Glen Newey


Free speech controversies are never far from the headlines. Online censure allows the viral spread of indignation against media columnists and other public figures who have voiced controversial views. After the Jyllands-Posten cartoons fracas and the commotion over the Pope’s quotation in a speech at Regensburg of anti-Mohammad remarks by a medieval Byzantine emperor, conflict surrounded the UK entry ban (now lifted) on the Dutch politician Geert Wilders by the British interior minister for alleged Islamo­phobia, and the appearance on a BBC TV political discussion programme by the leader of the far-right British National Party (BNP) Nick Griffin. Free speech remains keenly contested.


With the possible exception of the Danish cartoons, these are controversies not only about freedom of expression, but freedom of speech.1 In each case the advocates of a ban found themselves opposed by a strong lobby of free-speech liberals. But the arguments often looked like a sectarian dispute, where opposing fundamentalisms snipe at one another from entrenched positions. This should be more embarrassing to liberals than to the advocates of restriction, who quite often do not deny being dogmatic. But with liberalism, anti-dogmatism is supposed to be the point: it aims to be more than “just another sectarian doctrine”.2

But the reasons for holding to the doctrine – at least as regards free speech – often seem obscure. Why insist on an absolute entitlement to free speech or expression, regardless of the costs? The absoluteness of the position, coupled with the lack of visible support for it, compound the sense that the free speech principle rests on dogma rather than reason.

Also obscure are the bounds of the doctrine, especially as regards the relation between free speech and free expression. These phrases are often treated as though they were synonymous. But, on any showing, ‘expression’ must be more encompassing than ‘speech’. As the conveying of semantic and illocutionary content, speech is necessarily expressive; but not all expression is verbal – for example, most of the Jyllands-Posten cartoons had no verbal content. Defenders of some non-verbal expression, such as pictorial pornography, sometimes conflate speech and expression, hoping to benefit from the special treatment that many accord to speech.

But does speech deserve such treatment? In fact, there are several levels at which the free speech doctrine needs clarifying. A major distinction, insufficiently remarked on in the literature, is between content and opportunity. Sometimes proponents of free speech seem to say that nobody should be denied the opportunity to contribute to public discussion, so that no group should be silenced. Here, what is held to matter above all is whether or not certain people get a say in some matter. This, for example, seemed to be in the minds of those who saw an attack on free speech in the ban on Geert Wilders or the possible forced withdrawal of the UK BNP leader from a political discussion programme.

On the other hand, some free speech advocates are concerned to deregulate the content that is put into the public sphere. In this case, the salient task is to remove restrictions not on who gets a say, but on what gets said. Some kinds of content may be deemed to be so heinous or pernicious that it should be forbidden to utter them. The obvious example is the legal prohibition in a number of jurisdictions, such as Austria, Germany, Israel, Romania, Belgium and Spain, on denying the Holocaust. Here it is a particular content which is proscribed, rather than banning specific groups from uttering it, or a particular mode of expressing it (Liechtenstein’s Holocaust denial law makes explicit mention of non-verbal modes of dissemination).

Differently again, the question of restriction may focus not on the content of what is said, or who says it, but on how it is said or otherwise conveyed. For instance, during the Vietnam War, Paul Robert Cohen, was prosecuted and convicted (overturned on appeal by the US Supreme Court) for wearing a jacket bearing the legend “Fuck the draft” in a California courtroom.3 Here the legal point at issue was not whether Cohen was entitled to utter the opinion that conscription was unjustified; he clearly had that right under the First Amendment. The legal issue was whether the expletive violated section 415 of the California Penal Code, which prohibited the ‘malicious’ and ‘wilful’ disturbance of the peace.

So free speech may be taken to require freeing either: certain groups of people or individuals who wish to speak, or certain kinds of content, or certain modes of expressing content, or certain contexts in which it is expressed. These possibilities call for rather different kinds of justification, and correspondingly different free speech regimes result from them. Accordingly, it seems naïve and unwarranted to think that there is one thing, free speech, supported by one form of justification.


It may be said, even if one grants that this monistic picture is wrong, there is at least a widely-held view among defenders of free speech, as well as its detractors, that ‘free speech’ names a unitary principle. If so, someone who represents free speech in a more pluralistic light should be able to explain how this illusion has gained currency. I shall offer an explanation of this, which will also serve to justify a distinct principle of free speech. This will also suggest limits within which the principle has to operate.

Let’s begin with a couple of platitudes. First, in liberal societies such as the modern capitalist democracies – the Netherlands, the US, Japan, the UK – the principle of free speech is widely upheld in law, in civil society and in political rhetoric. Second, if we consider actual speech situations, formal or informal censorship is pervasive.

Only the second of these observations is likely to seem at all controversial. But even leaving aside self-censorship and the informal regulating of speech which runs through social life, an apparatus of restriction marks speech situations in clubs, debating forums like parliament, forensic, liturgical and educational contexts, entertainments such as films and plays, commercial interactions, sporting contests, and so on. Restriction may apply in any or all of the three areas identified – who speaks, what can be said, how it can be said. In general, the more formalised a discursive context, such as in a catechism, the heavier the restrictions.

Speech wholly unrestricted in each of these respects is in fact rather rare, though not unheard-of. Some internet sites, and graffiti, provide examples of discourse which approximates a free-for-all. Two features of graffiti bear out the wider point about speech. First, much graffiti makes use of taboo content and expressions, i.e. language which is restricted or censured elsewhere; second, it is often unintelligible, precisely because the contextual cues conferred (inter alia) by restriction are lacking.

Nor is it just that the regulation is superadded to some already-existing piece of social life. Rather the regulation is partly constitutive of it. Judge Wendell Holmes famously located the limits on free speech as being crossed by falsely shouting “Fire!” in a crowded theatre.4 But more striking is the usual embargo on shouting or saying anything in a theatre, at least during the performance. And if the audience could say whatever they liked, whenever they liked, the theatrical event itself would be quite different, which was why some experimental drama dispensed with the embargo. Similar remarks go for religious ceremonies, lectures, interviews, and so on.

Insofar as the fact of regulation is visible, it is presumably thought not to conflict with free speech. Why? My suggestion is that an underlying assumption of voluntarism applies, at least to autonomous and compos mentis adults. In seminars, for example, presence is taken to signify assent to the ground rules, which impose heavy limits on opportunity, content, and expression.

Thus it would generally be deemed unacceptable, while my colleague is delivering a seminar paper on Kant, for me to rise and deliver an epic poem I have written about baseball, whatever the poem’s merits. And the point is that this is not really a restriction on my freedom, given that I freely choose to attend the seminar and so, implicitly, agree to the ground rules. In agreeing, I accept the restrictions placed on what I can say, when, and how.


Let me sum up the argument so far. Free speech is treated by proponents (and often opponents too) as a principle with wide scope and absolute weight. But those who propound the principle seem to have rather different notions of what free speech means in practice. This suggests that the associated justifications may also be correspondingly diverse. Moreover, the absoluteness of free speech has to be squared with the widespread limitations on speech in social, political and professional life.

I have now indicated that these limitations can be reconciled with freedom by understanding them as freely accepted, along contractual lines. This leaves further questions unanswered. Why do people think there is a single principle? Why do they also often think that the principle admits of a single justification?

My suggestion is that the basic or Ur-principle of free speech can be understood as a corollary of a further principle, freedom of association. This follows from a form of transcendental argument about the presumptive purpose of speech. Admittedly, the argument from purpose has had a chequered history, particularly when employed by for instance Aquinas and Mill5 to argue from the supposed alethic function of speech.

I aim to provide a less controversial grounding for free speech than those which say that speech has an essential purpose, such as conveying truths, and then gearing the justification of free speech to that purpose. Part of the problem is that any justification which appeals to ‘the’ purpose of speech is likely to be confuted by the range of other uses to which speech can be put, of which assertion (truth-conveying) is but one. So, for example, speech is used to entreat, repudiate, deceive, flatter, joke, confess, narrate, promise, condemn, pray, impress, extol, ridicule, insinuate, cajole, hector, avow, declare, accuse, inform, and so on.6

It is tempting to conclude that speech lacks any generic purpose. But this is mistaken. The generic purpose of speech is communication, a point missed by the monological rather than dialogical focus of many discussions of free speech. This is not to say that all uses of speech are communicative: I may speak, for instance, just to render what someone else is saying inaudible, or when I am alone. But this doesn’t matter much. Free speech opposes censorship, and the aim of censorship is to prevent communication. This may be to stop certain content being communicated, or communicated in a certain way, or to deny certain persons the chance to communicate.

Hence a principle of free speech essentially opposes attempts to deny communication, as a form of joint agency. This is why what underlies free speech is the principle of free association: communication in all its forms embodies an association between speaker and hearer (as well as sometimes being prefatory to further association, as in negotiations over a contract). This also explains why despite the different dimensions in which speech-limitation can occur, it is widely thought that there is a unitary principle of free speech.

In specific cases the focus may be on a speaker’s entitlement to be heard, or on an audience’s entitlement to hear her. It may, more narrowly still, highlight what is said, how it is said, or who says it. Communicative acts, in prospect or retrospect, may face denial on any or all of these fronts. And so the denying of denial – the freeing of speech – may focus on a specific aspect of the communication.

Once the picture foregrounds censorship as denial, and free speech as denying denial, it becomes less compelling to think of free speech as a free-for-all, a free and unstructured glossolalia. What free speech attends to are specific limitations on communication not agreed between those who would be a party to communicative acts. The point of freeing speech in context is to permit communication by removing specific obstacles to it. That is, it removes obstacles to freedom of association – which, contextually, permits agreed restrictions on speech, as already discussed. Hence the apparent conflict between a wide-scope free speech principle and the contextual proliferation of curbs on speech when it is realised that both follow from freedom of association.


This is of course a very abstract and general account of free speech. It does not show how far free speech is a non-negotiable principle, either in the round or in specific cases. It does not say anything about such matters as nuisance, commercial speech, non-verbal modes of expression, and so forth. I shall finish, however, by indicating that the outline theory I have given does have teeth, in that it can provide a guide to action in cases that have aroused dispute.

Hate speech or ‘fighting words’. Contrary to some writers who have tried to extract rather substantial principles (such as equal respect for persons) as transcendental conditions of speech, the very existence of hate speech suggests that such a principle either cannot be validly derived as conditions of speech, or their being so derivable is consistent with gross violations of the principle in concrete speech. So no protection against hate speech is afforded simply by speech as association.

However, at the public level matters are different. If certain epithets systematically exclude certain groups from the opportunity to associate, they frustrate the point of speech as association. They do so by denying members of these groups a voice, or a hearing, not merely in the way that they prejudice negotiation over future association, but by denying them presence here and now.

That is not to say that no use of derogatory epithets would benefit from free-speech protection. Different jurisdictions might construct the burden of proof differently. But the principle would work asymmetrically as between relatively disadvantaged and other groups, as regards not just specific derogatory terms, but also who was using them. A case in point would be the term ‘nigger’ as used, on the one hand, by white supremacists and, on the other, by African Americans. To give the principle legal effect, the bench would need to venture a judgment about whether the effect of a term’s use, in context, would be such as to exacerbate extant disadvantage were its use to become current through judicial protection.

Pornography. Purveyors of pornography have claimed First-Amendment protection and have received qualified support from the judiciary.7 That states, of course, that Congress “shall make no law … abridging the freedom of speech”.

Judicial limits on the First Amendment’s scope, usually to exclude hard-core porn from its protection, seem to have been largely ad hoc. However, the most obvious point is that over­whelm­ingly, pornography is not speech. Nearly all of commercial porn nowadays is visual, i.e. still or moving images.

Pornographers may try to claim that even this material is speech. This is at least dubious, since even if it has semantic content, it is doubtful whether making a photograph counts as a speech act. Second, visual pornography is not in any obvious sense an act of association, or prefatory to one. Rather it puts forward the fantasy of association, at least in the stereotypical porn created for heterosexual male consumption.

There is no blanket protection for porn under the arguments set out above. Either porn is not speech, so not protected by any principle of free speech, or else it is speech, but bypasses speech’s generic point. To deny protection is not necessary to state that porn should be banned. But it is to cast doubt on whether there is a “right to pornography”, as some theorists have argued.8


I add one final thought about the relation between free speech and free expression. As I have already said, those with a commercial or other interest in the latter have often tried to ride on the coat-tails of the former. With regard to pornography, it is more defensible jurisprudentially to argue not from a hard-to-sustain definition of obscenity, which is then held to constitute an ad hoc exception to First Amendment protection, but simply to deny that visual pornography is speech, so that it falls outside the amendment’s scope.

The point applies more generally. On any plausible showing, free expression will have to fit in with the criminal law. Few would argue that gross violations of public decency, for example by public bestiality or copulation, are defensible on the plea of freedom of expression, let alone (say) murder carried out as an ‘expressive’ form of performance art. If so, however, the law runs less risk of discrediting itself by treating such prohibitions not as extemporaneous departures from an otherwise sacrosanct principle of free speech, but as acts to which that principle is strictly irrelevant.

* Glen Newey is Professor of Politics and International Relations at Keele University, UK, and in 2008-10 is a Visiting Fellow at the Helsinki Collegium for Advanced Studies.

1 I take it that ‘freedom of expression’ subsumes ‘freedom of speech’, so that any question involving freedom of speech necessarily involves freedom of expression (but not conversely). I treat ‘free speech’ and ‘freedom of speech’ as synonymous, and similarly ‘free expression’ and ‘freedom of expression’.
2 John Rawls’s phrase: see his ‘Justice as Fairness: Political not Metaphysical,’ Philosophy and Public Affairs, (14 iii) 1985, 223-251, at 246.
3 Cohen v. California, 403 U.S. 15 (1971).
4 Schenck v. United States, 249 US 47 (1919).
5 T. Aquinas, Summa Theologiae, II-II q. 110 a. 3, J.S. Mill, On Liberty, ch. 2.
6 See: J.L. Austin, How To Do Things With Words 2nd edn. (Oxford: Oxford University Press 1976).
7 See e.g. Jacobellis v. Ohio, 378 U.S. 184 (1964).
8 R. Dworkin, ‘Is There A Right To Pornography?’ Oxford Journal of Legal Studies 1 (1981), 177-214.

The Amsterdam Law Forum is an open access initiative supported by the VU University Library.