A self-regulating system, read through Wiener, Ashby, Maturana & Varela, and Pask
Each morning, in a tropical rainforest, the canopy begins to breathe. Water held in the soil is drawn up through millions of trees, released as vapour through the leaves, and rises on the warming air. By late morning the vapour has cooled, condensed, and clouded the sky. The clouds burst. Rain falls back into the canopy, runs into the soil, and is taken up by the roots again. The cycle restarts the next day.
The forest does not receive its weather. It makes its weather. And the weather it makes is the weather it needs.
This is one of the cleanest illustrations of cybernetic principles in nature. What follows is a teaching in three parts: first the loop itself; then the principles tagged where they actually do their work; then a Paskian entailment structure that shows what depends on what — with challenges that ask you to reproduce the understanding, not just recognise it.
Read the diagram clockwise. Start with the trees.
Notice three things about this loop. One: there is no first cause. You can start your description anywhere — at the leaves, at the cloud, at the soil — and you will arrive back where you began. Two: the only thing that enters from outside is sunlight. The water never leaves. Three: the loop is not just circular; it is regulatory. It actively maintains the conditions the forest needs to live.
Each of the following is a foundational concept in cybernetics. Each one is doing visible work somewhere in the loop above.
Cause does not run in a straight line; it runs in a circle. The trees produce the vapour that produces the cloud that produces the rain that produces the soil moisture that produces the trees. There is no upstream and downstream — every part is upstream of every other part by going round once. This is the foundational move of cybernetics: it is the discipline of loops, not chains.
Where in the loop: everywhere. It is the shape of the whole.
The loop does not just go round — it regulates. If the canopy gets too hot and dry, transpiration accelerates, vapour rises faster, the cloud forms sooner, the rain falls earlier, and the canopy is cooled and rewetted. If the morning is unusually cool and damp, transpiration slows and rain is delayed. The loop counteracts whatever pushes it away from its working range. This is what Ashby called an ultrastable system: one whose internal dynamics return it to the conditions its life requires. The forest is, structurally, a giant homeostat.
Where in the loop: in the rate at which trees transpire — they speed up when dry, slow down when wet.
Ashby's Law: only variety can absorb variety. A regulator must have at least as much internal variety as the disturbances it needs to handle. A single tree cannot regulate a forest's climate. But millions of trees, of thousands of species, with layered canopies and varied leaf shapes, can. The forest's diversity is not decorative — it is the regulatory mechanism. Strip out the variety (clear-fell a section) and the loop collapses: deforested patches stop making their own rain and dry out. The variety is the regulation.
Where in the loop: in the canopy itself — the population, not any individual.
No leaf knows it is making weather. The cloud that bursts at eleven in the morning is not in any tree's plan. It emerges from billions of micro-events — a stoma opening here, a droplet nucleating there — none of which contains a representation of the whole they are constituting. Pattern at one level is produced by interaction at another, with no central controller. Cybernetics takes this seriously: there is no manager of the rainforest's weather, only the loop.
Where in the loop: at the cloud — the level at which the local act of transpiration becomes a global event.
The forest produces the conditions of its own existence. The rain the trees create is the rain the trees need. The system's outputs become its own inputs. It is organisationally closed — the regulatory loop is internal — while remaining thermodynamically open (sunlight enters, heat radiates away). This is Maturana and Varela's autopoiesis: a system whose operation produces and maintains the very components and relationships that allow it to operate. The forest is, in this sense, alive at the scale of the forest, not merely at the scale of the trees.
Where in the loop: in the closure itself — the fact that the rain returns to the same trees that made it.
In Pask's Conversation Theory, understanding a topic means being able to reproduce it — to teach it back, to derive it from what it depends on, and to follow the why-paths that connect it to its neighbours. Here is the entailment structure for this teaching. Read each arrow as "you cannot fully understand the topic at the head of the arrow without first understanding the topic at the tail."
Pask distinguished serialist learners (who prefer to walk a single chain step by step) from holist learners (who prefer to see the whole pattern first and fill in detail later). Both paths are legitimate. Both arrive at the same understanding.
A → B → C → D → E → integration. Build one concept at a time. Each step depends only on the previous one. Slow but airtight.
Start at the integration node — "the forest is a self-regulating living system" — and work backwards, asking at each step: what would I need to be true for this to be true? The mesh fills in from the destination toward the foundations.
These are not multiple-choice questions. They are teachback challenges, in Pask's sense. Each one asks you to reproduce, derive, or transfer the understanding — not to recognise it. Click to expand each challenge and the kind of answer it is looking for.
Use the rainforest. Do not use the word "feedback" — you have not earned it yet at this stage of the mesh. In your own words, explain why the question "what causes the rain in a jungle?" does not have a single answer.
Imagine a circular loop that is not regulating. Describe what would happen to it. Then explain what the rainforest's loop has that this hypothetical loop lacks.
Use Ashby's Law to explain why a tree plantation of a single species fails to make its own weather even when planted at jungle density. What variety does the plantation lack, and what disturbances can it therefore not absorb?
No tree contains a cloud. No leaf intends one. Yet the cloud is real and falls as rain on the very trees that made it. Explain what kind of thing the cloud is. What level of description does it live at, and why can't it be reduced to the level below?
A single tree is alive in an obvious biological sense. Make the case that the forest, as a whole, exhibits the cybernetic structure of life — even though it is not a single organism. What does it do that qualifies, and what would have to break for it to stop being alive in this sense?
Choose a system that is not a rainforest — a city neighbourhood, a classroom, a heart, a small economy, a family at dinner. Walk through the five concepts (A–E) and identify, for each one, where in that system the principle is doing its work. If a principle is missing, say so — and ask whether the system can be self-regulating without it.
Did you read these sections in order (serialist), or did you scan ahead and let the diagram orient you first (holist)? When you got stuck on a concept, did you go back to its prerequisites, or did you reach forward to the integration to see what it was for? There is no right answer — but noticing your own strategy is part of learning to learn.
This module ends here, but the entailment continues. Three natural next steps:
Toward Ground Regulation. The rainforest loop sits inside something larger: the medium of soil, atmosphere, and biological tissue through which all the regulation passes. Reading the forest as a Pischinger/Heine ground-regulation system shifts attention from the trees-as-components to the canopy-as-medium — and exposes what's called the Virchow Error: the habit of treating components as primary while ignoring the medium that lets them communicate. The trees do not regulate the climate through each other directly; they regulate it through the air, soil, and water they share. That is a different reading of the same loop.
Toward Rosen's anticipatory systems. The rainforest's regulation is reactive — it corrects a disturbance after it begins. A more sophisticated cybernetic question is whether any system in the forest anticipates its environment via an internal model of the future. Some plants do this (circadian stomatal opening that precedes the dawn). Most do not. The distinction matters when we ask why human institutions, which can model the future, often regulate worse than forests, which cannot.
Toward design. If you wanted to build an organisation, an economy, or a school that regulated itself the way a rainforest does, what would you have to give it? Not a manager. A loop, with enough variety, that returns its outputs to its own inputs. Most institutions fail this test because they have learned to export their consequences. The forest cannot export its rain. That is precisely why it works.