SSI Thought Letter Vol. 1 Issue XX: Substrate, Field, and the Cost of Layer Collapse

Image: Gary Larson — Post Cards From the Far Side

Often, conceptual failures in systems thinking, ecology, and organizational theory share a common mechanism. They are not failures of observation or even of logic. They are failures of category discipline: the systematic collapse of ontologically distinct layers into a single undifferentiated object. The result is frameworks that appear explanatory but are, on inspection, circular; metaphors that seem clarifying but smuggle in the assumptions they were meant to examine; and models that attribute system behavior to structural foundations that cannot, by definition, produce it.

The correction begins with the substrate.

I. The Irreducible Definition

Substrate, properly defined, is the material, structural foundation that constrains what can exist, occur, or persist within or upon it. It is constituted by *physical composition and by the geometry and topology* that composition produces. Its properties include consolidation, grain size, layering, porosity, and chemical constitution. These determine what the substrate can bear, filter, or resist.

What substrate does not do is act, regulate, process, exchange, or produce. It does not provide refuge, cycle nutrients, or drive dynamics. Any attribution of function to substrate is a category error. Functions arise from interactions among contents under specific boundary conditions. The substrate is the condition of possibility for those interactions. It is not their cause, their medium, or their product.

Worth noting: The hyperlinked National Park Service article (**https://www.nps.gov/articles/estuary-landforms.htm) is instructive. It defines estuaries as landforms and then immediately redefines them as hydrologic processes, collapsing constraint and dynamics into a single category. It never separates them again. This is not incorrect empirically; it is imprecise ontologically. It is an ontological collapse of the landform substrate, estuarine dynamics, and boundaries, a classification phenomenon replicated throughout estuarine literature.

This distinction is precise: substrate permits and limits. It does not perform.

Between the substrate and the field sits a transition layer: sediment, soil, or engineered material. Compositionally mixed, positionally intermediate, partially substrate-derived, and partially content-derived. It is shaped by the substrate below and by the settled material from the dynamics above. It is neither fully one nor the other. The transition layer is defined by its mixed composition and positional role, not by the functions that happen within it. For example, there may be rich biochemical activity, or it may be an oxygen-depleted dead zone of silt and biodecay. But that activity does not define the layer; its structural position in the statistical transition defines it, even though its activity is what draws our attention to it.

II. The Markov Blanket

A system has an identity only when it can be distinguished from what it is not. That distinction is not merely geometric. It is statistical. A system is coherent to the degree that the patterns operating within it are recognizably its own: mutually conditioning, internally coupled, characteristic of that system and not of what surrounds it. Where those patterns begin to transition into something other than what they are within, the system is reaching its boundary. That zone of statistical transition is the Markov blanket.* A statistical pattern, in this sense, is simply what the system reliably does (whether or not they are measured): the characteristic behavior that recurs, that is recognizable as belonging to this system and not to its surroundings. Where that reliability ends, the blanket begins.

The Markov blanket is not a line, a surface, or a layer. It is the region where internal pattern coherence dissolves. On the inside, dynamics are coupled and mutually conditioning. At the blanket, those patterns begin to change into something the system neither generates nor governs. Beyond it, they are no longer the system’s patterns at all. The blanket has no fixed thickness. It is as wide as the transition takes.

The estuary has two such transitions.

Below, the statistical patterns of the estuarine field gradually give way to those of the substrate. The physical correlate of this transition is the lower boundary layer: compositionally mixed, partially substrate-derived, partially constituted by material settled from the dynamics above. It has thickness, material extent, and a gradient of pattern change from the field to the substrate. It is the material site where the transition occurs, not the transition itself.

Above, the statistical patterns of the estuarine field give way to the patterns of the exogenous regime: gravity, atmospheric pressure, solar heat, and a local topology that is itself an expression of the substrate. The physical correlate here is not a material layer but a functional surface, the air-water interface, at which the contained dynamics of the field meet the forces that drive them from outside. This transition is relatively sharp. What crosses it inward (tidal forcing, precipitation, solar energy) activates the field. What the field produces crosses back through it outward.

The Markov blanket (***as explained by Karl Friston***) is the complete statistical description of both transitions. The lower boundary layer and the upper functional surface are its physical correlates. They are where the transition is materially or functionally locatable. The blanket itself is defined by the change in pattern, not by the physical features that coincide with it.

This reframes what boundary conditions are. In systems literature, the term is used loosely to mean external modulating forces. Those are inputs that cross the blanket. They are not the blanket. The blanket is the statistical transition zone that marks where the system’s pattern coherence ends, what enters, what exits, and what remains internal. Tides and solar heat are what cross it. The blanket is what they cross.

It also restores precision to the estuary’s identity. The Chesapeake Bay, the Thames Estuary, the Gironde: each is a specific, geolocated, persistent landform whose Markov blanket is defined below by a gradual statistical transition into the substrate and above by a sharp transition into the exogenous regime. The estuarine system is what exists and operates within that blanket. The blanket is what makes it a system at all, distinguishable from the open coast, the river above, and the sea beyond.

A system without a Markov blanket is not a system, but a region.

III. The Container/Content Distinction

The clearest narrative test of the substrate definition is the estuary, or more precisely, the systematic misdescription of it.

**An estuary is a specific, geolocated landform. The Chesapeake Bay is an estuary. The Thames Estuary is the estuary. The Gironde is the estuary. Each is a named, bounded, stable, persistent, mappable geomorphic feature. The name attaches to the landform. It does not attach to the specific [**hydrological condition occurring within it](https://oceanservice.noaa.gov/education/tutorial_estuaries/est05_circulation.html), except by default. Hydrologists can classify the water movement, but that’s a secondary function of the geological landform that contains and permits it. The Chesapeake Bay does not cease to be the Chesapeake Bay when seasonal drought reduces freshwater inflow, the mixing regime shifts, or human intervention alters its tidal reach. The landform persists. The identity persists with it.

What occurs within that landform under specific conditions (the mixing of freshwater and saltwater under tidal influence, salinity gradients, and ecological productivity) constitutes the estuarine system. It depends on the landform. The landform does not depend on it. And the mixing itself is not, by definition, estuarine simply because it occurs. Freshwater and saltwater mix in river plumes, in fjords, and in submarine groundwater discharge zones. What makes mixing estuarine is that it occurs within a semi-enclosed basin whose geometry retains it. The substrate is not incidental to that condition. It is the condition.

Compare the river basin. Here, the naming convention has not been distorted: the river basin is identified by its catchment geometry, the ridges and slopes that define where water drains. That identity persists regardless of current flow. Former river basins exist in deserts. The substrate is the system because it was always defined by structural constraints rather than by the dynamics passing through it.

Both the estuary and the river basin are named for their landforms. The difference is that one has been systematically misdescribed: its dynamics promoted to the status of its identity, its substrate demoted to mere container. That inversion is not a semantic quibble. It is the category that this essay collapses into.

IV. Layer Collapse as Systematic Failure

The error is not occasional. It is the default.

Ecological literature routinely describes substrate as providing refuge, regulating exchange, processing nutrients, and supporting biodiversity. Standard practice in wetland science defines substrate correctly in its opening terms (sediment, soil, bedrock, the physical floor) and then immediately attributes to it the functional properties of the systems it hosts. Refuge, nutrient cycling, hydrological regulation, and ecosystem foundation are assigned to substrate within the same framework that defines it as non-living physical material.

This is not careless writing. It reflects a consistent and widespread category collapse: the container absorbs the properties of its contents, and the distinction that would allow clean causal analysis disappears.

The same collapse operates in organizational and systems theory. Complexity frameworks require practitioners to classify a system (simple, complicated, complex, chaotic) before selecting a response strategy. The classification is supposed to determine the method. But system classification in the relevant sense is only fully possible after emergent behavior has been observed. The result is a loop: you must know what kind of system you are dealing with in order to act, but you can only know that after acting. The loop is unavoidable because classification requires observing emergent behavior, but emergent behavior cannot be observed until after the field has already been altered by intervention.

The framework presents this loop as productive ambiguity. It is more precisely a consequence of failing to separate what a system is from what it does, and both of those from what the practitioner currently knows. The practical cost is overcommitment: resources, authority, and momentum are allocated based on a classification that the system has not yet confirmed. When the classification proves wrong, the intervention is already underway, and reversal is expensive.

The error in all three cases is structurally identical. A lower-order property (material constraint, geomorphic form, system classification) is expanded to include the behaviors and functions that properly belong to higher-order dynamics. Once that misattribution and expansion occur, the lower-order concept loses its analytical utility. You can no longer use it to separate constraint from causation, container from content, or structure from what has come to be called “emergence.” (Emergence has become the catchall term for weak explanatory power.)

V. The Field Above the Substrate

Once “substrate” is restored to its proper definition, the architecture above it becomes visible.

Kurt Lewin’s field theory describes behavior as a function of the person and the environment: B = f(P, E). The formulation is correct in what it asserts: behavior is not produced by the person alone, nor by the environment alone, but by their dynamic interaction within a field of forces. What it leaves implicit is the ground on which that field operates. Lewin’s environment tends to absorb everything external, including structural constraints that are properly substrate (the material conditions that precede and persist independently of any particular interaction). The distinction matters because, in Lewin’s usage, environment is dynamic: it changes with the situation. Substrate does not. Collapsing the two into a single term obscures the difference between what is given and what is contingent, and that difference is precisely where most organizational interventions go wrong.

Separating the substrate from the field corrects this. The substrate is not part of the field. It is what the field operates upon and within. It defines the possibility space; the field determines the trajectory through it.

Within the field, three frames operate and are irreducible to one another.

The individual frame encompasses the internal states, dispositions, and properties of agents: what each actor brings to the interaction, independent of the specific situation.

The relational frame encompasses the interactions among agents: what occurs between them, the feedback structures, and the reciprocal influences that neither agent produces alone.

The situational frame encompasses the contextual and environmental conditions that modulate both individual and relational dynamics: what is present externally, as well as the pressures and affordances the context provides at a given moment.

These three frames are jointly sufficient and individually necessary. Behavior cannot be derived from any one alone. Individual properties without relational context produce no interaction. Relational dynamics without situational grounding produce no traction. Situational conditions without agents produce no behavior. The three are irreducible and mutually conditioning.

VI. The Operative Synthesis

The complete structure is then as follows.

Substrate defines what is possible. The field (constituted by individual, relational, and situational frames) determines what actually occurs within that dynamic space of possibility. Behavior arises from the interaction among the three frames, constrained by the substrate.

This resolves the layer collapse problem. Substrate is no longer credited with what the field produces. The field is no longer treated as unconstrained. And the three frames within the field are kept analytically distinct rather than collapsed into an undifferentiated environment.

It also resolves the practical problem that complexity frameworks encounter. The question is not what type of system this is. The question is what the substrate permits, what the current field conditions are, and what the interaction among individual, relational, and situational forces is producing. That is answerable without prior classification. It is answerable by attending to constraints, to dynamics, and to the boundary conditions that modulate both.

When I have publicly asserted, “an estuary is just a substrate” (raising some hackles), it does not deny the well-known dynamics of mixing, salinity gradients, or ecological productivity. It isolates the ontological layer on which those dynamics depend. Properly defined, the estuary is, in actuality, the geomorphological landform: a coastal configuration of geometry and topology that constrains what can occur there, nothing more, nothing less.

Mixing dynamics and salinity gradients are not irrelevant. What people usually mean by “estuary” in hydrology or ecology, circulation, nutrient exchange, biological richness, etc., are precisely those contingent processes that arise within that landform under specific boundary conditions. They are not the estuary itself; they are what happens in it.

Framed this way, the distinction becomes clean. Estuary names the substrate; estuarine (not the brand, but the actual natural estuarine) names strictly the dynamics. The landform exists whether or not the characteristic flows are present, whereas the system that people usually describe exists only when coherent patterns occur within a definable boundary (a Markov blanket).

This isn’t a rejection of conventional usage so much as a tightening of it. It’s separating structure from programmable process to prevent the category collapse that treats conditions of possibility as if they were the dynamics they merely permit.

Lewin modeled the game. The substrate is the board. Neither is sufficient without the other, and neither is the other.

VII. Key Test at Scale

Image by Perrin Remonté, Cartographie et exploration — PerrinRemonte.com

Cartographer Perrin Remonté has captured, at a continental scale, the essence of multiple interacting fields whose boundaries sometimes align, creating the beautiful illusion of a single organizing principle. The two-color scheme presents a single binary, north vs. south drainage. But Europe doesn’t have one continental divide. It has multiple watershed boundaries serving multiple seas. The watershed divide doesn’t create two different substrates. It creates two different trajectories from the same substrate. The watershed boundary isn’t a domain classifiable as “complex” or “complicated.” The ridge is the structural feature that produces the classification.

The Danube drains east to the Black Sea. The Rhône drains south to the Mediterranean. The Rhine goes north to the North Sea. The Vistula goes north to the Baltic. These are distinct drainage systems with their own ridges, their own organizing geometries. The cartographer collapsed multiple interacting watershed systems into a two-color binary, which is imposed as a cartographic decision rather than a geographic fact, but the substrate is one. Though there are many drainage systems, the substrate is unified and continuous, with multiple organizing principles operating across it.

VIII. The Cost of Layer Collapse

More than merely a matter of conceptual imprecision, layer collapse is the loss of the analytical leverage that comes from knowing what kind of thing you are dealing with before you decide what it explains. Even in metaphor. Only a dead elephant can be eaten one bite at a time. Otherwise, there won’t be a second bite. Substrate constrains. Interaction generates. Context modulates. Behavior reveals itself. Maintain those distinctions, and the model works. Collapse them, and you are left with a framework that describes everything and explains nothing.

The art of mixing metaphors is one way to contain the collapse within a transition. Terms do migrate across domains with modified meanings. The mixing manifests when the terms straddle metaphorical domains or transition within a domain. The metaphor can then attempt to capture the convenience of both positions without the rigor of either.

In geomorphic landforms, the substrate is the physical structure that defines and contains the system. The estuary is the substrate that precedes the system it enables. Not what’s in it, not what lines it, not the sediment dynamics. It’s the fundamental form not easily modified without cascading systemic consequences. That maps to an organizational analogy: the substrate is the organization's structural form. If the metaphorical substrate is the geomorphic landform, the question becomes, what is the organizational equivalent of that landform in the framework?

However, if the metaphorical shift pushes “substrate” upward, through and beyond the sediment layer, to the conditions that determine what transformations are possible or impossible (i.e., the affordance structure of a system) rather than its physical structural base layer, it’s closer to the disposition space, the set of enabling and constraining conditions that make certain pathways viable (vs. possible) and others not. The geomorphic landform persists while the disposition of space shifts around and within it in a dynamic estuarine environment.

IX. Alternative Facts, and “Theoretaphors”

Estuarine environments are among the most dynamic, ephemerally transitional, and ephemerally unstable geomorphic forms on the planet. They’re defined by flux. Tidal, fluvial, sedimentary, and ecological, constantly negotiating between river and sea. On any timescale relevant to organizational strategy, human systems, the geomorphic landform is, for all practical purposes, fixed. Estuaries don’t reshape themselves or their flow patterns between quarterly reviews or even generational leadership cycles.

Thus, the “theoretaphor” (theory+metaphor) that becomes the substrate for alternative abduction contingent on the metaphorical overlay.

Opting out of the geomorphic substrate in favor of substrate dynamism isn’t a merge lane for two different definitions of the same idea. They’re functionally opposite, one defined by its persistence, constraints, and its possibility space, the other defined by transformability, affordance, and viability space. The constructor is the competing agent that performs the transformation. The [geomorphic] substrate is what gets transformed. An attractor sitting between the geomorphic substrate and the estuarine system is a dynamic layer that doesn’t mediate or sustain but delegitimizes the geomorphic substrate over time, turning the estuary into something entirely different, or into nothing.

Follow this series via: **Mercator and The Right to Draw Maps**

Copyright © 2026 — Robert D. Jones — All Rights Reserved

End Notes:

These are environmental scientists who did the foundational nonmetaphorical work on and in actual estuaries, names you will seldom see mentioned elsewhere.

Endnote 1. Odum, H.T. and Hoskin, C.M. (1958). Comparative studies on the metabolism of marine waters. Publications of the Institute of Marine Science, University of Texas, 5, 16–46. No **DOI**. Available through Texas ScholarWorks, University of Texas at Austin Libraries. repositories.lib.utexas.edu

Endnote 2. Levin, S.A. (1998). Ecosystems and the biosphere as complex adaptive systems. Ecosystems, 1(5), 431–436. https://doi.org/10.1007/s100219900037

Endnote 3. Odum, W.E. (1982). Environmental degradation and the tyranny of small decisions. BioScience, 32(9), 728–729. https://doi.org/10.2307/1308718

*A system has an identity only when it can be distinguished from what it is not. That boundary is not a line or a surface. It is the zone where the system’s internal pattern coherence transitions into something it neither generates nor governs. Below the estuarine field, that transition is material: compositionally mixed, partially substrate-derived, partially settled from the dynamics above. Above it, the transition is functional: the air-water interface where exogenous forces enter and the field’s outputs exit. The estuary is what exists and operates between those two transitions. The substrate is the hard boundary terminus. The Markov blanket is the soft perceptual boundary that defines the transition. They are not the same thing.

The hyperlinked NPS* article (https://www.nps.gov/articles/estuary-landforms.htm) defines estuaries as landforms and then immediately *redefines them as hydrologic processes, collapsing constraint and dynamics into a single category, a collapse of the landform substrate, estuarine dynamics, and boundaries repeated throughout the literature.