Every institution in sustainable farming defines the problem in terms it can solve. Payment schemes say farmers need money. Coordination bodies say they need connections. Data platforms say they need information about their own soil. None of these assumptions survive contact with the actual constraint.
British farming is not failing to transition because of insufficient funding, poor coordination, or lack of data. It is failing because the industrial model replaced a specific kind of knowledge, observational, place-specific, built through repetition on particular ground, that sustainable farming requires. Most farmers can see the problem. Many understand the principles. But the practical skill of acting on them, the kind that accumulates over years of reading soil and managing complexity, was made economically irrelevant, and a generation later it had largely disappeared. It is hard and no longer profitable to rebuild it.
The lock-in
A farmer with 500 acres of drained arable, £400,000 of machinery, and a bank loan underwritten against industrial yields was not born into that position. They were trained into it. The machinery only pays for itself on large, bare fields. The loan assumes industrial output. The agronomist is supplied by the input company. The processing contract specifies industrial product. Every component of the system reinforces every other component. This is not a market that can be corrected with a subsidy top-up. It is a lock-in.
The deeper problem is that the skills sustainable farming requires, reading soil, managing complex rotations, working with pest ecology rather than against it, gradually fell out of use. Thirty years of industrial practice removed the conditions under which they are learned. Nobody set out to delete this knowledge. It simply stopped being rewarded, and a generation later it was gone.
That turned out to matter more than anyone predicted. A farmer who wants to transition is not simply weighing up a financial risk. They are confronting the fact that the industrial model defined farming so narrowly that the skills a different approach requires, reading soil biology, managing complex rotations, working with ecological processes that take years to understand, were never passed on to them. The system trained them for one thing and is now asking them to do another, without ever having provided the means to learn it.
Nobody wants to admit that.
The wrong level
Every major intervention currently on offer treats the problem as financial, informational, or organisational. But the farmer is not short of information. Most are not short of understanding either, they can see their soil compacting, they know what continuous wheat does to disease pressure, they understand in principle that a rotation would help. What they are short of is the accumulated practical knowledge that would let them act on that understanding, and the infrastructure around them is designed for a different kind of action.
A soil report tells you your organic matter is low. It does not tell you what a healthy soil surface looks like when you push a spade in, or how to read the weeds that indicate compaction, or when a cover crop is ready to be grazed down rather than terminated. That knowledge lives in observation and repetition. It takes years to acquire. It cannot be downloaded.
Why nobody says this
A food company funds a think tank to diagnose the problem. The diagnosis is that coordination is lacking. The recommended solution is a coordination body. The food company is already partnered with one. The government department cites the report as evidence. Each actor’s output becomes another actor’s input.
Nobody in this arrangement is lying. But every institution at the table has a structural reason to define the problem in terms that require their continued involvement. The result is a consensus that has not been tested against reality. It has been agreed upon, which is different.
The person inside one of these institutions who suspects the frame is wrong faces a practical difficulty. Their career depends on the frame holding. They are not going to say what they see, yet they are perhaps the only one expert enough to see it.
What actually works
If the constraint is replaced knowledge, then the only thing that matters is rebuilding it. Not through a training programme. Not through a “train the trainer” model. Through farmers getting stuck into the land, learning what works on their soil, and sharing what they find with other farmers on similar ground.
This already happens. I have spoken to arable farmers who started visiting each other’s fields to compare direct-drilling results. No funding body. No facilitator. Just a WhatsApp group to arrange meetings. They walked the tramlines, dug holes, looked at root structures, and argued about cover crop mixes. Within a few seasons, several had significantly cut their herbicide use, not because someone told them to, but because they could see what was working on similar ground nearby. The knowledge transferred because both people were standing in the same mud.
But look at what happens when institutions try to scale this. A field day fills up with consultants selling products. A conference becomes a marketplace. A knowledge-sharing network becomes a funded programme with outputs, metrics, and a coordinator who has never farmed. The thing that made it work, that it was direct and unmediated, is the first thing removed.
This is not a detail. It is the whole problem in miniature. A coordination body will never recommend that farmers simply talk to each other without a facilitator, because that recommendation eliminates the coordination body. An advisory service will never conclude that the farmer already knows what to do, because that conclusion eliminates the advisory service.
Nobody gets paid for farmers learning from each other. So nobody proposes it.
Those farmers did not wait for a proposal. They were not short of intelligence or ambition. They were short of a system that valued what they were trying to learn. So they built their own, on their own terms, in their own fields.
The question is why that has to happen despite the institutions, rather than because of them.
Regenerative agriculture when you strip it down to its essence is farmers learning to listen to the land. Support services run the risk of diluting that process if they are not careful.
Further reading
↳ The pre-inclosure farming this article hints at is reconstructed in more detail in The Oldest Idea.
↳ For the question of whether this lost knowledge can be rebuilt in time, see When Your Best Solution Is Too Slow.
↳ Why the institutions that could have protected this knowledge structurally couldn’t is the subject of The Lost Architecture of Prohibition.
Garden notes
- Separated knowledge — what this essay shows in a field-level context
- Overoptimisation — the industrial model as a thirty-year overoptimisation
- Proxy capture — institutions defining the problem in terms they can solve
- Informed intervention — what the farmers’ WhatsApp group actually was
- When Your Best Solution Is Too Slow — companion essay: the temporal problem