Designed to Silence, Not to Enforce

Section 28 created no criminal offence. No criminal prosecution was ever available under it, because it belonged to civil, not criminal, law — and across its life in Britain, no court ever found that a local authority had breached it. By the ordinary measure we use to ask whether a law does anything — is it enforced against anyone? — it did almost nothing at all.

And yet many people educated while it was in force remember it as the thing that made a classroom unsafe to be honest in. They remember teachers who would not name what was happening to them, funding withdrawn from small support groups, a newspaper pulled from a library shelf. That is the paradox worth sitting with, and it is not a curiosity of legal history. It is a lesson about how a particular kind of harm operates, and about why we keep failing to recognise it until it has already done its work. The near-empty enforcement record is not the evidence against the harm. It is the harm’s signature.

This is not a history of Section 28. That history has been told carefully and often. This is an argument about mechanism: what the law actually did, how it did it without ever being successfully enforced, and why the same mechanism should concern anyone who cares about how trans health and education are governed today.

What it said, and what it did not

Section 28 of the Local Government Act 1988 inserted a new section into the Local Government Act 1986. It provided that a local authority “shall not intentionally promote homosexuality or publish material with the intention of promoting homosexuality”, nor “promote the teaching in any maintained school of the acceptability of homosexuality as a pretended family relationship”. The phrase “pretended family relationship” is the one people remember, and those it described found it insulting for an obvious reason: it told them their families were counterfeit.

Three facts about the text matter more than they are usually allowed to. The first is its legal character. Section 28 was civil law. It regulated what a local authority could do and created no criminal offence; it could not be invoked against an individual, a company, or even other public bodies. The only route was that a ratepayer might seek judicial review of a council’s action, and at most a court could grant an injunction. No damages, no fines, no penalties existed.

The second is its reach. It bound local authorities, not teachers. Responsibility for sex education already sat with school governing bodies under the Education (No. 2) Act 1986, and the government’s own guidance said so from the start: the 1988 circular accompanying the Act stated that Section 28 did not affect the activities of school governors or of teachers, and by 1994 official guidance stated plainly that Section 28 did not apply to schools at all. This is the fact to hold onto, because it is the whole argument in miniature: the law formally did not reach the classroom, and the classroom is where some of its best-documented damage occurred.

The third is what follows from the first two. Because no court ever found an authority in breach, it is tempting to call it toothless. That is a mistake — but so is its opposite, the belief that it was a criminal ban that could put a teacher in the dock. It was neither. Holding both false pictures in view at once is the only way to see what it actually was.

The enforcement record, taken at its strongest

Let me put the minimising case at full strength, because it is not stupid.

Research by Cardiff Law School in 1990 found that no council had been taken to court in relation to Section 28, and no criminal prosecution was available in any event. The litigation it did eventually produce was sparse and inconclusive. When Calderdale’s library service refused to stock a gay newspaper for fear of breaching the law, the civil-rights organisation Liberty applied for judicial review and the library reversed its decision — the section’s clearest day in court spent undoing an act of self-censorship committed in its name. The one occasion on which someone tried to use it the other way came in May 2000, when a Glasgow ratepayer, backed by the Christian Institute, took Glasgow City Council to the Court of Session over its funding of an HIV-prevention safe-sex guide; the case was settled out of court rather than decided, weeks before repeal reached Scotland. No court ever found that a local authority had breached Section 28, or endorsed an authority’s actions as justified by it.

So the minimiser asks a fair question: if a law is never enforced against anyone, in what sense did it harm anyone at all? An unenforced statute is, on its face, political theatre. Where is the injury?

I want to leave that question standing for a moment, because the whole argument turns on the answer.

Where the harm lived

The answer is that Section 28 did not need to be enforced. It worked the way a warning sign works. You do not need to take anyone to court if the mere existence of the prohibition changes how people behave, and Section 28 appears to have changed behaviour without ever being successfully used. To see how, it helps to take the mechanism apart.

The active ingredient was ambiguity. “Promote” was never defined. Nobody could say with confidence whether answering a pupil’s question, stocking a novel, or funding a youth group counted as promotion. When a prohibition is vague and the stakes feel real, the rational response is not to find the true line and walk up to it. It is to retreat well behind where the line might be. Vagueness does not need enforcement to control conduct; it does its controlling precisely by leaving people to guess, and to guess safe.

The risk was asymmetric, which is what made the retreat rational rather than merely timid. For an authority, the concrete exposure — a judicial review, an injunction, the political cost of a challenge — was real, even if remote. The cost of over-caution was diffuse and fell on other people, mostly on children who could not complain. For individual teachers the legal risk was, strictly, not there at all, since the law did not bind them; but the perception of risk was, and perception was enough. Faced with that asymmetry, institutions and individuals did what people do under uncertainty: they protected themselves.

And the effect diffused. There was no single national directive commanding silence. There did not need to be — and this is the part that should give anyone pause: the official circulars said the opposite, that the law did not bind teachers, and the silence descended anyway. A 1997 study of 307 schools in England and Wales found that most teachers who responded were unsure what the law required of them; over a quarter either wrongly believed it made discussing homosexuality illegal or were not sure; 44% said its continued existence made it harder to meet the needs of lesbian, gay and bisexual pupils; and the great majority wanted the position clarified. Alongside that ran a documented trail of withdrawals: a council banning a youth volunteering guide, another withdrawing an information booklet, grants to small support groups refused and then, after protest, restored.

That is the positive evidence, and it matters that it is positive. The harm of Section 28 is not inferred from the empty court record; if anything, the empty court record is the thing being explained. The harm is documented directly — in surveys of what teachers believed and did, and in the archive of what was withdrawn. When a harm takes the form of things that did not happen, this is the appropriate evidence, and treating it as second-rate is itself an error.

Here is the point that carries the whole piece, and clinicians will recognise its shape at once, because it is the reasoning we are trained to be most careful with. A chilling effect produces absence — the lesson not taught, the book not ordered, the bullied child not helped. Absence leaves no docket, because avoidance is what success looked like to those doing the avoiding. That is exactly why the enforcement record is empty, and exactly why the emptiness is not evidence of harmlessness. Absence of enforcement is not evidence of absence of harm — and the way to tell the two apart is to go looking for the positive evidence, which in this case is substantial.

Why both sides get it wrong

Once you can see the mechanism, the way Section 28 is fought over today comes into focus, and both familiar positions collapse the same distinction.

The minimisers read the empty enforcement record as proof of low harm: never used, therefore harmless. This is a clean instance of a specific fallacy — treating absence of evidence as evidence of absence, or here, the absence of any court action as the absence of injury. It is the same move I spend a great deal of time flagging in clinical claims, where “no evidence of harm” is quietly upgraded into “evidence of no harm.” Section 28 is that move in a wig.

The advocates make the mirror error. Invoking “Section 28” at every explicit restriction on what schools may teach dilutes the term until it means little more than “a policy I oppose”, and, worse, it misplaces the mechanism. It locates the danger in visible, nameable bans. But the whole lesson of Section 28 is that the danger lived in ambiguity and self-censorship, in a law that officially did not even apply to the people it silenced. If you only sound the alarm at explicit prohibitions, you will miss the quieter thing that actually reproduces Section 28 — because the quieter thing never announces itself as a ban.

Both errors share one root: they mistake the visible instrument for the operative one. The statute was visible. The operative instrument was fear under uncertainty.

The mechanism now

I want to be exact about the claim I am making, because this is the paragraph an honest reader should test hardest.

I am not saying any current policy “is Section 28 again.” I am saying something narrower and harder to dismiss: that the mechanism of harm — vague or shifting guidance, serious stakes, asymmetric risk, and defensive silence as the result — can be identical even when a policy’s intent and its legal form are entirely different.

You can see the mechanism wherever guidance is ambiguous enough that the safe professional move is to do less, and where the cost of doing less falls on someone who cannot answer back. The harm surfaces not as a documented decision but as an absence: the question not asked, the support not offered, the person quietly not treated. These are illustrations of a mechanism, not established causal claims; each would need its own primary-sourced case before it could be asserted as fact, and the association between ambiguous guidance and defensive practice is not proof that any given policy caused any given harm.

Two distinctions keep the analogy honest, and both come straight from the Section 28 record. The first: not every absence is a manufactured one. Empirical absence (a treatment genuinely lacking good evidence), methodological uncertainty (an open question badly studied), and politically produced absence (a silence created by fear) are three different things, and only the last is the Section 28 analogue. Reading “no evidence of benefit” as “evidence of suppression” is the same fallacy as reading it as “evidence of no benefit”, just pointed the other way. The second: defensive practice has many sources — litigation risk, safeguarding duty, professional regulation, commissioning constraints, plain evidentiary uncertainty — not only ideological chill. Even in the Section 28 case this was true: the Chief Inspector of Schools at the time disputed that the law harmed teachers’ handling of bullying at all, and the 1997 study itself named inexperienced staff and parental disapproval among the causes. A mechanism can be real and still be one cause among several. Claiming more than that is the over-reach that gets the whole comparison dismissed.

And one boundary this platform holds to strictly. Section 28 was about sexual orientation. Carrying its lessons into questions of gender identity is an analogy across two different populations, and it must be marked as one. The mechanism may transfer; the evidence does not. Nothing about the experience of Section 28’s targets can be quietly imported as though it were data about trans people specifically. The parallel earns its keep only if it stays structural.

What follows

If a law’s harm lives in its chilling effect rather than its enforcement, then the questions we use to judge policy are the wrong ones. “Has anyone been taken to court?” and “What does it intend?” both miss. The questions that track harm are: what does this make the cautious professional do, and who bears the cost when the guidance is unclear?

That reframing carries three commitments. Judge a policy by its chilling effects and by who is left holding the risk of its ambiguity, not by its enforcement record and not by its stated good intentions — Section 28 shows that a law can be near-unenforceable, and even formally inapplicable, and still do enormous work. Demand guidance specific enough that it does not induce defensive silence, because vagueness is never neutral: it transfers risk downward, onto the patient and the pupil who cannot answer back. And treat harm that takes the form of absence with the same seriousness we give harm that takes the form of an event, by going and looking for the positive evidence rather than waiting for a paper trail the mechanism is built never to produce.

This is, in the end, the problem this platform exists to work on: the distance between what the evidence actually supports and what people are actually told and actually given. Section 28 is the cleanest case I know of a harm that hid inside that distance for the whole of its life and left almost no fingerprints. It is worth understanding precisely, because the mechanism did not repeal itself in 2003. It sits underneath the debate on conversion practices too — a harm delivered not through any prosecutable act, but through ambiguity, pressure, and the silence they produce.

Section 28 was never once successfully used against a local authority. That was never how it worked. It worked through silence, and the silence is what it achieved.

Added 11 September 2026. This essay’s question — what a framework built around caution makes a school feel able to do — is taken up for the statutory guidance now in force in schools in England in our evidence appraisal, Grading the Guidance (GA-001), written as a companion to this piece. It concludes that KCSIE 2026 does not silence schools as Section 28 did, but makes caution the statutory default on evidence of low certainty.

Sources

• House of Commons Library, Research Paper 00/47, The Local Government Bill [HL]: the ‘Section 28’ debate (6 April 2000) — civil-law character and absence of criminal offence; the judicial-review/injunction route; Cardiff Law School (1990) on the absence of any court action; the Liberty judicial review of Calderdale’s library decision; the 1988 circular and 1994 guidance that the law did not apply to schools; the 1997 study of 307 schools.
• The National Archives, Section 28: Impact, fightback and repeal (2023) — no prosecutions across its life; the May 2000 Scottish court challenge, backed by the Christian Institute, settled out of court.
• Hansard, HC Deb, Section 28 Repeal: 20th Anniversary (29 November 2023) — never used to prosecute anyone; repeal dates; British Social Attitudes figures.
• University of Exeter, Section 28 and its afterlives — no one ever successfully prosecuted; the “silent, forbidding presence” framing.
• Local Government Act 1988, section 28 (legislation.gov.uk) — verbatim statutory wording and the civil-proceedings inference provision.
• Education (No. 2) Act 1986, section 18 — sex-education responsibility resting with school governing bodies.
• DfE Circular 5/94 — explicit statement that Section 28 did not apply to schools.
• “Unused but dangerous”, The Guardian, 11 October 1989 — the contemporaneous “self-censorship is the danger” framing.
• Ethical Standards in Public Life etc. (Scotland) Act 2000; Local Government Act 2003 — repeal in Scotland (Section 2A, 21 June 2000) and in England and Wales (18 November 2003).