Free speech & censorship
Can Hate Speech Laws Infringe on Freedom of Expression?
LNAT Section B · Model essay
The essay prompt
Is it possible for laws against hate speech to encroach on the right to free expression, and what might follow if they do? Set out a clear position and weigh the consequences.
The stance
Hate-speech laws can infringe freedom of expression, but whether that infringement is illegitimate depends on drafting and proportionality: narrow, harm-focused laws that catch incitement are a justified limit, while vague, offence-focused ones chill lawful speech and harm even the groups they aim to protect.
Defining the terms
- Hate-speech laws — Criminal or regulatory rules prohibiting expression that incites or stirs up hatred or hostility against groups defined by protected characteristics such as race, religion or sexual orientation (e.g. Public Order Act 1986, Part III, ss.18-23).
- Freedom of expression — The right to hold opinions and to receive and impart information and ideas without unjustified state interference, including ideas that offend, shock or disturb (Article 10 ECHR; First Amendment in the US).
- Infringe — To penalise, deter or narrow lawful speech beyond what is necessary to prevent a real harm such as incitement to violence; a limit only becomes an infringement when it is disproportionate or unforeseeable.
- Consequences — The legal (prosecution, prior restraint), democratic (chilled dissent), social (minority self-silencing) and cultural (paternalism, fragility) effects that follow from restricting expression.
Assumptions to interrogate
- That 'hatred' and 'grossly offensive' can be defined precisely enough that ordinary speakers can foresee liability and will not over-censor themselves.
- That the state can reliably separate genuine harm (incitement to violence, threats) from mere offence without overreaching.
- That equality and dignity are better advanced by criminalising expression than by protecting counterspeech and using narrower harassment and threat laws.
- That the long-run effects of restriction (institutional drift, self-censorship, a narrower public conversation) are negligible or manageable.
The case for
Vagueness and overbreadth chill lawful speech
Statutes that turn on indeterminate terms ('insulting', 'grossly offensive', 'stirring up') leave speakers unable to predict liability, so they self-censor before any court can narrow the law. Free-expression doctrine demands foreseeability; where it is absent, the chilling effect bites at the moment of speech. Parliament itself conceded this when it removed 'insulting' from s.5 of the Public Order Act 1986 in 2014 after sustained civil-liberties pressure.
Broad categories can be turned against dissenters and minorities
Once a 'hate' category exists it is enforced by officials with incentives, and the line can be stretched against journalists, protesters or reformers. Minority advocates often use sharp, transgressive rhetoric; if 'hate' is read expansively, counter-hegemonic speech is among the first to be chilled. A shield built for the vulnerable can, in the wrong hands, become a sword against them.
Criminalising offence inverts the function of free speech
On Mill's harm principle, the state may already punish incitement, threats and harassment. Hate-speech laws that go further and punish non-inciting expression because it offends a group's dignity redefine liberty as a permission contingent on others' feelings. Conflate harm with offence and you constitutionalise a heckler's veto over unpopular ideas.
The case against
Unchecked hate speech itself silences the vulnerable
Mill assumed a level playing field, but hostility and intimidation can drive targeted groups out of public debate altogether. On this view a narrow hate-speech law is a net enabler of expression: it sets a floor of safety so that more voices, not fewer, can participate. Regulation restores the marketplace of ideas rather than shrinking it.
Speech can cause real, systemic harm
Hate propaganda is not merely hurtful; at the extreme it legitimises violence, as the ICTR found when it convicted RTLM broadcasters for inciting genocide in Rwanda. Even short of that, sustained group vilification produces marginalisation and civic exclusion that fall squarely within a properly applied harm principle.
Liberal systems already contain the abuse risk through proportionality
Democracies routinely carve speech into categories (fraud, defamation, incitement) without collapsing into censorship. Independent courts, proportionality review under Article 10(2) ECHR and the Oakes test in Canada exist precisely to keep restrictions narrow. The danger of overreach is an argument for vigilance and tight drafting, not for abandoning the protection entirely.
The argument, step by step
- Define the terms so the question is whether the limit is legitimate, not whether 'hate is bad'; concede at once that any criminal speech rule is a limit on Article 10.
- Establish that vagueness is the decisive defect: foreseeability operates at the moment of speech, so an indeterminate standard chills lawful expression before appellate courts can rescue it.
- Show the enforcement risk: even well-drafted laws are stretched under pressure, and minority reformers using transgressive speech are disproportionately exposed.
- Engage the strongest counter directly: unchecked hostility also silences, and extreme cases (Rwanda) show speech can incite real violence, so some restriction is justified.
- Resolve the tension by distinguishing harm from offence: narrow, incitement-and-threat-focused laws are a justified limit; broad, offence-based ones are an illegitimate infringement.
- Conclude that hate-speech laws can infringe freedom of expression, and that the consequences (chill, selective enforcement, minority self-silencing) are real wherever drafting drifts from harm toward offence.
The model plan
Thesis: hate-speech laws CAN infringe Article 10; legitimacy turns on drafting and proportionality (harm vs offence). Intro (~90 words): define 'hate-speech laws' and 'infringe'; concede every such law limits Art.10; thesis. Para 1 (FOR): vagueness/overbreadth = chill at moment of speech; s.5 POA 1986 'insulting' repealed 2014; Handyside protects speech that offends/shocks/disturbs; Brandenburg shows a clean boundary. Para 2 (FOR): selective enforcement + minority self-silencing; E.S. v Austria shows dignity frames can curb critique. Para 3 (AGAINST, steel-man): hostility itself silences; RTLM/Nahimana shows incitement can produce atrocity; Keegstra and ICCPR Art.20(2) justify narrow bans. Para 4 (resolution): distinguish harm from offence; incitement, threats and harassment already punishable; narrow rule legitimate, broad rule infringes; comparative spread (US Skokie absolutism vs SA s.16(2) carve-out vs ECHR margin of appreciation). Conclusion (~70 words): yes they can infringe; harm-focused = justified limit, offence-focused = illegitimate; consequences are chill, capture and minority silencing. Link every paragraph back to '...therefore these laws can infringe freedom of expression.'
The model essay
Any law that criminalises what people may say is, by definition, a limit on freedom of expression. The real question is not whether hate-speech laws touch Article 10 of the European Convention, but whether they do so legitimately. My answer is that they can infringe it, and that whether the infringement is justified turns on a single distinction: between laws aimed at harm and laws aimed at offence.
The clearest danger is vagueness. Statutes that turn on words like 'insulting', 'grossly offensive' or 'stirring up' ask speakers to predict liability they cannot foresee. Because the chilling effect operates at the moment of speech, not after a sympathetic appeal court has narrowed the law, ordinary people fall silent rather than risk prosecution. This is not theoretical. In Handyside v United Kingdom (1976) the Court held that Article 10 protects ideas that 'offend, shock or disturb'; yet vague offence-based offences punish exactly that. Parliament effectively conceded the point in 2014 when it removed 'insulting' from section 5 of the Public Order Act 1986. By contrast the American rule in Brandenburg v Ohio (1969), which permits restriction only of incitement to imminent lawless action, shows that a sharp boundary chills far less. Therefore vague hate-speech laws can infringe freedom of expression.
A second danger is who enforces the category. Even carefully drafted laws are applied by officials with incentives, and the label 'hate' is elastic under political pressure. Minority reformers, who frequently use transgressive rhetoric against dominant religions or norms, are among the first to be caught: in E.S. v Austria (2018) a conviction for disparaging the Prophet Muhammad was held compatible with Article 10, showing how a dignity frame can be turned against critique. A shield built for the vulnerable can become a sword against them. Therefore broad hate-speech laws can infringe freedom of expression.
The strongest objection must be met head-on. Unchecked hostility also silences: intimidation drives targeted groups out of public debate, so a narrow law arguably widens participation rather than narrowing it. And speech can do catastrophic harm. The International Criminal Tribunal for Rwanda convicted RTLM broadcasters for inciting the 1994 genocide, proof that some expression is not mere offence but a trigger for violence. Canada's Supreme Court in R v Keegstra (1990) accepted that its hate-propaganda offence infringed free expression yet upheld it as a proportionate limit, and ICCPR Article 20(2) actually requires states to prohibit incitement to discrimination, hostility or violence.
This is where the harm-offence distinction resolves the debate. Incitement, threats and harassment are already punishable; a law confined to them is a justified limit, not an illegitimate infringement. A law that reaches beyond, to opinions that merely wound a group's feelings, conflates harm with offence and hands hecklers a veto. Comparative practice maps onto this line. The United States protects even a neo-Nazi march (National Socialist Party of America v Skokie, 1977), which many find corrosive; South Africa's Constitution excludes only 'advocacy of hatred that constitutes incitement to cause harm' (s.16(2)); the ECHR sits between, granting states a margin of appreciation. The lesson is consistent: the narrower and more harm-focused the rule, the safer both liberty and equality.
Hate-speech laws can therefore infringe freedom of expression, and the consequences of getting them wrong are concrete: lawful dissent is chilled by uncertainty, the category is stretched against reformers, and the very minorities the law protects can find their sharpest advocacy reclassified as hatred. Drafted narrowly around incitement and threats, such laws are a defensible price of pluralism. Drafted broadly around offence, they quietly erode the freedom they claim to balance.
Authorities worth knowing
Handyside v United Kingdom
(1976) 1 EHRR 737, App no 5493/72 (ECHR)
Article 10 protects not only inoffensive ideas but those that 'offend, shock or disturb'; restrictions must be necessary in a democratic society (here the obscenity ban was upheld).
Brandenburg v Ohio
395 U.S. 444 (1969)
Advocacy may be punished only where it is directed to inciting imminent lawless action and is likely to produce it; a clear boundary that minimises chilling effects.
DPP v Collins
[2006] UKHL 40
Interprets 'grossly offensive' under s.127(1) Communications Act 2003 by the standards of an open and just multiracial society; illustrates the interpretive difficulty of offence-based criminal standards.
E.S. v Austria
App no 38450/12 (ECHR, 25 October 2018)
Conviction for disparaging religious doctrine (calling the Prophet Muhammad a paedophile) did not violate Article 10; shows how dignity and 'religious peace' frames can curb critical speech.
R v Keegstra
[1990] 3 SCR 697 (Supreme Court of Canada)
The wilful-promotion-of-hatred offence (Criminal Code s.319(2)) infringes s.2(b) Charter free expression, but is a reasonable limit saved under s.1 on the Oakes test.
National Socialist Party of America v Village of Skokie
432 U.S. 43 (1977)
First Amendment procedure protected the right of a neo-Nazi group to march; exemplifies the US absolutist approach that tolerates even deeply offensive expression.
Public Order Act 1986, Part III (ss.18-23)
Public Order Act 1986 c.64, Part III
Creates offences of using threatening, abusive or insulting words or behaviour intending or likely to stir up racial hatred; the core UK hate-speech regime.
Prosecutor v Nahimana, Barayagwiza and Ngeze (the Media Case)
ICTR-99-52-T (ICTR, 3 December 2003)
RTLM radio and Kangura figures convicted of direct and public incitement to commit genocide; the strongest example that hate propaganda can produce mass violence.
How the law frames it
United Kingdom
The UK criminalises stirring up racial, religious and sexual-orientation hatred under Part III of the Public Order Act 1986 and 'grossly offensive' communications under s.127 Communications Act 2003. Speech is protected by Article 10 ECHR via the Human Rights Act 1998, subject to limits that are prescribed by law and necessary in a democratic society. Parliament's 2014 repeal of 'insulting' in s.5 shows live concern about overbreadth.
Canada
Section 319(2) of the Criminal Code criminalises wilfully promoting hatred. In R v Keegstra (1990) the Supreme Court held this infringes s.2(b) Charter free expression but is a justified limit under s.1 on the Oakes proportionality test, accepting that narrowly drawn hate-propaganda offences can coexist with a strong free-expression guarantee.
ECHR
Article 10 protects expression that offends, shocks or disturbs (Handyside), but Article 10(2) permits proportionate restrictions, and Article 17 excludes abuse of rights. The Strasbourg Court grants states a margin of appreciation, upholding some convictions (E.S. v Austria) while striking down disproportionate ones, producing a balancing model between US absolutism and outright prohibition.
Counter-arguments and how to defeat them
Counter. Unchecked hate speech itself silences the vulnerable, so a narrow law expands the marketplace of ideas rather than shrinking it.
Rebuttal. True for genuine intimidation, threats and harassment, which the law already reaches. The objection only justifies a narrow, harm-focused rule; it does not license offence-based offences whose vagueness chills lawful minority advocacy as readily as bigotry.
Counter. Speech can cause catastrophic harm, as Rwanda shows, so the harm principle clearly permits restriction.
Rebuttal. Agreed, and that is exactly why incitement to violence is and should be criminal. Rwanda is incitement, not mere offence; citing it to defend broad offence-based laws conflates the extreme case with the everyday one.
Counter. Independent courts and proportionality tests (Oakes, Article 10(2)) prevent abuse, so the slippery-slope worry is overstated.
Rebuttal. Safeguards reduce but do not remove the risk, because the chilling effect operates before any judge is involved. The right answer is to draft tightly around harm so the proportionality question rarely arises, not to trust ex-post review to undo self-censorship.
Counter. The US absolutist model shows resilience, so all hate-speech laws are unnecessary.
Rebuttal. Absolutism is the global outlier and tolerates real harms; the choice is not between Skokie and total prohibition. South Africa's s.16(2) and Canada's Keegstra show a defensible middle path that bans incitement to harm without criminalising opinion.
Conclusion
Hate-speech laws can and do infringe freedom of expression; the decisive question is whether the infringement is legitimate. Drafted narrowly around incitement, threats and harassment, they are a justified price of pluralism that may even widen participation. Drafted broadly around offence, they chill lawful dissent through vagueness, invite selective enforcement, and can silence the very minorities they claim to protect. The harm-offence distinction is the line that separates a proportionate limit from an illegitimate infringement.
Evidence you can cite
- Police in England and Wales recorded 140,561 hate crimes in the year ending March 2024, a 5% decrease on the previous year's 147,645.Home Office, 'Hate crime, England and Wales, year ending March 2024' — source
- The word 'insulting' was removed from the s.5 Public Order Act 1986 offence by s.57 of the Crime and Courts Act 2013, in force from 1 February 2014, narrowing a criticised overbroad standard.Crime and Courts Act 2013, explanatory notes; House of Commons Library briefing SN05760 — source
- ICCPR Article 20(2) obliges states to prohibit by law 'any advocacy of national, racial or religious hatred that constitutes incitement to discrimination, hostility or violence'.International Covenant on Civil and Political Rights, Article 20(2) (OHCHR) — source
Further reading
- J.S. Mill, On Liberty (1859), ch. 2 on the marketplace of ideas and the harm principle.
- Council of Europe, ECHR Factsheet on Hate Speech (Article 10 case law).
- House of Commons Library, 'Insulting words or behaviour: Section 5 of the Public Order Act 1986' (Briefing SN05760).
- OHCHR, Rabat Plan of Action on the prohibition of incitement to discrimination, hostility or violence.