Melbourne University Law Review
[Do the reasons that may justify restricting freedom of expression when a hate speech act is involved apply regardless of the particular race, religion, sexual orientation or other identity category that the speech act targets? In this article this question is answered by applying speech act theory and Foucauldian discourse analysis to two instances of hate speech: vilification on the grounds of race and sexuality. This article does not demonstrate that anti-vilification legislation targeting hate speech against such minorities as lesbians and gay men or Aboriginal people is justifiable, even if it assists understanding why it may be. But it does conclude that hate speech laws are unjustifiable if they outlaw vilification on such grounds as heterosexuality or Anglo-Australian identity. The argument is that vilification on these grounds cannot be understood as producing the kinds of harm that may justify limiting freedom of expression.]
Anti-vilification laws, or hate speech laws, are widely, though not universally, accepted as an important way of dealing with the problem of the dissemination of hatred. Different features characterise these laws in different jurisdictions. For example, some may provide for criminal remedies, others only for civil ones, and the hate speech acts targeted by the laws may differ from jurisdiction to jurisdiction.
One element, however, is common to most of these laws: they tend to be worded in neutral terms. This means that they proscribe the dissemination of, say, racial or religious hatred, but within the general category of race or religion they do not make further distinctions between one race and another or one religion and another.
A problem with this neutral approach is that the relevant law may be applied in a way that reinforces current social hierarchies. Thus, in the British context the relevant (neutrally worded) racial anti-vilification law ‘was used, at least in its first decade of operation, more effectively against Black Power leaders than against white racists.’
The point is not so much that we should wonder if it is possible for, say, racial hate speech laws to target racism, whatever its colour, without also impinging upon legitimate political protest. Rather, the question is whether or not laws proscribing racist hate speech should be colour-blind in the first place. Analogous questions could be asked in the context of laws proscribing the dissemination of hatred on other grounds such as nationality or religion.
Although such questions have not often (or have only cursorily) been discussed by legal scholars, a similar question has indeed been asked by Australian law-makers with respect to New South Wales’s sexuality anti-vilification legislation.
Increasingly, anti-vilification laws are forbidding the dissemination of bigotry when this is motivated by sexuality. In the Australian context, Western Australia and Queensland are reportedly interested in possibly passing sexuality anti-vilification legislation in the future, while Tasmania and New South Wales are already endowed with such laws.
Tasmania’s sexuality anti-vilification legislation is neutrally worded. A similar neutrality is contemplated in a private member’s Bill recently introduced in the Senate, directed to outlawing hate speech on the ground of sexuality across Australia. The sexuality anti-vilification legislation of New South Wales, however, has a feature that sets it apart from other anti-vilification laws. New South Wales stands alone in Australia, and quite possibly in the world, in prohibiting the dissemination of hatred specifically on the basis of homosexuality rather than on such neutral grounds as ‘sexuality’ or ‘sexual orientation’.
Last year, a Bill introduced in the New South Wales legislature proved unsuccessful in its attempt to change the relevant legislation so as to extend protection from vilification to heterosexuals. However, the justification provided by the government’s spokesperson for refusing to replace the category of homosexuality with the wider one of sexuality in New South Wales anti-vilification legislation was not very convincingly articulated. The New South Wales government expressed its view that the legislation was meant to promote ‘substantive equality for minority groups’ and that ‘[t]here is no ... evidence to suggest that heterosexuals are routinely vilified on the basis of their sexuality.’ But the government did not persuasively explain why legislation worded in neutral terms would not support substantive equality for lesbians and gay men. Nor did it dwell on the reasons why vilification, if it happens on an episodical basis rather than as a matter of routine, is unworthy of a legal response.
In Western Australia, a Ministerial Committee established to formulate recommendations on several aspects of that State’s sexual orientation law completely failed to address these issues when it recently proposed the adoption of an anti-vilification regime modelled on the New South Wales legislation. The Western Australian Committee did recommend adopting the New South Wales model, but its report does not make it clear whether or not the fact that New South Wales outlaws vilification on the ground of homosexuality contributed to the Committee making this recommendation.
In short, the question of whether or not anti-vilification laws should be neutrally worded has not yet been convincingly answered by either legal scholars or public authorities. On closer inspection, it may also be the case that that question is not the most appropriate for us to ask.
The experience of New South Wales in the area of racial vilification suggests that neutrally worded hate speech laws may be practically administered so as to shield only vulnerable minorities from vilification, and not other groups. Thus the crucial question is not the technical one of whether or not hate speech laws should be neutrally worded, but rather one more philosophical in nature. That question is: ‘do the reasons that may justify restricting freedom of expression when a hate speech act is involved apply regardless of the particular race, religion, sexual orientation or other identity category which the speech act targets?’
In the following analysis I will address this question using the examples of vilification on the grounds of race and sexuality. This article will not conclusively demonstrate that anti-vilification legislation targeting hate speech against such minorities as lesbians and gay men or Aboriginal people is justifiable, even if it contributes to understanding why it may be. But an important conclusion, however limited, will be reached: that it is unjustifiable for anti-vilification laws to outlaw hate speech targeting individuals or groups on such grounds as heterosexuality or Anglo-Australian identity.
A key point was conspicuous in its absence from the New South Wales government’s justification for not extending the protection afforded by anti-vilification legislation to heterosexuals: the importance of freedom of expression. If free expression is a fundamental value which presumptively applies to all — or virtually all — speech regardless of its content, heterosexual vilification can be outlawed only if sufficiently important justifications are adduced.
The relevance of freedom of expression to the Australian legal system has been theorised relatively recently by the High Court. On a number of occasions during the 1990s, the Court has read into the Australian Constitution an implied guarantee to freedom of political communication, which it argued is a necessary ingredient of representative government (the so-called ‘argument from democracy’).
There are good philosophical reasons for arguing that freedom of expression should presumptively cover hate speech, just as it should presumptively cover all speech regardless of its content. This does not mean that the regulation of hate speech is necessarily unwarranted. It only means that if we wish to justify such regulation, we need to point to countervailing considerations and explain why they overcome the presumption in favour of liberty of communication in that particular context.
The ‘argument from democracy’ alone — appropriated by the High Court’s freedom of political communication decisions — is sufficient to establish, philosophically, that people should, subject to countervailing considerations, be presumptively free to engage in all kinds of speech, regardless of its content (including acts of vilification). This is because the best version of the argument from democracy is not that self-government requires freedom of political speech. Rather it is that self-government, properly understood, involves a process of collective self-definition which requires freedom of (unqualified) speech.
Thus, in order to invoke the argument from democracy in the context of hate speech, we need not characterise vilification as political speech. Vilification would, by virtue of being speech of some kind, be covered anyway by the presumption in favour of freedom of expression established by the (best version of the) argument from democracy.
However, the version of the argument from democracy on which the High Court seems to have settled is one that ‘leaves no room for any suggestion ... that it might protect freedom of speech generally’ instead of the more limited concept of freedom of political communication. Thus, if we wish to make the present discussion relevant to the Australian context rather than an exercise in wishful thinking, we need to argue that vilification presumptively falls within the boundaries of the judicially identified category of ‘political communication’.
Michael Chesterman has already done this with respect to racial vilification. He has argued that ‘political communication’ may be interpreted broadly or narrowly, and that under the broad conception — which he defends as more balanced — racial vilification clearly qualifies as political communication for constitutional purposes.
As far as sexuality vilification is concerned, I would argue that both heterosexual and homosexual vilification can be considered ‘political communication’ even under the narrow interpretation of that concept. According to Chesterman, adopting the narrow conception of political communication requires that, in order for an expressive activity to qualify as political communication,
some sort of ‘match’ ... be identified between the specific subject-matter of the speech and some topic which falls or might fall for consideration by the legislative or executive branch of the Commonwealth Government or by some other agency (including State and local government bodies) which is politically or institutionally linked with either of these.
At least more often than not, the messages conveyed by both homosexual and heterosexual vilification will resonate with issues which may fall for consideration by the government. Australian laws still discriminate against lesbians and gay men in a variety of areas, and lesbian and gay lobbying ensures that such inequalities of treatment remain political issues and, indeed, high profile ones. Even the least articulate homophobic remark says a lot about how lesbians and gay men should be treated. As such, that remark clearly identifies a particular stance with regard to such issues as access to reproductive technology services for lesbians or superannuation rights for same-sex partners.
An act of heterosexual vilification will also generally evidence a certain political belief with respect to the same issues: that heterosexuals are being unfairly privileged by the legal system. If, for example, the message were to the effect that heterosexuals are ‘fascist’, it would be clearly relevant to such political issues as the adequacy of the current government’s responses to homophobic violence.
If this is true, then even under the least generous interpretation of the version of the argument from democracy appropriated by the High Court, both heterosexual and homosexual vilification qualify as ‘political communication’. Thus, they are presumptively protected unless weighty countervailing considerations can be put forward.
One may remain unconvinced that the argument from democracy means that people should be presumptively free to engage in acts of vilification. Even admitting that hate speech is ‘political communication’, does it not reflect the wrong kind of political commitments — the kind that right-thinking people have collectively rejected and that should have no citizenship in a democratic system properly understood? In other words, showing that vilification is ‘political communication’ may not be the same as showing that it contributes to democratic self-government. And if vilification does not contribute to democratic self-government, the presumption in favour of freedom of expression does not apply in the context of hate speech.
For our present purposes, if hate speech has no merit in terms of the argument from democracy, it should be distinctly easier to outlaw it irrespective of whether it targets vulnerable minorities or other groups.
Indeed, if all we focus on is the inherent value of hate speech, there is no doubt that hate speech is beyond redemption. However, hate speech may have instrumental value.
The instrumental value of (protecting) hate speech may be thought to be threefold. First, protection of hate speech may be a way of establishing a useful precedent. This can be used, once the need arises, to shield the political dissent of disadvantaged/unpopular minorities from the silencing practices that the mainstream would be willing to implement through the law. Secondly, the line between acceptable political dissent and hate speech can, on occasion, be thin. In this sense, protecting hate speech is a way of ensuring that no chilling effect interferes with the legitimate and valuable expression of such dissent. Finally, hate speech may contribute to the vitality of opinions emphasising egalitarianism, insofar as such opinions are compellingly reaffirmed as a response to the challenge levelled at them by hate speech.
In short, the fact that hate speech conveys a message which is false and does not directly add to rational democratic discourse should not disqualify hate speech from the realm of presumptively protected expression. This is because hate speech retains instrumental value. It is significant that the High Court has pointed out that political speech is protected even where the messages conveyed are ‘false, unreasoned and emotional’.
This emphasis on the importance of freedom of expression in the context of the dissemination of hate carries with it the risk that the value of equality will be downplayed. For example, civil libertarians often argue against regulating hate speech without adequately addressing the relationship between hate speech and discrimination. As a result, hate speech is treated like any other kind of unpopular expression, without paying sufficient attention to surrounding contexts of racism, (hetero)sexism and other forms of bigotry.
Even so, it would be just as flawed philosophically to subscribe to the position of those who discount the general soundness of the freedom of expression doctrine in order to support their position favouring hate speech laws.
Besides, a presumption in favour of free speech (or freedom of political communication) need not pre-empt the debate about whether or not hate speech should be regulated, in such a way as to make it impossible to argue in favour of hate speech restrictions. Such a presumption only means that restrictions need to be rigorously justified. Thus, even though they started from the same proposition that strong freedom of expression is essential to democracy, Canada and the United States were able to reach diverging conclusions on the issue of hate speech regulation.
In the process of developing justifications for the regulation of hate speech, we may discover that whilst some restrictions may be justifiable (for example, those against homophobic hate speech or against racist hate speech targeting racialised minorities), others may not be. This is precisely the conclusion to which Mari Matsuda came in the context of racist hate speech.
In ‘Public Response to Racist Speech: Considering the Victim’s Story’, Matsuda articulates an important perspective inasmuch as she argues in favour of hate speech laws even though her point of departure is the importance of freedom of expression.
Matsuda takes freedom of expression principles seriously, but looks for reasons to exempt hate speech from the general operation of those principles. Accordingly, the more contextual her approach, the better the arguments it can make for overcoming the scepticism of a perspective biased towards freedom of speech. With this principle in mind, Matsuda concentrates her analysis on the problem of racist hate speech rather than on hate speech more generally.
Matsuda highlights the structural presence of racism as a practice of subordination effected through the implementation of several techniques. She considers racist hate speech to be one of these techniques. Put into a context of historical oppression, racist hate speech adversely affects minorities’ ability to enjoy equality by undermining their own sense, as well as dominant groups’ perception, of their equal worth.
Matsuda holds that our historical experience should make us see the difference between racist hate speech directed at minorities and other forms of speech that are controversial or offensive but unrelated to the maintenance of subordination. Matsuda’s commitment to freedom of speech makes her highlight the contextual differences between speech vilifying minorities and speech vilifying dominant groups. In her analysis there is no justification for laws outlawing, in addition to hate propaganda directed against historically oppressed groups, hate speech against ‘dominant’ groups.
The aim of the next sections is to expand on this point made by Matsuda in a way that does not rely, as Matsuda’s analysis does, on an appeal to historical experience alone. In particular, in what follows I shall use the analytical framework of speech act theory. My goal will be to explain why anti-heterosexual or anti-white bigotry (as opposed to homophobic hate speech or vilification against racialised minorities) cannot be characterised as producing the sort of harm that justifies subjecting speech to legal regulation.
Saying that speech causes harm means that speech has an ability to do something. This ability to do things with words can be investigated within the analytical framework of speech act theory. This has developed the concept of the performativity of speech, which refers to the ability of words to bring about what they name.
Every speech act is a locutionary act. This means that whenever I speak, I am doing something, because speaking is, to begin with, a specific type of action, not unlike eating, sleeping or driving.
Every speech act is also always an illocutionary act. The illocutionary aspect of a speech act determines the way in which we use the utterance. That is, it determines whether we speak in order to advise, order, warn, promise, suggest and so on. Whether I am using my utterance to do one or another of these things will depend on what I say and the context in which I say it.
Utterances, in their capacity as illocutionary acts, have an ability to perform certain results. The specific illocutionary result will vary depending on the type of speech act in question. Thus, the illocutionary result of a minister’s uttering a marriage formula is that the spouses are bound in marriage; and the illocutionary result of my promising you something is that I am bound by my promise.
In order for an utterance to enact its illocutionary result, the illocutionary ‘uptake’ must be secured. This happens when the ‘force’ of the utterance is understood and, by virtue of being understood, established. In other words, the hearers need to understand in which way I want to use my speech act (for example, to promise or to order) if my speech act is to achieve its illocutionary result.
We can better understand the concepts of illocutionary force and illocutionary result by considering, rather than a speech act at random, specifically a performative — or, in Jürgen Habermas’s terminology, regulative — speech act. Even if every speech act is an illocutionary act, it is in the use of regulative utterances that the illocutionary force of the speech act is emphasised. This is the same as saying that in a regulative utterance the relationship entered by speaker and hearer is central, because such a relationship varies according to how the utterance is used, and the way in which an utterance is used reflects the particular illocutionary force of the utterance.
As an example of an illocutionary act, Austin refers to (the regulative speech act of) christening a ship. Provided certain conditions are met, in uttering the formula which mentions the name of the ship, that name becomes the ship’s name. These conditions are the so-called conditions for the happiness or felicity of performatives — including that there exist a procedure for the act and that the procedure be performed completely and correctly by the right person.
Illocutionary results must be distinguished from perlocutionary effects — that is, the consequential effects on the feelings, beliefs or actions of an audience. The production of perlocutionary effects is the third sense in which one does something with one’s words.
If the conditions for the felicity of the performative are met, the illocutionary result will take place quite independently of whether I manage to produce the perlocutionary effects by which I may have wished to see my audience affected. For example, I may wish to change my audience’s opinions: this would be my perlocutionary aim, and it could fail to come true even if all the conditions for the felicity of the performative were met. This is because the specific perlocutionary effects which are actually produced in any given case depend on the particular context in which my speech act is uttered. Thus, if my audience is very impressionable, it may be easier for my perlocutionary aim to be achieved. On the other hand, my perlocutionary aim may fail even if the utterance is an illocutionary success.
In the example above, it does not matter that the audience at the christening of the ship, despite the fact that they understood the meaning and the force of the locution, are unconvinced and will refer to the ship in the future as the Generalissimo Stalin. The ship’s name will still be the Queen Elizabeth. That is, the illocutionary result has been secured irrespective of the perlocutionary effects.
Habermas distinguishes between institutionally bound speech acts (such as christening a ship or performing a marriage ceremony) and institutionally unbound speech acts (such as promising or warning). It is only when the former are performed that the authority of certain institutions is directly engaged.
Institutionally bound speech acts derive their power to coordinate human relationships — that is, their illocutionary force — from the social force of norms. In other words, their ability to bring about their illocutionary result is authorised by ‘official’ norms. These are norms produced by certain social formations — institutions such as the state and the church — whose authority is generally recognised by the community to whom the norms purport to apply. As long as the norm is not repealed or otherwise deprived of its legitimate force, it will provide the authority for performing what the utterance names.
Thus, in the utterance ‘I proclaim the name of this ship to be the Queen Elizabeth’, there is a norm which underlies the ‘I proclaim’ formula. This norm authorises the speaker to name the ship while subjecting the hearers to that authority, hence providing support for the establishment of a particular relationship between speaker and hearer. In order to bring about the illocutionary result of the ship’s name being the Queen Elizabeth, this norm needs nothing but the felicity of the performative. The first condition of felicity is that there exists an accepted conventional procedure with a given conventional effect. This is the same as saying that there must exist a norm authorising, among other things, the speaker to perform the utterance and, in so doing, achieve the illocutionary result.
Institutionally bound speech acts are instances of regulative utterances. Regulative utterances, however, can also be institutionally unbound. Regulative utterances that are institutionally unbound (such as the orders not originating from an official authority) are related, not unlike institutionally bound speech acts, to certain norms. These norms justify the specific kind of relationship into which a speaker and a hearer enter as a result — indeed, as the illocutionary result — of the utterance being performed (for example, my obligation to do what you order). However, the norm invoked by an institutionally unbound regulative utterance cannot borrow its force from the institution officially producing the norm because, the act being institutionally unbound, there is no such institution. Thus, invoking the norm is not sufficient, in and of itself, to authorise an institutionally unbound regulative utterance to achieve its illocutionary result (for example, our entering a relationship whereby I must do as you have ordered). In order for the norm to provide that authority, the speaker must offer, even if only implicitly, to defend the authority of the norm (for example, I must have a sense that you are prepared to justify why it is right that you should have the authority to order me to do the thing you are ordering).
To use the vocabulary of speech act theory, the interpersonal relationship described in the illocutionary component of an institutionally unbound regulative utterance is justified by norms (implicitly invoked) whose rightfulness the speaker implicitly offers to vindicate. This is the same as saying that an implicit validity claim underlies institutionally unbound regulative utterances, and this very claim allows the speech act to perform its illocutionary result.
What has just been said with regard to institutionally unbound regulative speech acts illustrates a point which applies to institutionally unbound speech acts in general. That is to say, institutionally unbound speech acts have an ability to produce their illocutionary result because of the implicit offer speakers make to justify the validity claims they implicitly raise when speaking. These validity claims are:
These three different validity claims are implicitly raised, according to Habermas, in every (institutionally unbound) speech act directed to reaching understanding, but they are differently ‘thematised’ depending on the nature of the speech act at issue. Thus we have seen that when a regulative speech act is uttered, the speaker emphasises the third of these claims. That is to say, they ‘thematise’ the relationship that they intend to bring about in using their words in order to, for example, issue a warning or make a promise.
The problem in describing how racist hate speech acts can secure an illocutionary result (that is, enact what they name) is that they do not fit into this framework.
On the one hand, they are not officially and technically bound to any institution. Thus, they cannot produce their illocutionary result — that is, they cannot derive the authority to perform what they name — by referring to norms whose force is guaranteed by such organisations as the state, its agencies or the church.
On the other hand, Habermas has developed the doctrine illustrated above about the universal validity claims in order to explain how speech acts directed to reaching understanding can secure their illocutionary success. But a racist hate speech act is hardly describable as an utterance oriented towards reaching understanding.
In fact, racist hate-mongers are likely to desire the realisation of changes in the beliefs, thoughts and feelings of their audiences — that is, they desire the production of specific perlocutionary results. However, in Habermas’s view the presence of a speaker’s perlocutionary aim makes the utterance an act of what he calls ‘linguistically mediated strategic action’, as opposed to an act oriented towards reaching understanding. Thus, Habermas’s doctrine about the validity claims universally inhabiting speech acts oriented towards reaching understanding does not, strictly speaking, apply to the case of hate speech acts.
In order to explain how hate speech acts can achieve their illocutionary result, then, we have to add some nuances to the picture. I shall do so by combining elements of Foucault’s discourse analysis with the account just given of speech act theory.
Arguably, even though they are not technically bound to any specific institution, racist hate speech acts can draw from the social force of norms the authority they need to produce their illocutionary result — not unlike institutionally bound speech acts. In other words, there is no need for a racist hate-monger (as opposed to an author of institutionally unbound speech acts directed to reaching understanding) implicitly to offer to justify any validity claim of any type.
In fact, could we not dislodge the ‘institutionally bound/institutionally unbound’ dichotomy by referring to the Foucauldian notion of power/knowledge regimes? Suggesting, as Foucault does, that power is not only (and not even mainly) juridical, loosens the ties between power and juridical and quasi-juridical norms — that is, norms linked to some sort of official institution.
Juridical norms, as we know, have a force which is de facto. This means that the speaker who implicitly invokes such a norm (to make their regulative institutionally bound speech act produce its illocutionary result) need not implicitly offer to justify the norm’s rightfulness. This is because the norm’s authority is guaranteed by the power of the institution to which the norm is linked. For example, a minister officiating at a marriage ceremony need not implicitly offer to justify his authority to perform the marriage and he need not justify why the spouses will be married as a result of taking part in the ceremony he performs. It is the law that provides any relevant justification.
But, Foucault tells us, the dimensions of power are not exhausted by the juridical; power is relational and not concentrated in, or possessed by, any particular institution. Could we not think, then, of norms that have a de facto force — that is, a force that need not rely on somebody’s implicit offer to justify the norms’ rightfulness — even if these norms are not produced as the official norms of any juridical or quasi-juridical institution?
Power, from a Foucauldian perspective, ‘acts’ through discourse, and it is through discourse that truth effects are installed. Because the discursive truths thus produced are power-laden, they can be considered in some sense ‘normative’ rather than objective. But this is the same as saying that these normative truths have the status of ‘true’ norms — that is, norms that, to be binding, do not require the speaker who is implicitly invoking them to make an implicit offer to justify them.
If this is so, then a question arises. Do appropriate ‘true’ norms — that is, non-institutional, or non-juridical, norms with a special social force — exist that can secure the illocutionary result of enacting what is named by racist hate speech acts? We can put this question in another form. ‘True’ norms, as I have suggested, are produced — in the form of normative (that is authoritative, rather than objective) truths — by discourse, as discourse is coextensive with power. Does a recognisable, conspicuous discourse about blackness, Aboriginality, Asian-ness, and other such marginalised identities exist?
There is little doubt that the answer is ‘yes’, and it is a discourse not localised in any particular institution. Rather, it crosses transversally a number of sites, disciplines and institutions. The structural presence of racism to which Matsuda refers is maintained precisely by the totality of these discursively produced, racist normative truths about minorities.
The double structure of these normative truths, which, as I have suggested, are at the same time ‘true’ norms, allows them to be invoked to ensure the illocutionary success of racist hate speech acts in two cases. In their capacity as true norms, they authorise regulative racist hate speech acts to produce their illocutionary result. In their capacity as normative truths, they authorise constative racist hate speech acts to produce their illocutionary result.
But what is a constative speech act? Habermas says that when we use language in a cognitive way, we utter constative speech acts (such as statements) and concentrate on the propositional content of an utterance. Thus, if I say ‘I state that I am a teacher’ you will focus on the ‘I am a teacher’ part. In fact, I would normally say just that, without prefacing ‘I am a teacher’ with ‘I state’.
On the other hand, when we use language in an interactive way, we utter regulative speech acts (such as orders), bringing to the fore the interpersonal relationship established through the illocutionary component of the utterance — that is, the way in which we use the utterance. Thus, if I say ‘I order you to leave this room’ your focus will primarily be on the way in which I use my speech act — that is to say, to give you an order. Even if I just say ‘leave this room’, you will, to begin with, concentrate on whether or not I have the authority to order you about. In other words, you will focus on our interpersonal relationship.
Recall that three implicit validity claims underlie all speech acts directed to reaching understanding: the claim to the truth of what is said (that is, of the propositional content of the utterance); the claim to the truthfulness of what the speaker expresses with the speech act; and the claim to the rightfulness of the norms that govern the interpersonal relationships established in using the utterance with a certain force rather than another (that is, in order to warn, promise, bet, proclaim and so on). Whilst in regulative speech acts we stress, as we already know, the last of these validity claims, in constative speech acts we stress the first.
This distinction is drawn by Habermas in the context of speech acts directed to reaching understanding. But we know that hateful utterances, whether they be regulative or constative, are scarcely amenable to being defined as speech acts oriented towards reaching understanding. Therefore, their illocutionary success in bringing about what they name must rest on something other than the implicit offer to justify the validity claims invoked at the same time as the utterance is issued.
We have seen that the illocutionary force of institutionally bound (regulative) speech acts is secured by pointing to norms whose force is, so to speak, de facto. That is to say, the speaker does not have to implicitly offer to justify the rightfulness of these norms, as the link with the institution is enough to make these norms binding. I have further suggested that there are other norms that have a similar de facto force, even if they do not derive it from the juridical power of institutions, but rather from the power of social discourses. We have seen that such a powerful discourse exists about racial minorities. The conclusion was that it is precisely by referring to such norms that regulative racist hate speech acts (which are institutionally unbound) can secure their illocutionary success.
It can be argued that constative racist hate speech acts can likewise obtain their illocutionary force by pointing to truths whose authority is not rationally based, but de facto. These truths are the same norms that regulative hate speech acts rely on to ensure their illocutionary effectiveness. Remember that (power through) discourse produces truths which, being power-laden, become normative (rather than objective) truths; and which, by virtue of their being normative truths, are at the same time ‘true’ norms.
The double-structure of these effects of discourse is crucial. In their capacity as (true) norms, such discursive effects are implicitly invoked by regulative racist hate speech acts, because in regulative utterances our focus is on the rightfulness of the speaker/hearer relationship and that rightfulness is judged on the basis of norms. In their capacity as (normative) truths, these discursive effects implicitly authorise constative racist hate speech acts, because in constative speech acts our focus is on the truth of the propositional content of the utterance.
When a racial slur is uttered, it can assume the form of a regulative speech act. Made fully explicit, such an utterance would be:
‘I name you a — [racial slur]’
or, even more emphatically:
‘I hereby proclaim that you are a — [racial slur]’.
This use of language brings to the fore the relationship (here one of subordination) between speaker and hearer. The illocutionary effect in this case is precisely the enactment of such a relationship.
The speaker’s intention to bring about such a relationship of subordination is made clear in the formula ‘I hereby proclaim’ — that is, the illocutionary component of the utterance (of course, the aim would be clear even without expressly prefacing the racial slur with the illocutionary formula). This formula emphasises the position of authority assumed by the speaker. The speaker’s authority to degrade the addressee derives from the de facto (‘common sense’ and irrational) social force of norms produced by discourses about racial minorities. The speaker implicitly invokes these norms which codify the appropriateness of hierarchical relationships between races considered to be of different worth.
In addition to constituting the propositional content of a regulative speech act, the same racial slur could also be used in a constative speech act. This, in its fully explicit form, would be:
‘I state that you are a — [racial slur]’.
The propositional component (that is, the ‘you are a —’ part) of the utterance is here brought to the fore. In other words, the focus here is on the truth of what is stated (rather than on the relationship between interlocutors thematised by the illocutionary part of the speech act — that is, the ‘I state that’ formula).
The propositional content of this statement — that is, another person’s inferior worth because of their racial group membership — does not contain the speaker’s implicit offer to justify the truth claim raised in it. This is because this constative utterance, being a hate speech act, is not directed to reaching understanding. Thus, the illocutionary force of the utterance — the ability of the speech act to perform what it names — is here secured in another way.
In particular, the message of inferiority contained in the racist hate speech act fulfils itself owing to the congruence between this message and the ‘truths’ produced by social discourses about racial minorities. These construct a social subject of inferior worth on the basis of its racial group membership.
In practice, a hate speech act consisting merely of a racial slur is likely to have a regulative as well as a constative dimension. It follows that the final illocutionary result will be twofold. On the one hand, the illocutionary success of the regulative dimension of the slur will be actualised in the establishment (or rather, the reinstallation) of a relationship of subordination in which the speaker has the authority to degrade the addressee. On the other hand, the illocutionary result will also consist in the enactment of the propositional content of the utterance — that is, in the addressee (and, by extension, other members of their racial group) being tied to a social identity of inferior worth. This will be thanks to the illocutionary success of the constative dimension of the slur.
The latter result will ensue even if that social identity is not deemed to be ‘valid’ by the addressee and/or by a non-racist spectator, so that the speaker’s perlocutionary aim fails (because the speaker cannot convince the audience of the addressee’s inferior worth). The reason for this lies in the familiarity, both on the part of the addressee and the spectator, with the ‘truths’ produced by discourses about racial minority groups. This familiarity would work, in the case we are considering, in a way similar to the hearers’ knowledge of the existence of the norms authorising the officer to name the ship the Queen Elizabeth in Austin’s example. The hearers might decide, for political reasons, that those rules on the christening of ships are not legitimate because, for example, they are the expression of Western will to power. They might, therefore, subsequently refer to the ship as the Generalissimo Stalin. The ship’s name, however, would hauntingly remain the Queen Elizabeth, precisely because those norms that the officer followed in christening the ship do exist and have a certain social force, of which the hearers are aware even if they disavow it.
The illocutionary success of a racist hate speech act will generally also be secured when the hearers are the spectators to, but not the victims of, the racist utterance.
Let us take the case of a regulative hate speech act issued by a white hate-monger warning a white audience against the dangers supposedly created by a group of racialised immigrants. The audience would obviously understand the meaning of the words employed by the speaker (the locutionary aspect of the utterance). The utterance, however, would also ‘work’ at the illocutionary level, even if the whites listening to the warning made up their minds to reject it as unfounded and were not alerted by it — that is, even if the speaker’s perlocutionary aim failed. This is because the audience would have no problems in understanding the speech act as a warning.
In other words, the (illocutionary) force of the locution — the way in which it was meant to be used — would not be incommensurable with the discourses constituting the subjects in the audience, and it would not be lost on them. The regulative utterance (the warning) and the interpersonal relationship established through it between speaker and hearer (whereby the former appears to have a ‘right’ to warn the audience) would be implicitly justified by certain norms of intra-group solidarity. These norms are, in our example of a white speaker and a white audience, the counterpart of those norms of inter-group hostility produced by discourses about the racialised other which are implicitly invoked when the addressee of the racist utterance is also its victim.
In the article cited above, Matsuda chooses not to extend her analysis to anti-gay and anti-lesbian hate speech ‘because of ... the different way in which sex operates as a locus of oppression.’
There are indeed differences between the ways in which racist hate speech and hateful speech directed against gay men and lesbians operate. In particular, the illocutionary force of homophobic hate speech acts is only partly derived from the true norms/normative truths that are discursively produced about sexuality, homosexuality, compulsory heterosexuality, and the continuity among sex, gender and desire. Much of that force, on the contrary, comes from the law itself, as an institutionally produced discourse.
The role of the law in securing the illocutionary success of homophobic hate speech acts is complex. To begin with, the law, not unlike other social discourses, produces normative truths/true norms of the kind described so far, for example through the elaborations of judicial discourse on homosexuality. In these cases, the naturalness of heterosexuality is discursively produced both as a (true) norm and as a (normative) truth by the law, alone and in combination with a range of other social discourses. These truths/norms, in turn, sustain the illocutionary force of homophobic hate speech acts, both regulative and constative, and whether addressed to victimised or non-victimised hearers.
The process parallels that described for racist hate speech acts and allows the utterance to be successful at the illocutionary level (assuming the conditions of felicity are met) even when the speaker’s perlocutionary aim goes amiss. That is, the existence of these truths/norms about homosexuality allows the audience to relate to the utterance in a way that goes beyond mere understanding of the locutionary meaning of its propositional content. Let us consider the example of constative utterances.
Clearly, a statement such as ‘gay men have been unfairly treated’ could be successful at the illocutionary level. The reason is that, as an act directed to reaching understanding, this utterance implicitly contains the speaker’s offer to justify the claim to truth underlying the constative speech act.
On the contrary, despite its clear locutionary meaning, a statement affirming, say, that gay men are surrounded by rainbow-coloured halos would generally secure no illocutionary uptake. The reason is that the hearer would not understand the force of the utterance. As there is no recognisable, implicit offer to defend the truth of the propositional content of this speech act, the hearers would fail to perceive the statement as such and would rather think of it as something like raving or a joke.
Likewise, in a hateful statement to the effect that gay men are vessels of disease, there would also be a lack of a recognisable, implicit, serious offer to defend the truth claim underlying the statement. Such an utterance, however, would ‘click’ as a statement. The reason is that the audience is already conversant with the subtext of discursively produced, normative homophobic ‘truths’ regarding the supposedly necessary association between gay men and HIV/AIDS.
Even here, the (performative) success of the utterance at the illocutionary level is independent of the realisation of the speaker’s perlocutionary aim. Even if the speaker fails to convince the audience, at the same time as the latter rejects the validity of the implicitly invoked ‘truth’, its existence ends up being acknowledged and gay identity is consequently constructed as inferior.
So far, this analysis does not significantly differ from our discussion about racist hate speech utterances. However, homophobic hate speech acts are also, more clearly than hateful utterances against racialised minorities, institutionally bound speech acts.
The law, in its institutional dimension, produces legal norms through both statutory and case law. These norms are not, and need not be, produced as the normative truths/true norms that the law produces in its capacity as a social discourse. This is because it is generally accepted that these norms should be valid in their capacity as institutional (or juridical) norms (provided that, among other things, their production follows a certain procedure and emanates from certain social organisations).
In a range of situations in social life, the law — whether expressly or not — still excludes lesbians and gay men from the enjoyment of rights on a basis of equality with heterosexuals. Thus, homophobic hateful utterances can derive their illocutionary force and, as such, their power to subordinate in two ways. First, by seeking justification in true norms/normative truths that are produced by social discourses about homosexuality, of which the law, when not operating strictly as an institution, is one. Secondly, by pointing to official, institutional legal norms that discriminate against lesbians and gay men.
Of course, legal norms tend not to expressly and formally invest specific citizens with the power to disseminate homophobic hate. They do this, however, by indirection, to the extent that in many countries the subject of rights is heterosexual. For example, in almost all Australian jurisdictions, legal rules prevent lesbian and gay couples from adopting children. These rules provide the normative background against which the illocutionary success of hate speech acts identifying gay men with child abusers and lesbians with inadequate mothers can be secured.
At this point it is also apposite to qualify my previous claim that racist hate speech acts are institutionally unbound. The lack of even formal equality between heterosexuals and lesbians/gay men makes it easy to see the extent to which homophobic hate speech acts are institutionally bound. But racist assumptions pervade the law alongside homophobic ones, even if the former do not necessarily take the blatant form of a denial of formal equality. Inasmuch as racism finds expression in the law, racist hate speech acts are, not unlike homophobic hate speech acts, institutionally bound. This circumstance contributes to their illocutionary success, even if, in the context of homophobic hate speech acts, the role played by institutional norms in securing their illocutionary success is probably more crucial.
Finally, note that the normative truths produced by the law as a social discourse benefit from their contiguity with the legal rules (institutional norms) enacted by the law as an institution. Clearly, the power that the law has, as a social discourse, to create its own truths and disqualify other knowledge derives partly from the authority that the law enjoys as an institutionally produced discourse.
We have seen how a special kind of subordination is performed at the illocutionary level by acts of homophobic vilification and hate speech acts against racialised minorities. The illocutionary way of subordinating is a mechanism whereby, when a hateful speech act is uttered, certain subjects (first of all the interlocutors) end up inhabiting certain (hierarchical) social positions and/or reaffirming the existence of certain (deleterious) social norms/truths. This happens because (and at the same time as when) the speech act is understood as being used in the way in which it is used, and as endowed with the meaning that the speaker attaches to it.
Judith Butler observes that, according to Matsuda’s model, hate speech acts are one means through which domination and subordination are installed. She specifically calls Matsuda’s model illocutionary, and understands it to suggest that at the moment when the speech act occurs, certain subjects are constituted in a way that makes them inhabit a position of social subordination. Thus, in the performance of a hate speech act, lesbians, gay men or Aboriginal people, for example, are constituted (produced) as subjects of inferior worth.
The question of whether or not this kind of subordination is a kind of harm that may be taken to justify hate speech restrictions is perhaps open to debate. According to Matsuda’s illocutionary model, hate speech, as Butler puts it, constitutes the subject who is the target of hate. However, as far as the construction of the subject is concerned, that is not the end of the story. Counter-discourses exist in society that may undo the effects of hate speech.
The constitution of the subject, whether they be lesbian, gay or racialised, is never completed at a certain stage, as the subject is the very possibility of resignification. Any illocutionary effect produced by a hateful utterance should not be seen, in this sense, as being permanent, inescapable and unproblematically established. Thus, for instance, one may wonder whether the most appropriate response to homophobic hate speech would be to proactively bolster and promote egalitarian and empowering counter-discourses rather than to reactively coerce hate-mongers into silence.
Even so, the fact that in the context of hate speech against lesbians, gay men and racialised minorities it is at all possible to speak of (illocutionary) subordination means that it is not manifestly ill-founded to impose restrictions on freedom of expression in that context. This is because such restrictions can be made sense of in light of the objective of removing opportunities for the discursive production of lesbians, gay men and racialised individuals as subjects of inferior worth.
Let us now turn to vilification on such grounds as heterosexuality or Anglo-Australian identity. Does the model of illocutionary subordination apply in the context of heterosexual vilification? The answer is clear: an act of heterosexual vilification has no way of securing the illocutionary result of producing the heterosexual as a subject of inferior worth. This is because there is no normative/discursive subtext against which acts of heterosexual vilification can secure their illocutionary success.
The true norms/normative truths socially produced about heterosexuality are narratives about the superiority of heterosexuality. This is because, as Cheshire Calhoun puts it,
subordination depend[s] on the cultural articulation of basic social identities that are taken to be ... in polar opposition to an Other identity where polarity of evaluation (good–bad, respectable–unrespectable, superior–inferior, natural–unnatural) is central.
In other words, heterosexuality and homosexuality are identity categories, and ‘[i]dentity categories and nouns convey meaning according to a structure of binary oppositions, with one term of any pair valued more highly than the other.’ In the opposition ‘heterosexual/homosexual’, the former term is culturally associated with positive values and the latter with negative ones. Because of this sedimentation over time of culturally produced truths/norms about homosexuality and heterosexuality, hate speech acts against lesbians and gay men subordinate them at the illocutionary level in a way that is not experienced by the subjects of heterosexual vilification.
The same observations apply, for example, in the context of Anglo-Australian identity vis-à-vis Australian Aboriginality. Thus, a regulative hate speech act whereby an Aboriginal (or gay) man proclaims a white (or heterosexual) man deficient because of his whiteness (or heterosexuality) fails to bring about the enactment of the interpersonal relationship ‘thematised’ by the speech act. The speaker has no power to actually degrade the addressee because he lacks the cultural authority to do so. The Aboriginal (or gay) man, with his hate speech act, cannot implicitly invoke any cultural norm with a de facto social force which codifies the appropriateness of his assuming a position of superiority vis-à-vis a white (or heterosexual) man.
Thus, if what justifies imposing restrictions on hate speech is the power of hate speech to subordinate at the illocutionary level, the argument for outlawing vilification against heterosexuals or Anglo-Australians fails. This is because these kinds of vilification cannot secure the desired illocutionary uptake.
Ordinarily, speech acts also have perlocutionary effects — that is to say, they generally bring about consequences (changes) with respect to people’s beliefs, feelings and attitudes. Hate speech acts are no exception.
Thus, Matsuda describes specific effects of racist hate speech that cannot be understood by resorting to the illocutionary way of enacting subordination. Matsuda gives an account of how the individual victims of racist hate speech experience emotional distress, while even well-meaning non-victimised hearers end up, to some extent, internalising racist beliefs. Within the analytical framework of speech act theory, these effects, produced by racist utterances at the individual level (both in victimised and non-victimised hearers), are perlocutionary effects.
When one is considering whether or not (different forms of) hate speech should be restricted, three categories of perlocutionary effects possibly produced by hate speech acts may raise one’s concern. These are:
As far as harm (1) is concerned, offence and distress, although undeniably injurious, are not the kinds of harm that may ever justify, in and of themselves, speech restrictions. It is trite to say that if speech restrictions were justified every time somebody felt offended or distressed because of somebody else’s speech act, we would be left with very little freedom of expression. For example, Joel Bakan emphasises that the purpose of measures proscribing bigoted speech goes beyond that of protecting individuals from being injured by speech and is better understood as promoting the empowerment of silenced groups. We could say that emotional distress becomes relevant as a reason to single out hate speech for special regulation only to the extent that it is found to be linked with subordination.
With regard to harms (2) and (3), changes in feelings, beliefs, thoughts and attitudes are a normal consequence of any speech act. The fact that perlocutionary effects of some sort are normally produced by speech acts does not mean, however, that given perlocutionary consequences will normally, and predictably, ensue following a given type of locution, just by virtue of what is said. Of course, what is said will generally be relevant to identifying the specific perlocutionary effects that ensue in any given speech situation. But this is no truer with regard to what is said than in respect of any other contextual factor that characterises the speech situation. In the vocabulary of speech act theory, the relationship between perlocutionary (as opposed to illocutionary) effects and speech acts is not ‘conventional’.
This has a consequence for hate speech laws taking such perlocutionary effects as those in (2) and (3) as the basis on which to justify the regulation of hate speech. Given that effects (2) or (3), depending on the contextual speech situation, may or may not occur, hate speech laws will be more defensible if they outlaw speech acts not merely because of their message, but also having regard to other contextual factors. Apart from this, note that although what a speaker says does not determine the hearer’s intellectual, emotional and attitudinal response (the specific perlocutionary results), the locutionary content and illocutionary success of an utterance have a bearing on that response. In particular, it will normally be the case that a prerequisite for the production of the perlocutionary effects (2) and (3) — that is, for the success of the hate-monger’s perlocutionary aim — is the illocutionary success of the hate speech act.
However, whilst hate speech acts against lesbians, gay men and racialised minorities can implicitly appeal to the normative subtext which ensures their illocutionary success, I have suggested that acts of vilification against heterosexuals or Anglo-Australians will tend to be illocutionary failures. This is because acts of vilification against dominant groups are institutionally unbound, and hence cannot secure their illocutionary uptake from the social force of official norms. Because they are hate speech acts, they also do not contain any serious offer to implicitly defend the validity claims they make, and hence cannot rely on the same mechanisms which ensure the illocutionary success of acts directed to reaching understanding. Lastly, they are at odds with cultural narratives (true norms/normative truths) about heterosexuality or Anglo-Australian identity. Thus, a speaker vilifying heterosexuals or Anglo-Australians cannot implicitly invoke these narratives to secure an illocutionary uptake, as it is possible to do with respect to hate speech acts targeting lesbians, gay men and racialised minorities.
To summarise, vilification on such grounds as heterosexuality or Anglo-Australian identity is an illocutionary failure. As such, it is neither conducive to producing silence in the victimised hearers of the speech act nor to triggering a non-victimised hearer’s unfavourable treatment of the victims of the speech act. Thus, there is no case for outlawing vilification on such grounds as heterosexuality or Anglo-Australian identity with a view to preventing the production of the perlocutionary effects (2) and (3).
The point of departure for the analysis undertaken in the previous sections has been Matsuda’s model for the regulation of (racist) hate speech. Under this model, as Matsuda expressly states and as I have demonstrated using speech act theory, there is no room for outlawing hateful utterances targeting their victims on the ground of their membership of a ‘dominant’ group. Those wishing to justify outlawing hateful utterances targeting heterosexuals or Anglo-Australians must do so under an alternative model for the regulation of hate speech.
Such a model could not possibly be the one, which I have already criticised, that takes offence to sensitivities as a good reason for the regulation of hate speech. Here I shall briefly discuss another model on which legal systems that do not differentiate between hate speech targeting minorities and ‘dominant’ groups (either through the express wording of statutes or through judicial interpretation) are often predicated. I shall call this the ‘social harmony model’.
According to the social harmony model, when freedom of expression is at issue, the proper role of government is generally that of maintaining social harmony in society. This would be done by making it antecedently clear, in situations of group conflict, where one set of claims legitimately begins and the other disappears. The problem with the social harmony model is that (not unlike the view that offence to sensitivities alone is sufficient to justify speech restrictions) it legitimises governmental pre-emption of the political debate in a way that contradicts the argument from democracy. The state, in this view, is entitled to decide not only what political claim can be legitimately satisfied but also to control at any moment what can or cannot be claimed at all, because such a decision is thought to promote good relations among different groups. This is clearly at odds with the preservation of a meaningful system of freedom of political expression.
In fact, in order to maintain that the promotion of good relations among societal groups is sufficient to justify speech restrictions, the advocates of the social harmony model end up having to downplay the contribution of freedom of expression to democratic government. I know of no account which has convincingly managed to do so.
Matsuda’s subordination-centred model shifts the emphasis away from public order/social harmony considerations towards an overriding concern for substantive equality which is more compatible with freedom of expression. This does not mean that Matsuda considers inter-group violence of secondary importance. In fact, Matsuda’s analysis expressly ties racist hate speech to perpetuation of violence and degradation of minorities. But she does not take the role of hate speech regulation to be that of preventing inter-group violence irrespective of who commits the violence and who is victimised by it. This is clearly not because she condones minorities’ violent attacks against dominant groups. Rather, it is because it is implicit in Matsuda’s analysis that violence against dominant groups is not related to hate speech of minorities in such a way as violence against minorities is connected to hate speech of dominant groups.
This makes a great deal of sense in light of the analysis undertaken in the previous sections. We have seen that the enactment of violent behaviour, as a perlocutionary effect of hate speech depending on the illocutionary success of the speech act, is an extremely unlikely consequence of vilification against dominant groups.
Of course, this does not mean that speech inciting violence against, or ill-treatment of, heterosexuals or Anglo-Australians could escape regulation. But it would undergo regulation not in its capacity as vilifying speech, but rather inasmuch as it amounts to criminal incitement. And utterances need be neither hateful nor made on the ground of heterosexuality or white Australian identity in order for them to be proscribable under the offence of incitement, which is, and should remain, one of general application.
I have applied the model of speech act theory to the case of racist and homophobic vilification, incorporating some Foucauldian insight into my analysis. I have concluded that it may be argued that the illocutionary and perlocutionary results of hate speech targeting lesbians, gay men and racialised minorities justify singling out this kind of speech for special treatment. However, the same cannot be argued in the context of vilification on other grounds such as heterosexuality and Anglo-Australian identity. This is because this kind of vilification tends to be an illocutionary failure. As such, it does not subordinate heterosexuals or white Anglo-Australians, and its ability to produce the perlocutionary effects of silencing them and triggering their ill-treatment by other groups is seriously compromised.
In particular, laws outlawing heterosexual or Anglo-Australian vilification fail the test that statutory measures limiting freedom of political communication must pass in order to be constitutionally valid — even the wide version of that test set out in Lange. Vilification against dominant groups does not produce any (illocutionary) subordination. Preventing such subordination is not, therefore, a useful aim, let alone a legitimate one reasonably furthered by vilification legislation. Additionally, the illocutionary failure of vilification against ‘dominant’ groups stands in the way of the production of (perlocutionary) silencing and ill-treatment of such groups (including violence). Thus, there is no rational connection between targeting vilification against these groups and the aim of preventing these perlocutionary effects.
With regard to governmental promotion of social harmony and prevention of offence and distress, my arguments have been to the effect that fulfilling these ends through statutory restrictions on freedom of speech is not ‘compatible with the maintenance of the constitutionally prescribed system of representative and responsible government’. Laws proscribing vilification on such general grounds as sexuality or race can be deemed constitutionally valid only if they are understood to be geared towards avoiding the harm of sexual and racial minorities’ subordination. If anti-vilification provisions worded in neutral terms are interpreted as affording protection also on such grounds as heterosexuality and Anglo-Australian identity, it is hard to see how they can be reconciled with constitutional values of free speech.
With regard to anti-vilification laws which are not neutrally worded, their constitutionality also depends on their being open to an interpretation that understands them as directed to avoiding subordination (and not exclusively some lesser harm such as offence to sensitivities). For example, the New South Wales sexuality vilification regime may initially appear to be primarily concerned with preventing private distress/offence, but I think that it is best understood as doing so incidentally, for the overriding purpose of dealing with the harm of lesbian and gay subordination. The emotional distress that lesbians and gay men may suffer as a result of homophobic hate speech occurs as a result of utterances that also enact lesbian and gay subordination. In this context, the offence caused by homophobic hate speech has a different significance from that suffered by the victims of heterosexual vilification. It is open to the law to choose to serve the aim of avoiding the performative re-enactment of lesbian and gay subordination through remedying the harm suffered by the individual victim.
In fact, the New South Wales regime of homosexual vilification has features (in addition to its taking homosexuality as the ground on which to outlaw vilification) which can be made sense of in light of a central concern for the harm of (lesbian and gay) subordination. In particular, the subordination of lesbians and gay men produced at the illocutionary level is enough to raise the law’s concern, so that its remedies are triggered even without the production of deleterious perlocutionary effects compounding that subordination. Thus, ‘proof ... that anyone was in fact incited [to hatred] is not required.’ This seems a recognition that subordination, in its illocutionary dimension, may occur independently of the actual realisation of the speaker’s perlocutionary aim — that is, affecting the audience’s feelings and beliefs in such a way as to stir up hatred.
Instead of focusing on the actual perlocutionary effects, the law concentrates on the likelihood that the hate-monger’s presumed perlocutionary aim will carry itself out — that is, on whether ‘all or part of [the speaker’s] conduct was capable of urging on, stimulating, or prompting to action the ordinary reasonable person to the requisite feelings of ill will’. Note that the illocutionary success of the hate speech act is normally a prerequisite for the success of the hate-monger’s perlocutionary aim (even if the former alone will not bring about the latter). Thus, the likely success of the hate-monger’s presumed perlocutionary aim is a rough measure of the illocutionary success — that is, the subordinating power — of the utterance.
It is also centrally important that ‘the audience or potential audience of the public act should be assumed to be the “ordinary reasonable person”’ — that is, somebody sharing the true norms/normative truths socially produced on homosexuality. It is only if such norms are assumed to exist that a homophobic hate speech act can be understood as an illocutionary success. Besides, concentrating on this kind of audience renders the likely success of the speaker’s perlocutionary aim — on which the law focuses — a relatively accurate measure of the illocutionary success of the utterance.
Additionally, as was said in Burns:
It does not follow automatically that verbal abuse ... that includes words understood to be insulting of homosexuals, is capable of inciting the requisite ill-feeling required to establish a complaint of homosexual vilification. The circumstances in which the abuse occurred are critical.
This principle, too, makes it plausible to understand the law as preoccupied with the harm of lesbian and gay subordination. Recall that there are several conditions for the felicity of performatives: for example, the person pronouncing the formula must be the appropriate one and do it in the appropriate circumstances. Thus if a hate-monger called a male football player a ‘filthy dyke’, it would be unlikely that anybody would consequently be incited to hatred against the player. That is to say, in these circumstances, the speaker’s presumed perlocutionary aim would go amiss, and it would do so because the speech act would be an illocutionary failure. It would fail to attach to the football player an identity of inferior worth because only women are socially thought to be capable of inhabiting the lesbian role, and the speaker’s odd gender notions would undermine his authority to degrade the player on the ground of homosexuality. Anti-vilification laws insensitive to the circumstances which could cause even hate speech acts against racialised minorities and lesbians/gay men to be illocutionary failures would be difficult to defend in light of principles of free speech.
Homophobia and racism against minorities remain serious social ills. If anti-vilification legislation is appropriately tailored to the goal of tackling these ills, contending that such legislation is constitutional is not outright unreasonable. Bigoted or hostile speech against such groups as heterosexuals or Anglo-Australians may cause offence. But this offence is a much lesser evil than the subordination, violence and silencing that individuals belonging to stigmatised minorities may experience. The case for regarding such offence as a sufficient reason for limiting freedom of expression has yet to be made.
[*] JD cum laude (Milan), LLM (British Columbia), MSt (Oxon); PhD Candidate, Faculty of Law, The University of Sydney. Thank you to the two anonymous referees for their insightful comments, to Bill Black for reading and commenting on an early version of this piece, and to Roy Baker for his encouragement and for forcing me to render some complex ideas more accessible to the reader.
 See, eg, Anti-Discrimination Act 1977 (NSW) s 20D.
 See, eg, Racial Discrimination Act 1975 (Cth) pt IIA.
 See, eg, Anti-Discrimination Act 1977 (NSW) s 20C; Discrimination Act 1991 (ACT) s 66; Racial Vilification Act 1996 (SA) s 4; Racial and Religious Tolerance Act 2001 (Vic) s 7.
 Geoffrey Robertson and Andrew Nicol, Media Law (3rd ed, 1992) 169.
 See, eg, Anne Flahvin, ‘Can Legislation Prohibiting Hate Speech Be Justified in Light of Free Speech Principles?’  UNSWLawJl 17; (1995) 18 University of New South Wales Law Journal 327, 339.
 For example, the Canadian Human Rights Act, RSC 1985, c H-6, s 13(1) provides:
It is a discriminatory practice for a person or a group of persons acting in concert to communicate telephonically or to cause to be so communicated, repeatedly, in whole or in part by means of the facilities of a telecommunication undertaking within the legislative authority of Parliament, any matter that is likely to expose a person or persons to hatred or contempt by reason of the fact that that person or those persons are identifiable on the basis of a prohibited ground of discrimination.
Such grounds have included sexual orientation since 1996: Canadian Human Rights Act, RSC 1985, c H-6, s 3, as amended by Bill C-33, An Act to Amend the Canadian Human Rights Act, SC 1996, c 14, s 2. In the Netherlands, it is a criminal offence to publicly express discriminatory remarks, encourage discrimination, or incite hatred, discrimination or violence against a group of persons on the basis of sexual orientation: see Wetboek van Strafrecht (Netherlands) art 137(c)–(f). See also International Lesbian and Gay Association, World Legal Survey — Country: Netherlands (1999) <http://www.ilga.org/Information/legal_survey/Europe/
Netherlands.htm>. As the latest instance of hate speech legislation, sexuality anti-vilification laws have recently attracted scholarly attention: see, eg, Jonathan Cohen, ‘More Censorship or Less Discrimination? Sexual Orientation Hate Propaganda in Multiple Perspectives’ (2000) 46 McGill Law Journal 69, 76–91; David Marr, ‘How Can We Square Freedom with Anti-Vilification Laws?’ (2000) 9 Australasian Gay and Lesbian Law Journal 9.
 See Commonwealth, Parliamentary Debates, Senate, 27 March 2003, 10 328 (Brian Greig).
 The Anti-Discrimination Act 1998 (Tas) s 19 makes it unlawful for a person, by a public act, to ‘incite hatred towards, serious contempt for, or severe ridicule of, a person or a group of persons on the ground of’, inter alia, sexual orientation.
 Sexuality Anti-Vilification Bill 2003 (Cth).
 In New South Wales, it is unlawful to publicly ‘incite hatred towards, serious contempt for, or severe ridicule of’ individuals or groups because of their homosexuality, and it is a criminal offence to do so by means which include threats (or incitement of another to make threats) of physical harm towards the victims of such incitement or their property: see Anti-Discrimination Act 1977 (NSW) ss 49ZT, 49ZTA.
 See the second reading debate for the Anti-Discrimination (Heterosexual Discrimination) Amendment Bill 2002 (NSW): New South Wales, Parliamentary Debates, Legislative Council, 14 November 2002, 6742.
 New South Wales, Parliamentary Debates, Legislative Council, 14 November 2002, 6744 (Ian Macdonald, Parliamentary Secretary).
 Ministerial Committee, Parliament of Western Australia, Lesbian and Gay Law Reform (2001) 38–45. The recommendation has not yet been enshrined in legislation.
 See below n 94.
 See, eg, Theophanous v Herald & Weekly Times Ltd  HCA 46; (1994) 182 CLR 104, 121 (Mason CJ, Toohey and Gaudron JJ).
 The three most popular philosophical justifications for freedom of expression are that free speech is necessary to democratic self-government, to the discovery of truth and to individual self-fulfilment: see, eg, Frederick Schauer, Free Speech: A Philosophical Enquiry (1982) 15–59.
 The Supreme Court of Canada adopts precisely this kind of approach. Generally, when the Court is willing to accept the regulation of some speech as constitutionally permissible, it avoids claiming that the speech at issue is not protected expression under the constitutional free speech guarantee. Rather, the Court starts by recognising that virtually everything is constitutionally protected expression, and goes on to justify speech regulation (where it deems it appropriate) under the following constitutional provision: ‘[t]he Canadian Charter of Rights and Freedoms guarantees the rights and freedoms set out in it subject only to such reasonable limits prescribed by law as can be demonstrably justified in a free and democratic society’: Canadian Charter of Rights and Freedoms s 1, pt 1 of the Constitution Act 1982, being sch B to the Canada Act 1982 (UK) c 11 (‘Canadian Charter of Rights and Freedoms’). Violent expressive conduct is the exception to the principle that virtually all expressive activities are covered by the constitutional free speech guarantee: see A-G (Quebec) v Irwin Toy Ltd  1 SCR 927, 970 (Dickson CJ, Lamer and Wilson JJ).
 Cf Robert Post, ‘The Constitutional Concept of Public Discourse: Outrageous Opinion, Democratic Deliberation, and Hustler Magazine v Falwell’ (1990) 103 Harvard Law Review 601, 671, quoting Hanna Pitkin, ‘Justice: On Relating Private and Public’ (1981) 9 Political Theory 327, 343, 346; Roy Baker, ‘Extending Common Law Qualified Privilege to the Media: A Comparison of the Australian and English Approaches’ (2002) 7 Media and Arts Law Review 87, 93–4.
 Michael Chesterman, Freedom of Speech in Australian Law: A Delicate Plant (2000) 44.
 Ibid 62–3.
 Ibid 240.
 According to the test adopted by the High Court, a statutory measure restricting freedom of political communication will only be constitutional if it is ‘reasonably appropriate and adapted to serve a legitimate end the fulfilment of which is compatible with the maintenance of the constitutionally prescribed system of representative and responsible government’: Lange v Australian Broadcasting Corporation  HCA 25; (1997) 189 CLR 520, 567 (Brennan CJ, Dawson, Toohey, Gaudron, McHugh, Gummow and Kirby JJ) (‘Lange’). Given the importance of freedom of expression, this is a relatively loose standard and ‘the ground has ... been laid for significant modifications to it’: Chesterman, above n 19, 72. At any rate, laws outlawing vilification on such grounds as heterosexuality and Anglo-Australian identity, as I shall argue, would not pass even such a low-level scrutiny.
 Levy v Victoria  HCA 31; (1997) 189 CLR 579, 623 (McHugh J).
 See, eg, Franklyn Haiman, ‘Nazis in Skokie: Anatomy of the Heckler’s Veto’ in Thomas Tedford, John Makay and David Jamison (eds), Perspectives on Freedom of Speech — Selected Essays from the Journals of the Speech Communication Association (1987) 216; Laurence Maher, ‘Free Speech and Its Postmodern Adversaries’ (2001) 8(2) E Law — Murdoch University Electronic Journal of Law <http://www.murdoch.edu.au/elaw/issues/v8n2/maher82.html> .
 These advocates of hate speech laws criticise such categories of free speech doctrine known as the argument from truth, the argument from democracy, the clear and present danger test (see, eg, Kathleen Mahoney, ‘Hate Speech: Affirmation or Contradiction of Freedom of Expression’ (1996) 3 University of Illinois Law Review 789) and the notion that freedom of expression is a useful tool for minorities (see, eg, Richard Delgado and Jean Stefancic, Must We Defend Nazis? Hate Speech, Pornography, and the New First Amendment (1997) 102).
 The Canadian constitutional free speech guarantee ensures citizens’ ‘freedom of thought, belief, opinion and expression, including freedom of the press and other media of communication’: Canadian Charter of Rights and Freedoms s 2(b). The American free speech clause reads ‘Congress shall make no law ... abridging the freedom of speech, or of the press’: United States Constitution amend I.
 In RAV v City of St Paul,  USSC 99; 505 US 377, 391–3 (1992) (Scalia J) the Supreme Court of the United States declared unconstitutional a section of a local piece of legislation criminalising hate speech, finding viewpoint-based regulation of speech impermissible because of the risks such regulation creates of establishing a situation of government thought-control. In Canada, the constitutionality of the Criminal Code, RSC 1985, c C-46, s 319(2), which criminalises the wilful promotion of hatred against identifiable groups, was upheld by the Supreme Court of Canada in R v Keegstra  INSC 224;  3 SCR 697. The Court recognised that hate speech can cause harm by culturally legitimising the discrimination of minorities.
 Mari Matsuda, ‘Public Response to Racist Speech: Considering the Victim’s Story’ (1989) 87 Michigan Law Review 2320.
 Ibid 2326–41.
 Ibid 2361. Thus, a racist hate speech proscribable under Matsuda’s approach will employ ‘language that is, and is intended as, persecutorial, hateful, and degrading’ (at 2358) and will convey a message of racial inferiority directed against a historically oppressed group (at
 See John Austin, How to Do Things with Words (1962). Judith Butler, Excitable Speech: A Politics of the Performative (1997) applies notions of speech act theory to the case of hate speech.
 See Judith Butler, ‘Burning Acts — Injurious Speech’ in Andrew Parker and Eve Kosofsky Sedgwick (eds), Performativity and Performance (1995) 197.
 Austin, above n 32, 94.
 Ibid 98–9.
 Ibid 116.
 In order for an illocutionary act to be carried out, an audience must understand the meaning and the force of the locution: ibid 115–16.
 Ibid 145.
 See Jürgen Habermas, ‘What is Universal Pragmatics?’ in Thomas McCarthy (ed), Communication and the Evolution of Society (Thomas McCarthy trans, 1979 ed) 1, 53.
 Cf ibid 40–4.
 Austin, above n 32, 5, 116.
 Ibid 12–52, esp 14–15.
 Ibid 101.
 See also Butler, ‘Burning Acts’, above n 33, 197–8.
 Austin, above n 32, 116.
 See Habermas, Communication and the Evolution of Society, above n 39, 38.
 See Jürgen Habermas, The Theory of Communicative Action — Volume 1: Reason and the Rationalization of Society (Thomas McCarthy trans, 1984 ed) 296–7 [trans of: Theorie des Kommunikativen Handelns — Band I: Handlungsrationalität und Gesellschaftliche Rationalisierung].
 Austin, above n 32, 14.
 There are also other categories of institutionally unbound speech acts. Constative utterances, for example, are discussed later: see below n 63 and accompanying text.
 The illocutionary result of an utterance can be generally characterised as the enactment of that interpersonal relationship described in the illocutionary component of the utterance. Thus, in the context of institutionally bound speech acts, the illocutionary result that a ship’s name is the Queen Elizabeth can also be described, emphasising the interpersonal relation between the interlocutors, as the fact of the hearers being bound to calling the ship the Queen Elizabeth owing to their being subject to the legitimate authority of the speaker who christened the ship.
 See Habermas, The Theory of Communicative Action, above n 47, 307.
 Ibid 295–309, esp 308–9.
 Or so I want to assume at this stage, for the sake of simplicity. The statement that racist hate speech acts are institutionally unbound will be qualified later in this article: see below n 71 and accompanying text.
 A speech act oriented towards reaching understanding is an utterance acceptable on the basis of a rationally testable validity claim raised implicitly by the speaker.
 See Habermas, The Theory of Communicative Action, above n 47, 295.
 Racist speech acts making a serious offer to justify the validity claims raised in the utterance would not be criminalised by Matsuda (as confirmed by her exemption for the wrong social scientist): see Matsuda, above n 29, 2364–5. When the speech act is directed to reaching understanding it is possible to make the validity claims raised by the utterance become the object of serious theoretical argumentation: see Habermas, Communication and the Evolution of Society, above n 39, 63–4.
 Nevertheless, Katharine Gelber, Speaking Back: The Free Speech versus Hate Speech Debate (2002) 70–80 usefully applies Habermas’s doctrine of the validity claims to the case of hate speech acts.
 See Michel Foucault, ‘The Subject and Power’ in Hubert Dreyfus and Paul Rabinow (eds), Michel Foucault: Beyond Structuralism and Hermeneutics (1982) 208.
 Ibid 219–26.
 Ibid 222.
 See, eg, Delgado and Stefancic, above n 26, 72–82.
 See, eg, Matsuda, above n 29, 2331–5. The fact that this discourse can be found in connection with certain institutions does make the hate speech acts pointing to the ‘true’ norms produced by such a discourse become institutionally bound in a technical sense. This is because nowadays the norms on racialised minorities giving those speech acts their illocutionary force tend not to be the correct outcome of the regulated process through which a legitimate institution produces its official norms. For example, the narratives on racialised minorities produced by a judicial body will tend to be found in obiter dicta, cursory remarks or specific use of language and at most find an inexplicit, ambiguous expression in the official (as opposed to the discursive, ‘true’) norms produced by the judicial institution. (For an analysis of how the decision in RAV v City of St Paul,  USSC 99; 505 US 377 (1992) adds in this way to the discourse about American blackness, see Butler, ‘Burning Acts’, above n 33, 212–14. For the way in which Indian-ness is discursively constructed by judicial institutions, see Marlee Kline, ‘The Colour of Law: Ideological Representations of First Nations in Legal Discourse’ (1994) 3 Social & Legal Studies 451). Having said this, historically there have been clear cases of institutional norms denying racial minorities equal respect: see, eg, Ian McKenna, ‘Canada’s Hate Propaganda Laws — A Critique’ (1994) 26 Ottawa Law Review 159, esp 163–7. See also below n 71 and accompanying text.
 See Habermas, Communication and the Evolution of Society, above n 39, 55.
 Thus a racial slur uttered by a member of the same minority group as the victim’s cannot have the same illocutionary effect as one in which the speaker and the addressee belong, respectively, to a dominant and a subordinated group. An intra-group racist utterance would not allow the illocutionary result of subordination to take place because, in Austin’s terms, a ‘misapplication’, that is a particular type of infelicity (Austin, above n 32, 17), would have occurred.
 Robert Post characterises the harm identified by Matsuda as harm to identifiable groups that contributes to maintaining the oppression of disadvantaged minorities (rather than harm to individuals or harm to the marketplace of ideas): see Robert Post, ‘Racist Speech, Democracy, and the First Amendment’ in Henry Louis Gates Jr et al (eds), Speaking of Race, Speaking of Sex (1994) 115, 119. However, the example provided in the text illustrates that the dichotomy between individual harm and harm to groups is not clear-cut.
 Matsuda, above n 29, 2332.
 See, eg, QFG v J M  EOC 92-902. In that case, note the rhetorical effect of such recurrent expressions used by Ambrose J as ‘lifestyle’ (at 77 420, 77 425–6, 77 428, 77 430–1, 77 438–9) — which in the judgment is invariably lesbian and never heterosexual — contrasted with ‘normal’ (at 77 419–21, 77 424–30, 77 436–7, 77 439, 77 442) — which consistently appears as an attribute of heterosexuality.
 For example, several Australian states do not allow lesbian and gay couples to adopt: see, eg, Adoption Act 2000 (NSW) s 26; and a valid marriage for the purposes of Australian law must be between a man and a woman: see Re Kevin (2001) 28 Fam LR 158, 161 (Chisholm J).
 See Adoption Act 1993 (ACT) s 18(1); Adoption Act 2000 (NSW) s 26; Adoption of Children Act 1994 (NT) s 13(1)(a); Adoption of Children Act 1964 (Qld) s 12(1); Adoption Act 1988 (SA) s 12(1); Adoption Act 1988 (Tas) s 20(1); Adoption Act 1984 (Vic) s 11(1). Adoption by same sex de facto couples is permitted in Western Australia: Adoption Act 1994 (WA) ss 39, 67 and Interpretation Act 1984 (WA) ss 5, 13A.
 For example, the Native Title Amendment Act 1998 (Cth), in unduly restricting the recognition of Aboriginal people’s interests in land, failed to treat Aboriginal Australians with equal respect.
 See Butler, Excitable Speech, above n 32, 18.
 Discourses make the subject possible. The way in which discourses interact in constituting subjects gives rise to a combination that allows for other discourses to be elaborated and to re-constitute the subject. Because no discourse is complete in itself and seamless, the subject can turn the discourse against itself and in the process undergo resignification: see generally Judith Butler, ‘Contingent Foundations: Feminism and the Question of “Postmodernism”’ in Judith Butler and Joan Scott (eds), Feminists Theorize the Political (1992) 3.
 Gelber, above n 57, 88–91 identifies other reasons why punitive or restrictive responses to hate speech are not the least problematic.
 Cheshire Calhoun, Feminism, the Family, and the Politics of the Closet: Lesbian and Gay Displacement (2000) 5.
 William Turner, A Genealogy of Queer Theory (2000) 32–3.
 For analogous reasons, the illocutionary force of a constative ‘heterophobic’ hate speech act is more similar to that of the utterance that lesbians and gay men are surrounded by rainbow-coloured halos than to that of the homophobic statement that gay men are vessels of disease.
 Gelber, above n 57, 80–3 reaches a similar conclusion, but does so by applying Habermas’s doctrine of the validity claims to the case of anti-Anglo-Australian ‘vilification’.
 See Matsuda, above n 29, 2335–41.
 The harm to the marketplace of ideas, if taken on its own, may be a problematic basis on which to justify hate speech laws. For example, in the marketplace of ideas, queer expression has been ostracised, marginalised, ghettoised and essentialised. But these attempts have been less than successful at least in countries where lesbians and gay men, although not enjoying equal rights, are not officially and positively per-/prosecuted because of their sexual orientation. The notion of intersectionality suggests that gay men and lesbians are neither a homogeneous group nor subjected to the same sort of discrimination and the same degree of silencing practices. Race and gender meet sexual preference and fasten around bodies, intersecting with class relations and a myriad of other factors, to create peculiar experiences of discrimination and empowerment. The same discourses that silence and constrain us in certain contexts incite us to speak up in others. In this sense, the silenced minority of lesbians and gay men has been less than silent. However, this does not mean that lesbians and gay men have been able to participate on equal terms in the process of society’s self-definition; they have been silenced, if not as a group, at least as individual histories.
 See also Gelber, above n 57, 81–7.
 In the context of New South Wales homosexual vilification legislation it has been noted that in order for speech to unlawfully vilify under the Anti-Discrimination Act 1977 (NSW) it is not ‘sufficient to prove that the victim was deeply wounded or concerned for their privacy, or indeed safety’: Burns v Dye  NSWADT 32 (Unreported, Judicial Member Britton, Members Silva and Toltz, 12 March 2002)  (Judicial Member Britton and Member Toltz) (‘Burns’).
 Joel Bakan, Just Words: Constitutional Rights and Social Wrongs (1997) 74.
 See Habermas, The Theory of Communicative Action, above n 47, 290.
 Ibid 292.
 Cf Gelber, above n 57, 87.
 See Mahoney, above n 26, 797.
 See, eg, ibid 796.
 Consider, for example, Kathleen Mahoney’s contention that whilst the argument from democracy was appropriate in the 18th century, it no longer fits the circumstances of 20th (and 21st) century Western democracies: ibid. A response to her contention is that it was precisely in the 20th century that free speech jurisprudence justifying protection for political speech as a vital ingredient of democratic self-government originated in the United States. During the same period of time there developed a greater appreciation of the function of the free speech guarantee as largely the democratic one of protecting political dissent. These phenomena started exactly as a response to 20th century governmental suppression of political (in particular communist) speech. See also Henry Louis Gates Jr, ‘War of Words: Critical Race Theory and the First Amendment’ in Henry Louis Gates Jr et al (eds), Speaking of Race, Speaking of Sex (1994) 17, 21.
 See Matsuda, above n 29, 2357.
 For example, the Crimes Act 1958 (Vic) s 321G, in connection with s 2A, makes it an indictable offence to command, request, encourage, propose, advise, authorise or otherwise incite ‘any other person to pursue a course of conduct which will involve the commission of an offence’. The law rightly provides that it is not necessary for the incited offence to occur in order for the inciter’s speech to be punishable: s 321G(3). However, it goes too far in providing that even if it was impossible for the incited offence to be committed, the speech is punishable so long as the inciter thought the offence could take place. As the Supreme Court of the United States noted in Brandenburg v Ohio USSC 141; , 395 US 444, 447 (1969) (Warren CJ, Black, Douglas, Harlan, Brennan, Stewart, White and Marshall JJ):
the constitutional guarantees of free speech and free press do not permit a State to forbid or proscribe advocacy 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.
However, the shortcomings of incitement laws is an issue different from the one at hand. The fact that the offence of incitement is not ideally defined in Victoria clearly does not detract from the point that anti-vilification legislation directed to preventing violence against ‘dominant’ groups would be an unjustifiable free speech infringement.
  HCA 25; (1997) 189 CLR 520, 567. See above n 23.
 Lange  HCA 25; (1997) 189 CLR 520, 567.
 From this perspective, the failure of Kazak v John Fairfax Publications Ltd  NSWADT 77 (Unreported, Hennessy DP, Members Farmer and Jowett, 22 June 2000)  (Hennessy DP, Members Farmer and Jowett) to distinguish between vilification of ‘dominant’ groups and of minorities while examining the constitutionality of racial anti-vilification legislation is analytically unfortunate. However, interestingly, the New South Wales Anti-Discrimination Board declined all racial vilification complaints (not formally withdrawn or otherwise failing to be pursued) lodged with it by Anglo/British people between 1993 and 1995: see Luke McNamara, ‘Research Report: A Profile of Racial Vilification Complaints Lodged with the New South Wales Anti-Discrimination Board’ (1997) 2 International Journal of Discrimination and the Law 349, 350, 369. This may indicate that, at least in its operation, the law does not protect ‘dominant’ groups.
 Cf Chesterman, above n 19, 244 (making this observation with respect to Australian racial vilification laws).
 Even so, Gelber, above n 57, 88–90 points out that the New South Wales anti-vilification legal regime has shortcomings of its own.
 Burns  NSWADT 32 (Unreported, Judicial Member Britton, Members Silva and Toltz, 12 March 2002)  (Judicial Member Britton and Member Toltz).
 Ibid .
 Ibid .
 Certain circumstances could make a speaker’s perlocutionary aim fail even without compromising the illocutionary success of the utterance. Thus a gay-friendly audience would be highly unlikely to be incited to hatred by a hate speech act successful at the illocutionary level. However, if the hate speech act is an illocutionary success, it is more difficult to think of circumstances making a hate-monger’s perlocutionary aim fail on the hypothetical audience that the law specifically focuses on (ie, the ordinary reasonable person ‘not immune from susceptibility to incitement or prejudice’: ibid ). Thus, the law’s attention to the likely success of the hate-monger’s presumed perlocutionary aim (incitement to hatred) can be viewed as a concern for the subordination of lesbians and gay men occurring at the illocutionary level.
 Ibid .
 Is the slur — albeit an infelicity as a regulative as well as constative (see above n 66 and accompanying text) utterance regarding the football player — still an illocutionary success as a constative speech act about lesbians? I am inclined to think that the felicity of the slur as a constative utterance on lesbians would be disturbed by the illocutionary failure of the slur as a regulative/constative utterance on the football player (the slur is not a statement about lesbian identity independent of its quality as an utterance on the football player).