A position paper on the safety net, the stories it makes people tell, and a different door.
August 2026
Twenty five years ago I was a fresh psychology graduate working in rehabilitation. Part of my job was to sit beside my clients in their meetings, as their advocate, and they all said variations of the same sentence: it's like I have to spend my life proving how sick I am to get any help. I never stopped hearing it. I just kept finding more evidence.
A quarter of a century later, I stand in their shoes. Complex PTSD, and a host of other letters to my name that I never wanted. And the sentence is there waiting for me, word for word. To survive this stretch of recovery I must prove that I am sick, and I must stay provably sick for exactly as long as recovery takes... because the day I look recovered is the day the support that recovery stands on is withdrawn.
Look at the shape of that. It is not a complaint about paperwork. It is an oxymoron with a caseload of millions: get better, but do not dare get better. Twenty five years of welfare reform sit between the advocate's chair and the claimant's chair, and the sentence spoken in both is the same.
This paper is about that sentence. Where it comes from. What it does to the brain and body of a person required to keep saying it. And what it would take, at last, to retire it.
Medicine has an old name for wasting away on adequate provision: failure to thrive. Millions of people are currently recorded, in effect, as failing to thrive. This paper argues they are being failed to thrive... and that the cure begins where the illness began, with changing the words.
Position papers usually bury their proposal behind the evidence. This one leads with it, so the reader knows where every page is going.
Replace the incapacity assessment system, for people long-term out of work through ill health, with a participation income: a secure income floor, paid without any test of brokenness, carrying an invitation... around five hours a week of contribution, defined as generously as life itself. Art counts. Care counts. A tended riverbank counts. Learning counts. The hours are invited and self-declared, never policed, and the income does not depend on them: there is no new assessment hiding in this design, because this paper's whole argument is that the assessment is the pathology.
The arithmetic is less radical than it sounds. At £1,600 a month, the rate of Britain's current basic income pilots, covering the entire incapacity caseload of 3.4 million people would cost about £65 billion a year: roughly what the existing working age health and disability benefit system already spends, £62.7 billion in 2025/26, rising to £74.9 billion by 2028/29 on current forecasts.³² And that is before counting the assessment machinery, the tribunals that overturn two decisions in three, or any part of the £212 billion a year that health-related worklessness costs the country.²
The recommendation is one honest experiment: one town, one thousand people currently inside the incapacity system, a participation income in place of the assessment cycle, and rigorous measurement of health, wellbeing, activity and cost against a matched comparison group.
Everything that follows is the evidence. Part One shows the scale, and the people standing at the door. Part Two shows the mechanism: how the current system's language manufactures the very sickness it pays for. Part Three shows the design, the philosophy, and the answers to the six obvious objections, including the loudest one: no, everyone will not quit their job, and the trials that tested exactly that fear are in the footnotes. The short version fits in a sentence: the safety net asks millions of people to prove they are broken, and people become what they are paid to rehearse... so pay them to rehearse something else.
There is one door.
Whoever you are, whatever has actually happened to you, if you cannot work in Britain today there is one door to the help that keeps you alive, and the sign on it says broken. You do not knock. You fill in a form that asks, page after page, what you cannot do. Then you say it out loud to a stranger. Then, in time, you say it again.
Three people are standing at that door. They are composites, made of true things: every detail is drawn from the statistics and studies cited in this paper. No one in these portraits is a real individual. All of them are real.
The first is a care worker, sixty one years old. Thirty years of lifting other people out of beds and baths have worn out her back, her knees and her sleep. Her pension is six years away and she cannot get there. There is no door marked worn out and nearly there. There is no bridge for the last stretch of a working life spent working. There is one door, and to walk through it she must stop being a woman who cared for three decades and become a list of the things she can no longer lift.
The second is a school leaver, nineteen. He is autistic. He is also precise, honest, pattern-brilliant and funny, though no form has ever asked about any of that. He has waited two years for the assessment that decides what support exists for him. Every page of every application wants his deficits, itemised. He is learning, at nineteen, to introduce himself by what he cannot do. The lesson is landing.
The third is a woman in her forties. For twenty years someone made her rehearse her own worthlessness until she could recite it fluently. The abuse left no scan, no fracture, no witness. She escaped. And to eat, she must now describe herself as broken, to strangers, on schedule, in writing. She got out of a house where she had to perform her smallness to stay safe, and arrived at a system that requires the same performance, notarised.
Three different lives. One shared instruction: prove you are broken, and keep proving it.
This is not a corner of the system. This is the system.
2.8 million people in Britain are out of work through long-term sickness: 800,000 more than in 2019, with another 600,000 projected by 2030 if nothing changes.¹ Hold the denominator alongside it: the UK has about 42.5 million people of working age, so this is 7 per cent of them, a record... one working-age person in every fourteen. And more than half of the 2.8 million, around 1.5 million people, are aged 50 to 64: the last stretch of a working life, with the pension out of reach. The cost of health-related economic inactivity now stands at £212 billion a year, roughly seven per cent of GDP.²
Four million working-age people in England and Wales claim disability or incapacity benefits: one in ten, up from one in thirteen in 2019.³ Of those, 1.3 million claim primarily for mental health or behavioural conditions. That is 44 per cent of all claimants, and more than half of the entire rise since the pandemic.³
Around 30 per cent of autistic adults are in work. For non-disabled people the figure is 82 per cent.⁴
And for those near the end of a working life: of the 330,000 rise in incapacity claims between 2008-09 and 2023-24, roughly 250,000 tracks the raising of the female State Pension age.⁵ One in ten people in their early sixties claims an incapacity benefit. People who leave work sick in their fifties mostly never work again.⁶ When retirement moves out of reach, sickness becomes the only remaining safety net, because it is the only door there is.
The debate about these numbers is loud and largely about money: who deserves it, who is gaming it, how to spend less of it. This paper is about something quieter that sits underneath all of it. It is about what the door does to the people who walk through it.
Follow one person through the system and count the number of times they are required to produce an account of their own incapacity.
The initial claim form, fifty-plus pages, structured entirely around loss of function. The assessment, face to face or by phone, in which the account is repeated to a stranger with a checklist. The reassessment, months or years later, in which it is repeated again. The mandatory reconsideration if refused. The tribunal after that. Each stage asks for the same document: your worst day, in your own words. And each stage carries the same unspoken warning: describe yourself too capably and the money that keeps you alive is withdrawn.
Psychology has a clear view of what repetition like this does. We become the stories we practise. A narrative rehearsed often enough stops being a report and becomes an identity; this is not a weakness of damaged minds, it is how every mind works. The system takes people at their lowest and pays them, in the only currency that matters when you cannot work, to rehearse their lowest... indefinitely.
This is not speculation about subtle influence. Researchers studying the Work Capability Assessment found that the system's narrowly medical framing forces claimants and even their advisers to reshape complex lives into the language and constructs of illness, because that is the only language the system can hear.⁷ Difficulty becomes diagnosis. Circumstance becomes symptom. A person becomes a condition, in paperwork first and then, slowly, in person.
Here is the detail that gives the game away. When refused claimants take their case to tribunal, where the criteria are identical but a human panel simply lets them talk, around two thirds of decisions are overturned in the claimant's favour.⁸ Same person. Same rules. Same facts. Different language ritual, opposite result. The form is not measuring incapacity. The form is manufacturing incapacity narratives, and then the tribunal, given a conversation instead of a script, discovers the person underneath.
Now look harder at the largest group behind that door: the 1.3 million whose primary condition is mental health.
Ask where much of that injury comes from, and the answer is one the system is structurally unable to process: abuse. Psychological abuse from partners. From families. From institutions, workplaces, churches. Decades of research on adverse childhood experiences show early psychological harm predicting illness across a whole adult life.⁹ The wound is real, common and disabling. It is also, by its nature, the injury that produces no evidence. No scan shows it. No X-ray finds it. The perpetrators do not write letters of confirmation.
The system, though, runs on evidence. Its first language is proof. So the survivor faces a double bind that would be comic if it were not lethal: prove the unprovable, or starve. And there is exactly one form of proof available to someone whose injury is invisible: symptoms, performed, on demand. Be visibly unwell in the assessment. Describe the panic convincingly. Cry in front of the right stranger. The person becomes the evidence, because nothing else is admissible.
The wound itself predates all the paperwork. A University of Glasgow study found GP consultations for mental health rising steadily for three full years before a person first claims incapacity benefit.¹⁰ The injury arrives long before the label. The system just cannot see it until it has been converted, by the person who carries it, into fluent sickness.
Coercive control has a grammar, and it is the same grammar at every scale.
Make them doubt their own account of events. Make them perform distress before they are believed. Keep them financially dependent. Punish every attempt at autonomy. Whether the controller is a partner, a family, a religious community or an institution, the pattern is the pattern; it is well documented and, since 2015, in the domestic context, it is a crime.
Now read the benefits system through that lens, slowly.
Disbelief is the default setting; the entire apparatus exists to filter out the insufficiently broken. Suffering must be performed, on demand, to be credited. Support is conditional on continued incapacity, so that a claimant who tries a few hours of work risks triggering reassessment and losing everything... which any rational person learns to fear, and any traumatised person learns to dread. The state does not intend to be an abuser. Intent is not the point. The point is that a woman who escapes a controlling house and turns to the state for refuge meets the same grammar again, this time holding the money.
The research record here is stark. The Welfare Conditionality project, five years of fieldwork across the UK, concluded that sanctions and conditionality were largely ineffective at moving people into work and routinely produced destitution, distress and worse health.¹¹ And a landmark study of the reassessment programme, titled with the first words of medical ethics, First, do no harm, found the process independently associated with measurable increases in mental ill-health across whole populations, including additional suicides.¹²
Set that beside what every trauma-informed clinical framework on earth says recovery requires: safety first, stabilisation first, identity rebuilt in the present tense, and no forced reopening of traumatic memory without a trained companion and consent. The assessment regime requires the precise opposite: repeated re-narration of the worst material, to strangers, under threat of destitution. For its single largest client group, the safety net administers the reverse of treatment... as the condition of entry.
A note, because this page may land close to home for some readers: Samaritans are free, any hour, on 116 123.
None of this is new. The one door has an ancestor, and the label has a lineage.
The ancestor is the workhouse. The Poor Law of 1834 built British welfare on the principle of less eligibility: relief must be grimmer than the meanest independent living, so that only the truly desperate would submit to it. The workhouse test did not assess need. It priced need, in dignity. Prove your destitution by walking through the door and surrendering everything at the step. The modern assessment is that test's direct descendant: the currency has changed from labour to narrative, and the stone building has become a call centre, but the transaction is the same. Prove your brokenness, and the proving is the price.
The label's lineage is clinical, and it is worse. In 1945 the psychoanalyst René Spitz studied infants in foundling homes who were fed, warmed and kept clean, and who wasted away regardless, some to death, because provision without relationship is not enough for a human to live on.³⁸ Medicine gave their condition a name: failure to thrive. Now notice where that name was written. On the child's chart. The institution failed; the infant got the diagnosis. Eighty years later the same clerical move runs at national scale: a system withholds belonging, purpose and safety, watches people wither on adequate provision, and records the withering under the person's name. The caseload is not failing to thrive. It is being failed to thrive, and the difference between those two sentences is the subject of this paper.
There is a final turn of the trap, and it is the cruellest.
The philosopher Miranda Fricker gave it a name: testimonial injustice, the process by which prejudice deflates a speaker's credibility until their word simply stops counting.¹³ Few labels deflate credibility as efficiently as a psychiatric one. Once anxiety, depression, personality disorder or PTSD is attached to a person's file, everything they say arrives pre-discounted. Their account of events becomes a symptom of their condition.
Watch what this does to the survivor. The label she was required to accept in order to eat is the same label that now ensures she is not believed... about anything, including the abuse itself. Her testimony becomes evidence of her illness rather than evidence against her abuser. And the asymmetry is not accidental: the party with resources can defend their account with solicitors, appeals and reputation management, while the labelled party's account arrives pre-discredited, a pattern familiar to any family court practitioner who has watched a mental health history deployed as a weapon in proceedings.
You lose your income to the injury. Then you lose your voice to the label the injury forced you to wear. That is the bottom of the pit.
None of this stays metaphorical. Chronic coercion and complex trauma leave measurable neurological change: in the stress response system, in the hippocampus, in the amygdala, in the prefrontal circuitry that regulates both.¹⁴ The science is solid and the implication is plain: prolonged psychological abuse is a physical injury with no visible wound. The rehearsals the system demands are rehearsals on injured tissue.
And then comes the quietest injustice of the lot, so normalised that it is hard to see. Someone else inflicts the injury. The labour of repair lands entirely on the injured: the waiting lists, the therapy costs, the self-help homework, the years. The perpetrator's life continues intact. The victim gets a workbook. And when recovery is slow, the system's framing is ready: failure to engage. Failure to progress. Not-recovering is recorded as the patient's personal shortcoming, on forms that will be read at her next assessment.
So name the pit properly, because it is not what the newspapers say it is. It is not that people will not climb. It is that every ladder on offer requires speaking the language that keeps you at the bottom. To eat, say you are broken. To be believed, perform it. To stay safe, stay incapable. The pit is made of words.
If Part Two is true, its mirror should be true as well: if deficit language measurably sickens, then safety, kindness and the other language should measurably heal. They do, and this section is the evidence for the ladder.
Start with the folk version, because it is the one everybody knows. The Japanese researcher Masaru Emoto made it famous: two jars of rice, one spoken to kindly, one cursed at, and the kind jar keeps. As science it does not replicate, and this paper will not lean on it. But pause on why millions of people found it instantly believable: everyone who has lived in a body already knows the finding. The replicated literature says their bones are right.
Expectation alone changes people. In 1968 Rosenthal and Jacobson told teachers that certain randomly chosen children were about to bloom, and by the end of the year those children's measured IQ had risen: nothing had changed but the words in the teachers' heads.¹⁵ The effect runs the other way too, and the reverse has its own name, the Golem effect: expect less of a person, and less reliably arrives. An assessment regime is an expectation machine, and it only turns one way.
Words change stress chemistry. People who spend a few minutes writing about what they value show measurably lower cortisol when placed under stress afterwards, and brief self-affirmation exercises have shifted school trajectories for years.¹⁶ Sentences about capability are not compliments. They are instructions to a nervous system.
Underneath all language sits the body's accounting. Poverty itself taxes the mind: the same Indian sugarcane farmers, tested before and after harvest, lose roughly thirteen IQ points to pre-harvest money worry, an effect the researchers compared to losing a full night's sleep.¹⁷ Worrying about your next meal is a full time cognitive job. And when the worry is lifted, the biology answers: recipients of unconditional cash transfers in Kenya showed large gains in psychological wellbeing, with early analyses finding lower cortisol.¹⁸ One large trial has run the comparison this paper implies, head to head: thousands of people in rural Kenya randomly given either cash, a five week psychotherapy programme, or both. A year later the cash group were psychologically and economically better off; the therapy alone had moved nothing measurable, and adding it to the cash added nothing more.³⁷ That is one context and one programme, not a verdict on therapy, whose trauma-focused forms are well evidenced. It is a verdict on sequence. Therapy delivered into ongoing insecurity is asked to outrun the rent, and for the poverty-shaped share of mental illness, money is not a support alongside treatment. Money is treatment. Security is not a luxury feeling. It is the chemical precondition for planning a future.
Contribution, finally, gives back more than it takes: across dozens of studies, people who volunteer show lower depression and, in the meta-analyses, lower mortality.¹⁹ Which reframes the five hours at the centre of this paper's proposal. They are not the fee for the income. They are part of the medicine.
None of this is mysterious. A nervous system that is safe, seen and useful stands down from survival mode, and a mind that stands down can think in years instead of hours. The pit is made of words. So is the way a brain climbs.
The pit is also heritable, and it matters to be precise about how, because the tabloid version is wrong.
Researchers went looking for the famous families in which three generations had never worked. They could not find them: a Joseph Rowntree Foundation study hunted for such households in some of the most deprived neighbourhoods in Britain and concluded that the "cultures of worklessness" story is a myth.²⁰ What passes down is not idleness. What passes down is the pit itself: poverty's cognitive tax, adverse childhood experience,⁹ and the expectations of a system that has already met your mother. The biology may follow: children exposed to the Dutch Hunger Winter of 1944 to 1945 still carried measurable epigenetic marks six decades on, and early work suggests trauma may echo in the stress systems of the next generation... young science, held here lightly.²¹ What is not young science is this: a child raised inside the door marked broken learns the family language long before any assessment teaches it. Some people are written off before they have written anything.
And for the pre-rejected there is a second door, the one polite policy least likes to discuss. Criminology named this paper's mechanism decades ago. Labelling theory: calling someone deviant is among the most reliable known ways to produce deviance.²² Strain theory: offer people society's goals without society's means, and some will take the means on offer.²² The bond tradition: what actually holds people to the rules is not fear but attachment, involvement, belief... belonging.²² Then the British data: 42 per cent of adult prisoners were permanently excluded from school.²³ Excluded first. Offending second. In that order.
Seen from inside, crime is not a mystery and not a moral collapse. A gang offers exactly what this paper's proposal offers: income, purpose, protection, membership, a name. It is ubuntu's dark twin, belonging run by the wrong people, recruiting daily from the queue outside the one door. "Why care about a society that does not care about you?" is not nihilism. It is an accurate reading of the contract currently on offer. And if the reading is accurate, the answer is not louder moral instruction. The answer is a better contract, and it has to arrive first.
There is a Southern African word for the thing the current system has mislaid: ubuntu. It is usually translated as I am because we are: personhood as membership, existence as belonging. Set it against the creed implied by the current door, which is you are because you are broken: you exist to the state only insofar as you can prove damage.
Here is a different door.
Replace the incapacity maze, for those long-term out of work through ill health, with a basic income paid on participation rather than on proof of damage: around five hours a week of contribution, defined as generously as life itself. Making a piece of art. Tending a stretch of riverbank. Sitting with someone lonely. Learning something. Teaching something. Caring for someone, which millions already do, unpaid and uncounted.
This is not a new idea, and that is its strength. The economist Sir Tony Atkinson proposed a participation income in 1996 and defended it for the rest of his life: a citizen's income conditional not on unemployment or incapacity but on social contribution in the widest sense.²⁴ Atkinson's case was political arithmetic; he judged that the public would fund contribution where it would not fund unconditional cash. What his proposal always lacked was a psychological mechanism, a reason it would change people and not just budgets.
This paper has spent its whole length building that mechanism. If a rehearsed story becomes an identity, then everything depends on which story the system pays people to rehearse. The current door pays 2.8 million people to rehearse what is wrong with me, indefinitely, and we are surprised the numbers only grow. A participation income pays the same people to rehearse what I made this week... and identity follows the rehearsal, because that is how identity works. From sickness to artists. Not as a slogan: as a mechanism.
Notice, too, what the proof trap does inside this design. It dissolves. A contribution-based system never asks anyone to prove the unprovable. Nobody testifies to their damage to qualify for their dinner. The survivor is never again required to perform her worst day for a stranger with a checklist. She is asked a different question, one the system has never asked her before: what would you like to make?
And the daily sentence changes with the question. The current system's liturgy is I am sick, I am broken, it hurts, repeated to strangers until it hardens into a self. The participation income replaces the liturgy: I make. I create. I take part. Same mouth, same mechanism, opposite direction.
One safeguard belongs in the foundations, because this paper's own logic demands it: the invitation must never harden into a new required identity. A system that pressed people to perform wellness, to declare themselves artists on schedule, would be the fit note again in brighter paint... someone else holding the pen. So, the pen clause. The system stops requiring the broken story and opens the door; the person does the naming, including the right to name nothing. Some will say I make. Some will say I am caring for my mother. And some, the most ill, the dying, will say nothing at all, and the floor arrives anyway, because the deepest form of I am because we are is this: you owe us no story whatsoever. The art path is a door held open, never a name pinned on. There is a name for a principle like that, and it is the quiet engine of this whole proposal: radical kindness. Not the soft kind that looks away, but the structural kind... kindness with architecture, built into the design so deeply that nobody has to hope for a compassionate assessor, because there is nothing to assess.
The nearest experiments point the right way.
Finland ran a national basic income trial across two years. Employment effects were mildly positive; the unambiguous result was wellbeing. Recipients reported markedly higher life satisfaction, autonomy, financial security and confidence in the future than the control group.²⁵ Wales has paid care leavers £1,600 a month and found reduced financial stress and better transitions, with the sharpest harm arriving at the cliff edge back onto Universal Credit.²⁶ An English pilot is running now in Jarrow and East Finchley.²⁷
And the oldest fear about all such schemes, that people given money will do the minimum or nothing, happens to be the best tested fear in this literature, because every trial measures it. It keeps failing. Finland's recipients worked slightly more than the control group, not less.²⁵ In Stockton, California, recipients of a guaranteed income moved into full time work at roughly double the rate of the control group in the first year.³⁶ And Britain already runs on voluntary over-contribution: millions of people volunteer regularly, and around five million more provide unpaid care, with no payment and no invitation at all.³⁶ People do not contribute because systems force them. They contribute when they can afford to. Five hours is a floor, and floors are for standing on: the evidence says that most people, once steady, build.
What no one has yet trialled is the participation version: contribution framed as creation, offered to the long-term sick themselves. That is the missing experiment, and Britain, with 2.8 million people behind the one door, is where it is missing most. The recommendation of this paper is exactly that pilot: one town, one thousand people currently in the incapacity system, a participation income in place of the assessment cycle, and honest measurement of health, wellbeing, activity and cost against a matched comparison group.
Who checks the five hours? Nobody, and this is a design principle, not an oversight. The strongest academic critique of participation income has always been administrative: policing participation recreates the assessment bureaucracy in new clothes, with clipboards at the community garden.²⁸ So participation is invitational and self-declared, celebrated rather than audited... a monthly what I made, shared if the person chooses, never inspected. Belonging is not the kind of thing that survives auditing, and ubuntu is not a compliance framework.
Some people cannot manage five hours. Then the income arrives anyway. The five hours are not a condition on the money; they are what the money makes possible. Anything else quietly rebuilds the Work Capability Assessment through the back door, and this paper's entire argument is that the assessment is the pathology. There is no test. There is an invitation, and invitations can be declined by people having the kind of year in which five hours is a mountain.
It costs too much. Run the numbers before deciding. Working age health and disability benefits already cost £62.7 billion in 2025/26, forecast to reach £74.9 billion by 2028/29 as caseloads climb towards four million.³² A participation income at £1,600 a month, the rate of the current British pilots, paid to the entire incapacity caseload of 3.4 million people, comes to about £65 billion a year: the same order of money, minus the machinery.
And the machinery is not small. Assessment contracts have paid private providers more than £100 million a year, and defending refused claims cost the Department a further £121 million in just two years, £58 million on mandatory reconsiderations and £63 million on tribunal appeals... money spent losing two cases in three.³³ Downstream sit the pharmacy and the therapy waiting list: England dispensed more than 91 million antidepressant prescriptions in 2024, the highest annual total on record³⁴, and some share of all of it is treating the illness the process itself generates.¹² Behind every figure stands the real bill: £212 billion a year in lost health and work.² This is static arithmetic, not a Treasury costing; tax, housing support and earnings interactions need proper modelling, which is precisely what the pilot is for. But the order of magnitude is the point. The status quo is not the cheap option. It only looks cheap because its largest costs are booked to other budgets, and to other people's bodies.
It is workfare. / It is money for nothing. It is neither, and the two accusations cancel each other out rather neatly. There is no coercion in it, which answers the first; there is contribution at the heart of it, which answers the second. That balance point is precisely where Atkinson placed the idea thirty years ago, and it is where a divided public might actually meet.
£1,600 for five hours of art? Everyone will quit their jobs. Take this one seriously, because it is the loudest objection there is, and begin with the fact that dissolves most of it: this is not a universal basic income, and it is not on offer to the working population. It replaces incapacity support for people already long-term out of work through ill health... the 2.8 million, who already receive support, through the door this paper has spent its length describing. Nobody in work can quit into it, because entry runs through boundaries that already exist and do not require performance: months of ordinary GP certification, not an afternoon of rehearsed despair. What this paper abolishes is the adversarial worst-day assessment, not the existence of a gate. Next, disarm the arithmetic the objection smuggles in: the income is not a wage for the hours, any more than the state pension is a wage for gardening. It is a floor, and the hours are an invitation resting on top of it. Then the evidence, because the great quitting has been hunted through every experiment since the negative income tax trials of the 1970s and has never been found: hours dipped modestly there, concentrated among new mothers and students³⁹, Finland's recipients worked slightly more²⁵, and Stockton's moved into full time work at double the control group's rate.³⁶ And notice, finally, which system actually destroys the incentive to work. The current one, where support is all or nothing and a few brave hours of work can trigger reassessment and the loss of everything: a cliff every claimant learns to stand well back from. A floor that stays under a person as they step back into work, tapering through ordinary tax, is not the threat to work. It is the first design in forty years that makes trying safe.
We already tried changing the language. The fit note changed nothing. In 2010 the sick note became the fit note, and GPs gained the option of recording that a person "may be fit for work" with adjustments. The result is on the record: around 94 per cent of fit notes still say "not fit for work", the new option is used roughly 6 per cent of the time, and the government's own review conceded the reform was not achieving what it set out to do.⁴⁰ But look at what was actually tried, because it was not this paper's proposal. The fit note changed the vocabulary and kept the grammar. Someone else still writes down what you are, and money still hangs on the word; all the reform did was flip which word carried the threat, so that people now feared being declared fit before they were ready, and doctors, protecting their patients' incomes, rationally retreated to the safe words. Ninety four per cent is not the failure of the idea that language matters. It is a live demonstration of this paper's mechanism: people produce whatever language the incentives demand. The opposite of forced sick language is not forced fit language. It is the person's own language... I make, I create, I take part... the only sentence in this system that would carry no threat, because nothing is withdrawn when it is said.
One more thing, and almost nobody talks about it. Section 20 of the Equality Act 2010 places a duty on providers of services to make reasonable adjustments for disabled people, and the duty is anticipatory: it is owed to disabled people at large, in advance, before any individual arrives to struggle and to ask.³⁵ British law already says it plainly. Do not wait for the disabled person to prove their difference and request the ramp. Anticipate the difference. Redesign the building first.
Who does this? Almost nobody. The anticipatory duty is among the least discussed and least enforced provisions in British equality law; adjustment in practice is reactive, individual, evidence first... which is to say, the one door again, wearing a legal costume.
But read the duty at system scale and it points somewhere remarkable. The benefits system is a service. Its users are, in very large part, disabled people. An anticipatory adjustment to that service, made in advance, for everyone, so that no individual ever has to prove and plead: that is precisely what a participation income is. The proposal in this paper is not a departure from British law's deepest principle on disability. It is that principle, finally taken at its word.
One more shelf of evidence belongs here, because artists is not a flourish. It is the oldest fact about us.
A recent essay in Aeon puts the deep version of the claim: humans did not invent art; art invented humanity.³⁰ When early people pressed pigment to cave walls they were not expressing a self that already existed. The self arrived through the making. The image came first; the I followed... a creature reaching for a wall in the dark, leaving a mark, and meeting itself for the first time in the meeting of hand and stone. If that is even half right, then making things is not what recovered people do. It is how people recover; it is how people become at all. And the modern evidence behaves exactly as that account predicts. The World Health Organization's scoping review gathered over nine hundred publications and found arts engagement associated with prevention, management and treatment across both mental and physical health.³¹ In Britain, arts on prescription schemes report reduced GP visits and improved wellbeing in their evaluations, which is why social prescribing exists at all. The one door asks a person to document who they have stopped being. A wall in the dark asks the older question, and the species has been answering it for sixty five thousand years.
Permit one paragraph of imagination, strictly evidence-adjacent.
A country where the 2.8 million are asked what they would like to make. Where the care worker of sixty one spends five hours a week teaching teenage apprentices how to lift a frightened person with dignity, and is paid to reach her pension whole. Where the school leaver of nineteen is not a case file learning his deficits but the volunteer who rebuilt the community centre's booking system, because pattern-brilliance is like that when you stop itemising what it cannot do. Where the survivor makes mosaics in a church hall on Tuesdays, and nobody anywhere holds a form that requires her to be broken.
None of that is utopia. All of it is five hours a week, an income floor, and a different question at the door.
One more piece of science, and it is the hopeful one.
The same property of the brain that makes coercion so damaging is the property recovery runs on: plasticity. Nothing about the rewiring is a life sentence; tissue that adapted to threat adapts again to safety, and the clinical literature on post-traumatic growth documents people not merely restored but enlarged. The brain that language talked into a pit is a brain that language can walk out of it. Slowly. In company. In the present tense.
So ask the question this whole paper has been circling: what might a brain do if we abandoned the sickness model of survival? Psychology has an answer, and it has a name: broaden and build. Barbara Fredrickson's research programme showed that threat states narrow the mind to the width of the threat, which is survival doing its job, while states of safety and positive emotion do the opposite: they broaden attention, widen the range of thoughts and actions a person can even conceive of, and build durable resources... skills, ideas, relationships, resilience.²⁹ Fear spends the mind. Safety compounds it. A brain released from performing sickness does not simply feel better; it becomes structurally more capable of curiosity, play, connection and invention, because those are what a nervous system does when nothing is hunting it. We do not actually know what 2.8 million broadened minds would build, and that is the most honest argument in this paper. Nobody has ever run the experiment.
And consider, finally, what renaming might do, because the system's most powerful document was never the payment. It is the letter that tells a person what they are: claimant, case, condition. Imagine the same envelope carrying a different noun. Artist. Architect of a corner of the world. This is not sentimentality; Part Two of this paper is one long demonstration that the nouns a system issues become the people who receive them, and machinery that programs downward can program upward without a single new part.
The deepest change is what the new noun does to a person's history. Under the sickness model, everything that happened to you is evidence: a liability to be documented, dated and kept provable, a wound held open for inspection at the next assessment. In the artist's frame the same history becomes material. The Japanese repair tradition of kintsugi mends broken pottery with seams of gold, on the understanding that the break is part of the object's story and not the end of it: the crack is where the gold goes. A person whose past is evidence must keep it open. A person whose past is material gets to close it, shape it, and make it speak on their own terms, which is close to what the recovery literature means by meaning made. Nothing that happened is wasted. The years in the pit turn out to have been an apprenticeship... and nobody else knows the dark that well.
Which returns this paper to its title, and turns it around. Sick to death is what the 2.8 million are: sick to death of forms, of performances, of being asked, professionally and permanently, what is wrong with them. And how language programs us is not, in the end, a complaint. It is an instruction manual, read in the wrong direction for forty years.
The pit is made of words. So is the ladder.
1. ONS, economic inactivity by reason, long-term sickness series (2.83 million aged 16 to 64, a record 7 per cent of the working-age population; around 1.54 million aged 50 to 64, 655,000 aged 35 to 49, 359,000 aged 25 to 34, 254,000 aged 16 to 24); DWP analysis for the Keep Britain Working review, update March 2026. 2. Analysis of the cost of health-related economic inactivity, 2026 (£212bn per year, approximately 7% of GDP). 3. Institute for Fiscal Studies, The role of changing health in rising health-related benefit claims, 2026. 4. House of Commons Library, Supporting neurodivergent people into employment, CDP-2025-0179; The Buckland Review of Autism Employment, 2024. 5. Office for Budget Responsibility, The effects of ageing and a rising state pension age on incapacity benefits caseloads. 6. The Health Foundation, analysis of rising State Pension age, ill health and poverty risk; Resolution Foundation, Revisiting the State Pension age. 7. Conceptualising the social in mental health and work capability: implications of medicalised framing in the UK welfare system, Social Psychiatry and Psychiatric Epidemiology, 2023. 8. Ministry of Justice, tribunal statistics quarterly: social security and child support appeals, claimant success rates at hearing; Resolution Foundation, Reassessing the Work Capability Assessment. 9. Felitti, V. et al., Relationship of childhood abuse and household dysfunction to many of the leading causes of death in adults (the ACE study), American Journal of Preventive Medicine, 1998. 10. University of Glasgow, study of GP consultation patterns in the three years preceding incapacity benefit claims, 2010. 11. Welfare Conditionality project, University of York and partners, final findings report, 2018. 12. Barr, B. et al., 'First, do no harm': are disability assessments associated with adverse trends in mental health?, Journal of Epidemiology and Community Health, 2016. 13. Fricker, M., Epistemic Injustice: Power and the Ethics of Knowing, Oxford University Press, 2007. 14. See the collected neuroscience citations in the author's evidence file for Prism Break (Teicher et al. on maltreatment and brain development; the complex trauma literature): available on request. 15. Rosenthal, R. and Jacobson, L., Pygmalion in the Classroom, 1968; on the reverse (Golem) effect, Babad, E., Inbar, J. and Rosenthal, R., Journal of Educational Psychology, 1982. 16. Creswell, J.D. et al., Affirmation of personal values buffers neuroendocrine and psychological stress responses, Psychological Science, 2005; Cohen, G. and Sherman, D., The psychology of change: self-affirmation and social psychological intervention, Annual Review of Psychology, 2014. 17. Mani, A., Mullainathan, S., Shafir, E. and Zhao, J., Poverty impedes cognitive function, Science, 2013. 18. Haushofer, J. and Shapiro, J., The short-term impact of unconditional cash transfers to the poor: experimental evidence from Kenya, Quarterly Journal of Economics, 2016. 19. Jenkinson, C. et al., Is volunteering a public health intervention? A systematic review and meta-analysis, BMC Public Health, 2013; Okun, M., Yeung, E.W. and Brown, S., Volunteering by older adults and risk of mortality, Psychology and Aging, 2013. 20. Shildrick, T., MacDonald, R., Furlong, A. et al., Are 'cultures of worklessness' passed down the generations?, Joseph Rowntree Foundation, 2012. 21. Heijmans, B. et al., Persistent epigenetic differences associated with prenatal exposure to famine in humans, PNAS, 2008; Yehuda, R. et al., on intergenerational stress effects, held as young science. 22. Becker, H., Outsiders: Studies in the Sociology of Deviance, 1963; Merton, R., Social structure and anomie, American Sociological Review, 1938; Hirschi, T., Causes of Delinquency, 1969. 23. Ministry of Justice, Prisoners' childhood and family backgrounds, 2012 (42 per cent of prisoners reported permanent exclusion from school); see also the Timpson Review of School Exclusion, 2019. 24. Atkinson, A.B., The Case for a Participation Income, Political Quarterly, 1996; Inequality: What Can Be Done?, Harvard University Press, 2015. 25. Kela, Results of Finland's basic income experiment, final report, 2020. 26. Welsh Government, Basic income for care leavers in Wales pilot evaluation, annual report 2025 to 2026. 27. Basic income pilot, Jarrow and East Finchley, 2026 onwards. 28. De Wispelaere, J. and Stirton, L., The public administration case against participation income, Social Service Review, 2007; Political Quarterly symposium on Atkinson's participation income, 2018. 29. Fredrickson, B., The role of positive emotions in positive psychology: the broaden-and-build theory, American Psychologist, 2001. 30. Martel, J.F., Humans did not invent art. It was the other way around, Aeon, July 2026. 31. Fancourt, D. and Finn, S., What is the evidence on the role of the arts in improving health and well-being? A scoping review, WHO Regional Office for Europe, Health Evidence Network synthesis report 67, 2019. 32. Office for Budget Responsibility, Welfare trends report, June 2026: working age health and disability benefit spending of £62.7 billion in 2025/26, forecast £74.9 billion by 2028/29; incapacity caseload 3.4 million in 2024-25, forecast 4.0 million by 2030-31. The £65 billion figure is the author's static calculation: £1,600 x 12 x 3.4 million. 33. House of Commons Work and Pensions Committee, Health assessments for benefits, 2023, on assessment contract costs (the Atos contract alone was reported at £112 million a year); Disability Rights UK, 2020: DWP spent £121 million across 2017-18 and 2018-19 defending ESA and PIP decisions, £58 million on mandatory reconsiderations and £63 million on tribunal appeals.34. NHS Business Services Authority prescribing data: more than 91 million antidepressant items dispensed in England in 2024, the highest annual total on record; Stedman et al., National prescribing trends and cost analysis of antidepressants, anxiolytics, and hypnotics in England, Human Psychopharmacology, 2025. 35. Equality Act 2010, section 20; Equality and Human Rights Commission, statutory Code of Practice on services, public functions and associations, on the anticipatory nature of the reasonable adjustments duty. 36. West, S. and Castro, A., preliminary analysis of the Stockton Economic Empowerment Demonstration (SEED), 2021: full time employment among recipients rose from 28 per cent to 40 per cent in the first year, against a rise from 32 to 37 per cent in the control group; NCVO, UK Civil Society Almanac, on regular volunteering; Census 2021, around five million unpaid carers in England and Wales. 37. Haushofer, J., Mudida, R. and Shapiro, J., The comparative impact of cash transfers and a psychotherapy program on psychological and economic well-being, NBER Working Paper 28106, 2020; Ridley, M., Rao, G., Schilbach, F. and Patel, V., Poverty, depression, and anxiety: causal evidence and mechanisms, Science, 2020. 38. Spitz, R., Hospitalism: an inquiry into the genesis of psychiatric conditions in early childhood, The Psychoanalytic Study of the Child, 1945; on the workhouse test and less eligibility, the Poor Law Amendment Act 1834. 39. On the United States negative income tax experiments (New Jersey, Seattle and Denver, 1968 to 1980) and their modest, concentrated labour supply effects, see Widerquist, K., A failure to communicate: what (if anything) can we learn from the negative income tax experiments?, Journal of Socio-Economics, 2005. 40. NHS England Digital, Fit notes issued by GP practices, England (around 94 per cent recorded "not fit for work" in recent years; the "may be fit for work" option used in roughly 6 to 7 per cent of cases); DWP, An evaluation of the Statement of Fitness for Work, 2016 ("not fully achieving what it set out to do"); Policy Exchange, Not Fit for Purpose, on the fit note's operation in practice.
The author is a psychology graduate and freelance researcher who spent twenty years delivering and managing employment support programmes, including DWP back to work contracts and European Social Fund provision.
Published anonymously, August 2026. If any of this lands close to home: Samaritans are free, any hour, on 116 123.