A Cybernetic Teaching · Module Two

The Salmon, the Bear, and the Tree

System Zero, Ground Regulation, and the question Stafford Beer left underneath the VSM

Each autumn, on the rivers of British Columbia, Alaska, and Kamchatka, salmon return from the open ocean to the streams of their birth. They are met by bears. The bears eat what they can, drag carcasses into the forest, abandon the rest. The carcasses rot. Their nitrogen — nitrogen the salmon accumulated by feeding for years on krill and herring at sea — enters the soil. The trees lining the riverbank take it up. The trees grow. Their roots hold the riverbank in place. The riverbank shapes the stream where the next generation of salmon will spawn.

Up to 50% of the nitrogen in old-growth Sitka spruce trees beside salmon rivers is of marine origin. The trees holding the bank are, materially, partly made of fish. The salmon are, in turn, made of trees that fed the insects that fed the fry. The bear is made of both. Each of them is mostly made of each other.

This is the loop. But this module is not, primarily, about the loop. The rainforest module showed how a loop regulates. This module asks the question that the loop forces but classical cybernetics defers: what is the medium through which the loop runs? The salmon does not pass nitrogen to the bear directly; nitrogen passes through digestion, defecation, decomposition, soil chemistry, root uptake. The bear does not "regulate" the tree directly; it does so through the riverbank, the carcass, the soil. There is something under the loop — a connective medium — and the loop only works because that medium does its silent, continuous work.

That medium is what Peter Tuddenham calls System Zero: the ground that comes before Stafford Beer's VSM begins counting. This module is a teaching about it.

1. Watch the loop

Before the analysis, watch the imagery. The footage below is from Planet Earth (BBC, 2006), Episode 3: Fresh Water. Note: as you watch, do not focus only on the bears or only on the salmon. Watch the river. Watch the banks. Watch the forest in the background. Most of what matters is what the camera is not centred on.

Grizzly bears fishing salmon — BBC Earth, from the Planet Earth series. Watch what's off-frame: the riverbank, the forest, the carcasses left behind.

And a second clip — the longer ecological view, including the way bears carry salmon away from the river and into the forest. This is the move that closes the loop. Without bears (or wolves, or eagles) doing this transport, salmon nitrogen would stay in the water and never reach the trees.

The bears as nutrient pump — BBC Earth. The bear is not just a predator. It is the mechanism that lifts marine nitrogen into the forest. Without it, the loop opens.
"Around salmon-rich rivers, 40 to 80 percent of the nitrogen in shrubs and trees originates in the open ocean. The salmon signature isotope can be found as far inland as the Rocky Mountains." — Tom Reimchen, forest ecologist, University of Victoria

2. The loop, drawn

The salmon-bear-forest nutrient loop, with the connective medium of System Zero shown as the ground SYSTEM ZERO · THE CONNECTIVE MEDIUM soil · water · air · microbiome · time OCEAN marine N reservoir SALMON nitrogen carrier BEAR nutrient pump RIPARIAN FOREST ~50% N from salmon salmon return caught carcass dragged inland N enters soil → riverbank holds → next spawn fry return to ocean THE VISIBLE LOOP — what Beer's VSM would call System One salmon → bear → forest → riverbank → next salmon
ocean (marine N) salmon (carrier) bear (pump) forest (sink & bank) System Zero (medium)

Notice what the diagram does. The arrows around the top — salmon to bear to forest to river to next salmon — are the visible loop. They are what a documentary camera films. They are what Beer's System One would identify: the operating units doing the work. But the band along the bottom — soil, water, air, microbiome, time — is what every one of those operating units is standing on, moving through, and made of. The dashed lines connecting each operating unit downward into that band are not optional. They are how the loop actually runs.

You cannot photograph System Zero. The salmon's nitrogen doesn't appear on camera as it leaches from a carcass into the soil over a winter, gets fixed by microbes, gets taken up by a Sitka spruce root, and ends up — three centuries later — in a tree-ring. But it is in that invisible work that the loop actually closes.

3. The principles

0

System Zero — the connective medium

System Zero is the ground that comes before counting begins. Stafford Beer's Viable System Model identifies five systems within any viable organism or organisation: System One (operations), System Two (coordination), System Three (internal management), System Four (intelligence/environment), System Five (identity/policy). The numbering implies a complete count. It is not. The numbering begins one level too high. System Zero is the medium through which Systems 1–5 operate at all.

In the salmon-bear-forest loop, System Zero is the soil chemistry, the river hydrology, the microbial community in the forest floor, the atmospheric deposition, the slow time of decomposition. Beer's VSM would describe the salmon, the bear, and the forest as the operating units. But none of them can perform their function except through a medium that the model leaves silent. The salmon does not transmit nitrogen to the tree by handing it over; it transmits nitrogen by dying into a medium that does the rest. The medium is not a sixth system. It is the prior condition for there being any system at all.

This is the diagnostic move of ground regulation: to refuse what Pischinger and Heine, working in connective-tissue biology, called the Virchow Error — treating components as primary while ignoring the connective medium that makes communication between them possible. Cells are not primary; the extracellular matrix is. Trees are not primary; the soil-water-microbiome ground is. Beer's Systems 1–5 are not primary; the medium they regulate through is.

A

The boundary is a choice

Where does the salmon system end? At the fish? At the bear that ate it? At the tree whose nitrogen came from a carcass? At the soil microbe that fixed the nitrogen between bear and tree? Every answer is defensible and every answer is partial. The boundary of a "system" is not a fact in the world — it is a cut made by an observer for a purpose. This is Spencer-Brown's first move: draw a distinction. The cybernetic mistake is to forget that the cut was made and treat the system inside the boundary as if it could exist without what the boundary excluded. The salmon, the bear, and the tree cannot be understood as separate systems linked by transactions. They can only be understood as a single system whose components have been provisionally distinguished for analytical convenience.

B

Recursion across scales

Beer's most important architectural claim was that viable systems are recursive: each viable system contains, and is contained within, other viable systems of the same structure. A cell is viable. The organ is viable. The organism is viable. Each of these has its own Systems 1–5. The salmon is a viable system. So is the bear. So is the forest. So is the watershed. So is the coast. The same regulatory architecture appears at every level. This is what Beer meant by recursion, and it is on full display in the salmon-bear-forest loop. But — and this is the System Zero addendum — at every level of that recursion, there is a System Zero. The bear's gut microbiome is a System Zero for the bear. The soil microbiome is a System Zero for the forest. The ocean's chemistry is a System Zero for the salmon. Recursion goes all the way down, and at every level, the medium goes underneath.

C

Prehension and process

Whitehead's word for how one entity takes account of another is prehension — a "grasping" that need not be conscious or intentional. The tree prehends the salmon when its roots take up the nitrogen that was the salmon's body. The salmon prehended its parents' watershed when it returned to spawn. The bear prehends the river when it positions itself at the falls. None of these is a transaction between separate substances. Each is a process by which one occasion of experience inherits the past of another. This is what makes the loop more than a flow chart: every node is constituted by its relationships to the others, not merely connected to them. Whitehead would say the salmon does not have a relationship to the tree; the salmon and the tree are their relationships.

D

Time, slow and fast

The rainforest module showed a 24-hour loop. This loop runs on a different clock. The bear catches the salmon in a second. The carcass decomposes over months. The nitrogen enters the soil over a year. The tree takes it up over decades. The riverbank holds shape over centuries. The forest's structural memory of past salmon runs is recorded in the isotope ratios of tree-rings going back centuries. System Zero has its own time, slower than any individual organism's life, faster than geology. A regulator that operates only at the speed of its operating units will miss the regulation that happens through the medium. This is why human institutions, which think in fiscal quarters, regulate ecosystems badly — the medium operates on a clock the institution does not see.

E

What it means to "export your consequences"

The rainforest module ended with a design principle: the forest works because it cannot export its consequences — the rain it makes is the rain it gets. The salmon-bear-forest loop says the same thing in a different key. The salmon dies into the medium that grows the tree that holds the bank that shapes the stream where its descendants spawn. Nothing is exported. Everything that leaves a node arrives at another node, eventually, through System Zero. Industrial fisheries break this loop precisely by exporting it: salmon caught at sea never return to spawn, never feed bears, never fertilise trees. The visible loop continues — humans eat the fish — but the System Zero loop opens. Marine nitrogen no longer reaches the forest. Decades later, the trees are subtly poorer; the banks subtly weaker; the next generation of salmon subtly fewer. The loss is invisible to anyone watching only the visible system.

4. System Zero in relation to Beer's VSM

Stafford Beer's Viable System Model is one of the great architectural achievements of cybernetics. It says any viable system — a cell, a body, a company, a country — has the same five interlocking sub-systems. Below is the standard VSM stack with System Zero added underneath, where the salmon-bear-forest loop says it has to be.

Beer's VSM Systems 1–5 with System Zero as the prior ground SYSTEM 5 — IDENTITY / POLICY "who are we?" SYSTEM 4 — INTELLIGENCE / ENVIRONMENT "what's coming?" SYSTEM 3 — INTERNAL MANAGEMENT "here-and-now coordination" SYSTEM 2 — COORDINATION "don't trip over each other" SYSTEM 1 — OPERATIONS "the actual doing" Beer's VSM (1972, 1979) ▲ where Beer stops counting ▼ where the prior ground begins SYSTEM ZERO — THE CONNECTIVE MEDIUM soil, water, air, microbiome, language, trust, time — the ground through which Systems 1–5 operate at all — Tuddenham (2024)

Beer's question was: what does a system need to be viable? His answer was Systems 1–5. The System Zero question is prior: what does a system need in order to exist as a system at all? The answer is a connective medium — soil for a forest, an extracellular matrix for a body, a shared language and a shared trust for a human organisation. Without that medium, the operating units have nothing to operate through. Beer assumed it. Ground regulation insists on naming it.

The salmon-bear-forest loop makes the case visible. You can describe the loop in pure VSM terms: each animal is a viable system; the watershed is a viable system at the next level of recursion; the regulatory feedback loops are clean. But that description leaves out the soil, the water, the microbial decomposition, the slow nitrogen chemistry. And it is in those that the loop actually runs. Subtract the medium, and Beer's tidy diagram regulates nothing.

5. The entailment mesh

The Paskian structure for this module connects the rainforest concepts you already know to the new ones, and shows how System Zero sits underneath all of them.

Entailment mesh for the salmon-bear-forest module showing dependencies and the System Zero foundation A. THE BOUNDARY IS A CHOICE (Spencer-Brown) B. RECURSION ACROSS SCALES (Beer) C. PREHENSION & PROCESS (Whitehead) D. SLOW TIME OF THE MEDIUM (ground regulation) E. NOTHING EXPORTED: DESIGN IMPLICATION (the integration) ★ SYSTEM ZERO — THE PRIOR GROUND every concept above only operates through the medium soil · water · air · microbiome · time · trust · language ENTAILMENT MESH solid arrows = conceptual entailment · dashed arrows = grounding in System Zero
Why these arrows. You cannot grasp recursion until you accept that boundaries are choices (without that, every level looks the same and recursion is invisible — A→B). Prehension only makes sense once you accept boundaries are porous (A→C). Slow time is the temporal face of recursion — different levels have different clocks (B→D). The integrating insight — that nothing is exported, the design implication — depends on holding both prehension (everything depends on everything) and slow time (the consequences arrive eventually). Underneath all five, the dashed gold arrows: each concept only operates because the medium does.

6. Challenges

Reproduction · 0Explain System Zero to someone who knows Beer's VSM

Imagine you are talking to a manager who has just learned about the Viable System Model and wants to apply it to their organisation. They are about to draw the standard five boxes. Stop them, and explain why they need to draw a sixth thing first — beneath the others. Do not use the word "rainforest" or "salmon"; the explanation has to land for someone whose system is an organisation, not an ecosystem.

What a good answer reproduces The answerer should articulate that Beer's five systems are all operating systems — they do work. But operating requires a medium. In an organisation, the medium is shared language, shared trust, shared assumptions, shared physical and digital infrastructure, shared time. None of this is in any of Beer's boxes. All of it is required for any of his boxes to function. A strong answer will note that the medium is usually invisible until it breaks (when trust collapses, when the language people use stops meaning the same thing, when the shared infrastructure fails) — and that, by then, the VSM diagram still looks fine on paper while the organisation is no longer regulating.
Derivation · AWhere does the salmon end and the bear begin?

A salmon swims up a river. A bear catches it and eats it. The bear digests it. Months later, the bear defecates. A microbe in the soil mineralises the nitrogen. A spruce root takes it up. Five hundred years later, the wood of that spruce contains a recognisable marine isotope signature. At what point did this nitrogen stop being "the salmon" and start being "the tree"? Defend your answer — and then defend a different answer.

What a good answer reproduces The point of the challenge is not to find the right place to draw the line — there isn't one — but to expose that any place you draw it is a choice. The answerer should be able to articulate at least two defensible cuts (e.g. "when the salmon stops moving" vs "when the nitrogen is incorporated into living tree tissue") and recognise that these are different distinctions made for different purposes. A strong answer will note that the inability to draw a non-arbitrary line is not a failure of analysis but a feature of the world: the entities are continuous through the medium, and our distinctions are operational, not ontological.
Derivation · B → 0The bear's gut is a System Zero. Whose?

A bear's gut microbiome is the medium through which the bear digests salmon. So for the bear, the gut microbiome is part of System Zero. But for an individual gut microbe, the bear's gut is the world — the medium and the operating unit are different from one level to the next. Use this to explain how System Zero behaves under recursion. At every level of Beer's recursive nesting, what counts as "operating system" and what counts as "medium"?

What a good answer reproduces System Zero is level-relative. What is medium at one level of recursion is operating at the level below; what is operating at one level is medium at the level above. The bear's gut microbiome is medium for the bear, but at the microbial level it is itself a population of operating units operating through their own medium (mucus, pH, oxygen gradients). A strong answer will note that this is not a problem for the model — it is the deeper sense of recursion that Beer's VSM points toward but does not fully articulate. Recursion is not just nested operating systems; it is nested operating systems with their nested grounds.
Integration · D → EWhy do industrial fisheries break the loop without anyone noticing?

A fishing fleet catches salmon at sea. The visible System One — fish caught, fish sold, fish eaten — works. The economic VSM looks healthy. But the salmon-bear-forest loop has been opened: marine nitrogen no longer reaches the forest. Use the concept of slow time of the medium to explain why the consequences are invisible to the institutions that caused them, and what kind of regulatory architecture would be needed to see them.

What a good answer reproduces The fishery's regulatory clock runs in fiscal years; the medium's clock runs in tree-decades and forest-centuries. The fishery cannot see what it is breaking because the consequences will not register on any sensor it possesses, on any timescale it monitors, in any of the variables it has chosen to track. This is not an information problem (more data won't fix it); it is a System Zero blindness — the institution doesn't even have a category for what it is harming. A strong answer will articulate what kind of regulator would not have this blindness: one whose System Four (intelligence/environment) reaches into the medium itself, and one whose System Five (identity) is defined by belonging to a watershed rather than by extracting from it.
Transfer · whole meshIdentify a System Zero in your own work or organisation

Pick an organisation, community, or system you know well. Apply the VSM if you can — identify what its System One operations are, what coordinates them, what manages them, what scans the environment, what holds its identity. Then ask: what is the System Zero? What is the medium through which all of those Systems 1–5 operate? Is it being attended to? When was the last time it was repaired? What would happen if it failed?

What a good answer reproduces A strong answer will identify a specific medium — usually shared trust, shared language, shared physical infrastructure, shared cultural memory, or shared physical place — and notice something about its current condition. Often the diagnostic is uncomfortable: "we have not maintained this in years; we don't have a budget line for it; nobody's job is to attend to it; it is starting to degrade and our visible systems still look fine." The point of the challenge is not to fix the System Zero (that is much harder) but to develop the capacity to see it. Many organisational problems are System Zero problems misdiagnosed as System One problems.
Meta · cross-moduleWhat did the rainforest module not show you that this one does?

If you have read the rainforest module first, name at least one thing that this module brings into view that the rainforest didn't. And — harder — name something the rainforest showed that this module did not.

What this challenge is for The rainforest module showed regulation as a fast, internal loop. It did not foreground the medium because the rainforest's medium is partly itself (canopy and soil). This module foregrounds the medium because the salmon-bear-forest loop crosses radically different domains (ocean, land, river, forest) and the medium is what carries continuity across those domains. Conversely, the rainforest module showed autopoiesis — a system producing its own conditions — more cleanly than this one does, because the salmon loop spans the ocean, which the forest does not produce. A strong answer will notice that the two modules are not redundant; they pick out different things, and the cross-module entailment is itself a teaching.

7. Where this leads

The salmon module brings two threads forward into territory the rainforest left untouched.

First, toward groundregulation.com. The Pischinger/Heine connective-tissue work, the extracellular matrix as a regulatory organ, the Virchow Error and its diagnosis — all of this transfers cleanly from biology to organisations. A future module could read a clinical case and an organisational case side by side: the body that fails because its connective tissue can no longer mediate, and the organisation that fails because its trust-medium can no longer mediate. The structural pattern is the same, and the diagnostic — look at the ground, not the components — is the same.

Second, toward the System Zero addition to Beer's VSM. The next module in this sequence would build out the formal extension: how does System Zero appear in System One? In System Two? Up the recursion? What does it mean for an organisation's System Five (identity) to be defined by its System Zero (the place, the people, the language) rather than abstractly above it? Stafford Beer's Cybersyn project in Allende's Chile is, in retrospect, often read as a System One/Two/Three triumph. It might be more honest to read its tragedy as a System Zero failure: the political and cultural medium through which it would have had to operate was being systematically destroyed around it, and the model could not see this because it had no name for it.

Third, toward what the camera doesn't show. Both modules so far have used Planet Earth footage — but a recurring lesson is that the cybernetics is largely off-camera. The bear, the salmon, and the cloudburst are the spectacle. The microbiome, the soil chemistry, the slow nitrogen flux are the regulation. A discipline that takes its examples only from what can be photographed will systematically under-theorise what can't. That, too, is a System Zero point.