Freedom of Expression and the Advocacy of Violence
| DOI | 10.1177/016934411503300105 |
| Published date | 01 March 2015 |
| Author | Andrew Dyer |
| Date | 01 March 2015 |
| Subject Matter | Part B: Article |
Netherlands Qu arterly of Human Ri ghts, Vol. 33/1, 78–107, 2014.
78 © Netherlands I nstitute of Human Rig hts (SIM), Printed in the Net herlands.
FREEDOM OF EXPRESSION AND
THE ADVOCACY OF VIOLENCE
WHICH TEST SHOULD THE EUROPEAN COURT OF
HUMAN RIGHTS ADOPT?
A D*
Abstract
is article examines the European Court of Human Rights’ violent advocacy
jurisprudence . It observes that, since the decision in Leroy v France was delivered in
October 2008, there ha s emerged some confusion concerning what the applicable tes t is
in cases where it is alleged that a Member State’s proscription of such speech amounts
to a breach of Article10 of the European Convention on Human Rights. In particular,
there has been growing suppor t in the Strasbourg case law for tests that are similar to the
US Supreme Court’s ‘clear and present danger’ stand ard. e argume nt presented here
is that the European Court should adopt a test under which there is but one enquiry:
‘did the impugned speech create a real risk of violence?’ However, in answering that
question, the Court, in every such case, should consider ve further matters: ‘did the
speech create an imminent risk of harm?’; ‘did the applicant intend to cause violence
or foresee the possibility that his/her speech would cause violence?’; ‘did he/she occupy
a position of in uence in society?’; ‘was the advocacy disseminated widely?’; and
(where applicable) ‘did the speech occur close to the ‘centre of violence’?’ If faithfully
applied, this test would achieve a prope r balance between, on one hand, protecting the
community from violence, and , on the other, preventing states from placing unnecessary
restrictions on those who impart and rec eive information relevant to political and other
decision-making. at is, it would en sure that the Court upholds only those restrictions
on violent advocacy that fu l l a ‘pressing social need’.
Keywords: advocacy of violence; European Court of Human Rights; Freedom of
expression; test for
* Andrew Dyer is t he Colin Phegan Lecturer i n Legal Reas oning at the Universit y of Sydney Law
School, Australia.
Freedom of Expression a nd the Advocacy of Violence
Netherlands Qu arterly of Human Ri ghts, Vol. 33/1 (2015) 79
1. INTRODUCTION
In Leroy v France1 the European Cou rt of Human Rights (ECtHR or C ourt) held that
France had not violated Art icle10 of the European Convention for the Protection
of Human Rights and Fundamental Freedoms (ECHR) by convicting a newspaper
cartoon ist of a glori cation of terrori sm o ence. Some commentators considered that
the Court in Leroy departed from the approach that it took in its earlier case law
concerning the advocac y of violence.2 In those ea rlier cases, the C ourt had insisted
that a conviction for advocating v iolent activity would only be compatible with
Article 10 if the applicant’s expression was ‘liable to’ result in v iolence/‘capable of’
resulting in v iolence.3
is article does four things. First , it discusses why the ECtHR should give any
protection to speech that advocates acts of violence. It is argued that some protection
is due to such advocacy mainly because of its ability to a id democratic and other
decision-making and to allow individuals to ach ieve self-ful lment. If a ll such
advocacy were prohibited, the public’s ability to receive inform ation about matters of
general interest would be unduly compromised, leg itimate speech would be ‘chil led,’
and counter-productive results might be produced . Moreover, if governments were
able to prohibit all such speech, there would be a d anger of them abusing such a power.
Second, the art icle contends that, although commentators’ claims that Leroy
departed from the ECt HR’s previous case law must be quali ed slightly, this does
not mean that that decision wa s a desirable one. Rather, the fact that the Leroy Court
purported fait hfully to apply the ‘capable of’ test i ndicates how open to manipulation
that test is. It gives the Str asbourg Court too much scope, in times of pa nic, to uphold
prohibitions on allegedly ‘violence-c onducive’4 spe ech that is, in fact, innocuous.
1 Leroy v France App no 36109/03 (ECtHR, 2O ctober 2008).
2 See for example Stefa n Sottiau x, ‘Leroy v France: Apolog y of Terrorism and the Mal aise of the
European Cou rt of Human Rig hts’ free sp eech jurispr udence’ (2009) 3 European Human Right s
Law Review 415, 420; Dirk Voorhoof, ‘European Court of Hu man Rights: W here is the Chilling
E ec t?’ (15October 2008)
en.pdf> acce ssed 12 March 2014; Peter Noorla nder, ‘Recent Trends in Strasb ourg Case Law –
Freedom of Expression i n Decline’ (Free Speech L aw 22October 2008)
blogspot.com.au/2008/10/recent-trends-in-strasbourg-caselaw.html> accessed 12 March 2014;
and Committ ee of Experts on Terrorism (CODEXTER), ‘Free dom of Expression and Apologie Du
Ter ro ri sm e’ (15th meeting , Strasbourg, 24 –25November 2008) 2.
3 Arslan v Turkey (2001) 31 EHRR 9; Baskaya and Ok çuoglu v Turkey (2001) 31 EHRR 10; Ceylan v
Turk e y (2000) 30 EHR R 73; Erdoğdu and Ince v Turkey App nos 25067/94 and 25068/9 4 (ECtHR,
8July 1999); Gerger v Turkey App no 24919/94 (ECtHR, 8 July 1999); Karataş v Turkey App no
23168/94 (ECtHR, 8July 1999); Ok çuoğlu v Turkey App no 24246/94 (ECtHR , 8July 1999); Polat v
Turk e y App no 23500/94 (ECt HR, 8July 1999); Sürek v Turkey (No. 1) App no 26682/95 (ECtHR,
8July 1999); Sürek v Turkey (No. 3) App no 24735/94 (ECtHR, 8Ju ly 1999); Sürek v Turkey (No. 4)
App no 24762/94 (ECtHR, 8July 1999); Sürek and Őz demir v Turkey App nos 23927/94 and 24277/94
(ECtHR, 8July 1999).
4 Stefan Sotti aux, ‘“Bad Tendencies” in the E CtHR’s “Hate Speech” Jur isprudence’ (2011) 7(1)
European Constitutional Law Review 40, 60; Stefa n Sottiaux, Terrori sm and the Limitation of Rig hts:
Andrew Dyer
80 Intersentia
ird, the art icle examines the ECt HR’s approach in cases concerning the
advocacy of violence that have been decided since Leroy.5 Interestingly, in some
cases the Cour t has not applied the ‘capable of’/‘liable to’ test. Rather, there has been
growing support for tests t hat are similar to the US Supreme Cour t’s ‘clear and present
danger’ standard. i s is a noteworthy development: as recently as 2003, Sottiaux was
able to observe that, although the ‘clear and present danger’ test had in uenced the
Court’s case law, ‘the European Cou rt has never adopted a “clear and present danger”
test or anythi ng like it.’6
Fourth, the ar ticle considers which test the Court should apply in advocacy of
violence cases. It is arg ued that the Court should adopt a test under which t here is only
one inquiry: ‘d id the impugned speech create a real risk of violence?’ In determining
whether a ‘real risk ’ existed, the C ourt should be required, in all such cases, to take
into account both the imminence of the threatened harm a nd whether the speaker
either intended or foresaw the possibility of it. It should al so consider three questions
identi ed by Judge Palm in a nu mber of cases decided on 8July 1999,7 namely, whether:
the speaker occupied a position of in uence i n society; the advocacy was d isseminated
widely; and (where this is applicable) the speech occurred close to the ‘centre of
violence.’ is test is better able than ot hers – including the version of the ‘clear and
present danger’ test that is cur rently in favour in the US8 – to achieve a proper balance
between, on one hand, the need to protect t he community from violence, a nd, on
the other, the interests in ensur ing that: (a) the public receives as much information
that is relevant to democratic a nd other decision-making as is possible; (b) indiv idual
autonomy is not unduly compromised; and (c) there is proper scrutiny of states’ claim s
that their restr ictions on the advocacy of violence ful l a ‘pressing socia l need.’
5 Demirel and Ates v Turkey (No. 3) App no 11976/03 (ECtHR, 9 December 2008); Saygili and
Falakaoglu v Turkey (No. 2) App no 38991/02 (ECtHR, 17February 2 009); Gul and Others v Turkey
(2011) 52 EHRR 38; Kilic and Eren v Turkey App no 4 3807/07 (ECtHR, 29 November 2011); and
Faber v Hungary App no 40 721/08 (ECtHR, 24July 2012). And see text acc ompanying nn 135–152.
6 Stefan Sott iaux, ‘“ e Clear and Pre sent Danger” Test in the Case Law of the Euro pean Court of
Human Right s’ (2003) 64(3) Zeitschri für Rec ht und ausländis ches ö entliches Völkerrecht 653, 678.
7 See cases li sted at n 3.
8 Brandenburg v Ohio 395 US 4 44, 447 (1969) where, as noted below (n 123), the US Supreme Court
held that ‘the const itutional guara ntees of free speech and f ree press do not permit a State to forbid
or proscribe advoc acy of the use of force or of law violation except where such advocacy is directed
to inciting or producing imminent lawless action and is likely to incite or produce such action.’
[emphasis added] A very sim ilar test was accepted by t he dra ers of t he ‘Johannesburg Principles
on National Secu rity, Freedom of Expres sion and Access to In formation’ (November 1996) U N
Doc E/CN.4/1996/39, which, sig ni cantly, was adopted wel l before the events of 11September 2001.
Principle 6 of the Joha nnesburg Principles relev antly provides that ‘… expression may be pu nished
as a threat to nation al securit y only if a gover nment can demonstr ate that: (a) the expression i s
intended to incite im minent violence; (b) it is likely to i ncite such violence; and (c) there is a direc t
and immediat e connection bet ween the expres sion and the like lihood or occu rrence of such
violence:’ w.article19.org/data/ les/pdfs/standard s/joburgprinciples .pdf> access ed 12M arch
2014.
Freedom of Expression a nd the Advocacy of Violence
Netherlands Qu arterly of Human Ri ghts, Vol. 33/1 (2015) 81
2. WHY SHOULD THE ADVOCACY OF VIOLENCE/
TERRORISM BE PROTECTED AT ALL?
Before discussing the level of protection that the ECtHR should gra nt speech
that advocates violence, it is necessa ry to consider why such advocacy should be
protected at all. For, as a number of scholars have noted,
9
it might appear at rst
glance that, if the government may prohibit particular conduct, it should also be
able to silence all those who counsel such conduct.
A number of justi cations for
permitting such speech have been advanced. e four main such rationales are the
same as those that a re used to justif y the right to free speech more generally:
10
the
‘truth rationale’; the ‘democr acy rationale’; the ‘self-ful l ment/autonomy rationale’;
and the ‘dist rust of government rationale.’ It is necessar y to discuss t hese in turn,
because, as wil l be seen, they have each in uenced particular tests t hat di erent
members of the Strasbourg Cou rt have applied – or suggested should apply – in
advocacy of violence cases . In other words, only by determining whether such
justi cations, or any of them, provide sound reasons for protecting speech of this
nature, can we properly evaluate t he tests that ECtHR judges so fa r advanced and
make recommendations about the standa rd that should be adopted. A er discussing
these rationales, some pragmatic reasons for allowing at least some advocacy of
violence will b e considered.
2.1. THE ARGUMENT FROM TRUTH
As presented by J.S. Mill, t he argument from truth has two prongs. First, Mi ll held
that, because the government is fallible, if it were to prohibit speech simply because
it disagreed wit h it, the government would – at least somet imes – be suppressing
ideas that are true .11 Second, Mil l contended that even where the government can be
sure that an idea is false, it should not prohibit its expression. is is because if the
government’s competing true idea is not ‘f ully, frequently and fearlessly’ questioned,
the reasons why it is true will be obscured or lost: t he idea will come to be ‘held as a
9 See for example T.R.S . Allan, ‘Citizenship and O bligation: Civil Disobedienc e and Civil Dissent’
(199 6) 55(1) C ambridge Law Journal 89, 111; and omas Healy, ‘Brandenburg in a Time of Terror’
(200 9) 84 Notre Dame Law Re view 655, 682.
10 ese and (in some c ases) other, less commonly a dvanced free sp eech rationale s are noted in
Eric Barendt, Freedom of Speech (2nd edn, Oxford Unive rsity Press 2005) 6–23; omas Emerson,
‘Toward a General e ory of the First Amendment’ (1963) 72(5) Yale Law Journal 877, 878–886;
Kent Greenawalt, ‘Free Spe ech Justi c atio ns’ (19 89) 89 (1) Columbia Law Rev iew 119, 130 –155; an d
Frederick Schauer, Free Spe ech: a Philosophical Enquir y (Cambridge University Press 1982) 15–86 .
See also R v Sec retary of State for the Home Depa rtment, Ex parte Simms a nd Another [2000] 2 AC
115, 126 (HL, Lord Steyn) and t he Canadian Supreme Cour t case of R v Keegstra [1990] 3 SCR 697,
[92] – [94] (Dickson CJC).
11 John Stuar t Mill, On Libert y (Penguin Books 2010) 27–28.
Andrew Dyer
82 Intersentia
dead dogma, not as a livi ng truth.’12 e c lear implication is that, where true and fa lse
ideas compete with one another, the tr ue ideas will always be accepted.
In a series of US Supreme Court decisions, Holmes J expressed a simi lar
philosophy. In so doing, his Honour created the ‘clear and present danger’ test , which,
as noted above and discussed b elow, has recently gained increa sing support from the
Strasbourg Court. Holmes J rst stated this test in Schenck v United States,13 holding
that in every cas e concerning the advocacy of unlawfu l activity:14
… [t]he question… is whether the words used a re used in such circu mstances a nd are
of such a nature as to create a clea r and present danger th at they wil l bring about the
substantive ev ils that Congre ss has a right to prevent. It is a quest ion of proximity and
degree.
Certain ly, assuming that Holmes J even applied t he ‘clear and present danger’ test
in Schenck,15 he did so in no very rigorous way: it was far from clear that the speech
suppressed by the Court in t hat case presented a danger of the k ind to which his
Honour referre d.16 How eve r, by the ti me o f hi s fa mou s d iss ent in g ju dgm ent in Ab rams
v United States,17 Holmes J had d eveloped a full-blow n theory of spe ech that matched
his earlier rhetoric. A er noting t hat ‘time has upset many g hting faiths,’ his Honour
deployed his famous ‘marketplace of ideas’ metaphor, stating that ‘the ultimate good
desired is better reached by free t rade in ideas… [T]he best test of truth is the power of
the thought to get itself accepted i n the competition of the market.’18 He then added:19
I think that we should be eternally vigilant against attempts to check the expression of
opinions that we loathe and believe to be fraught with death, unless they so imminently
threaten immed iate interference with the lawf ul and pressing purposes of t he law that an
immediate chec k is required to save the countr y.
In the later case of Whitney v C alifornia,20 Brandeis J (in an opinion in which Holmes J
joined) adopted this reasoning. Spe ech that advocates ‘serious violence,’21 he thought,
may only be prohibited where ‘the incidence of the evi l apprehended is so immi nent
12 Ibid at 52.
13 Schenck v United States 24 9 US 47 (1919).
14 Ibid at 52.
15 David G. Barnum, ‘ e Clear and Present Dan ger Test in Anglo-Americ an and Europea n Law’
(2006) 7 San Diego Inte rnational Law Journ al 263, 272.
16 See for example Ma rtin Redi sh, ‘Advocacy of Unlawfu l Conduct and the Fi rst Amendment: I n
Defense of Clear a nd Present Danger’ (1982) 70(5) California Law Review 1159, 1166 –1167.
17 Abrams v United States 25 0 US 616 (1919).
18 Ibid at 630.
19 Ibid.
20 Whitney v California 274 US 357, 376–377 (1927).
21 Ibid at 376.
Freedom of Expression a nd the Advocacy of Violence
Netherlands Qu arterly of Human Ri ghts, Vol. 33/1 (2015) 83
that it may befall before there is an opportunity for fu ll discussion.’22 Brandei s J
implied that, where there is such an opportunity, all ‘falsehood and fallacies’ will be
exposed.23 Accordingly, his Honour approved the ‘clear and present danger’ test f rom
Schenck (though he subst ituted the word ‘immi nent’ for ‘present’). It was ‘settled,’
according to Brandeis J, that speech may only be restricted, ‘if it would produce, or
would be intended to produce, a clear and im minent danger of some substantive evil
which the State constitutiona lly may seek to prevent’.24
e aws in Mil l and Holmes and Brandeis JJ’s reasoning have been w idely
noted.25 e greatest problem with the argument from truth is its assumption that,
if socially worthwhile and dangerous ideas compete with one another, the socially
worthwhi le ideas will win acceptance. As Barendt notes, there are h istorical examples
of pernicious ideas winning favour in societies where freedom of speech was given
reasonably st rong protection.26 Accordingly, it is impossible to accept the idea that,
because full discussion wil l reveal the ‘violence-conducive speech ’s’ falsehood, the
advocacy of violence should be permit ted unless there is no opport unity for such
discussion. It follows that it is also impossible to accept the idea that the Strasbourg
Court should adopt a ‘clear and present danger’ test of the type favoured by Holmes
and Brandeis JJ – and, increasingly, in the ECtHR’s case law. is point will be
developed more fully below.27
2.2. THE DEMOCRACY RATIONALE
e second main justi cation for free speech – the democrac y rationale – is associated
mainly with t he writings of the American philosopher, Alexander Meiklejohn.28
Meiklejohn based his argument on the text of the US Constitution. According to
him, when the First Amendment of that Constitution provides that ‘Congress sha ll
make no law… abridging the freedom of speech,’ it means what it says: this freedom
is absolute.29 is does not, however, mean that Americans may say whatever they
want.30 Rather, for Meiklejohn, because t he Constitution’s purpose is to create a
22 Ibid at 377.
23 Ibid. Speci c ally, he said: ‘[i]f there be time to ex pose through d iscussion the fa lsehood and fal lacies,
to avert the evi l by the processes of educ ation, the remedy to be applied i s more speech, not enforced
silence.’
24 Ibid.
25 See for example Barendt (n 10) at 8–13; Redish (n 16) at 1162; Schauer (n 10) at 25–34; and Wojciech
Sadurski, Freedom of Speech and it s Limits (Kluwer 1999) 8–16.
26 Barendt (n 10) at 9.
27 See text accompany ing nn 179–184.
28 Alexander Meiklejoh n, Political Freed om: e Constituti onal Powers of the Peopl e (Harper and
Brothers 1960) and Alex ander Meiklej ohn, ‘ e First Amendment is a n Absolute’ (1961) e
Supreme Court R eview 245. See als o Cass Sunstein, Democracy and the Probl em of Free Speech ( e
Free Press 1993).
29 Meiklejohn, Politica l Freedom (n 28) at 20.
30 Meiklejohn, ‘ e First Amendme nt is an Absolute’ (n 28) at 249.
Andrew Dyer
84 Intersentia
system of government by the people, and because t he guarantee of freedom of speech
is an essential pa rt of this scheme, only politica l speech is covered by the absolute free
speech guarantee. at is, no speech that is relevant to a public policy issue may be
prohibited. is is so irrespective of how wrong-headed the speech might appear to
be or how likely the speech is to cause v iolence. For to prohibit such speech would be
to prevent citizens from part icipating as ful ly as they could in democratic decision-
making:31
… [t]he welfare of the community requ ires that those who dec ide issues should understa nd
them. ey must know what t hey are voting about. And this, i n turn, requires that… al l
facts and interest s relevant to the problem shall be fu lly and fairly presented…
As noted below,32 the ECtHR has rightly accepted the democracy rationale for free
speech. It follows that, when framing a test for advocacy of violence cas es, the Court
should take into account the need to faci litate discussions of political issues by the
people. However, two points must be made. First, this does not mea n that the Court
must prov ide absolute protection to advocacy of violence that is also political speech:
apart from anything else, un like the First Amendment of the US Constitution,
Article10(2) of the ECHR provides clearly t hat the right to freedom of expression is
a quali ed one. Second, the need to facilitate democracy is not the only matter that
the ECtHR should consider when developing its test. is is because the democracy
rationale is too narrow to provide a comprehensive explanat ion for the right to free
speech.33 In other words, if the f reedom of expression guarantee’s purpose were merely
to ensure that citizens are provided with all i nformation relevant to the discha rge of
their democratic decision-mak ing responsibilities, there would be no justi cation for
protecting (for example) artistic or commercial sp eech.34 It was partly to address this
unsatisfactory aspect of Meiklejoh n’s theory (of which Scanlon generally approved)
that Scanlon developed a broader, though simi larly absolute, theory of free speech.35
2.3. THE SELFFULFILMENT/AUTONOMY RATIONALE
Scanlon’s theory nds its basis in the self-ful lment/autonomy rationale.36 Unlike
some such theories, though, it emphasises the autonomy not of the speaker, but of
those who hear the speech. Scanlon stresses t hat – leaving aside presently irrelevant
31 Meiklejoh n, Political Freedom (n 28) at 20.
32 See text accompany ing n 79.
33 omas Scanlon, ‘Freedom of Express ion and Categories of E xpression’ (1979) 40 Universit y of
Pittsburgh Law Review 519, 529.
34 Meiklejohn did ulti mately accept that , under his the ory, the First Amend ment covers arti stic
speech; however, his reas oning is unpersuasive: se e Alexander Meiklej ohn, ‘ e Firs t Amendment
is an Absolute’ (n 28) at 263.
35 Scanlon (n 33) at 530–531.
36 Barendt (n 10) at 15–18.
Freedom of Expression a nd the Advocacy of Violence
Netherlands Qu arterly of Human Ri ghts, Vol. 33/1 (2015) 85
exceptions such a s where there is ‘a situation of near catastrophe’37 – expression may
not be prohibited simply to prevent people from: (a) acquiring false beliefs as a resu lt
of this speech; or (b) performing harmful acts t hat they would not perform but for
the expression, in circumstances where the only connection between the ex pression
and the o ence would be that the expression led, or inclined, the o ender to believe
that he/she should commit t he o ence.38 So, his argument, if acc epted, would prevent
the government from prohibiting the advoc acy of violence even where such advocacy
made violence likely, or even certain. is is necessary, Scanlon thought, partly because
to allow the state to wit hhold from citizens the right to hear such spee ch would be to
grant it ‘the right to deprive citizens of the grounds for arriving at an independent
judgment as to whether the law should be obeyed .’39 is , he believed, is a power that
autonomous individuals would be u nwilling to grant to the government.40
Strauss has advanced a ver y similar a rgument,41 which, however, is slightly less
absolute than Scanlon’s position: he argues, for instance, that t he government may
suppress speech that creates an im minent risk of violence, for the person who responds
violently immediately upon bei ng exposed to the advocac y of such conduct has
engaged in behaviour that, b ecause it is ‘ill-considered,’ is not properly autonomous.42
But both Scanlon’s and Strauss’s theories must be rejected. As Amdur notes, it is
true that our autonomy is compromised when we are deprived of the oppor tunity to
consider arguments about what we should believe.43 However, while this means t hat
autonomous people would be reluctant to give the government the rig ht to deprive
them of access to such arguments, such people have concerns other than in being
autonomous: ‘[i]n particular, they are certai n to want to remain free from physica l
assaults and th reats to their lives.’44 Because autonomous people have an interest not
only in hearing a broad ra nge of views, but also in not being victimised by crimes
that result from acts of expression, t hey would ‘almost certainly’ ag ree that, while the
government generally may not suppress speech, it may do so where that speech results
in serious harms.45
37 omas S canl on, ‘A eory of Free dom of Expression’ (1972) 1(2) Philosophy and Public A airs 204,
226.
38 Ibid at 213.
39 Ibid at 218.
40 Ibid at 215.
41 David Strauss, ‘Persu asion, Autonomy and Freedom of Ex pression’ (1991) 91(2) Columbia Law
Review 334. For other, similar, autonomy-based free s peech theories, s ee Allan (n 9); oma s Nagel,
‘Personal Right s and Public Space’ (1995) 24(2) Philosophy and Pu blic A airs 83, 96 –99; and Sheldon
L. Leader, ‘Free Speec h and the Advocacy of Il legal Action in Law a nd Political eory’ (1982) 82(3)
Columbia Law Rev iew 412.
42 Strauss (n 41) at 365.
43 Robert Amdur, ‘Scanlon on Freedom of Expression,’ (1980) 9(3) Philosophy and Public A airs 287,
298.
44 Ibid.
45 Ibid at 299.
Andrew Dyer
86 Intersentia
Other self-ful lment/autonomy theories emphasise not only the interests of the
hearer of speech, but also of the spea ker. ese theories derive from the idea ‘that
the proper end of man is the realisation of his own character and potentialities as a
human being.’46 is ‘self-realisat ion’ can only occ ur if people have the right not only
to hear other people express t heir opinions, but also to express their own opinions.47
One such theory is provided by Redish, who correctly notes that – despite claims
to the contrary48 – there is no ‘ logically necessary lin k’ between a bel ief in the self-
ful lment rationale for free speech and absolute free speech protection.49 Redish
draws a similar conclusion to that drawn by Amdur, contending that he ‘refuse[s]
to believe’ that the rig ht to free speech ‘requires us to protect the statement of a mob
leader, outside a poorly defended prison, urging his torch-carr ying compatriots to
lynch a prisoner inside.’50
As noted below,51 and as with the democracy rationale, the ECH R has rightly
accepted the self-ful lment/autonomy rationale for free speech. Accordingly, when
developing its test for advocacy of violence cases , the ECtHR should consider the need
not only to facilitate democratic decision-making, but also to f urther spea ker and
hearer autonomy – at least insofar as t his is possible without unjusti ably endangering
the community.
2.4. THE DISTRUST OF GOVERNMENT R ATIONALE
In a subsequent article,52 Scan lon repudiated the theory of his that has been dis cussed
above. is he d id partly because, consistently with A mdur’s (later) criticisms, ‘there
are… limits to the sacri ces we are willing to make to en hance our decision-making
capacity.’53 But Scanlon maintained t hat the government may not suppress any
discussion of politica l issues – and, therefore, for example, seeming ly much advocacy
of violence54 – as, accordi ng to him, ‘where political issues are involved governments
are notoriously partisa n and unreliable.’55 Schauer, too, is a prominent advocate of this
46 Emerson (n 10) at 879.
47 Ibid.
48 See for example Vincent Blasi, ‘ e Check ing Value in First A mendment eory’ (1977) 2(3)
American Bar Foundation Research Journal 521, 547 cited in Martin Re dish, ‘ e Value of Free
Speech’ (1982) 130(3) University of Pennsylvania L aw Review 591, 623.
49 Redish (n 48) at 623.
50 Ibid.
51 See text accompa nying n 79.
52 Scanlon, ‘Freedom of E xpression and Categories of E xpression’ (n 33).
53 Ibid at 533.
54 Eric Barendt, ‘Incitement to, and Glori cation of, Terrorism,’ in Iv an Hare and James Weins tein
(eds) , Ex treme Speech and Democrac y (Oxford University Press 2009) 452; se e also Sürek v Turkey
(No. 1) (n 3) [59], where the ECtHR ex pressly noted that, w hen the media convey s to the public
statements that mi ght be interpreted as condon ing or encouraging v iolent acts, its ‘essentia l role …
in ensuring t he proper functioni ng of political democra cy’ must be considered.
55 Scanlon, ‘Freedom of Ex pression and Categories of E xpression’ (n 33) at 534.
Freedom of Expression a nd the Advocacy of Violence
Netherlands Qu arterly of Human Ri ghts, Vol. 33/1 (2015) 87
‘distru st of government’ rationale for free speech.56 Among other arg uments, he cites
historical exa mples of ‘fairly plain errors’ by the government and other authoritie s in
regulating speech. 57 Like Scanlon, he argues that such errors might have been made
because of governmental ‘bias or self-interest’: the responsibility of regulating the
relevant speech, he states, was ‘entruste d to those very people who … ha[d] the most
to lose from arguments aga inst their authority.’58
But, although there is considerable force in this argument,59 it is very questi onable
to say the least whether there should be a limitation on government power of the
sort suggested by Scanlon.60 To prevent the government from prohibiting political
speech that makes v iolence certain or likely, because of a mere concern that, if it were
allowed to prohibit such speech, the government mig ht in other cases suppress views
that carry no rea l risk of harm, seems to pay insu cient regard to individuals’ i nterest
in being free from the harms that result from acts of ex pression. e argument also
ignores the supervis ory role played by the cour ts. When Scanlon refers to the bias of
the ‘government,’ he surely means the legislatu re and executive, not the judiciar y61
(he refers, for instance, to ‘governments’ ‘mak[ing] polic y’,62 a function that is usually
associated with the political branches). Accordingly, what his argument overlooks is
that, when the legislatu re and executive do abuse their power to silence dangerous
political speech, cour ts have the power to declare that such action breaches free
speech rights. It is not as though t he ‘government’s’ biased view of the limits of free
speech is nal.
is last point is of considerable importance to the argument that is advanced in
this article. It is true that Leader doubts courts’ ability to check legislative excesses in
this area: he points to Fra nkfurter J’s deferential approach to the legislative judg ment
in Dennis v United States,63 and states t hat this provides some support for Sc anlon’s
view that decisions concerning the suppression of political speech w ill, in fact , be
‘in the partisan hands of the leg islature and not the Cour t’ even where a system of
judicial rev iew exists.64 However, in some cases,65 the ECtHR has rightly shown a
willingness to hold governments to greater scrut iny than the US Supreme Cour t
56 Schauer (n 10) at 81.
57 Ibid.
58 Ibid.
59 As noted by Barendt (n 10) at 21.
60 Schauer falls short of draw ing the conclusion that Sc anlon does. Whi le he thinks that ‘ it is possible
to use… [distrust of gove rnment] arguments ‘ to justify a gene ral limit ation on government,’ he
then adds that, ‘I m ake no such argument here. I a m arguing only that t he power of government to
regulate spee ch should… be more limited t han are its powers in othe r areas of governance’: Schauer
(n 10) at 86.
61 is is certa inly Leader’s interpret ation of Scanlon’s views: Lea der (n 41) at 442.
62 Scanlon, ‘Freedom of Ex pression and Categories of Ex pression’ (n 33) at 534.
63 341 US 494 (1951).
64 Leader (n 41) at 442.
65 See for example the case s listed at n 3.
Andrew Dyer
88 Intersentia
did in Dennis. e ECtHR has o en stated that, while it grants Contracti ng States a
‘margin of appreciation’ in assessi ng whether interferences with freedom of speech
are proportionate to one of the aims identi ed by A rti cle10 (2), th is ‘g oes h and in ha nd
with European supervision.’66 Further, and importantly, it appears that one factor
that the Court has t aken into account when determining the breadth of the margi n
of appreciation in violent advocacy cas es is the potential for governments to si lence
speech for reasons of self-interest. is point is developed below.67
In short, the ‘d istrust of government’ a rgument does not establ ish that governments
may never suppress discussion of politica l issues. What it does establish is that the
ECtHR, when it considers whether a state’s proscription of politica l speech that is said
to advocate violence amounts to a breach of Art icle10, should carefu lly scrutinise the
state’s claim that there exist s a ‘pressing social need’ for such an interference. In ot her
words, the ECtHR’s test for advocacy of violence cases should be one that requires
states to substantiate their a llegations that the relevant speech was d angerous, and not
merely o ensive.
2.5. PR AGMATIC JUSTIFICATIONS
Finally, it is necessary to note t hat there are at least two other, pragmatic, reasons why
at least some speech that advocates violence should be protected. e rst reason is
that the excessive suppression of such speech might wel l be counter-productive. Many
of those who have commented on the ‘glori cation of terrorism’ o ence in s 1 of the
Terrorism Act 2006 (UK) have questioned whet her this measure – which cri minalises
statements that are ‘l ikely to be understood… as a direct or ind irect encouragement or
other inducement…’ to commit terrorist o ences, whether or not there was a ris k that
the speech would be acted upon – wi ll lead to feelings of resentment among minority
groups and, consequently, an increased tendency on their members’ part to become
involved in terrorist activit y.68
e second reason is that, if all advocacy of violence were to be prohibited, it is
likely that much legitimate speech would be ‘chilled.’ e reason why it is likely that
there would be a ‘chilling e ect ’ on leg it im ate spe ech is t ha t so me s pee ch tha t ad voc at es
66 Handyside v United Kingdom (1976) 1 EHRR 737 [49].
67 See text accompany ing nn 80–82.
68 See for example Clive Walker, ‘Milit ant Speech About Terrorism in a Smart Mi litant Democracy’
(2011) 80(4) Mi ssissippi L aw Journal 1395, 1443; Tufya l Choudhury, ‘ e Terrorism Act 20 06:
Discourag ing Terrorism,’ in Ivan Hare and James Weinste in (eds) Extreme Speech and Democ racy
(Oxford University Press 2 009) 481–487; and Ellen Park er, ‘Implementation of the UK Terrorism Act
2006 – e Relationsh ip Between Counter terrorism Law, Free Speech, a nd the Muslim Com munity
in the United Ki ngdom versus the United States’ (2007) 21 Emory Inte rnational Law Review 711,
753. Van Ginkel makes a sim ilar point: s ee Bibi van Ginke l, ‘Incitement to Terrorism: A Mat ter
for Prevention or Repression’ (Inter national Centre for Counte r-Terrorism Resea rch Paper, August
2011) 7–8 and 24.
Freedom of Expression a nd the Advocacy of Violence
Netherlands Qu arterly of Human Ri ghts, Vol. 33/1 (2015) 89
violence does so indirectly.69 Consequently, it is necessary to prohibit indirect as well
as direct advocacy.70 However, because speakers cannot be sure of how their words
will be interpreted, it is possible that a restriction on all such advocacy would lead
individuals to refrain from expressi ng views about, for example, terrorism for fear
that their words would be interpreted as condoning or i ndirect ly advocating terrorist
o ences. For example, as has repeated ly been noted in the literature,71 the Joint
Committee on Human Rig hts reported that it was unclea r whether remarks made
by Cherie Booth Q.C. and Jenny Tonge M.P. that expressed some understanding of
suicide bombers’ mot ives, amounted to the ‘g lori cati on of te rror ism’ c ontr ary to s 1.72
2.6. CONCLUSIONS
Despite the shortcomings of some of the theories mentioned above, the democracy
and self-ful lment/autonomy rationales for free speec h do provide persuasive reasons
why some advocacy of violence should be protected. Such advocacy is o en political
speech,73 the prohibition of which plausibly inhibits democratic decision-making.
Even where this is not the case, such spee ch’s proscription would certainly amount to
an interference with bot h speaker and hearer autonomy. Nevertheless, as noted above,
the failure of absolutist f ree speech theories such as Sca nlon’s and Meik lejohn’s means
that the right to engage in ‘v iolence-conducive’ speech may be restric ted where there
is a su cient risk t hat it will be acted upon. Democracies are not required to com mit
suicide.74 Nor are governments requ ired to tolerate speech that creates a danger of
violence,75 however much such speech might contribute to public debate and/or
promote speaker and hearer autonomy.
It follows that the test that the Cour t should adopt in cases involving the advocacy
of violence is one that achieves a proper balance bet ween, on one hand, speaker a nd
he a re r a u to no my a nd t he fa c i li t at io n of de m oc r at i c d ec i si on -m a k i ng ; a n d , o n t h e o t he r,
69 Healy (n 9) at 686; see also David G . Barnum, ‘Indire ct incitement and freedom of s peech in Anglo-
American law’ (2006) 3European Human Rights Law Review 258. And note the fa cts of Zana v
Turk e y (1999) 27 EHR R 667: see text accompany ing nn 88–91.
70 ere is furt her discussion below conc erning the desi rability of prohibit ing both direc t and indirect
advocacy: se e text accompanyin g nn 155–158.
71 See for example Choudhur y (n 68) at 469–470; Darrag h Murray, ‘Freedom of Expression, Cou nter-
Terrorism and the Inter net in Light of the U K Terrorist Act 2006 a nd the Jurispr udence of the
European Cou rt of Human Rights’ (2009) 27(3) Netherland s Quarterly of Human Rights 331, 34 4;
and Adrian Hunt , ‘Crimina l Prohibitions on Direc t and Indirec t Encouragement of Terrorism’
[200 7] Criminal Law Review 441, 450.
72 Joint Committee on Huma n Rights, Cou nter-Terrorism Policy and Hum an Rights: Terrori sm Bill
and related mat ters (2005–06, H L 75-I, HC 561-I) 19.
73 See Barendt (n 54) at 452.
74 A point made by Conor G earty, Libert y and Securit y (Polity Press 2013) 95; and Eric Ba rendt,
‘ reats to Freedom of Speech in the United Kingdom’ (2005) 28 University of Ne w South Wales
Law Review 895, 899.
75 For a discussion of how great t hat danger must be see tex t accompanying nn 160 –171.
Andrew Dyer
90 Intersentia
the interests of citizens i n being free from the harms that potentia lly result from such
acts of expression. Furt her, because there are reasons to doubt governments’ ability
to exercise unbiased judgment in this area, the test should be one that is su ciently
demanding to allow t he ECtHR carefully to scruti nise states’ claims that the releva nt
restrictions are ‘necessa ry in a democratic society.’ Finally, any test that is adopted
must be unlikely to lead to cou nter-productive results or to ‘chill ’ legitimate speech.
3. THE ECtHR’S ADVOCACY OF VIOLENCE
JURISPRUDENCE
3.1. ‘NECESSARY IN A DEMOCRATIC SOCIETY’
e ECtHR has broadly accepted that the matters referred to in t he paragraph
directly above are t he pertinent considerations for it: (a) when determining whether a
particula r interference with ‘violence-conducive speech’ is ‘necessa ry in a democratic
society’ for one of the purposes l isted in Article10(2) ECHR; and, therefore, (b) when
framing a test to b e used in advocacy of violence cases.
It is well-established that t he word ‘necessary’ in Ar ticle10(2) ‘implies the exi stence
of a “pressing social need”’ for the releva nt interference and, ‘[i]n particular,’ that it is
‘proportionate to the legitimate aim pursued.’76 It is also well-established that, while
it is for Contracting States to ‘ma ke the initial assessment’ of whet her such a ‘pressing
social need’ exists, there are l imits on their ‘power of appreciation.’77 is point has
been noted already.78 How great these limits are – that is, how broad or narrow the
state’s margin of appreciation is – depends on a number of factors.
First, the importance of the interests interfered with narrows the margin of
appreciation in advocacy of violence cases. at is, because ‘freedom of expression
constitutes one of the essentia l foundations of a democratic society and one of the
basic conditions for its progress and for each indiv idual’s self-ful lment,’ restrictions
on the enjoyment of that right ‘must be established convincingly.’79 In so holding,
of course, the Strasbourg Court has accepted bot h the democracy and autonomy
rationales for free speech. In other words, it is because freedom of expression is such
an important interest t hat the margin of appreciation contrac ts; in turn, f reedom of
expression is such an important interest because of the contribution that it makes to
democratic and other decision-mak ing and individual ‘self-ful lment.’
Second, the narrowing of t he margin of appreciation will especia lly occur in cases
where the relevant speech is ‘political speech or … debate on questions of public
76 See for example Zana (n 69) at 689.
77 Handyside (n 66) at [48]-[49].
78 See text accompany ing n 66; see also for ex ample Sürek (No. 1) (n 3) at [58].
79 See for example Zana (n 69) at 689.
Freedom of Expression a nd the Advocacy of Violence
Netherlands Qu arterly of Human Ri ghts, Vol. 33/1 (2015) 91
interest.’80 is is not only because of the need, ‘i n a democratic system’, to subject
‘the actions or omissions of government … to … close scrut iny.’81 It is also because
governmental proscriptions of anti-govern ment speech are suspect. ‘[T]he domina nt
position which the government occupies’82 mea ns that, if its decisions to prohibit such
speech were not exposed to close scruti ny, there would be a danger that it would abuse
its power by self-interestedly using the cri minal law to suppress speech that is, in fact ,
harmless.
ird , however, the margin of appreciation is not so li mited as to prevent the state
from placing any restric tions on the advocacy of violence. e Court has stated that ,
in determining whether the interference ful ls a ‘p ressi ng s ocia l nee d,’ it wil l ta ke in to
account ‘problems linked to the prevention of public disorder and terrorism.’83 is
makes it clear that, a s Gearty puts it, the Ar ticle10 right will in some cases ‘y ield to the
greater good of the politica l community as a whole.’84 Of course, as argued ab ove, this
is unobjectionable; however, the Court has gone fur ther than this. It has a lso provided
that the margin of appreciat ion expands ‘where [the relevant] … remarks incite to
violence against a n individual or a public o ci al or a sector of the popu lation.’85 is
last p ropositio n causes s ome di culties. e C ourt’s words here are ambiguous. Does
it mean that the margi n is wider in cases where there has been a direc t call for violence?
Or, when it uses the word ‘incite,’ is the Court a lso referring to indirect incitements?
If the Court mea ns the former, we must approach this statement of principle with
some caution. It might be that allowi ng a wider margin in cases of direct , as opposed
to indirect, advocacy is justi ed, to an extent, by the pos sibly greater scope for the
state, in the latter t ype of case, to exa ggerate the dangers of the releva nt speech. But
the principle should not be taken too far. If not appropriately quali ed, it tends to
encourage future cour ts to focus unduly on the content of the message rather t han its
possible e ects.
Indeed, as we will see, the main criticism of the test that the Cour t developed in
the cases that will now be discussed, has been precisely this. at is, according to a
minority of judges in the Turkish ca ses, the Court should have, to use one member
of the minority’s words, placed less emphasis ‘on the vehemence a nd outrageous tone
of the words employed and more on the di erent elements of the contextual setti ng
in which the speech was uttered.’86 Is this charge justi ed? Or does the test that
the majority of the Court developed achieve a proper balance between t he relevant
factors?
80 See for example Sürek (No. 1) (n 3) at [61].
81 Ibid.
82 Ibid.
83 See for example Saygili and Fa lakaoglu (n 5) at [25].
84 Conor Gearty, Principle s of Human Rights Adjudicat ion (Oxford University Press 20 05) 42.
85 See for example Sürek (No. 1) (n 3) at [61].
86 Sürek (No. 1) (n 3) (Judge Palm).
Andrew Dyer
92 Intersentia
3.2. THE TURK ISH CASES
As Sottiaux obser ves, it was not until the late 1990s, in a series of case s from Tu rkey,
that the ECtHR developed a detailed theory concer ning the advocacy of v iolence.87
In those cases, t he applicants alleged t hat the authorities had breached Ar ticle 10
ECHR by convicting them of cr iminal o ence s for expressing support for the cause of
Kurdish self-determi nation and/or the Workers’ Party of Kurd istan (the PKK).
e rst such case to reach the Court was Zana v Turkey.88 e applicant had
been the Mayor of Diyarbak ir, a large city in sout h-eastern Turkey, but was serving a
prison sentence at the time that he made the relevant remarks. In an interview with
journalists , he made the following statement, which was then published i n a national
newspaper:89
I support the PKK nat ional liberation movement; on the other ha nd, I am not in favour of
massacres. A nyone can make mistak es, and the PKK kil l women and children by mista ke.
e applicant was convicted of breaching A rticle312 of the Turkish Crimina l Code,
which prov ided at the re levant time that it was a n o ence ‘publicly to praise or defend
an act punishable by law as a serious cri me’ or ‘publicly to incite hatred or hostility
between the di erent classes in society, thereby creating discrimi nation based on
membership of a social clas s, race, religion, sect or region [and] endanger[ing]… public
safety.’90 A majority of the Grand Chamber of the ECtHR held that t his conviction
amounted to no breach of Article10. In doing so, it did not li mit itself to considering
the applicant’s words – which, of course, did not di rectly advocate violence – but took
account also of the risk, g iven the context in which those words were spoken, that they
would cause harm. A er noting t hat the statement had coincided with murderous
PKK violence in south-easter n Turkey, the Court held that it ‘had to be regarded as
likely to exacerbate an a lready explosive situation in that region.’91
In Incal v Turkey,92 decided in 1998, and 12 fur ther Turkish cases decided on 8July
1999,93 the Court con rmed its approach to the advocacy of violence. For although
in all but two of these c ases the Court, unlike i n Zana, held that there had been a
violation of Article 10, its statement of principle was not substantially d i erent from
that which it issued in t hat earlier authority. Typical of these case s is Sürek and Őzdemir
v Turkey.94 In that case, the applicants were the major shareholder in a company that
87 Sottiaux, ‘ e “clear and present danger” tes t’ (n 6) at 669; although see t he earlier subversive spe ech
cases – such as Arrowsmith v Unite d Kingdom (1981) 3 EHRR 218 – that Sot tiaux notes at 666 –668.
88 Zana (n 69).
89 Ibid at 673.
90 Ibid at 675.
91 Ibid at 691.
92 (2000) 29 EHRR 449.
93 See cases listed at n 3.
94 Sürek and Őzdemir (n 3).
Freedom of Expression a nd the Advocacy of Violence
Netherlands Qu arterly of Human Ri ghts, Vol. 33/1 (2015) 93
owned a weekly review, e Truth of News and Comments, and the ed itor-in-chief of
that review. In 1992, there was published in the review an interview wit h a leader of
the PKK, in which he stated :95
is war w ill continue as long as the Turkish state re fuses to accept the wil l of the people
of Kurdistan: there will not be one single step backwards. e war wi ll go on until there is
only one single indiv idual le on our sid e.
Furthermore, t he review published a joint statement from four soci alist organisations,
in which call s were made to ‘unite agains t State terrorism, against the repression and
oppression of the Kurdish people.’96
A majority of the Grand Chamb er of the ECtHR held that the applicants’ result ing
convictions for o ences of d isseminating propaganda against t he indivisibility of the
state, contrary to se ctions 6 and 8 of the Prevention of Terrorism Act 1991, amounted
to a breach of Article 10.97 e Court observed that t he words in the interview
conveyed a message of ‘intra nsigence,’ and it noted the Turkish authorities’ concern
‘about words or deeds which have the potential to exacerbate the sec urity situation
in… [south-eastern Turkey], where since approximately 1985 serious disturbances
have raged between members of the security forces and the members of t he PKK.’98
However – as it i s wont to do in Art icle10 cases99 – the Cour t also acknowledged that
freedom of expression ‘constitutes one of the essential foundations of a democratic
society and one of the basic conditions for its progress and for each individual ’s self-
ful lment.’100 And it emphasised the press’s ‘essential role… in ensuring the proper
functioning of political democracy ’; it is, the Court said, ‘ incumbent on the press to
impart informat ion and ideas on political issues, i ncluding divisive ones.’101 e Court
did stress that the pres s must exercise ‘part icular caution’ when considering whether
to publish statements from members of organ isations that engage in violence against
the state ‘lest the media become a vehicle for the dissemination of… the promotion
of violence.’102 But it concluded that ‘the v iews expressed in the interviews could not
be read as an incitement to violence; nor could they be construed as liable to incite
violence.’103 e Court t hen had regard to the joint statement, nding that it could not
be ‘construed as an incitement to v iolence.’104
95 Ibid at [10].
96 Ibid.
97 Ibid at [64].
98 Ibid at [61].
99 See text accompanyin g n 79.
100 Sürek and Őz demir (n 3) at [57].
101 Ibid at [58].
102 Ibid at [63].
103 Ibid at [61].
104 Ibid.
Andrew Dyer
94 Intersentia
Sottiaux has suggested, tentatively, that in the 8 July 1999 cases, the majority –
contrary to the approach taken in Zana – placed greater stress on whether the words
used were in ammatory than on the risk that the impugned speech would result in
harm.105 But he is right not to ta ke this conclusion too far.106 It i s tr ue t hat t he C our t in
the 8July cases placed greater emphasis than t he Zana Court did on t he virulence of the
speakers’ words: in the case of Sürek (No. 1) v Turkey,107 for instance, the Court noted
that letters published i n the applicant’s review had demonstrated a ‘clear intention to
stigmatise the ot her side of the con ict by the use of lab els such as ‘the fascist Turkish
army.’108 But t here is an obvious explanat ion for this. In Zana, unlike i n the 8July cases,
the speaker’s words were ambiguous; they were not intemperate or violent. Had Zana’s
words advocated violence more directly, it is quite conceivable that the Court would
have emphasised their maligna nt content. Conversely, had the words considered in a
case such as Sürek (No. 1) be en less direct but equally dangerous, it is ha rd to imagine
the Court nding that t here had been no violation of Article10.109 It is also true that
in Zana the Court considered t hat the speech was ‘likely to exacerbate’110 the situation
in south-eastern Turkey, whereas in the 8 July cases the Court required merely that
the speech be ‘liable to’ incite, or ‘capable of’111 i nciting, violence. However, it is not
clear that, by using t he word ‘likely’ in Zana, the C ourt was laying down a formal
legal test from which it then departed in the 8 July cases. In short, the better view
seems to be that the 8 July cases entrenched t he approach established in Zana that,
in determining whether a prohibition on ‘violence-conducive speech’ a mounts to a
breach of Article10, the Cour t will consider not merely whether the speaker’s words
advocate violence, but also whether t hey might have the e ec t of causing it. Indeed, in
a later article, Sott iaux has acknowledged as much.112
But whether or not Zana and the 8 July cases ca n be reconciled in this way, the
minority of the Cour t in the 8 July cases arg ued that the majority’s ‘liable to incite
violence’/‘capable of inciting… violence’ risk th reshold was insu ciently demanding.
Two minority views are particu larly worth noting. First, in each of the cases Judge
Palm (someti mes joined by Judges Tulkens , Fischbach, Casade vall and Greve) arg ued
that, ‘the majority has at tached too much weight to the admittedly harsh and vit riolic
language used i n the impugned letters and insu cient attention to the genera l context
105 Sottiaux, ‘ e “clear and present d anger” test’ (n 6) at 674–675.
106 Ibid at 675.
107 Sürek (No. 1) v Turkey (n 3).
108 Ibid at [62].
109 Sürek (No. 1) was one of the t wo 8July cases in which t he Court held that the re had been a breach of
Article10; the ot her was Sürek v Turkey (No. 3) (n 3).
110 Zana (n 69) at 691 [emphasis added].
111 In some of t he 8July cases, the C ourt used the latter for m of words: see for example Sürek (No. 1) v
Turk e y (n 3) at [62].
112 Sot tiaux, ‘Leroy v France’ (n 2) at 419–420. D avis appears to provide a si milar analy sis: see Howard
Davis, ‘Less ons from Turkey: Anti-Terrorism Legisl ation and Free Speech ’ (2005) 1 European
Human Rights Law Re view 75, 82.
Freedom of Expression a nd the Advocacy of Violence
Netherlands Qu arterly of Human Ri ghts, Vol. 33/1 (2015) 95
in which the words were used and their l ikely impact.’113 Accordingly, she proposed
the following test: in de termining whether the releva nt restrictions ful l led a ‘pressing
social need,’ the Cour t should have asked whether (a) ‘the lang uage [was] intended to
in ame or incite to violence’ and (b) ‘there was a rea l and genuine risk that it might
actually do so.’114 Judge Palm attached especial importance to the second inquiry: in
opinions written jointly w ith the judges mentioned above, she stated that, in every
case, the Cour t should ask itself three questions:115
Did the author of the o en ding t ext oc cupy a po sition of in uence i n society of a sort like ly
to amplify t he impact of his words? Was the pu blication given a deg ree of prominence
either in an impor tant newspaper or throug h another medium which w as likely to enhance
the in uence of the i mpugned speech? Were the words far away from the c entre of violence
or on its doorstep?
Appl yi ng th is a pproa ch t o the fac ts, J udge Pal m hel d in Sürek (No. 1), for instance, t hat
– because neither the applicant nor the authors were ‘prominent gure[s] in Turkish
life capable of exercising an in uence on public opinion’; the impugned material did
not ‘occupy a central or headline position in [the] … review’; and the review was
published in Istanbu l, ‘far away from the zone of the con ict in south-e ast Tu rkey’ –
there was no ‘real and genuine risk ’ of the speech resulting in violence.116
Second, Judge Bonello argued in all of the cases that the Court should have
applied the US Supreme Court’s ‘clear and present danger’ test .117 In so doing, he
cited with approval the remarks of Holmes J in Abrams t hat I have quoted above;118
his Honour’s statement in Schenck that the relevant question is one of ‘proximity and
degree;’119 and Brandeis J’s judgment in Whitney.120 While there have been a number
of manifestations of the ‘clear and present danger’ test ,121 Judge Bonello made it clear
that the version that he supported was the one accepted by the Supreme Court in
113 Sürek (No. 1) (n 3) (Judge Palm).
114 See c ases listed at n 3 (Judge Pal m/ Judges Palm, Tulkens, Fisc hbach, Casadeval l and Greve).
115 Arslan (n 3) (Judges Palm, Tulkens, Fischb ach, Casadevall and Greve); Baskaya a nd Okçuoglu (n
3) (Judges Palm, Tulkens, Fis chbach, Casadevall and Gre ve); Ceylan (n 3) (Judges Palm, Tulkens,
Fischbach, Cas adevall a nd Greve); Erdoğdu and Ince (n 3) (Judges Palm, Tulk ens, Fischbach,
Casadeva ll and Greve); Gerger (n 3) (Judges Palm, Tulkens, Fischbach, Ca sadeval l and Greve);
Karataş (n 3) (Judges Palm, Tulkens , Fischbach, Casad evall and Gre ve); Okçuoğlu (n 3) (Ju dges
Palm, Tulkens, Fis chbach, Casade vall and Greve); Pola t (n 3) (Judges Palm, Tulkens , Fischbach,
Casadeva ll and Greve); Sürek (No. 4) (n 3) (Judges Palm, Tulkens, Fischbach, Ca sadevall a nd Greve);
Sürek and Őzdemi r (n 3) (Judges Palm, Tulkens, Fischbach, Ca sadevall and Greve).
116 Sürek (No. 1) (n 3) (Judge Palm).
117 See c ases listed at n 3 (Judge Bonel lo).
118 See te xt accompanying n 19.
119 See te xt accompanying n 14.
120 See cases li sted at n 3 (Judge Bonello).
121 As noted by Barnum, ‘ e Clear and Pres ent Danger Test’ (n 15) at 271–280; see also Sotti aux, ‘“ e
Clear and Prese nt Danger” Test’ (n 6) at 658–666.
Andrew Dyer
96 Intersentia
Brandenburg v Ohio.122 In Brandenburg, the Court held that the First Amendment
allowed for the prohibition of the advocacy of un lawful activity only where such
advocacy is ‘directed to inciting or producing imminent lawless action and is likely
to incite or produce such action.’123 In Judge Bonello’s opinion, although the various
8July appl icants’ words might have appeared to be ‘pregna nt with mortality,’ they
‘created no peril, let alone a clear and present one.’124
ere is discussion below of the desirability of t he ECtHR adopting a test wit h
a higher risk threshold than that accepted by the majority in the Turkish cases. For
now it is su cient to note that the majorit y’s ‘capable of’/‘l iable to’ test was applied in
sub sequ ent S tra sbo urg c ase s. I n Hogefeld v Germany,125 for example, the Cour t held to
be manifestly i ll-founded the applicant’s complaint that Article10 had been breached
in circumsta nces where she, a former member of the Red Army Faction (RA F) terrorist
group, had been prevented from, among other things , being interviewed by press and
radio journalists. e Court found that ‘the words of the applicant could possibly be
understood by supporters as an appeal to continue the activities of the RAF.’126 e
implication is that any such words were ‘capable of ’ resulting in violence.
3.3. LEROY v FRANCE127
In Leroy, the applicant was a cartoonist who worked for a Basque week ly newspaper.
Two d ay s a er t he 11 September 2001 terrorist attacks, the newspaper published
the applicant’s cartoon, which depicted the destruct ion of the World Trade Centre,
accompanied by the caption ‘We have all drea mt of it… Hamas did it.’ e applicant
was convicted of complicity in apologie du ter rorisme. e ECtHR dismis sed his
complaint that France, by convict ing him of this o ence, had breached A rticle10.
Some commentators have argued that in Leroy the ECtHR departed from the
position that it adopted in the Turkish cases;128 however, these claims must be
quali ed slightly. e Leroy Court did purport to apply the test from those earlier
cases, paying ‘specia l attention to the words used [in the cartoon]… and the context in
which they were published.’129 Further, the Cour t examined whether the expression
was ‘capable of’ result ing in violence: it observed t hat the cartoon ‘elicited react ions
which could have stirred up violence and which demonstrated that it cou ld have
a ected public order in the region.’130 Nevertheless, the result in Leroy is certainly
122 Brandenburg v Ohio (n 8).
123 Ibid at 447.
124 See cases li sted at n 3 (Judge Bonello).
125 (2000) 29 EHRR CD173.
126 Ibid at 178 [emphasis added].
127 Leroy (n 1).
128 See n 2.
129 Leroy (n 1) at [38].
130 Ibid at [45].
Freedom of Expression a nd the Advocacy of Violence
Netherlands Qu arterly of Human Ri ghts, Vol. 33/1 (2015) 97
very di cu lt to reconcile with those i n the Turkish cases.131 Whereas Turkey had
‘concrete reasons’ to fear that pro-PKK speech could lead to further blood shed,
the Court’s fears that L eroy’s cartoon could ‘stir up’ violence lacked any evidential
foundation.132 Ac cordin gly, the most ac curate interp retati on of Leroy seems to be that
the Court applied t he test from its previous case law, but in the process demonstr ated
that that test is a vague one that allows judges considerable scope to prohibit speech
merely because they consider it to be highly o ensive. e Leroy Court noted that
the applicant’s cartoon wa s published ‘at a time when the whole world was in a state
of shock about’ the 11September terrorist attack s, and stated that ‘the timing should
have prompted the applicant to act more responsibly.’133 ese comments make
almost irresis tible the conclusion that the Court’s decision was heavily i n uenced by
concerns that the expression wa s ‘gratuitously o ensive’.134
3.4. CASES DECIDED SINCE LEROY
Interestingly, in advocacy of violence ca ses decided since Leroy the ECtHR has
not uniformly applied the ‘capable of’ test. Cer tainly, in some post-Leroy cases the
Court has applied that test. For instance, in Demirel and Ates v Turkey (No. 3),135
the applicants’ newspaper had published the transcript of an interview i n which the
PKK’s leader, Abdullah Ocalan, had provided a jour nalist with his views about issues
such as education and cultu ral development, and responded to claims that another
political gure had made about the PKK’s establishment and development. e Court
held that, by convicting the applicants of the o ence of publishing the declarations
of an illegal orga nisation, Turkey had breached Article 10. is was because, ‘despite
particula rly libellous and acerbic passages… the article… was not capable of inc iting
to violence.’136 Similarly, in Saygili and Falakaoglu v Turkey (No. 2) (Saygili),137 the
Court applied the ‘capable of ’ test, nding – controversially138 – that a newspaper’s
publication of a declaration by Turkish prisoners of their i ntention to go on a hunger
strike unti l a particular ty pe of prison was abolished was ‘capable of inc iting violence
in the prisons.’139
In other cases, t hough, the Court has applied tests that are ver y reminiscent of
the US Supreme Court’s ‘clear and present danger’ standard. is case law appears
131 See Shawn M. Boy ne, ‘Free Speech, Terrorism, and E uropean Secur ity: De ning a nd Defending the
Political Com munity’ (2010) 30 Pace Law Review 417, 472.
132 Ibid at 473.
133 Leroy (n 1) at [45].
134 Aidan Wh ite, ‘Ethical Jou rnalism and Hum an Rights’ in C ommissioner for Human R ights, Human
Rights and a Chang ing Media Landscape , (Council of Europe Publicat ions 2011) 67.
135 Demirel and Ate s v Turkey App no 11976/03 (ECtHR, 9December 2008).
136 Ibid at [26].
137 Saygili and Fal akaoglu (n 5).
138 See text acc ompanying nn 162–166.
139 Saygili and Fal akaoglu (n 5) at [28].
Andrew Dyer
98 Intersentia
to have been in uenced by two pre-Leroy decisions. Over two yea rs before Leroy
was decided, the Court, in Ergin v Turkey (No. 6),140 held that Turkey had breached
Article 10 by convicting a newspaper editor of incitement to evade mi litary ser vice
a er he published in his newspaper a story in which war and its celebration were
criticised. e Court found that the article ‘d id not seek… to precipitate immediate
desertion.’141 Simila rly, in Vajna i v Hu ng ary ,142 decided three months before Leroy,
the Court concluded that t he applicant’s conviction for publicly wearing a totalit arian
symbol (he had worn a ve-pointed red star as a symbol of the internationa l workers’
movement) amounted to a breach of Article 10. is was because there was ‘no
evidence to suggest that t here is a real and present danger of a ny politica l movement
restoring the Communist d ictatorship.’143
e C ourt has applied similar tests in post-Leroy cases. In Gul and Others v
Tur ke y,144 the majority held t hat Turkey had breached Article10 by convicti ng the
applicants of an incitement o ence for shouting slogans such as ‘Political power
grows out of the barrel of a gun’ at May Day demonstrations. It found that t here
was no indication that there was ‘a clear and imminent danger which requ ired an
interference such as the lengthy cr iminal prosecution faced by the applica nts.’145 e
dissentients agreed t hat ‘in cases where v iolence… is advocated, the danger of such
consequences has to be clear and i mminent,’ but considered that the applicants’ words
did create a ‘clear and immi nent’ danger of violence.146 In Kilic and Eren v Turkey,147
another case in wh ich the applicants had been convicted a er publicly shouting
revolutionary slogans, t he Court applied the Gul Court’s ‘clear and immi nent danger’
standard. Finally, in Faber v Hungary,148 the majority held that the applicant, by
holding up a ag that was associated w ith the Arrow Cross regime while he was
engaging in an anti-Socialist counter-demonstration, had created no ‘clear threat or
present danger of violence.’149 Accordingly, the Hungar ian authorities had breached
Article10 (read with Art icle11) by convicting him of the o ence of refusing to obey
police instructions, a er he had failed to comply with a police request to remove the
banner. But, as Judge Pinto du Albuquerque noted in his concur ring judgment,150 the
majority also, somewhat conf usingly, considered two other tests when determining
whether there had been a breach of Art icle10. One of those tests was t he ‘capable of’
test from the Turkish cases: apply ing Sürek (No. 1), the Court found that the ag’s
140 Erg in v Turkey (No. 6) App no 47533/99 (ECtHR, 4May 2006) [34].
141 Ibid [empha sis added].
142 Vajnai v Hungar y (2010) 50 EHRR 44.
143 Ibid [49] [emphasis add ed].
144 Gul an d Others (n 5).
145 Ibid at [42] [emphasis a dded].
146 Ibid at [0 -I1] (Judges Sajo and Tsotsoria).
147 Kili c and Eren (n 5) [29].
148 Fabe r v Hungary (n 5).
149 Ibid at [44].
150 Ibid (Judge Pinto du Al buquerque).
Freedom of Expression a nd the Advocacy of Violence
Netherlands Qu arterly of Human Ri ghts, Vol. 33/1 (2015) 99
display ‘was not capable of disturbing public order.’151 Judge Pinto du Albuquerque
made clear his belief that t he correct sta ndard was the ‘clear and im minent danger’
one.152
4. WHICH TEST SHOULD THE ECtHR ADOPT?
ere is, then, some confusion about what the applicable test is when the ECtHR
determines whether a prohibition on the advocacy of violence is a violation of
Article10. Accordingly, the question arises: which te st should the Court adopt? Should
the Court adopt the ‘capable of ’ test from the Turkish cases or one like it? Altern atively,
should the Court accept t he Brandenburg test or one like it? A s suggested above,153 it
has rightly been accepted in the ECtHR’s case law that the a nswer to these questions
hinges largely upon which test wi ll best balance, on one hand, speaker and hearer
autonomy and the facilitation of democrat ic decision-maki ng; and, on the other, the
interests of citizens in bei ng free from harms resulting from t he advocacy of violence.
Moreover, this case law has also correct ly implied154 that the test must be one that
ensures that states are requi red to substantiate their claims that the relevant speech
is dangerous. Finally, the Court must ensure, as far a s possible, that the test t hat it
adopts does not produce counter-productive results and/or deter individuals from
engaging in legiti mate discussion.
4.1. THE BRAN DENBURG TEST
It will be recalled that in the Turkish cases, Judge Bonello held that, for a state’s
restriction on violent advocac y to be ‘necessary in a democratic societ y’ for one of the
aims identi ed in Ar ticle10(2), the Brandenburg test had to be satis ed. Although
later Strasbourg decisions endorsing ‘clear an i mminent’/‘real and present’ danger
tests have not so explicitly approved Brandenburg, the holdings i n those cases appear
largely to be consistent with Judge Bonello’s view.
It will also be recalled that the Brandenburg test has three components. First, it
states that, to be va lidly prohibited, the advocacy of violence must have been ‘directe d
to producing or inciting… lawless action’; in other words, the question is whether
the speaker had a speci c intention to incite unlawf ul activity. Second, the advocacy
must have created an imminent risk of unlawful activity. ird, it is necessary that
the advocacy was ‘likely to incite or produce such action.’ It has been suggested that
the Brandenburg test has a fourth element, namely, that the speech must have directly
151 Ibid at [56].
152 Ibid (Judge Pinto du A lbuquerque).
153 See Sect ion 3.1 above.
154 See text acc ompanying nn 80 –82.
Andrew Dyer
100 Intersentia
advocated the unlawf ul activity.155 But Alexander (among others) rightly argues t hat
the Court seems not to have intended that there be any such requirement.156 e
words ‘directed to’ in t he Court’s test appear to refer to the need for the prosecution to
prove that the speaker had an intention to incite u nlawful activity, not to whether the
advocacy is direc t. Even if Alexander is wrong, it is clearly unde sirable that only direct
advocacy of violence be punishable: as i s repeatedly noted in t he literature,157 speech
of a similar nature to Ma rc Antony’s ironic funeral oration should surely be as open
to proscription as the words of a speaker who advocates v iolence more explicitly.158
e three criteria that the Brandenburg Court did identif y will now be discussed
in turn. e purpose in so doi ng is to establish whether the views ex pressed by Judge
Bonello in the July 8 cases and the ECtHR in later ca ses,159 are sustainable. In other
words, is there a ‘pressing social need ’ for restrictions on the advocacy of v iolence only
when each of the Brandenburg criteria is satis ed?
4.2. THE ‘LIKELY TO INCITE OR PRODUCE SUCH ACTION’
REQUIREMENT
A very important considerat ion is whether the Court should hold that, for prohibitions
on the advocacy of violence to be compatible with Article 10, the advocacy must be
likely to result in violence. A lternatively, should it be enough that – as the C ourt held
in the 8July 1999 cases – the advoc acy be ‘capable of’ causing violence? e Council of
Europe Convention on the Prevention of Terrorism (CEC PT)16 0 has approved the latter
standard. Ar ticle5 of the Convention obliges Council of Europe states to cr iminalise
speech that, ‘whether or not di rectly advocating terrorist o ences,’ is both intended to
incite a terrorist o ence and ‘causes a danger that one or more such o ences may be
155 Gerald Gunt her, ‘Learned Hand a nd the Origi ns of Modern First Am endment Doctri ne: Some
Fragments from Hi story’ (1975) 27 Stanford Law Review 719, 754–755.
156 Larry A lexander, ‘Incitement and Freedom of Spee ch’ in David Kretzme r and Francine Kershman
Hazan (eds), Freedom of Spe ech and Inciteme nt Against Democra cy (Kluwer Law Inter national,
2000) 101, 106; see also Red ish, ‘Advocacy of Unlawful Conduct a nd the First Amendment’ (n 16)
at 1176; and James Weinstein, ‘ e Story of Masses Publishing Co. v Patten: Judge Learned Hand,
First Amendment Prophe t’ in Richard W. Garnet t and Andrew Koppelman (eds .) First Amendment
Stories (Foundation Press 2 012) 80.
157 See for example A lexander (n 156) at 106; Barendt, ‘Incitement to, a nd Glori c ation of, Terrorism,’
(n 54) at 456–457; and Redis h, ‘Advocacy of Unlawful Conduct a nd the First Amendment’ (n 16) at
1189. For similar views, s ee Yael Ronen, ‘Incitement to Terrorist Acts and Inter national Law’ (2010)
23(3) Leiden Journal of Internat ional Law 645, 663–666.
158 One of the main c riticisms of t he test famous ly proposed by Judge Lea rned Hand in Masses Publishing
Co. v Patten 244 Fed 535 (SDNY 1917) was that it app eared not to cover indi rect incitement of
unlawfu l action: see Redish, ‘Advocacy of Unlaw ful Conduct and the First A mendment’ (n 16) at
1189; and Gunther (n 123) at 729, 773. But see also Berna rd Schwartz, ‘Holmes v Ha nd: Clear and
Present Danger or Advocacy of Unlawful Action’ (1994) 1994 Supreme Cour t Review 209, 21 3–214,
who argues t hat the Masses test could apply to indire ct as well as direct ad vocacy.
159 See text acc ompanying nn 140–152.
160 CETS No. 196 .
Freedom of Expression a nd the Advocacy of Violence
Netherlands Qu arterly of Human Ri ghts, Vol. 33/1 (2015) 101
committed’ [emphasis added]. As to this lat ter requirement, the Expla natory Report
to the Convention states that:161
When considering whet her such a danger is cause d, the nature of t he author and of the
addressee of the mes sage, as well as the context in wh ich the o ence is comm itted shall be
taken into account i n the sense establ ished by the case-law of t he European Cour t of Human
Rights. e signi cance a nd the credible nature of the da nger should be considered…
e problem with the di rection that states should consu lt the ECtHR’s case law is
that the ‘capable of’ st andard provided for in that case law is too vague to prevent
the Court from suppressing harmless speech. As Leroy shows, it is not di cult for
Courts to nd reasons to justify a conclusion that speech could result in violence.
Saygili162 als o demon strat es th is.163 In th at case, the Court justi ed its nding that the
pris oners’ decla rati ons were ‘capab le of in citin g viol ence i n the pr isons ’ by st ating that
the ‘overall message… where [the]… authors state that they wil l rather die than enter
the cells and ca ll on others to take act ion in support of their general resistanc e’ was ‘not
a peaceful one’ and could not be seen a s a ‘mere criticism of the new prison system.’164
Although Saygili is not as obviously wrong as Leroy, there is certainly much force in
the minority’s view (echoing Judge Palm in t he 8July cases) that the majority placed
too much emphasis on the intemperate manner in which t he message was expressed,
and failed to subject t he expression’s possible e ects to su ciently searching scrutiny:
speech, the minority thought, should not be suppressed simply ‘because it is neither
“peaceful” nor con ned to “mere criticism .”’165
Moreover, both Leroy and Saygili demonstrate the ‘chilling e ect’ that the ‘capable
of’ test ca n have on legitimate expression – including of ideas that are relevant to
democratic decision-making. e minority in Saygili recognised this, implying that
the crimina l penalties enforced against the applicants in that case had a tendency to
deter press discussion of matters of ‘obvious public concern’ (prison conditions).166
Similarly, as Voorhoof noted at the time, Leroy had a clear tendency to cause
‘cartoonists and colu mnists all over Europe’167 t o hesitat e before ex pressi ng any v iews
concerning terrorism.
However, as is suggested by the second sentence of the quotation from the CECPT
Explanatory Report set out above, we must weigh these considerations against
the grave harm caused by v iolent o ences.168 Should the government only be able
161 Counc il of Europe, ‘Expla natory Report to the Cou ncil of Europe Convention on t he Prevention of
Terrorism’ (CETS No. 196) [100].
162 Saygi li and Falakaoglu (n 5): see tex t accompanying nn 137–139.
163 Boyne (n 131) at 473–477.
164 Saygi li and Falakaoglu (n 5) at [28].
165 Ibid (diss enting judgment of Judges Power and Gyu lumyan).
166 Ibid.
167 Voorhoof (n 2).
168 e Explanator y Report refers to the need to c onsider the ‘signi canc e’ of the threatened dan ger.
Andrew Dyer
102 Intersentia
to prohibit the advocacy of violence when such advocac y is likely to cause violent
activity; t hat is, should it only be entitled to proscribe ‘violence-conducive’ speech
where there is a greater than 50 per cent chance of the speech resulting in harm? To
ask this question is to suggest the answer to it: the ‘likely to’ stand ard seems to be
unsuitable, because it would protect spee ch that makes violence a real possibility.
e solution to the problem of devi sing a test that: (a) does not suppress too much
speech that i s relevant to democratic (and othe r) decision-ma king or ‘chill’ le gitimate
speech; (b) allows for scrutiny of government clai ms that there was a ‘pressing social
need’ for the relevant rest riction; but also (c) protects the commu nity against violent
crime, appears to lie i n the words of the CECPT Explanatory Report. A s noted above,
the Report states that in determining whet her the ‘a danger’ risk threshold has been
met, ‘the credible nature of the danger must be considered.’ [emphasis added] ere
seems to be nothing preventing t his idea from being incor porated into the ECtHR’s
test; indeed, it is perfectly consistent with the test t hat Judge Palm suggested in t he
8July ca ses.169 at is, one way of achieving a bet ter balance between the interests
identi ed above (and therefore of ensuring that the only restr ictions permitted are
those that are ‘necessary in a democratic society’) is to require not merely that the
advocacy creates ‘a danger,’ but instead a ‘real risk,’ of resulting i n violence. While all
legal tests are open to man ipulation,170 the word ‘real ’ might focus judicial minds on
the need to provide more concrete reasons tha n were given in Leroy and Saygili for
conclusions that speech is su ciently dangerous to make its prohibition compatible
with Article10. is would especial ly be so if, in each advocacy of violence case, the
Court, when determi ning whether a ‘real ’ danger existed, were required to address
the three questions noted by Judge Pal m in the 8July c ases.171 It is submitted that a
Court that considered: whet her the speaker occupied a position of in uence such as
to ‘amplify the impact of his words’; the prominence of the publication in which t he
advocacy appeared; a nd (where applicable) the speech’s proximity to the ‘centre of
violence’, would be far less likely to uphold disproport ionate restrictions on expression.
4.3. THE IMMINENCE REQUIREMENT
e next question is whether it is necessar y that the advocacy of v iolence create an
imminent – as well as a real – risk of resulting in violence, before its prohibition is
held to be compatible with Article10. Of course, in the Turkish cases, Judge Bonello
placed great emphasis on the need for such a requirement, a nd cited with approval the
statements of Holmes and Brandeis JJ to similar e ect. Was he right to suggest that
169 See te xt accompanying n n 114–115.
170 See t he above discussion of Leroy v France i n the text accompanying n n 127–134; see al so the US
Supreme Court’s purp orted application of the ‘clear and pre sent danger’ test in Dennis (n 63). e
decision in Dennis has been w idely-criticised: see for e xample Schwartz (n 158) at 231–236; Geart y
(n 74) at 17; and Barendt (n 54) at 457–458 .
171 See tex t accompanying n 115.
Freedom of Expression a nd the Advocacy of Violence
Netherlands Qu arterly of Human Ri ghts, Vol. 33/1 (2015) 103
a state’s prohibition of speech of the kind considered in this a rticle ful ls a ‘pressing
social need’ only if the relevant expression creates such an immediate danger of
violence?
Barendt o ers two reasons why an imminence requirement ‘can be justi ed’;172
neither supports the inclusion in t he ECtHR’s test of such a condition. First, he argues
that:173
… the advocacy of d isorder, even terrorism, at some unspeci ed, perhaps inde nite, time
in the future i s not su ciently closely connected wit h any likely or intended act; abstr act
advocacy of this sort – the preaching of revolution by academic Marxists is an obvious
example – has more in com mon with archetypical pol itical speech, than it does w ith the
incit ement to c ommit a speci c act of terrorism or insur rection immediately or e ven in the
near future .
It is true that the mere abstract advocacy of violence should be lawful. But there is
no prospect that its prohibition wil l be upheld if, as was suggested above, the ECtHR
adopts a test tha t provides that ‘vi olence-con ducive’ advo cacy w ill be co mpatible w ith
Article10 only if that advocacy creates a real risk of violence. Such a test would not
allow for the proscription of academic Marx ists’ calls for revolution.
Second, Barendt argues t hat ‘the advocacy or encouragement of terrorism can be
countered in the long term by other speech: t he advocacy of other courses of act ion,
the dissu asion of prospec tive terror ists from t heir course.’174 at is, like Brandeis
J in Whitney175 – and some academic commentators176 – Barendt expresses faith
in the ability of ‘counter-speech’ to dissuade those who, because of the advocacy of
violence, are incli ned to commit such a n o ence. But he is not altogether clear about
this, because he proceeds both to concede that ‘counter-speech may be ine ective’177
and to observ e that we s hould be con dent that ‘better, more liberal ideas will prevail’
once there has been time for fu ll discussion – ‘in the absence of evidence that [the
speech]… is likely to cause a terrorist atrocit y.”’178 is is confusing because it suggests,
contrary to what Barendt has expressly st ated, that if advocacy creates a likelihood of
terrorist crime, there is no need for the threatened violence to be imminent.
Redish persuasively arg ues that there should be no imminence requirement of
the type envisaged by Brandenburg;179 his a rgument applies wit h equal force in the
European context. Redish objects to this requi rement partly because, ‘it harks back
to the ‘marketplace of ideas’ rationale for protecti ng unlawfu l advocacy… [T]here is
172 Barendt (n 54) at 457.
173 Ibid at 457–458.
174 Ibid at 458.
175 See tex t accompanying nn 2 0–23.
176 See for ex ample Healy (n 9) at 708.
177 Barendt (n 54) at 458 cit ing Alexander (n 156) at 109–110.
178 Barendt (n 54) at 458 [emphasi s added].
179 Redis h (n 16) at 1180–1182.
Andrew Dyer
104 Intersentia
simply no basis for the conclusion that the opportunity for reasoned response wil l
always defuse unlawful activity.’
180 As argued above,
181 this is undoubtedly true.
In other words, the fai lure of the argument from truth necessitates t he conclusion
that a ‘clear and present danger’ test of the type advanced by Judge Bonello in the
8July cases can not be accepted. Even a scholar such as Barendt, who expresses much
faith in the abil ity of ‘counter-speech’ to ex pose ‘falsehood and fallacies ,’ is forced to
accept the limitations of its ability to achieve this result.182 Moreover, Redish argues
that, ‘requiring tr ue imminence in ever y case is unrea listic and unduly insensitive
to society’s legitimate interest in self-protection.’183 is too should be accepted. To
apply the ‘pressing social need ’ test to an example that he provides, it is di cult to see
why a Contracting State’s punishment of a racist whose advocacy created a real risk
that other racists would kill black people on a sign i cant date mont hs hence, should
be held to be disproportionate to one of the Art icle10(2) aims.184 To put it in the terms
used above, individua ls’ autonomy and ability to make democrat ic decisions would
not be unduly compromised – and nor does it seem that legitimate speech would be
unnecessarily ‘chilled’, counter-productive results produced, or states’ proscriptions
inadequately scrutini sed – if the ECtHR were to adopt a test for ‘violence-conducive’
speech that lacked an i mminence requirement.
Nevertheless, the les s imminent the threatened violence, the les s able a Court will
be to nd wit h any assurance that speech creates a ‘real ri sk’ of violence. Accordingly,
in every ‘violence-conducive’ speech ca se, the ECtHR should be required to consider
the imminence of any danger threatened, when determining whether the impugned
speech creates such a real risk. It is perhaps not too optimistic to suggest that, if the
Cour t were req uired t o take i mmine nce into a ccount i n this way, it woul d be force d to
focu s more t han it has, i n cas es suc h as Leroy and Saygili, on justify ing any conclusion
that the relevant speech is da ngerous.
4.4. THE SPECIFIC INTENT REQUIREMENT
e nal matter is whether the ECtHR’s test should require that t he speaker had a
speci c intention that hi s/her expression result in violence. As noted above, Article5
of the CECPT obliges Council of Eu rope states to criminalise speech that not only
‘causes a danger’ that a terrorist o enc e w i ll be co mm it te d , bu t a l so is int en de d t o i nc it e
such an o ence . We have also seen that Judges Bonel lo and Palm both e ectively held
in the 8July c ases that a speci c intent component to the ECtHR’s test is necessary
180 Ibid at 1181.
181 See tex t accompanying n 26.
182 See tex t accompanyi ng nn 26 and 177–178; this is tru e also of Ben Saul , ‘Speaki ng of Terror:
Criminalising Incitement to Violence’ (2005) 28(3) University of New South Wales Law Jou rnal 868,
885.
183 Redis h (n 16) at 1180.
184 Ibid at 1181.
Freedom of Expression a nd the Advocacy of Violence
Netherlands Qu arterly of Human Ri ghts, Vol. 33/1 (2015) 105
to ensure that the Court upholds only those restrictions t hat ful l a ‘pressing social
need.’185
One argument in favour of having such a requ irement is that, without it, legitimate
speech would be ‘chilled’: individuals would be deterred from discussing particular
matters – including politica l issues – for fear that they might unintentionally s ay
something that creates a real risk of violent activity.186 Another is that t here is a
danger that a test that treats the speaker’s mental state as irrelevant, and considers
merely the risk of violence that t he speech causes, is l iable to be abused in time s of
panic.187 However, one major problem with having a speci c intention requirement is
that , if it w ere adopt ed, som e speec h carr yin g a high risk of harm w ould be p ermit ted.
For instance, such a requirement would ma ke it practically i mpossible to proscribe
journalistic spe ech about violence. Because journalists rarely i ntend to cause violent
acts, very litt le of their commentary about such matters could be proscribed –
irrespective of the ris k of harm that such material created.188
is indicates that the test for ‘violence-conducive’ speech should contain at t he
most an advertent recklessness mens rea standard (under such a standa rd, it would
need to be established, at least, that the speaker realised that it was possible that
his/her speech would result i n violence).189 But it is questionable whether even a
recklessness mens rea standard is warranted. It has been argued that a subjective mens
rea requirement is necessary bec ause this ‘ensures that no innoc ent speaker risks being
punished for unpredictable react ions by members of the audience.’190 is, however, is
to confus e the question of when it is proper to crimi nalise conduct, with that of when
speech may rightly be proscribed.191 It might well be objec tionable to hold criminally
liable a person who has neither intended nor foreseen the harm c aused by his/her act
of advocacy. is does not mea n, though, that it is wrong to use mean s other than the
criminal law to prevent such a person from publishing such speech when t here is a
real risk of it resulti ng in violent acts.
Nevertheless, as w ith the immi nence requirement, so here: where a speaker does
not even foresee the possibility t hat his/her speech wi ll result in violence, there will
o en b e no ‘real risk’ that violence w ill ensue. erefore, in every ‘violence-conducive’
185 See tex t accompanying nn 114 and 122–123.
186 Healy (n 9) at 709–710.
187 Frederick L awrence, ‘Violence-Conduci ve Speech: Punishable Verbal As sault or Protected Politic al
Speech?’ in David K retzmer and Fra ncine Kershman Ha zan (eds), Freedom of Speech an d Incitement
Against Democracy (Kluwer Law Intern ational 2000) 11, 22–23.
188 Sottiaux, Ter ro ri s m (n 4) at 122, 148.
189 R v G [2004] 1 AC 1034, 1057.
190 Unite d Nation s O ce on Drugs a nd Crime, ‘Preventi ng Terrorist Acts: A Cri minal Just ice
Strategy Inte grating Rule of Law Standard s in Implementation of United Nations Anti-Terrorism
Instrument s’ (United Nations 2006) 20. A simi lar concern appears to underl ie Ronen’s contention
that there should be a speci c intention requirement where a person ha s indirectly inc ited terrorist
activit y: Ronen (n 157) at 669.
191 Barendt (n 54) at 457; see also Ad rian Hunt, ‘ e Council of Europ e Convention on the Prevention
of Terrorism’ (2006) 12(4) European Public Law 6 03, 621–622.
Andrew Dyer
106 Intersentia
speech case, the ECtHR should consider whether the applica nt intended to cause
violence or foresaw that violence might result. I f either (especially the former) is the
case, the Cour t might be able more safely to nd that the impugned expression created
a ‘real risk’ of v iolence. It was argued above192 that, if the ECtHR has sug gested that a
state’s margin of appreciation is wider in cases where t here has been direc t advocacy
of violence, this statement of principle must be approached with some caution.
But there is another way of interpreting the Court’s statement that the margin of
appreciation is wider where the applica nt’s remarks ‘inc ite to violence.’193 at is, t he
Court might merely have meant t hat there is more likely to be a ‘pressing social need ’
for a restriction where the relevant speech w as intended to cause harm. If that was t he
case, its remarks were unobjec tionable.
5. CONCLUSION
In cases where an applica nt claims that a member state’s proscription of his/her
‘violence-conducive’ speech is a breach of Article 10, the ECtHR should adopt a test
that requires consideration merely of whether there is a ‘real risk ’ that the advocacy
would cause such violence. In other words, it is only where the relevant speech does
create such a genuine risk that there ex ists a ‘pressing social need’ for restrictions to
be placed on it. But while the Cou rt should require neither that the applica nt intended
to cause violence/foresaw violence as possible, nor t hat there was an im minent risk
of harm, it should consider both of these matters when deter mining whether t he
applicant’s speech created a ‘real risk ’ of violence. It should also consider the three
questions noted by Judge Palm in the 8July cas es. at i s, the following inquirie s – did
the speaker occupy a position of in uence in society such as to i ncrease the impact
of his or her words?; was the advocacy disseminated widely?; and (where applicable)
‘did the speech occ ur close to the ‘centre of violence’? – are all relevant when deciding
whether the risk of har m was a genuine one.
is test, i f faithfu lly applied, would achieve a proper balance between, on one
hand, protecting the community from violence, a nd, on the other, preventing states
from placing unnecessa ry restrictions on those who impart and receive informat ion
relevant to political a nd other decision-making. It would also ensure that the ECt HR
scrutinises w ith su cient care Contracting States’ clai ms that such restric tions are
‘necessary in a democr atic society’ for one of the purposes listed by Art icle 10(2).
Finally, it would prevent Contracting States from unduly deterring legitimate
discussion about violence, and might well avoid the counter-productive results that
would possibly ow from the adoption of a test that placed greater restrictions on
speech.
192 See text ac companying nn 85–8 6.
193 See for example S ürek (No. 1) (n 3) at [61].
Freedom of Expression a nd the Advocacy of Violence
Netherlands Qu arterly of Human Ri ghts, Vol. 33/1 (2015) 107
e decisions in Leroy and Saygili were objectionable mainly because the Court
provided no, or little, evidence to substa ntiate its claims that the impugned press
comment was ‘capable of’ causing v iolence. In other words, the Strasbourg judges
placed too much emphasis on the tone of the relevant advocacy a nd not enough on its
possible e ects. In so doing, they upheld disproportionate restrictions on speech. If a
‘real risk’ sta ndard were adopted in place of the vague and pliant ‘capable of’ te st, it is
quite possible that the Cour t would be forced in violent advocacy cases to substa ntiate
more fully their c onclusions that the requisite danger existed.
To continue reading
Request your trial