Can a Peaceful Protest Ever be Banned Because Others Threaten It with Violence?

Written by

What do you think, dear readers, that the answer to the question above should be under international human rights law? Any comments would be very much appreciated. To start us off, let me write out a hypo; imagine we are in happier times before (or after) Covid:

Krakatowia is a country in which LGBT rights are still a matter of huge public controversy. A local NGO devoted to the protection of LGBT rights announces publicly that a week from now it plans to organize the first Pride parade on the streets of the Krakatowian capital.

Sigmund is a retired member of the Krakatowian special forces. He has a long history of virulently homophobic posts on social media and online forums. Upon seeing the news about the planned Pride parade, Sigmund posts on various digital platforms that he will use a sniper rifle – which he has an extreme level of proficiency in – to kill people joining the parade. He then goes into hiding.

The Krakatowian police attempt to locate Sigmund, but despite their best efforts are unable to find him before the Pride parade was to take place. Despite that, the organizers of the parade wish to proceed with the protest, saying that they are willing to take the risk. On the evening before the event, the minister of justice publishes a decision prohibiting the parade from taking place, noting that while the assembly was peaceful as such, the state was unable to guarantee the safety of the participants and any passers-by and that therefore it was necessary to prohibit it.

So, in the scenario above there is a (1) peaceful assembly, in which none of the participants themselves wish to resort to violence; (2) others threaten it with violence; (3) that threat is grave and real, beyond any doubt; (4) the state did all it reasonably could have done to mitigate the threat, but was ultimately unable to do so. In such circumstances, is it necessary and proportionate to prohibit the event, because the state cannot guarantee the safety of the participants?

There are two reasons why this question came to mind. First, because the Human Rights Committee has now directly pronounced on it, in its new General Comment No. 37 on the right to peaceful assembly, which I happened to read yesterday (by the way, Just Security ran a couple of pieces on it last month). Second, because I was actually counsel, on behalf of the Belgrade Centre for Human Rights, in a series of cases before the Constitutional Court of Serbia and the European Court of Human Rights that posed this exact question of principle, which human rights bodies were (at the time) yet to answer clearly enough.

Here are the Committee’s views in GC 37, para 52:

The fact that an assembly provokes or may provoke a hostile reaction from members of the public against participants, as a general rule, does not justify restriction; the assembly must be allowed to go ahead and its participants must be protected.[fn 64] However, in the exceptional case where the State is manifestly unable to protect the participants from a severe threat to their safety, restrictions on participation in the assembly may be imposed. Any such restrictions must be able to withstand strict scrutiny. An unspecified risk of violence, or the mere possibility that the authorities will not have the capacity to prevent or neutralize the violence emanating from those opposed to the assembly, is not enough; the State must be able to show, based on a concrete risk assessment, that it would not be able to contain the situation, even if significant law enforcement capability were to be deployed.[fn 65] Less-intrusive restrictions, such as postponement or relocation of the assembly, must be considered before resort to prohibition.

So, in the Committee’s view, a peaceful assembly can be banned – exceptionally, but even so, because others threaten it with violence. The two footnotes refer to the same paragraph (9.6) of a 2013 decision of the Committee, Alekseyev v. Russia, which does not actually say, at least not in so many words, what the Committee says it does:

The State party has not provided  the Committee with  any information in  the  present  case  supporting  the  claim  that  the  “negative  reaction”  to  the  author’s proposed  picket  by  other  members  of  the  public  would  involve  violence or that  the  police would  be  unable  to  prevent  if  they  properly  performed  their  duty.  In such  circumstances, the obligation of the State party was to protect the author in the exercise of his rights under the  Covenant,  and  not  to  assist  in  suppressing  them.

This brings me to the cases we litigated regarding the 2009, 2011, 2012 and 2013 Pride parades in Belgrade, the first of which was ‘relocated’ by the interior ministry to a middle of nowhere in the city, while the others were banned outright, on the basis of threats of violence by counter-protesters. We ‘won’ the cases before the Serbian Constitutional Court on purely procedural grounds (because the decisions were made on the eve of the protests with no possibility of recourse to a judicial remedy). The Serbian legislature amended the law on assemblies after the Constitutional Court indicated that it should do so; in more recent times, under the increasingly authoritarian Vucic regime, Pride parades have been taking place in Belgrade regularly, with no violence. And in a rather horrid decision, the European Court dismissed, by a majority, the case we filed in Strasbourg because the matter had been resolved domestically, despite the fact that adequate redress was not provided and that the only violations found were procedural. Indeed the Court even dismissed our non-discrimination claim as manifestly ill-founded, despite the fact (among others) that the minister of the interior who made the decisions prohibiting the parades later accepted a medal (seriously, he did) for refusing to allow the gays to march in Belgrade.

But I digress – let’s go back to the question of principle. In the Serbian cases we first made a categorical argument, saying that no peaceful assembly could ever be prohibited simply because others threaten it with violence that the state cannot completely prevent. Holding otherwise would effectively introduce a hecklers’ veto into human rights law, even in situations where the protesters were willing to assume risk to themselves (as they often do) from violent counter-demonstrators. This is not, however, the position that the Committee took in GC 37 – far from it.

In the alternative we did argue, in alignment with the Committee’s position, that the state could only prohibit a peaceful assembly on safety grounds if it took all feasible proactive measures to ensure the safety of the assembly, but was ultimately unable to do so. We thus argued that much of the risk assessments that the Serbian police had done were unreliable and that insufficient preventative measures were taken. For example, the police claimed to have had intelligence that anti-gay football hooligans were amassing stones and crowbars in the basement of an apartment block in the city centre, without actually having gone there themselves to check whether this was true or not. Or, and this is my absolute favourite, the police claimed to have had intelligence that the hooligans were ‘collecting honeybees into bottles and jars, which they would then at some point release into the assembly area’ against the peaceful protesters and the police. The mental image of football hooligans prancing in the meadows to collect bees is priceless – kudos are due to the Serbian police for creative writing, if nothing else.

But, as I explained, the Court never reached any of these questions, and I don’t think it has done so since in some other case (please do correct me if I am wrong). Under the Committee’s approach, the bans on the Serbian Pride parades would have been unjustified because the Serbian authorities did not take adequate preventative actions. But in the hypo I gave at the beginning, the position is more stark. My concern with the Committee’s position is, I hope, clear – it enables the shutdown of peaceful protests due to anticipated violence by others. Maybe that’s right, maybe that’s wrong, but the issue requires a bit more consideration, I would say, than that one overly emphatic and insufficiently footnoted paragraph in GC 37. But perhaps I am being too harsh? Again, any comments are very welcome.

Print Friendly, PDF & Email

Leave a Comment

Your comment will be revised by the site if needed.


N. Mopey says

August 26, 2020

Nice post, thank you. I don't have a full answer to your larger question on the GC's approach to this issue, save to say that you've perhaps baked so much into your hypothetical that it's really a limit case, which boils down to the question of whether a state can *ever* prevent people from assuming some risk to themselves (and, in your hypo, perhaps also to passersby) that they are happy to assume in the pursuit of some important good (or exercise of some important right), where the risk is sufficiently grave and certain. To which the answer surely must be yes (but immediately less so as we move away from limit cases - and the question here I suppose is whether the GC tends in that direction at the right gradient).

Kirsty Hughes says

August 26, 2020

Thanks for this interesting hypothetical Marco! Here are my initial reactions. First, the HRC comment if anything advances a far stricter level of scrutiny than the Strasbourg Court. The Strasbourg Court tends to be deferential in the face of security arguments in the protest case law, woefully so in some cases. Indeed as you note in some cases it doesn’t even address the central issues. Thus in the scenario you present I am sure the Strasbourg Court would accept the shutdown of the protest if the criteria established by the HRC were met, and probably with a far lesser degree of scrutiny than proposed in the Comment.

I too, am in general concerned about the prospect of shutting down peaceful protest due to the threat of violence from others, but the hypothetical you map out is a scenario in which I would accept that this is necessary, provided that a strict level of scrutiny was in place, alternatives had been duly considered, and that there were other avenues available for the demonstrators. In other words the whole context would need to be considered to assess the legitimacy of the state’s argument here. I am not sure that I can see a way in which the Court could rule that a peaceful protest could never be prohibited given in the scenario that you present it would need to consider Article 11 ECHR vis-à-vis the states positive obligations under Article 2 ECHR. In fact your scenario neatly encapsulates why the Court should not do that. Moreover, as you note in your scenario the protestors are not the only ones potentially at risk here, passersby are also implicated, and thus any ‘consent’ to ‘run the risk’ on the part of the protestors cannot be determinative here.

Returning to the HRC GC though I think there are various statements that reassure me that an appropriately high standard of scrutiny is intended:
e.g. at [37] that ‘the prohibition of a specific assembly can be considered only as a measure of last resort’.
At [38] that ‘any restrictions on participation in peaceful assemblies should be based on a differentiated or individualized assessment of the conduct of the participants and the assembly concerned.’
At [40] that and ‘measures must also be the least intrusive among the measures that might serve the relevant protective function’.
At [42] that the “interests of national security” may serve as a ground for restrictions if such restrictions are necessary to preserve the State’s capacity to protect the existence of the nation, its territorial integrity or political independence against a credible threat or use of force. This threshold will only exceptionally be met by assemblies that are “peaceful”.
This all seems to me stronger than I have seen in the ECHR case law where it is often difficult to see how protestors could ever really challenge assertions of security needs by the state - apart from in cases where the state’s argument is so obviously false that it falls apart in the face of obviously inconsistent evidence e.g. where another demonstration has been permitted instead.

So to sum up my answer to what I think are the questions that you pose: (i) as a matter of law yes there are exceptional cases where states can ban peaceful protests where they are threatened with violence by others; (ii) on a normative level – yes I think there are exceptional circumstances where that is justified as you map out in your hypothetical; (iii) I am not as immediately concerned about this aspect of the HRC GC as you are as I think the broader ethos of the GC is stronger than the stance of the Strasbourg Court, of course it comes down to interpretation and application but it is certainly more reassuring than some of the cases (iv) I do have reservations though as to how the stricter standards that the GC suggests may be watered down in practice.

Best wishes, Kirsty

PS – the absurdity of the bees is reminiscent of the claim in Disk that the police officer did not deliberately set off a gas bomb outside a hospital, he just sat on it by mistake.

Marko Milanovic says

August 26, 2020

Fantastic, thanks a lot both for your comments. Really helpful. Can I push back a bit on the categorical point by changing the right that's at stake?

Let's imagine a scenario in which a newspaper wants to publish a provocative piece, knowing with certainty (based on past experience) that if they did so third parties will engage in violence, either against the journalists themselves or against by-standers? Think, say, the cartoons of the Prophet Muhammad. There's no doubt about it, if the newspaper publishes them, people will die. Would you, following the same logic as in the assembly context, accept that in extremis the freedom of speech could be restricted and publication prohibited?

Or, to give another variant, imagine the Sigmund sniper guy makes a threat to the life of one specific person. Is it okay for the police to deprive that person of their liberty and put them in protective custody against their will, on the basis that there's no other way of guaranteeing that the sniper wouldn't kill them, a member of their family, or a passer-by?

Thanks again!

Goran says

August 26, 2020

In general, I don't think you're being too harsh. Taking into account the place of the Committee in the human rights protection system and its expected normative role, it would be reasonable to expect it not to give ANY way for states to limit peaceful protests without extreme and certain danger. The states should not be relaxed about limiting the freedom of peaceful assembly. On the other hand, the preferences of the protesters assuming risks are not solely a legal question, but a great question for political philosophy.

Kirsty Hughes says

August 26, 2020

I am just going to tackle the first of those scenarios for now as I am in the middle of something else. My initial reaction here is that there are some important issues that need to be teased out further.
In your protest hypothetical the prohibition was permissible only as a temporary measure and there were important temporal constraints – the demonstration was constrained in time, and the threat to life was limited to that – the problem was that the known threat could not be located in time for the event to go ahead as planned. There were also presumably other means available to the protestors to express their views such that there was no complete prohibition on expression even in the interim whilst the demonstration itself was prohibited. In the freedom of expression hypothetical, however it seems that there is potentially a complete prohibition on expression, moreover, that this is not temporally limited. I am not sure in this example whether or not the third parties are identifiable or not. I am also not sure in this scenario whether their response will be directly targeted at the journalists (albeit that others may be caught up in this) or whether their response may be anywhere, any time?

Kirsty Hughes says

August 26, 2020

PS on the latter scenario is it intended that there is supposed to be some sort of legal framework operating here to confer such powers on the police and that this purports to be compatible in some way with article 5 ECHR?

Marko Milanovic says

August 26, 2020

Hi Kirsty,

Thanks for this. I'm happy to stipulate whatever additional facts are necessary, so long as the core problem remains. So in the cartoons hypo we could (but need not) stipulate a temporary ban on publication, due (say) to heightened tensions in the country and greater risk of violence. But deaths could also result abroad (as they did with the Danish cartoons) without changing anything fundamentally about the basic dilemma, in my view. Similarly, we could stipulate some limited duration for protective custody in the second hypo, say a month. And yes assume whatever other elements you need (e.g. there is a legal framework that enables some kind of protective order to be made). Again the core dilemma is whether state should decide on what level of risk is tolerable due to violent actions of others, and on that basis limit the rights of a non-violent person against their own assessment of acceptable risk.

Ed Robinson says

August 26, 2020

Hi Marko, fascinating piece, thank you!
I see Kirsty has touched on positive obligations, which seems like an interesting angle to flesh out further. If the HRC position is correct I wonder whether on your hypothetical the state is obliged to postpone, relocate or even ban the assembly, given the last resort conditions which make such restrictions permissible from a right to assembly perspective. I.e. if direct measures against the perpetrator and protective measures around the protestor would both be ineffective then might the authorities actually have to take the only reasonably effective option open to them and stop the assembly (maybe compare the facts with Tagayeva and others v Russia regarding protection of the ‘Day of Knowledge’ celebrations)? I wonder if states would find that outcome congenial when applied to state-sponsored public assemblies, parades etc under threat of e.g. terrorist violence…
Of course much depends on the transparency and rigour of the risk assessment process, and that’s a concern in ECtHR-world for the reasons Kirsty has noted. But in principle if one were to say that such extreme restrictions on assembly require a positive obligation-triggering ‘real and immediate risk’ that adds a bit of control, since we have some jurisprudence on what that can and can’t be. It also forces states to accept the positive obligation ‘burden’ which follows from the permissive (ability to restrict assemblies) ‘benefit’. Positive obligations aren’t a complete answer here but might be part of the answer.
Looked at through the positive obligations lens I’d then argue that the first new hypothetical can be distinguished since there isn’t the specificity as to the target or perpetrator which would trigger an obligation to take specific preventive measures. Again that needn’t be dispositive as to the permissibility of the intrusion on free expression, but it does seem a legitimate basis to distinguish the publication scenario from the protest scenario.
Thanks again all

Kirsty Hughes says

August 26, 2020

As Ed comments positive obligations require the state to make its own assessment of the level of risk, and potentially to take measures, I think this addresses the core dilemma you identify Marko, namely ‘whether state should decide on what level of risk is tolerable due to violent actions of others, and on that basis limit the rights of a non-violent person against their own assessment of acceptable risk’. Of course that assessment has to be subject to oversight, and the non-violent person may contest both this assessment and seek to advance their own view as to risk that they personally are willing to tolerate (although as noted it is unlikely that the risk is limited to them in any event).

Where I think the temporal element and specificity comes in is in determining (i) the scope of the potential interference with the right and (ii) the clarity of the need for that interference – i.e. the likelihood and the extent of the interference with the rights of others – and the potential positive obligations that may attach to that. Indeed I think that in the way that the scenarios are formulated that it is the degree of the interference, the nature of the threat and the need for the measures that distinguishes them. I think what you are really wanting to get at though is whether the underlying nature of the right or our perceptions of the rights are different such that we would regard an equivalent restriction on freedom of speech as unacceptable, and therefore whether that in turn requires us to go back to consider whether such a restriction on freedom of assembly should also be a red line.

I am trying to think whether an adaptation of your first hypothetical to freedom of expression might aid with this perhaps something like:

Krakatowia is a country in which LGBT rights are still a matter of huge public controversy. A local NGO devoted to the protection of LGBT rights announces publicly that a week from now it plans to publish a book on LGBT rights.

Sigmund is a retired member of the Krakatowian special forces. He has a long history of virulently homophobic posts on social media and online forums. Upon seeing the news about the planned book, Sigmund posts on various digital platforms that he will use a sniper rifle – which he has an extreme level of proficiency in – to kill
(a) the author
(b) those connected to the publication and sale of the book
(c) anyone who purchases or read the book
(d) members of the public
He then goes into hiding.

The Krakatowian police attempt to locate Sigmund, but despite their best efforts are unable to find him before the book is due to be published. Despite that, the author and publishers wish to proceed, saying that they are willing to take the risk. On the evening before the event, the minister of justice publishes a decision prohibiting the book, noting that the state is unable to guarantee the safety of the public and that therefore it was necessary to prohibit it.

Incidentally I think your earlier third liberty scenario provides an interesting example for considering alongside covid liberty restrictions.

Florian Kriener says

August 26, 2020

Hi Marko,
thanks a lot for this hypothetical and the thought provoking post.

I would agree that the threat to the physical integrity and life of the protesters in the hypothetical is real, grave, and beyond doubt. I would, however, disagree on whether the state has done everything it can to prevent the threat from becoming real. As GC 37 states, less intrusive measures such as the relocation is preferable. In the hypothetical,protests could be moved to a place with few vantage points and receive adequate protection through police forces. This could involve a very comprehensive police detail screening the assembly point before the assembly commences. This would, of course, not eliminate the threat. However, all participants of assemblies face certain threats to their physical integrity and life, even without outside threats to the assembly. This involves the possibilty of mass panics that are inherent when masses of people assemble. This risk can be mitigated to a far extent. It nonetheless cannot be banned.
What I want to say with this comparison is that the right to physical integrity and life as proscribed in Art. 2 ECHR is not absolute and requires balancing with the fundamental importance of Art. 10 (or Art. 21 ICCPR). By consegrating the freedom of assembly in these treaties its drafters acknowledged this importance and proscribed the pertinence of assemblies in democratic societies. States should go to great lengths to protect assemblies and accept a certain threat to life and physical integrity. Only if the threat is overwhelming banning an assembly should be an option. In your hypothetical I would not see this threshold met. It would be different if multiple retired special forces threatened the assembly in a concerted effort.

In November 2015, the German Federal Interior Minister cancelled a football match
in Hannover between the German and Dutch national teams because of terrorist threats. The Bataclan attacks had devastated Paris only few days earlier and there were specific threats by a terrorist group against the match. I think that such circumstances would likewise justify the prohibition of an assembly.

Marko Milanovic says

August 26, 2020

Many thanks again to everybody for the super helpful comments.

(1) I'm very happy to vary the immediacy or nature of the threat aspects of any hypothetical. But this only goes to test our intuitions as to whether we're comfortable with some kind of in extremis restriction - the issue is how extreme the danger has to be. The bigger issue of principle, however, is whether a categorical rule is appropriate or not even in an extreme situation.

(2) While the positive obligations angle is obviously relevant, I think it only gets us so far. Imagine a hypo where there is a direct and immediate threat to the lives of a group of people in SOME OTHER COUNTRY, say in a town just across the border. (Again the violence in response to the Danish cartoons took place all over the world). Mainstream doctrine does not generally argue that positive obligations would attach in such circumstances, i.e. that state A would have the duty to protect the people of state B against violence. But surely the question of principle should not be answered any differently. If it is justified to restrict the freedom of assembly, speech or liberty of person to protect a segment of the state's own population from immediate harm, it must also be justified to do so to protect some other people elsewhere. Their lives must have equal value in this context.

(3) I guess my central problem with this is when is state paternalism appropriate. It does seem open to abuse for the state to be able to 'protect' people against third parties when they are aware of the risk and willing to take it for achieving some greater public good, especially in the context of political protest and speech (let's leave aside football games for the moment). I totally agree Kirsty that this has relevance with regard to Covid too. The difference here is that the harm is being caused by an independent moral actor who engages in violence, not by a virus, and again the issue in principle is whether the rights of the VICTIM can be curtailed because a third party is threatening THEM with violence. On the other hand I also agree that the position of by-standers is more complex in the sense that they do not decide willingly to take a risk upon themselves, and therefore state paternalism becomes more appropriate.

(4) Just a final note on the whole relocation business, which sounds like such a reasonable human rights proportionality line of argument to take - sometimes a message can only be conveyed effectively in a certain time and place. This is exactly what happened to our LGBT clients in the first Belgrade Pride case, where the government (with no legal basis domestically to do so) 'relocated' the protest from the city centre to a location where no public assembly had ever taken place, in which the Pride message would have completely lost its valence (indeed the clients chose not to proceed with the parade at all under those conditions).

Marvin Hips-Jacobs says

August 30, 2020

My thinking is:
The prevention of crime and disorder are legitimate grounds on which to interfere with the right to protest. Nowhere does it say that the crime and disorder to be prevented must have been at the hands of the protesters. Restrictions are not punitive measures directed at the protesters but preventative measures for the public good.

It follows that there can be no volenti argument. It is not exclusively in the gift of any peaceful protesters to demand an unfettered right to protest as if the prospect of crime and disorder only affected them (and not society as a whole, a fortiori more serious offences).

Incidentally, even in protests which are restricted because of the prospect of disorder within the ranks of those participating (rather than from outsiders reacting to them), the disorder risk inevitably stems from a small minority of fringe protesters - often condemned by their peers - meaning that the majority of protesters are also subjected to restrictions not causally related to their own conduct.