints channeling their own release of energy that constitutes further work. So surely, no constraints, no work. And often, no work, no constraints. Call this the Constraint Work cycle. [...] It takes constraints on the release of energy to get work—and the work done can construct yet more constraints![...] Thus, a set of constraints on a set of non-equilibrium processes can achieve a work task closure that constructs the very same set of constraints. The constraints do work tasks that construct the same constraints, or boundary conditions. The system can literally build itself! This is the amazing concept of Constraint Closure. (Kauffman, 2019, pp. 19-22, emphasis added).

This concept of constraint closure is similar to the concept of second order contextual constraints outlined by Evan Thompson. 

A “second-order contextual constraint” is established when the organization of the whole system emerges as a constraint on the system’s components. [...] downward causation corresponds to the second-order contextual constraint of a system’s organization restricting or limiting the degrees of freedom of the system’s components. (2007, p. 384).

Examples of second order contextual constraints include a cellular membrane on the scale of unicellular closure or individuation on the scale of multicellular closure. Second order contextual constraints and the developmentally-constrained are circularly causal. The organism generates boundary conditions, constraining the release of energy into a few degrees of freedom to enable useful work, which generates further boundary conditions (i.e. the constraint work cycle). The constrained actively produces the constrainer which shapes its form in circular return. Water does not self organize a cup to give it shape, but life does self organize its boundaries. Second order contextual constraints shape the organization of life, as life circularly organizes its own constraints. This type of circular operational closure is an example of constraint closure.

Evan Thompson gives a general definition of constraint as follows, applied to complex systems science:

What exactly does “constraint” mean in this context? In complex systems theory, constraints can be understood as relational properties that the parts possess in virtue of their being integrated or unified (not aggregated) into a systemic network. “Constraint”  is therefore a formal or topological notion. (2007, p. 382).

In this way, constraint closure contributes to operationalizing a Kantian whole, in which interdependent parts exist for and by means of the whole. For organisms, independent parts do not aggregate to make a whole. Instead the unified whole integrates and differentiates, and parts are abstractions. 

Coevolution between Constrainers and the Constrained

Processes of evolution, development, and niche construction can be understood as coupled, flowing and in coevolution. The coevolution of these processes constitutes second order- or meta- processes, such as evolvability and metaplasticity.

Together, governing and enabling constraints pressure and shape their flowing evolutionary and developmental content. Constraint, content and flow itself all coevolve together as processes in states-of-becoming. Enabling constraints expand the space of potential behaviors. In contrast, governing constraints narrow the space of potential behaviors (Juarrero, 1999). A balance between enabling constraints and governing constraints stabilizes a space of possible behaviors that is not absolutely free and not absolutely deterministic. Constraint closure enables work cycles to generate second order contextual constraints and these constraints enable more work cycles. 

Enabling constraints, governing constraints and enabled work cycles coevolve together. The pressured, shaped contents coevolve together with the shaping constraints, with flows of coevolution occurring at different temporal and spatial scales. Constraints, contents and flow are each plastic over deep time on their own scales in distinct but interacting ways. Constrainers shape and selectively pressure their metaphorically contained evolutionary and developmental flows. The shaped content reciprocally presses-back, shaping the constrainer’s second order contextual constraints. Processes of work cycles, constraints and second order contextual constraints coevolve. This coevolution is a meta process. Over time, meta processes evolve and this is exemplified by evolvability, metaplasticity and major evolutionary transitions. Evolution, evolutionary subjects and their flows are coupled processes that undergo second order processes of meta-evolution. 

Constraints, contents and the possible flow of evolutionary flow itself interact as a higher-order complex system of meta-processes. These processes undergo meta-coevolution yielding second order evolutionary processes, not previously possible. Under process ontology, the biological world is process “all the way down” and “all the way up.” Scaffolded, evolutionary-transitioned and macro-scale  processes become in ever new, previously un-predictable (un-prestatable) ways. Their processual coupling and circular causality leads to dynamic co-emergence of new possible evolutionary processes over time. The plasticity of evolutionary flow relates to evolutionary transitions and enables metaplasticity and evolvability.

To use Stuart Kaufman’s terminology: “Adjacent possibles” open-up and enable (not entail) “unprestatable” new state-spaces of “becoming” in the nonergodic universe above the level of atoms (Longo et. al, 2012). Constraints enable, not cause, pressures to shape the flow of ever new processes-of-becoming. 

Enabling constrainers shape and pressure their evolutionary and developmentally flowing content (modulation). The flowing content shapes the constraints in reciprocal coevolution (transduction). In many ways, the evolution and phylogenesis of life can be conceived as patterns of novel processual structures imposed on flow. Niche-constructions take the role of second order contextual constraint, pressuring the possible structures of phylogenesis. In turn, “contents” of phylogenesis and adaptive radiation reciprocally shape their “constrainers” of niche and environment. Speciation leads to niche-construction in a temporally-lagging coevolutionary effect, even as niche construction pressures speciation. This is an operationally closed circular causality of evolutionary processes undergoing constraint closure. A circular meta process of ecological evolutionary developmental science can be profiled, hence the study of “eco-evo-devo.” 

Similarly, the developmental ontogeny of one’s life can be conceived as patterns of process imposed on flow. Developmental constraints enable ontogeny, and the developing organism reciprocally affects its constraining world. The developmental constrainer, the ontogenetically constrained and the ontogenetic flow itself all coevolve together, albeit on different time scales and in different ways. Development is neither nature nor nuture alone. Development involves the complex coevolution of an organism-environment system and constraints, all processes themselves on different scales of space, time and organizational closure. 

Again, the co-evolution of flow itself relates to the concept of evolutionary transitions and enables evolvability, i.e. evolutionary flow is enabled in new ways. A general example of evolvability is the Baldwin effect combined with genetic assimilation (Kaas, 2009).  Examples of evolutionary transitions regarding biological scales of organization include unicellularity, multicellularity, sociality and cumulative culture (Smith & Szathmáry, 1995). Analogous to the concept of macrostates from statistical mechanics, higher scales of biological organization can be considered as dynamically constrained sets that satisfy conditions of higher-order formal organization. Similarly, the higher macroscale orders of biological organization enclose evolved, developed, and ordered constituent microscales. The highly ordered macroscale set space is dynamically maintained by the organism's operational closures, using constrained free energy to do relevant work cycles, reciprocally generating novel enabling constraints. Major transitions of evolutionary processes enable major meta-transitions of second-order processes, i.e. major transitions of evolvability and metaplasticity. Transitioned ways of evolving are scaffolded and ratcheted, thus macroscale orders of evolutionary flows-upon-flows emerge. This is a picture of basal (prim-ordial) complex systems of ontical processes self-organizing novel (higher-order) complex systems of ontical processes. Macro scales of ontical processes emerge as “the nonergodic universe above the level of atoms,” coevolving with their micro scale processes in constraint closure. This conception is coherent under process ontology, as process is enabled by process “all the way down” and “up.” The coherence is facilitated by constraint closure, and such circularly-causal operational closures are able to internally differentiate as long as the networked integration of closure obtains over time. 

This conception helps frame the “chicken-egg” problem. Constraints and content are holistically and nonlinearly co-evolutionary, and their synthesis yields dynamic co-emergence of ontically novel processes. Enabling and governing constraints pressure the shape of their flowing contents, and the content reciprocally shapes the constrainer in temporal delay. This process is a holistic concrescence as constraint, content, and flow-itself coevolve together. Constraints and content come-to-be in holistic nonlinear coevolution and co-ontogeny, albeit on different time scales, with temporal delays relative to each other’s changing state. The nonlinear coevolution of evolutionary processes yields dynamic co-emergence of second order evolutionary processes. 

Embedded in niches of complex processes, organisms emerge. Beings are thrown into becoming, not background-independent from their co-extensive environment. Their existence is always-already co-attuned, co-attuning and becoming. A scientific observer designates abstract and bounded entities, contingent to their pragmatic lines of questioning. They intersubjectively compare notes with methodological rigor, scaffolded by sociocultural niches of highly evolved linguistic, scientific and mathematical practices. Through an enactive process of sense-making, organisms and observers objectify stable patterns from their flowing environment. They abstract forms of meaning according to norms of their idiosyncratic pragmatics, with the possible space of objectifications constrained by physicality (i.e. objectifications are not freely enacted ex nihilo). Objectifications and significations are enacted via facultative interactions (structural coupling) with environmental flow. Organisms objectify and enact significance in coupling with their embedded environmental flows. Significance is relative to their enacted norms and the precariousness of coupling with an environment to obtain operational closure. The possible space of enaction is constrained by an organism-environment system’s spatiotemporally embodied attunement. Objectifications are pragmatically enacted stabilizations, but this apparent stability belies a metaphysical nature of flow and process. The ontology of ontically stable, observer-independent objects with determinate properties is the claim of substance metaphysics. Under process ontology, this proposition is an illusory reification. The proposition of substance ontology is a mistaken conclusion made by an observer reflecting upon the abstracted reality of their enacted objectifications. Under process ontology, to recognize process as metaphysically primary and objectification as enactively derived is to see through the fallacy of substance ontology. 

Constraint, content, and flow are coevolutionary. Alternatively: Stricture, structure and flow are coevolutionary. Enabling constraints serve as a “constrainer” that pressures the temporally flowing evolutionary and developmental content, and the pressured content shapes the constrainer back. This eco-evo-devo scene of phylogenesis, ontogenesis and niche construction is conceivable as generating metastable patterns of higher-order processes imposed on the flow of primordial processes. Different abstract structures of process become apparent on different scales and granularity of time, space and order of evolutionary transition. Ordered processual-structures emerge on different abstract orders of scale. These ordered forms of process require input and output of energy, external constraints and work cycles to persist over time. Ordered systems are thermodynamically open and operationally closed. The negentropic, ordered forms of processes require input of free energy into constraint closures. This enables ordered processes to propagate their work cycles and thereby stably persist over time, imposed upon a continuous flow of entropy, matter and energy.

Patterns impose on flow: flow through constrainers yields metastable activity patterns, formal patterns, and material patterns. Flow itself is coevolvable through evolutionary transitions, enabling metaplasticity and evolvability. 

Dialectical tensions in this complex coevolutionary path-dependent walk through state-space yield novel syntheses, but not necessarily resolutions, equilibriums, nor golden-mean solutions. This dynamic co-emergence accounts for synthesis of previously un-prestatable novel processes, ever becoming in the “non-ergodic universe above the level of atoms” per Stuart Kauffman.

The process of evolvability emerges from a complex adaptive system involving the interactive coevolution between 1. evolutionary constraints, 2. flowing evolutionary content, and 3. the possible structure of evolutionary flow-itself (e.g. via evolutionary transitions). Evolvability implies that the evolutionary process itself evolves in its own metaplastic state space. This framework is a meta-complex system yielding second-order evolutionary processes. 

The flow of evolutionary content is mediated by natural selection, but to end here and consider natural selection as the “grand unified theory” of evolutionary science is premature and myopic. To reiterate, evolvability itself is enabled by a meta- complex adaptive system involving the reciprocally circular coevolution between: 1. enabling constraints, 2. flowing evolutionary content, and 3. the flow of flow itself (e.g. evolutionary transitions). Second-order evolutionary processes are generated via dynamic co-emergence when primordial processes (including but not limited to natural selection) holistically interact and change each other, over time.

A conceptually continuous flow of water forms a river. The river erodes earth, yielding a coevolving structurally-coupled form of ground-with-water. The river water and the riverbed coevolve their form over time as a pattern imposed on flow. Living creatures afford meaning, purpose, life and use from the river. The river is unsignified until incorporated into the autonomy of signifying organism- environment systems.  

Enabling constraints imposed-upon flowing energy and matter yield ordered processual forms over time. The path of least energetic resistance is to passively let things flow. Input of energy is required for an organism to regulate its interactive coupling with the environment through agency. In a strict sense, free will is unconstrained and thus unable to do useful work. Instead, agentive will is enabled by constraints. Work is defined by Stuart Kauffman as energy channeled into a few degrees of freedom. Thus, will is unformed if totally free, i.e. with unbounded degrees of freedom. Will requires constraint closure of both governing and enabling constraints to be operationalized. Freedom of will can be a misnomer or a nonstarter if its generation requires an unconstrained, unprecedented prime mover. Degrees of freedom are instead generated by constraints. Agential will has freedom though a dialectical tension of governing and enabling constraints, simultaneously limited and freed to move via these structuring limits. 

To place the river metaphor in another way: flowing water changes the shape of sandstone into a groove. The groove of sandstone changes the shape of reiterated water flow. A stream lays a path in its flowing, and that laid-path shapes the direction of the shaping flow, in return; structured patterns form over time from previously unpatterned processes. This is a picture of coevolution between the niche constructors and their niche constructions, each affecting each other on different time scales and rates of delayed return. 

Living “containers” (niche constructions) and living “contents” (niche constructors) reciprocally shape each other in dynamic co-emergence, in nonlinear circular causality. Each is plastic and meta-plastic on different scales of time, space, and in their own transitioning ways. The paths laid-in-walking by living contents upon their living containers shape the future flows of contents in temporally delayed hysteresis. To walk-and-lay a path is to have the path walk you on your reiterated return. The walked-on path walks-on you, shaping your course even as you still further lay the path in walking. In this way, the living contents (organisms) and their also-living containers (niches and biomes themselves being composed of other living organisms) are structurally coupled and co-attuned. Their interrelationships are describable by nonlinear coevolution, dynamic co-emergence, dialectical co-generation, or circular causality of eco-evo-devo. Evolution changes developmental processes as development reciprocally changes evolutionary processes. Evolution and development change an ecological niche, as the niche construction reciprocally changes further evolutionary and developmental processes. 

The ecosystem itself is of living organisms and holistically coevolves with the subject organism. It is not an independent background for decoupled objects, but an integrating and differentiating gestalt (like the figure rising from-and-with its ground, like the form of a rising wave to the ocean). The niche is non-background independent to the niche constructors. 

The figures and grounds therein are structurally coupled on different levels and timescales, obtaining various facultative or obligate relationships. Structurally coupled figures and grounds are not dichotomous nor background-independent.