The Steelman — willpower and education, in the voice of the people who hold the position

Rules of engagement

Three rules for reading this page properly.

First, read it in her voice, not the voice you would use to mock her. The composite speaker on this page is a senior NHS biomedical scientist — twenty-two years in pathology, eight of those in liver histology. She has read the BMJ for thirty years. She knows what cirrhosis looks like under a microscope. She is not a Daily Mail caricature; she is the person who would have run your father's diabetes panel.

Second, the steelman is not a strawman in a hat. A real steelman is the argument made better than its actual proponents typically make it — with the most defensible evidence, the most coherent logic, and the least amount of unforced error. If, after reading the page below, the willpower position sounds more persuasive than it did before, the page is doing its job.

Third, the answer comes at the end, not in the middle. No fact-checking footnotes interrupting her sentences. Let her finish.

Her voice — the full statement

I have been on the wards for twenty-two years. I have processed the bloods of every alcoholic who came through A&E in my catchment. I have looked at hundreds of livers at autopsy. I have weighed the hearts of patients who died at 56 because their body weight was 140 kilos and their blood pressure was 180 over 110. I am not unsympathetic. I have given more bedside time to dying drinkers than anyone in your charity will ever see.

And what I have learned — what twenty-two years of looking at the actual organs has taught me — is that the people who get better are the ones who decide to. I have watched two patients with the same diagnosis, the same age, the same social background, the same family. One stops drinking and lives twenty more years. One does not, and dies. The difference between them is not biology. It is character. It is what they choose to do at six o'clock on a Tuesday evening when the bottle is on the counter and the wife is at her sister's. We used to call that self-discipline. Now we call it a brain disease, as if the words make a difference to the outcome.

I am not opposed to medicine. I have spent my life in it. I prescribe — under supervision — for diabetes, for hypertension, for depression. I have no objection to medicating a real disease. My objection is to the redefinition of normal human weakness as disease, so that the patient is relieved of responsibility for managing it. Hunger is not a disease. Boredom is not a disease. Loneliness is not a disease. Looking at pornography on your phone for four hours after midnight is not a disease. These are the ordinary frictions of being a person, and a culture that cannot name them as such has lost something important about adulthood.

Mounjaro is the example I will use because it is the one in front of us. We have a population, by the government's own data, where two-thirds of adults are now overweight or obese. We have decided — somehow, without a debate — that the answer is to put as many of them as possible on a permanent weekly injection. We will spend hundreds of millions of pounds doing it. The official rationale is that obesity is a disease. The actual mechanism is that we have lost the cultural authority to ask the patient to eat less and move more, because asking has been re-coded as cruelty. So we medicate instead. We are buying social peace with a needle. The fact that the drug works is not the point. Insulin works too. The question is whether the underlying condition is genuinely a disease — in which case medicate it — or whether it is a learned behaviour in a population that has been deprived of the social pressure that used to correct it.

The same goes for drink. The same goes for drugs. The same will, soon, go for pornography and gambling and social media. Each one will arrive with a clinical name, a screening questionnaire, a recommended pharmacotherapy and a public-health campaign. Each one will be presented as inarguable science. Each one will quietly replace the older, ruder, more accurate language of self-control, conscience, and personal responsibility.

The argument against me, when it comes, will be that I am stigmatising people. It is not stigma to tell a patient the truth. It is not stigma to expect a competent adult to manage their own behaviour. The stigma is in the other direction. To call someone diseased when they are not is to tell them, on the authority of a doctor in a white coat, that they cannot help themselves. That is the cruellest thing you can say to a person who could in fact help themselves, if you would only have the courage to ask them to. My grandmother lived through a war. She did not need an injection to control her appetite. She needed less food, more work, and the dignity of being trusted to manage herself. That is the dignity we are taking away from people in the name of compassion. I think it is a mistake. I do not think it is a kindness.

Addiction is real for a small handful of people whose lives are organised entirely around a substance. For the rest — the overwhelming majority of the people now being given that label — it is the medical name we give to behaviour we have decided not to challenge any more. Willpower works. Education works. We have stopped using both because we have lost the nerve. Bringing the nerve back is not cruelty. It is the basic act of treating an adult as an adult.

Let it sit for a moment before you answer it. If your first instinct is to argue with sentence three, you read too fast.

What she gets right

Six things — and they are not small.

One: most people who use addictive substances are not addicted. The DSM-5 epidemiology bears this out. Of US adults who drink, only a minority meet criteria for any alcohol use disorder; of those who do, only a fraction are severe. Of US adults who have tried heroin — a non-trivial number — most do not become addicted. Carl Hart's whole career rests on exactly this point: addiction is the exception even among users of "addictive" drugs, not the rule. To talk about "the obesity epidemic" or "the porn epidemic" as if everyone on the wrong side of a screening line were addicted is loose use of language. She is right to push back on it.

Two: the disease label is expanding faster than the underlying science. The DSM declined to recognise pornography addiction in 2013. The ICD-11 in 2018 put compulsive sexual behaviour disorder in the impulse-control bucket, not the addiction bucket. "Food addiction" is contested even within the field; Gearhardt's Yale group has been pushing the ultra-processed-food addiction frame for two decades and still does not have institutional acceptance. She is right that the diagnostic perimeter has expanded faster than the evidence — and right to be suspicious about who benefits from each expansion.

Three: behaviour is, in fact, responsive to social pressure and to information. Per-capita cigarette consumption in the UK is roughly a fifth of what it was in 1974. This is not because of pharmacotherapy — patches and gum are not magic. It is because smoking became unattractive at the level of culture: less status, less attractive, more expensive, less visible, less convenient. Cultural change is a real lever and we know it works. She is right that we have, in some areas, stopped pulling that lever.

Four: medicalisation has costs. A diagnosis is not free. It carries identity effects, insurance implications, downstream pharmaceutical exposure, and a particular set of expectations about agency that can be self-fulfilling. Marc Lewis — a neuroscientist sympathetic to harm reduction, hardly her natural ally — has spent a decade arguing exactly her point: the disease label can produce learned helplessness in the people it labels. She is right that the question "is this useful as a clinical category?" is separate from "is this real as a biological phenomenon?" and that public-health institutions routinely conflate the two.

Five: the "willpower is a myth" line is itself sloppy. Self-regulation is a measurable construct. It correlates with life outcomes. It is partly heritable, partly trainable, and unevenly distributed. The popular claim that "willpower has been debunked" is an over-reading of the ego-depletion replication failures — those failures called into question one specific mechanism of self-control, not the existence of self-control as a faculty. She is right that there is a real thing called self-discipline, that it varies between people, and that pretending otherwise is its own form of bad faith.

Six: pharma is not a neutral party. Eli Lilly's market capitalisation crossed $700bn during the GLP-1 boom. Novo Nordisk became, briefly, the most valuable company in Europe. The same global health architecture that gave us OxyContin marketed by Purdue is now giving us tirzepatide marketed by Lilly. She is right to be suspicious. The Frontier Economics report putting UK obesity costs at £98bn was commissioned by Novo Nordisk; the Tony Blair Institute report at £126bn was published as the GLP-1 NHS rollout was being negotiated. Funding bias is not an ad hominem when the figures keep rising and the funder keeps benefiting.

What she gets wrong

Three things — and they are load-bearing.

One: the inference from "behaviour responds to pressure" to "addiction is just a learning gap" does not hold. Smoking rates fell across the population, including among addicted smokers, when the price went up and the social cost rose. That tells us cultural levers move averages. It does not tell us that the dependent smoker has the same internal experience as the casual one. Population-level data and individual-level mechanism are different questions, and the willpower argument routinely confuses them. Most people on a cruise ship do not need to be tied to the mast; that does not mean Odysseus was making a fuss.

Two: the set-point physiology is not optional. Jules Hirsch and Rudy Leibel spent forty years at Rockefeller documenting what happens metabolically when a person loses weight. The body responds with a measurable, persistent reduction in resting energy expenditure of 300 to 400 kilocalories per day below what mass loss alone predicts, alongside elevated hunger signalling, that does not fade with time. This is not a story about character. It is a homeostatic system defending its prior state. Asking a reduced-obese patient to "just eat less" is asking them to override a hypothalamically defended set-point indefinitely. A few people can; most cannot; this distribution is not a moral ranking. When she says "we could give a fat patient the same outcome with a calorie deficit and a pair of trainers", she is describing an intervention that, on the published evidence, produces clinically meaningful sustained weight loss in approximately single-digit percentages of patients at five years. The drug does it for the majority. The drug is not cheating biology. It is meeting biology where biology is.

Three: "education" as a policy lever has the weakest evidence base of any addiction intervention. DARE — the largest school-based drug-education programme in American history — was studied to death and found to produce no measurable reduction in subsequent drug use, and in some studies a small increase. "Just Say No" did not reduce adolescent drug use. The American abstinence-only sex-education curricula did not reduce teenage pregnancy. The British government's repeated five-a-day, change-for-life, eatwell-plate campaigns have moved obesity rates in the wrong direction over twenty years. People know smoking is bad for them; they smoke anyway. People know the calorie content of a Big Mac; they eat it anyway. Knowledge is a necessary condition for behaviour change; it is empirically not a sufficient one. The whole policy page is about this. If you take "more education and stricter standards" as your central policy proposal, the evidence base for that proposal is among the weakest in the field.

"Mounjaro is cheating"

The specific moral judgement deserves its own treatment. The lab tech's claim is that pharmacological appetite suppression is somehow a shortcut, that it bypasses the work the patient should be doing, that it leaves the underlying lifestyle disorder untreated.

The consistency test settles this. We do not say insulin "cheats" Type 1 diabetes. We do not say statins "cheat" familial hypercholesterolaemia. We do not say SSRIs "cheat" major depression — or if we do, we do so as an outlier, not in the canonical position of a healthcare professional. We do not say methadone "cheats" opioid dependence. We do not say warfarin "cheats" atrial fibrillation. The "cheating" frame is reserved, almost exclusively, for weight. It is also reserved, almost exclusively, for fat women. The same lab tech who would be appalled at refusing insulin to a diabetic on the grounds that "they need to learn discipline" is, on the same shift, comfortable saying that exact thing about a 110-kilo woman asking for a tirzepatide prescription. The asymmetry is the giveaway. The objection is not medical. It is moral, and the morality being enforced is a quite specific one about who deserves to be thin and how thinness should be earned.

The actuarial maths is also one-directional. A 50-year-old with BMI 32 and hypertension is, on the Prospective Studies Collaboration data, losing 3 to 5 life-years to the composite. Compliant tirzepatide use, extrapolating from SELECT's 19% all-cause mortality hazard reduction, gives back perhaps 1.5 to 3 of those life-years. The drug has real side effects — gallbladder disease at relative risk ~1.5, GI symptoms in ~40% of patients, lean-mass loss requiring resistance training and protein, supply constraints, retail cost of £1,800 to £3,000 a year. None of those are reasons to refuse the prescription to a patient for whom the benefit-risk arithmetic favours it. They are reasons to manage the prescription competently. The full case is on the Costs page.

The concession we owe her

There is no honest answer to her that does not start by saying: you are right that medicalisation has expanded its perimeter, you are right that the abstinence-rebuilding-character story has gone underground in clinical settings, you are right that willpower is not a myth, you are right that pharma is not your friend, and you are right that asking a patient to do the harder thing is not always cruelty.

The answer is that after all of that is conceded, addiction is still real. The mechanism is shared across substances and behaviours, well-documented, and not reducible to willpower failure. The minority who are genuinely addicted are owed a treatment apparatus that meets them where their biology is, which is not where their conscience is. The majority who are not addicted are owed a public-health architecture that does not pretend they are, and does not extract revenue or compliance from them on that pretence. Both can be true at once. The willpower-only view forces a single answer to a question that has two layers, and that is where it ends up doing damage.

The next pages walk the evidence one layer at a time.