A map of how grown arrangements became designed ones — the companion to Cloth, The Long Way Home, and A Good Man.
This is a description, not an indictment. It maps a direction a society has travelled across the last century or more — from arrangements that grew to arrangements that were designed — and it tries to do so without naming a villain, asserting a conspiracy, or validating any cause's account of itself. It names mechanisms and incentives, not culprits, because the thing it describes does not require bad actors to operate. A tide can be described without accusing the moon.
It is the cold companion to the warm work. The three stories console; they do not diagnose. This explains the why the stories deliberately leave out, so that the stories can do their gentler work unburdened by argument. Nothing here belongs inside the stories. It is the skeleton beneath them, set down separately so the flesh can stay warm.
The single discipline running through all of it: take things in proportion and never in isolation, and remove the framing — including one's own. An observation that cannot survive being applied to itself is not an observation; it is a prior wearing the costume of one. Wherever this document reaches a claim, it tries to hold it to the same test it sets for everything else.
There are two ways a human arrangement can come to exist. It can grow — accrete over generations, one correction at a time, shaped by people who lived among the results of what they did — or it can be designed, conceived in a mind, to a plan, and installed. Call the first evolved, or Type 1; the second designed, or Type 2. The distinction is not old-versus-new and not left-versus-right; both contain their own monopolies, their own untouchable theories. It is a distinction of origin, and origin turns out to govern almost everything else.
One question separates them in practice: does the consequence of a decision return to the one who made it? Where decision and consequence stay joined — where the decider lives among the results — the arrangement can feel its own errors and correct them. Where they are severed — where the result falls on people the decider never meets — the arrangement cannot feel what it is doing, and the error runs. This is not a moral claim. It is mechanical. Accountability, in this sense, is not a virtue; it is the means by which a system perceives itself.
This is the foundational difference, and it explains almost everything that follows.
The grown arrangement arose through proximity. The people making adjustments lived where the adjustments landed. The neighbour who brought a meal before it was asked for was not implementing a policy about community resilience — she was a neighbour who could see the door hadn't opened. The correction was local, immediate, and felt by the one making it. The knowledge this generated was embodied — it lived in people, in habits, in the unspoken understanding of a street. It could not be written down, because it was the knowing: the knowledge that a particular door not opening for a day means something cannot be separated from knowing the person behind the door. This knowledge is dense, specific, and untransferable. It cannot travel. It can only be lived.
The designed arrangement, and the critique that so often precedes and licenses it, arose through distance. The institution built specifically for thinking at distance from consequences is the university. That distance is not a flaw — it is the function. Academia exists to allow people to think about things they do not personally experience, to find patterns a local cannot see, to generate knowledge that is abstract, recordable, and transmissible. This is genuinely useful, and the assessments it produces are often correct. But the knowledge it generates is of a different kind from the embodied knowledge of proximity — and a different kind of knowledge produces a different kind of solution: one that can be written down, taught, replicated, and scaled. The designed thing is what you get when the people solving a problem are not the people who must live in the solution.
Once an assessment is made at the level of academic knowledge, the mechanism that carries it through a whole society is already in place and requires no coordination to operate.
Consider how malaria is addressed by modern science. The anopheles mosquito is not evil. Malaria is not evil. The parasite evolved to use the mosquito; the mosquito evolved to feed and breed; both are doing precisely what they grew to do in their own ecosystem. The harm to humans is real, and the assessment of it as a problem requiring intervention is legitimate. But the mechanism of that intervention — releasing modified mosquitoes into the wild population, which interact naturally with the existing population, wave after wave, across decades, until the population is altered at scale — requires no conspiracy. The released mosquito does not know it carries a modification. It simply does what mosquitoes do, and the change happens at the population level over time.
Academic frameworks spread the same way, and for the same reasons. Once an assessment is made — that a component of the existing arrangement constitutes a harm requiring a designed solution — it does not need to reach everyone directly. It needs only to reach the training institutions: the universities and professional schools that produce the people who will run everything else. From there it propagates through careers, into every institution that takes its lead from trained expertise. Governments follow the research. Corporations align with the regulatory frameworks the research shapes. Social workers are trained on curricula built from theory. Surgeons operate within clinical guidelines derived from study. Engineers design to standards set by bodies whose assumptions came from somewhere. Think tanks translate the framework into policy language. None of these people are activists. They are professionals doing what professionals do — following the best available framework for their field. The framework got there first, through training, through accreditation, through what counts as evidence in each professional domain. That is the propagation. And it is the ordinary operation of how knowledge moves through a society that has decided to be evidence-based.
The mechanism is entirely neutral. It would carry a framework that restored the grown arrangement just as efficiently as one that dismantled it, if that framework had reached the training institutions first. The correctness of the original assessment — was the identified harm genuinely universal, or only incidental in some instances? — is a separate question, answered only by the variance test in Part III. The transmission mechanism adjudicates nothing. It only carries.
This is where the proximity-distance distinction and the transmission mechanism join to produce the cruelest feature of the whole transition.
The designed arrangement always arrives fully articulate. Being designed, and having been generated by the institution built for recordable, transmissible knowledge, it comes with its own vocabulary, its own justifications, its own metrics of success. It can argue for itself the moment it appears, in the language of evidence and expertise that every professional institution is trained to respect.
The grown arrangement is mute. It was never designed, so it never had to argue for itself; it simply worked, unnamed, the way a thing you have always had goes undescribed. The knowledge that sustained it — the embodied, proximity-built, untransferable knowledge of people living among results — cannot be converted into the language the transmission mechanism carries, because that language requires abstraction and the knowledge resists it.
So every contest between the two is between a fluent newcomer and a silent incumbent — and fluency wins by default, not because it is right but because it can speak. The grown thing has no defenders not because it has no value, but because it has no vocabulary — and a thing without a vocabulary cannot enter the transmission mechanism that shapes professional and institutional life. You cannot defend what you cannot describe, and you cannot describe what was never designed to be described.
There is a recurring interpretive move by which a real harm inside an arrangement becomes a warrant to dismantle the whole arrangement rather than reform it. It has a consistent five-step shape, and the same shape appears whether the domain is family, law, education, sex, nation, or work. It is set out here as structure, not accusation: each step can be taken by sincere people who believe they are doing good, which is precisely why the move spreads without anyone coordinating it.
The move begins in truth, not in a lie, which is why it has purchase. Every institution being critiqued did contain genuine injustice — the family contained tyranny, the old settlement contained exclusion, the nation contained cruelty to outsiders. Denying the harm fails, and deserves to, because the harm is real. This step must be honoured.
This is the hinge on which everything turns. Any harm inside an institution can be read two ways: as a corruption of the thing (a sound purpose, failed in this instance) or as an expression of the thing (the harm is what it is secretly for). Reform follows from the first reading; deconstruction follows from the second. The entire pivot from "fix it" to "abolish it" happens at this single interpretive choice — and the choice is almost never argued. It is assumed, smuggled in as if it were observation rather than interpretation. Once the harm is read as definitional, reform is recoded as complicity: to repair the thing is to preserve its essence, which has just been defined as the harm.
If the harm is the essence, every instance carries it, including the instances that show no harm at all. The contented case is not a counterexample; it is the harm successfully concealed — and the more contented it looks, the more effective the concealment is judged to be, so that visible happiness becomes evidence for the indictment rather than against it. This is the step that makes the framework unfalsifiable. Once the absence of visible harm is read as proof the harm is well-hidden, no observation can count against the thesis. The point here is not that this is dishonest — it is that it is structurally self-sealing, and operates whether or not anyone intends it to.
Since the institution is now essentially corrupt and its participants are either its victims or its unwitting agents — including those who claim to be content, who are held to have false consciousness — the correction cannot come from within. The people inside are compromised by definition. The remedy must come from outside and above: an expert, a policy, a designed replacement administered by those who have seen through the thing. This is the step where the loop severs. Authority moves from the people who bear the consequences to people who do not — which is to say, from the proximity-built arrangement to the distance-built one — and so the replacement is reliably Type 2, not by accident but by the logic of the move, which requires that the locals be distrusted and the non-local be installed. From here, the transmission mechanism carries it: the designed replacement enters the professional training infrastructure and propagates through every institution that defers to assessed expertise.
The exit is now closed. Because the arrangement has been defined as oppression, dismantling it is framed as freeing the oppressed; and anyone who defends it — including those who lived in it contentedly — is reframed as either an oppressor or a captive defending their own cage. There is no longer any standpoint from which the thing can be defended in good faith. The testimony of the content is inadmissible by construction, because the framework has already classified their contentment as the symptom.
The five steps require no coordination and no malice. The move is available in every domain; it is emotionally and morally rewarding to perform — the one who makes it is the one who sees the hidden harm, who stands with the victim; and it reliably produces the Type 2 outcome that enters the transmission mechanism already in place. A move that is available everywhere, rewarding to perform, and carried automatically by the existing professional infrastructure will be selected for, repeated, taught, and rewarded across institutions, until the pattern looks coordinated though it never was. A conspiracy without conspirators is simply a direction that pays, carried by a mechanism that was already running.
The move can be assessed without validating any claim and without ruling on whether any given act of dismantling was right or wrong. The instrument is a single question, asked identically in every domain, weighing only probability of reading, never rightness of action.
Given what can be observed, is it more probable that the identified harm was the institution's essence, or a failure within it?
The strongest evidence available is the same in every domain: does the harm appear wherever the institution appears, or does it vary? An essential harm should be near-universal across all instances of the thing, because an essence does not switch off. A harm that varies widely — present in some instances, absent in others, tracking the character of the people rather than the structure — is more probably incidental, a medium rather than a cause. This test is observational. It validates nothing. It only weighs frequency.
Applied across domains, the variance test makes "incidental, reformable harm" the more probable reading in most, and "mixed" in the hardest cases. The tyranny in families varied household to household; the exclusion in the old legal settlement moved its boundary repeatedly while the institution persisted; the cruelty of nations spans the whole range while they remain nations; the exploitation in work tracks distance and arrangement, not enterprise as such. In each, the harm varies — which makes essence the less likely reading. One or two domains (see Part V on the shared moral frame) read as genuinely mixed, and honesty requires marking them rather than smoothing them into the rest.
The observation that survives all of this is narrow and unanswerable: the move applied a single reading — "essential, therefore dismantle" — uniformly across domains, regardless of what the variance in each domain supported. A method that returns the same answer everywhere is not reading the evidence; it is applying a prior. That is the tell — not that harms were identified (they were real), not that deconstruction was sometimes chosen (sometimes that is the probable-correct call), but the uniformity. And a prior is not a crime. It is simply not evidence. Naming it accuses no one; it only distinguishes a reading of the world from a decision made before the world was consulted.
This method must survive its own test, or it becomes the thing it describes. It does not claim deconstruction is always wrong — that would be its own uniform prior, the mirror error. It allows that in any given domain the essential reading might be the more probable one. It reports, per domain, which way the variance points, and offers its own falsification: where a harm genuinely is universal across all instances of a thing, the essential reading is the better one, and dismantling may be right. The honest assessor and the deconstructing one diverge at exactly step two — one asks the question and can imagine either answer; the other has pre-decided that the answer is always "definitional," because that is the answer that licenses the move already intended.
The discipline that makes the whole map observation rather than counter-accusation is proportion: take each thing at its true scale, against all comparable things, and remove the framing that isolates one instance to carry a charge the whole would dissolve.
Taken in proportion rather than isolation, slavery is not a thing one civilisation did to another at one moment. It is a near-universal of human history, practised on every continent, by every kind of people, against every kind of people — and it did not end, it changed form: from chattel to bonded to trafficked to the wage paid at the margin where the wage is not truly a choice. The observation that survives all framing: a harm universal across time, place, and group cannot be the property of one group as accusation. To attribute it to one is not to describe it but to select it, and the selection is the framing. Removing the framing does not deny the harm; it restores the harm to its actual scale, which is everyone's. It indicts no group precisely because it indicts the species.
As the designed replaces the evolved, it cannot reproduce the organic function — the neighbour who knew, the judgement that lived in a person — so it builds a procedural proxy: a form, a policy, a reporting line, a metric that performs the shape of the thing without its substance. The safeguarding referral replaces the watchful street; the statement replaces the actual mixing; the care plan replaces the daughter at the bedside. And because the proxy measures itself by its own procedure rather than by the outcome, it can report success while the function it replaced quietly dies. This is the tell of the transition, visible everywhere once named, and it is a mechanism, not an accusation.
The replacement frequently reproduces, in proceduralised form, the harm it named: the anti-hierarchy move builds new hierarchies of purity and authorisation; the anti-exclusion frame excludes the unconverted; the liberation forecloses the choice it claimed to widen. This is real and recurring — and it must be held at often, not always, and shown case by case, because "everything it touches becomes its opposite" is itself the uniform prior the method exists to refuse. Shown in proportion, it is devastating. Asserted as law, it overplays, and an overplayed true thing invites its own dismissal.
A restraint is hated most by the thing it restrains, and the loudest voice against a constraint is often the appetite the constraint was holding. This is observable — and guarded by the same word, often. The protest against a restraint tells you nothing, by itself, about whether the restraint was just; sometimes the one railing is genuinely caged by an unjust limit. You tell from whether removing the restraint returns more harm or less. The moment this becomes always, it would let a real cage be defended by calling its prisoner's protest "the restrained railing against restraint" — which is the cruelty the whole project refuses.
A cause's founding grievance is usually real — the seed of truth. But the people who carry the cause forward are frequently not the people who suffered the harm; the carrier is often one for whom the grievance is useful rather than lived, and a grievance useful to its carrier will be played, amplified, and kept alive past its resolution, because the carrier's standing depends on its persistence. Stated as a structural incentive, this is unanswerable: causes are carried less by those who suffered the harm than by those whose standing requires the harm to continue. Stated as an accusation it smuggles a culprit back in and invites the fight the method avoids. Name the incentive, never the culprit. The incentive can be weighed; the culprit can only be denied.
The transition from evolved to designed can be mapped domain by domain, against both the structural change (what replaced what) and the lived change (how it felt to be inside it). They are ordered here roughly by how far the transition has gone — which, as Part VI explains, correlates inversely with how much record the evolved version left.
Structural: from the multigenerational household, where the old were kept and the keeping was assumed, to the care home, the funded package, care relocated from kin to institution. Lived: from dying in a known room among one's own to dying among paid strangers, a journey from the family's centre to a managed elsewhere. This is the furthest-gone domain, and the one where the few remaining witnesses are the oldest — there are still people who remember a grandmother dying at home, but their grandchildren cannot picture it as a real option, only as a logistics problem. The vocabulary is already mostly gone: "care" survives as a service-word, and has nearly lost its older sense as a thing a family simply is.
Structural: from the neighbourhood that knew its own and acted before crisis — the meal left on the step, the quiet word on a Wednesday — to the safeguarding threshold, the referral, the food bank: help proceduralised, triggered by crisis rather than proximity, and administered by people who arrive after the fact and do not stay. Lived: from being known to being a case. Most people under a certain age have never experienced the thing, and so cannot miss it specifically — only feel the ambient atomisation as "people don't know their neighbours anymore," the ache without the name. The word "community" survives, hollowed into an administrative or demographic category — the word outliving the thing and repurposed, which is itself a symptom of the transition.
Structural: from the assumed early sequence — marry, house, children, the long unpaid work — made affordable on one income, to the deferred sequence of education, then career, then family last, on two incomes, against the biological clock. Not banned. Priced out, and the pricing-out called progress. Lived: from family-as-foundation to family-as-capstone-if-time-permits. This is the domain where the ratchet bites most visibly right now, because women are alive in both conditions — the grandmother who married young and the granddaughter who wanted the same and cannot now reach it, and half-cannot say so, because the available language codes the wanting as backwardness. The witness and the bereaved are sometimes the same family, one generation apart.
Structural: from the trade, the family firm, the foreman who knew his men, the shop where you learned what no school teaches, to the scheme paid by results, the uniform that hides the wearer, the metric that skips the person it cannot measure. Lived: from work as where a person spent one life among known others to work as a function performed for a distant optimiser. Still partly sayable, because the older form is within living memory and survives at the edges as a sentimental category — but the proceduralisation is near-total in the large institutions.
Structural: from order kept close — the neighbour who intervened, the shame that governed, the consequence returned to the doer through people who knew him — to order kept distant: the formal complaint, the policy, the process that arrives long after and resolves nothing locally. Lived: from being answerable to your own street to being unanswerable to anyone until a threshold is crossed. This one is going fast and has reached an advanced stage of the ratchet: the informal mechanism is now often actively distrusted — the neighbour who intervenes is "taking the law into his own hands" — so the lost function is not merely gone but recoded as itself a wrong, remembered only as a danger.
Structural: from an inherited communal frame — the shared calendar, the common moral vocabulary, the institution as a node of mutual aid as much as of belief — to the privatised, individually-assembled frame. Lived: from a moral language held in common to one each person builds alone. This domain is flagged rather than asserted, because here the variance test gives a less clean reading: the evolved version genuinely contained more essential constraint — it bound partly by excluding dissent in a way the family or the trade did not. Honesty requires marking the mixed case rather than smoothing it into the others. The discipline shows itself in the awkward domain, not the easy ones.
The order of the transition correlates with how little record the evolved version left. Care, the street, and neighbourly order went furthest and fastest precisely because they were the least articulated — the most purely embodied, proximity-built knowledge — and so had no defenders, because they had no vocabulary that could enter the transmission mechanism. The more a domain had some formal existence — work with its contracts, faith with its texts — the slower and more partial its transition, because formality leaves a record, and a record can travel. The quiet things did not merely go unrecorded. Going unrecorded is why they went first — they could not enter the only channel through which defence was possible.
This is the cruelest feature of the whole transition, and the one that turns the map from a description into a diagnosis. It is a one-way epistemic trap.
The evolved arrangement is invisible while it functions. It leaves no record, because it is never designed, never funded, never named — it is simply how things are done. It becomes visible only as it fails: the absence is felt where the presence was never noticed. You do not notice the watching until the rubbish appears on the verge; you do not notice the kept household until the home forty minutes away.
The longer the thing is gone, the fewer people remain who can name what was lost — because naming it requires having lived inside it, and the witnesses age out. The loss therefore accelerates away from articulation. Each generation is less able to express the thing than the one before — not because it mattered less, but because the lived memory and the vocabulary thin together. The evolved thing left no description behind and no record in the transmission mechanism, so when the last witness goes, the thing becomes unsayable, and an unsayable loss cannot be argued for, mourned properly, or chosen again.
The end state of the ratchet is a generation that feels the cumulative absence across all the domains at once — uncared-for in age, unknown on its street, unmoored in the sequence of its life, unanswerable in its conduct, unmet in its work — and that has, for the sum of it, almost no language. It reaches instead for the vocabulary of symptoms: loneliness, burnout, anxiety, disconnection. These are the words that remain, because the words for the cause are gone. A hunger that cannot name its food.
This is the asymmetry of Part I, run to its conclusion. The designed replacement always arrives fully spoken, because it was generated by the institution built for transmissible knowledge. The evolved thing it replaces is mute, because its knowledge was proximity-built and untransferable. The contest is never fair — a fluent newcomer against a silent incumbent — and fluency wins by default. You cannot defend a thing you can no longer describe, and you cannot describe a thing whose knowledge lived only in the people who held it and is gone when they go.
The terminal generation — the one with the ache and no name for it — is exactly who the three stories are for. Not to argue them back to anything; argument is for the armoured, and these are the conscripted. The stories do one thing the ratchet had made almost impossible: they hand back, in the form of lives a reader recognises, the words for the thing that is missing. Because a named absence can be mourned, and a mourned thing can sometimes be chosen again — and an unnamed one can only ache. The cold map explains the loss. The warm stories return the language. That division of labour is the whole design: the diagnosis must never enter the parcel, because a person in that much absence needs company and vocabulary, not a lecture on what was done to them.
Everything above is held to one standard, and the standard is the same one the work itself runs on: name directions and incentives, never villains; weigh probabilities, never assert verdicts; say often, never always; and remove the framing, including one's own. The map accuses no group, because the harms it discusses are universal and attributable to no one without selection. It validates no cause, neither the critique's account nor its opposite. It offers its own falsification in every domain — where a harm is genuinely universal, the essential reading is the better one — which is the test the thing it describes never offers itself. It survives being applied to itself.
The transmission mechanism is neutral. It would carry a framework restoring the grown arrangement just as efficiently as one dismantling it, given the same access to the training institutions. The mosquito does not know what it carries. The professional does not know they are the vector. The assessment — whether the identified harm was genuinely essential — is the only question the mechanism leaves open, and the variance test is the only honest tool for answering it.
What the map finds is not that anyone was evil, nor even that any particular act of dismantling was wrong. It finds only this: that a society moved comprehensively from grown arrangements to designed ones, across every domain, by a move that read every harm as essential when the evidence usually supported reform; that the designed replacements simulate the organic functions they cannot reproduce, and often recreate the harms they named; that the loss is hardest to see precisely where it was deepest, because the deepest things were the most proximity-built and the least transmissible; and that we have arrived at a generation feeling the whole of that absence with almost no words for any of it.
The map is the cold account of why. The stories are the warm return of the words. Neither is an accusation. Both are, in the end, an attempt to make a nameless thing nameable again — which is the first condition of mourning it, and the only condition under which it might ever be chosen back.