Photo courtesy of Dawn

Words can, and often do, harm others. A society is usually justified in restricting speech that causes harm. The nature and extent of such restrictions depends on the type of harm involved. States are then expected to strike the right balance between the interests of the individual engaging in the speech, and the interests of others affected by the speech. Much depends on how much trust we can place in the state to strike that balance in good faith. This article explores the role of the state in countering harmful speech. It presents a conceptual model that sets out the different layers of harmful speech, and defines the specific role the state should play in curbing speech that falls into each layer. The article then makes a case for minimising the Sri Lankan state’s authority to impose criminal sanctions for harmful speech.

The concentric circles of harmful speech


Figure 1

Figure 1 depicts four layers of harmful speech. The first (outermost) layer represents all types of harmful speech, including speech that defames, stereotypes, or insults others. Falsely accusing someone of a crime, or insulting a person’s faith, can fall into this category.

The second layer represents ‘hate speech’, which is a subset of harmful speech. Speech that conveys hatred towards a particular ethnic, racial, religious, gender or other similar group falls into this category. Of course, the term ‘hatred’ needs to be defined carefully. It generally means ‘intense and irrational emotions of opprobrium, enmity and detestation towards the target group’.[1] For example, calling a community inherently ‘evil’ constitutes hate speech.

The third layer represents ‘incitement’, which is a subset of hate speech. While the broader category of hate speech involves conveying hatred, incitement additionally involves compelling others to act in a hateful manner. Such acts include discriminating the target group. For example, compelling others to boycott Muslim-owned businesses constitutes incitement to discrimination.

The final (innermost) layer represents incitement to violence, which is a subset of incitement, and the worst form of harmful speech. Incitement to violence contains two elements: first, it involves the act of compelling others to perpetrate violence against a specific person or group; second, it involves a context where such violence is ‘imminent’, i.e. there is a real risk of violence ensuing because of the speech.

Society has a legitimate role to play in interfering with each layer of harmful speech. Notice that I refer to ‘society’ here and not ‘the state’. Individuals in society can interfere with someone’s speech without necessarily involving the state. For example, rebuking someone for making a derogatory statement online, and deterring that person from repeating such a statement, is one simple example of interfering with the harmful speech of another.

The type and intensity of interference may depend on the classification of the harmful speech. The intensity of the interference can be placed on a spectrum, and should correspond to the intensity of the harm that the speech causes. At one end of the spectrum, there can be interference through private plaints and counter-messaging. Next, there can be civil sanctions of some kind, including boycotts, dismissals from employment, and the revocation of licences. At the other end of the spectrum, there can be criminal sanctions such as fines and imprisonment. It is natural for the state to play a role in how society curbs harmful speech. Yet there may be good reasons for minimising this role – particularly when it comes  to imposing criminal sanctions. One of the main reasons – and this reason is particularly relevant to the Sri Lankan context – is the loss of trust in the state.

Why distrust the state?

Criminal sanctions ought to be imposed sparingly, and reserved only for speech that causes the most serious harm. Leaving aside the philosophical reasons for adopting this position, there are compelling political reasons for constraining the authority of the state to restrict speech. The Sri Lankan experience offers an important lesson in this regard.

The historical and contemporary realities surrounding the use of criminal sanctions against speech in Sri Lanka reveals an alarming legacy of abuse. The state has routinely abused its power to impose criminal sanctions on persons for speech-related offences. In many cases, the so-called ‘harm’ in the speech has been interpreted through the lens of the state’s own interests.

The Sri Lankan state has targeted outspoken journalists, political opponents, and writers. For example, in 2009, it convicted journalist J.S. Tissainayagam under section 2(1)(h) of the Prevention of Terrorism Act of 1979 for writing an article that accused the armed forces of committing war crimes.[2] The section in question criminalises causing ‘communal disharmony’ through words. Prosecutors advanced the peculiar argument that a Tamil journalist criticising a predominantly Sinhalese military of committing war crimes would cause ‘disharmony’ between Sinhalese and Tamils. Then, in 2010, the Sri Lankan state punished political opponent Sarath Fonseka, who contested and lost the presidential election that year. He was convicted under regulation 28 of the 2005 Emergency Regulations for an allegedly false statement that certain members of the Liberation Tigers of Tamil Eelam were executed while surrendering to the armed forces.[3] The specific regulation criminalised causing ‘public alarm’ through a false statement. More recently, writer Shakthika Sathkumara was arrested for writing a short story that allegedly offended certain members of the Buddhist clergy.[4] He was arrested under section 3 of Sri Lanka’s International Covenant on Civil and Political Rights (ICCPR) Act of 2007, which criminalises the advocacy of ‘national, racial and religious hatred that constitutes incitement to discrimination, hostility or violence’.

None of these cases appear to warrant criminal sanctions. In each case, the speech did not involve the conveyance of hatred, nor did it involve incitement of any kind. Instead, the speech entailed expressions of dissent, criticism, or provocative literature – all within the domain of an individual’s freedom of expression. The Sri Lankan state appears to have wantonly used criminal sanctions to curb speech that undermines its own interests. The military, the political establishment, and the clergy often define and dictate those interests. The state has been quick to prohibit speech that criticises or offends those specific actors. By contrast, at least in the recent past, the Sri Lankan state has not prosecuted a single person for inciting violence against a minority community. For example, in the thirteen years since its enactment, the ICCPR Act is yet to be used to actually convict any of the instigators of violence against Muslims or Christians in Sri Lanka. This experience suggests that there are good reasons to distrust the state when it seeks to impose criminal sanctions for harmful speech. It makes political sense to constrain the authority of the state to impose such sanctions. If the state is permitted a wide terrain of authority to punish harmful speech, it is likely – as evident in the Sri Lankan case – to target critical speech and political dissent, which are genuinely within the domain of the freedom of expression.

The role of the state

It is tempting to remain cynical about the role the Sri Lankan state can play in countering harmful speech. But this role should be, at least in theory, properly defined so that there are clear benchmarks against which state responses can be assessed. We could frame four types of state responses to harmful speech: (1) abstaining, (2) countering, (3) regulating, and (4) prohibiting.

First, in cases of harm to reputation or personal sensibilities, the state should abstain from playing any active role. It should only establish the legal infrastructure for individuals to resolve their private disputes. Civil courts must be established and equipped to hear private plaints. For example, if a statement insults a political or religious actor, the state should not intervene, but should instead provide the necessary legal channels for such actor to seek a remedy in the civil courts.

Second, in the case of hate speech falling short of incitement, the state should play a role in countering it. State officials and political leaders have a duty to promote social cohesion and equal citizenship. For instance, they must not remain silent when ideas of racial, religious or gender-based superiority are being publicly disseminated. They must play an active role in reassuring the targeted groups that the state does not condone such ideas.

Third, when incitement of some kind takes place, the state should play a role in regulating such speech. It could engage in such regulation by setting out rules of conduct, and by imposing civil sanctions on those who violate such rules. For example, the state has a duty to dismiss public officials who engage in incitement. It could even suspend or revoke the licences of businesses or media institutions that disseminate inciteful speech. Such action may be particularly appropriate when public airwaves are utilised to engage in incitement. However, widening the scope of the state’s regulatory powers over the media is extremely dangerous in the Sri Lankan context, as the state is yet to demonstrate good faith in this domain. State actors have selectively targeted media institutions and news websites that are critical of the state, while permitting pro-state media to disseminate harmful speech with impunity. For instance, the state has yet to take any regulatory action against media channels that spread highly prejudicial and often blatantly false information on detained lawyer Hejaaz Hizbullah. At the same time, the government has announced that it wishes to regulate news websites and social media in an attempt to curb hate speech and disinformation. The announcement appears to be disingenuous given the double standards that are currently in play.

Finally, the state has a very clear remit to prohibit incitement to violence. It has a duty to investigate, prosecute, and punish those who engage in such incitement. The state’s role in this domain is crucial to ensuring the protection of vulnerable communities from acts of physical violence. In essence, the use of criminal sanctions should be confined to this domain, and the state should not be permitted to impose criminal penalties for any other form of harmful speech. The reason for such constraints is simple. The Sri Lankan state has thus far failed to demonstrate good faith in exercising power over speech. Therefore, it should be entrusted with punishing only the narrowest domain of harmful speech, and only when actual physical violence is imminent as a direct result of the speech. In all other cases, the role of the state should be to abstain from intervention, counter the speech, or regulate the speech, depending on the circumstances, and the nature of the harmful speech. In a context where public trust in the state’s good faith has broken down, confining the state’s role in this manner may be the only bulwark against abuse.


Society has an important and legitimate role to play in responding to harmful speech. By extension, the state too has a crucial role to play in this domain. However, there are serious risks of abuse that need to be confronted when defining the state’s role. One way to confront this challenge is to understand harmful speech through a model of concentric circles, and properly locate ‘hate speech’, ‘incitement’ and ‘incitement to violence’ within the remit of harmful speech. The type of interference that is warranted would then depend on the type of harmful speech in question. Within this scheme, criminal sanctions ought to be reserved only for incitement of the most harmful kind, i.e. incitement to violence. It is of course clear that many countries possess laws that criminalise speech that is nowhere near as harmful as incitement to violence. For instance, some European countries prohibit speech that denies the Holocaust.[5] Leaving aside the desirability of such laws in those specific country contexts, the Sri Lankan legacy of abuse offers good reasons to adopt a minimalist approach to criminal sanctions.

We ought to recognise the state’s role in countering harmful speech through civil remedies, counter-messaging, and certain limited forms of regulation. However, the state’s ultimate authority to prosecute and punish individuals must be confined to a narrowly defined domain of harmful speech – i.e. incitement to violence. At least in contexts such as in Sri Lanka, there is a compelling case for a minimalist approach – an approach that simply does not entrust the state with extensive criminal jurisdiction over speech.

[1] See Article 19, The Camden Principles on Freedom of Expression and Equality, April 2009.

[2] ‘A travesty of justice: The sentencing of J.S. Tissainayagam’, Groundviews, 1 September 2009, at

[3] Manopriya Gunasekara, ‘The White Flag case: The reasons behind the two verdicts’, Sunday Times, 20 November 2011, at

[4] ‘Arrest of writer Sathkumara sparks debate on freedom of expression’, Daily Mirror, 12 April 2019, at–freedom-of-expression/131-165392.

[5] See Laurent Pech, ‘The Law of Holocaust Denial in Europe: Towards a (qualified) EU-wide Criminal Prohibition’ (2009) Jean Monnet Working Paper 10/09,