Millian Liberalism and Extreme Pornography

: How sexuality should be regulated in a liberal political community is an important, controversial theoretical and empirical question—as shown by the recent criminalization of possession of some adult pornography in the United Kingdom. Supporters of criminalization argue that Mill, often considered a staunch opponent of censorship, would support prohibition due to his feminist commitments. I argue that this account underestimates the strengths of the Millian account of private conduct and free expression, and the consistency of Millian anticensorship with feminist values. A Millian contextual defense of liberty, however, suggests several other policy approaches to addressing the harms of pornography.

that had specialized in investigating corruption within British police forces, and public life, as an aide to the mayor of London (Attwood and Walters 2013, 975).
The Crown Prosecution Service maintains that the acts depicted were "extreme" even if the jury disagreed in this case, suggesting they see this kind of prosecution as in line with government policy. This illustrates the difficulty that prosecutors have had interpreting the law. There are particular legal vulnerabilities for sexual minorities, including individuals with a lesbian, gay, bisexual or transgender orientation, and practitioners of bondage, domination and sadomasochism. More than a thousand prosecutions happen annually (Crown Prosecution Service 2015, 92). In contrast to these observed harms arising from prosecution, there is comparatively little evidence that allowing access to extreme pornography causes or encourages violence against women, or is associated with specific social harms (Diamond 2009;Ferguson and Hartley 2009;cf. Itzin, Taket, and Kelly 2007). Nevertheless, feminist supporters of the law, though critical of its implementation in particular cases, including that of Walsh, argue for the principle of censorship using Millian liberal premises.

The Millian Case for Censorship
The Williams report made Millian principles of free expression (1979, 53) and the harm principle (1979, 57) a core part of the liberal defense of sexually explicit material. The report criticized the application of the legal notion of "obscenity" as unworkable and subjective (1979,11), and drew attention to the lack of evidence of specific harms, especially sexual offenses, associated with the availability of pornography (1979,79). The report did not advocate a laissez-faire policy, instead endorsing regulations on displays of pornography to prevent public offense and to protect children from inappropriate material (1979,130). However, the report argued that criminal prohibition should be restricted to material created by inflicting actual physical harm or by exploiting underage actors (1979,161).
The pro-censorship case argues that this liberal approach endorses metaphysically unjustified "absolute" (McGlynn and Ward 2014, 520), "abstract" (McGlynn and Ward 2009, 339), "a priori" (Dyzenhaus 1992b, 546) rights to privacy and free expression and that the liberal approach relies on an impoverished conception of "direct" harm that denies the real harms and injustice of pornographic expression beyond that directly associated with its production (McGlynn and Ward 2014, 503).
Once these abstractions are dismissed, those in favor of censorship suggest that liberals can approach questions of free expression and individual liberty on a more contingent, case-by-case basis (McGlynn and Ward 2009, 341). Liberals should acknowledge the existence of "cultural harm," namely, harms that contribute to a "social environment in which sexual violence is marginalised, in which rape conviction rates are at an all time low and in which pornography is becoming (if possible) even more ubiquitous" (McGlynn and Rackley 2010, 9). A Millian, on this account, should endorse this approach because Mill himself does not systematically define harm, nor a way of dealing with conflicts between liberties (McGlynn and Ward 2014, 506). Moreover, reading the "Applications" section of On Liberty, and The Subjection of Women provides plenty of exceptions to any notional "harm principle" (Dyzenhaus 1992b, 534;McGlynn and Ward 2014, 506). The pro-censorship case adds that pornography, especially extreme pornography, cannot conceivably be valuable expression, and so liberals should commend its prohibition.
If successful, this case justifies censorship on a broad basis, impacting not just on expression made in public but also ideas shared in private (indeed intimate) settings. This could render prosecutions such as Walsh's an unfortunate but, potentially, necessary part of the process of a political community deliberating and discovering the shared boundaries of acceptable behavior (cf. Johnson 2010). If the case fails, then the ban loses an important normative justification and may appear more like a sheer exercise of state power with the reactionary function of punishing alleged sexual deviancy (Carline 2011) and reinforcing some sexual minority practices as taboo (Tebble 2011).
My case for the latter is as follows. Millian 2 liberalism sees rights as political, not metaphysical. Critiquing the ontological status of rights does not impact straightforwardly on the content of the rights that a Millian defends. The harm principle affirms a tractable set of rights that includes possession of extreme pornography. Rather than rendering the harm principle indeterminate, the "Applications" section of On Liberty helps to establish its boundaries by explaining what counts as private conduct to be protected from state intrusion. Moreover, Mill's argument in The Subjection of Women does not support censorship.
A Millian anticensorship position stands not on affirming rights in the abstract, but on critical observations of what happens when governments censor. In the case of pornography, regulation addressing "cultural harm" leads authorities to punish arbitrarily members of sexual minority groups for the crimes and social problems of the rest of the community. Moreover, far from being valueless, queer feminist accounts of pornography, even extreme pornography, acknowledge its role in education and self-development, including the affirmation of alternative sexual identities. These accounts suggest that Millian defenses of free expression are applicable to sexually explicit expression. The anticensorship position does not affirm unlimited rights to free expression, but proposes boundaries that rule out certain kinds of state intervention, including the ban on extreme pornography as presently constituted.

Mill's Political Conception of Rights
Rejecting both the practical possibility and legitimacy of rights to privacy and free expression are critical elements of the pro-censorship case. In making their case, McGlynn and Ward establish a dichotomy between liberal fundamentalists and liberal humanists (or pragmatists). For them, the liberal fundamentalist position rests on shaky foundations, conceptual and historical, and affirms ontologically suspect "abstract rights" (2009,339). Pragmatists, by contrast, will not let such "abstruse distractions" get in the way of addressing "real harms and real injustices" in a liberal community (2009,350). Endorsing Rorty's (1991) stance that "to say something is right or wrong, just or unjust, is merely to say that it does or does not conform to current social practice" (McGlynn and Ward 2009, 339), they propose: [The pragmatist has] little time for those who peddle the insinuation that the proscription of pornography might have a "chilling effect" on some sort of allusive "right" of free speech. Neither will she fall prey to the juristic miasma of causal harm. The pragmatics of making life better should never be sacrificed on the altar of juristic abstraction. (2009,343) McGlynn and Ward claim Mill to the pragmatist side of their distinction. They suggest that liberal fundamentalists apply "a crude version of Mill's Harm Principle" (2009, 344), and that instead "Mill was never dogmatic in his thinking; he embraced complexity and compromise," and "it is to the subtle, accommodating and pragmatic Mill to whom we should turn, not the Mill whose car-icature is commonly discerned in so much libertarian thought" (2014,522). This account has compelling aspects, but the distinction between "fundamentalism" and "pragmatism" has its own critical weaknesses. Mill is not a libertarian, and the thesis in On Liberty simultaneously restricts and legitimates government action: "The interference of government is, with about equal frequency, improperly invoked and improperly condemned" (1977,223). McGlynn and Ward correctly differentiate the Millian approach from Dworkin's (2013), which conceptualizes "rights as trumps," essentially antimajoritarian claims against a democratic conception of the good. Zivi finds that both supporters and detractors alike have mistakenly assumed this approach to Mill is the only possible interpretation (2006,52). McGlynn and Ward's weakness is, having attached rights to liberal fundamentalism, a failure to recognize alternative approaches to conceptualizing rights. While Mill explicitly rejects "the idea of abstract right" (1977a, 224), he nevertheless has a key role for rights within his framework: "a recognition of certain immunities, called political liberties or rights, which it was to be regarded as a breach of duty in the ruler to infringe" (1977a, 218).
In contrast to McGlynn and Ward, Zivi (2006) suggests that rights claims are not the captives of apolitical abstractions but can act instead as an effective discursive strategy for addressing real harms and injustices. She argues that as well as risks, there are opportunities and benefits to using rights claims because of their persuasive means in solving practical social problems, and not because of their capacity to impose certain metaphysical assumptions on political discussion (Zivi 2012, 115;cf. Mill 1977a, 224). Rights claiming need not be antidemocratic, but instead a part of democratic politics (Zivi 2012, 59). Against this account, it is not enough, even as a pragmatist, to dismiss rights as mere "contingencies," disposable social fictions. For a pragmatist, every concept is a social fiction, so there is nothing especially problematic about invoking rights as part of a political argument.
Similarly, Dyzenhaus argues that even if a Millian wants to protect a right to privacy, she has no a priori definition of "private action" definitively exempted from the state's evaluation (1992b, 546). This is true but trivial once we acknowledge rights as political, not metaphysical. A Millian has few, if any, a priori claims to make. It is only within that framework that a Millian argues for acknowledging rights such as privacy as a fundamental aspect of human development and security. As a result, it is possible for a Millian to disavow some interpretations that rely on unjustifiable metaphysical claims and yet, on practical grounds, advocate rights to privacy and free expression. In the next section, I examine the practical implications of those rights.

Privacy, Free Expression, and the Harm Principle
A Millian liberal framework defends tractable notions of private conduct and free expression. While Mill himself did not comment on the issue of pornography, this framework implies a right to possess and share pornography that consenting adults produce (at least in private, voluntary, noncommercial settings).
Mill advanced his famous statement of legitimate government action in the form of what was later coined by Feinberg (1987) as "the harm principle." That is, "the only purpose for which power can be rightfully exercised over any member of a civilized community, against his will, is to prevent harm to others" (Mill 1977a, 223). This phrase seems to permit many interpretations, to the extent that "Mill is associated . . . with a wide range of views, to wit, everything from a conventional, and potentially illiberal, defence of 'reasonable' public dialogue to a radically libertarian defence of absolute freedom of expressing and publishing opinions" (Riley 2005, 148-49).
As a result, interpreters differ on the clarity in Mill's thinking. Reeves claims that "Mill's philosophical vision derives both its power and its weakness from his various attempts to knit together a number of diverse threads" and, as a result, attempts "to construct a coherent system from [Mill's] voluminous writings" are fruitless (2007,50). By contrast, Gray argues that "Mill's writings contain a coherent and forceful utilitarian defence of liberal principles about the right to liberty" (1996,12) and that while [Mill's doctrine] cannot mechanically resolve all questions to do with interference with liberty, [it] does supply a framework of considerations in terms of which such questions may be discussed. More incisively, it rules out from the discussion a whole range of considerations still widely invoked as germane to it. (1996,18) McGlynn and Ward come down decisively against this interpretation. They argue, following Bellamy (1992, 26), that Mill provides no means of reconciling competing interests "where there is a conflict of liberties" (McGlynn and Ward 2014, 506), and that he "provided no definition as to what might constitute 'harm'" (2014,505). I suggest that, while the philosophical grounds for Mill's position are inevitably contested, the practical outcomes are easier to discern.

Private Conduct
"Harm" can indeed include almost any behavior that disadvantages someone in some way. What critics tend to miss is the crucial condition of harm to others. The harm principle is about self-regarding action and personal conduct between consenting adults: "There is no room for entertaining any such question [of state interference] when a person's conduct affects the interests of no persons besides himself, or needs not affect them unless they like (all the persons concerned being of full age, and the ordinary amount of understanding)" (1977a, 276). This sets out a consistent limit on state action, applying to things like the possession of pornography as well as consensual sexual activities used to create it. It is equivalent, as Guest (2008, 118) proposes, to "cases where someone has produced a drug (e.g., growing cannabis) for their own use," where more complex questions surrounding social and commercial interactions are not present.
By contrast, when Dyzenhaus argues for censorship of pornography on Millian grounds, he offers support for "any coercion whether state initiated or by dint of informal public pressure, aimed at suppressing production, distribution, and consumption of pornography" (1992b, 534). This is a general description, both in terms of the methods justified, as well as the activities to be discouraged or suppressed. The implication is that these methods are all, in principle, legitimate actions according to a Millian account if they help to eradicate pornography. His justification is that Mill's concerns extend to the intimate, including relationships within households (1992b, 546-47). This is true. Instead of the threshold of the household, it is the harm principle itself that establishes the legitimate scope of government action, which Mill notes is sadly absent in the case of domestic relationships: The State, while it respects the liberty of each in what specially regards himself, is bound to maintain a vigilant control over his exercise of any power which it allows him to possess over others. This obligation is almost entirely disregarded in the case of the family relations, a case, in its direct influence on human happiness, more important than all others taken together. (Mill 1977a, 301) This is consistent with Mill's concerns about the domestic subordination of women, as familial partners and dependents are often exposed to abusive situations. However, it remains consistent with a principle of privacy. This goes beyond, as Gaus (1983) explains, a classical liberalism that sees individual rights as inextricably bound up with private property rights (Medearis 2005). Instead, more in line with contemporary liberal theorists, Mill sees individual interests as important but separate from claims on private property (Levy 2003, 278). This is a flexible, contingent conception of the division of the public and the private. For this reason, a Millian can view the "personal as political" in some domains precisely to criticize existing legal definitions of the private. She nevertheless believes that private conduct should be an acknowledged political category and protected from state intrusion.
We could reject this distinction between the public and the private and claim that even when viewing pornography looks purely self-regarding, it inevitably impacts on the social environment in a way that could be harmful. McGlynn and Ward suggest we "recall the lessons which can be drawn from the 'applications of Mill's 'harm principle' , particularly his willingness to recommend the regulation of behaviour and activity which raise 'only' the risk of future harm" (2014, 520). However, these policies refer specifically to cases where the interests of dependent family members, especially children, are in the balance (Mill 1977a, 304). This is a qualitatively different case from adults watching or engaging in consensual sex acts. By describing these scenarios, Mill shows that rights do not release individuals from positive obligations that they may have incurred (e.g., by getting married or having children). Mill distinguishes here between protecting civil liberties and "misplaced notions of liberty" (1977a, 304) that permit vice regardless of harm to others. Another example of this intuition could be that a personal right to smoke marijuana does not extend to the right to drive a car while under the influence of marijuana.
The problem with extending the case for regulation beyond applications regarding dependents is that it does not secure much space for private conduct at all. By contrast, Mill is keen to differentiate between "acts and habits which are not social, but individual" (1977a, 288). He places, as an example, drinking alcohol in the category of individual acts but the act of selling alcohol in the category of social acts. He explicitly rejects those who wish to subject private actions to the nebulous "social rights" of others: . . . the like of which probably never before found its way into distinct language: being nothing short of this-that it is the absolute social right of every individual, that every other individual shall act in every respect exactly as he ought; that whosoever fails thereof in the smallest particular, violates my social right, and entitles me to demand from the legislature the removal of the grievance. So monstrous a principle is far more dangerous than any single interference with liberty. (Mill 1977a, 288) So Mill does not accept current distinctions between the private and the public as natural, nor does he reject all notions of privacy altogether, but argues for the legitimate scope of private actions as those specific activities that do not, practically, impose any behavior on anyone who has not explicitly chosen to participate (Vernon 1996, 628). Drinking alcohol is protected even if many related activities, including the sale of alcohol, are subject to legitimate regulation. McGlynn and Ward point out that Mill supports severe and escalating penalties for violent drunks (2014, 509). Of course, as soon as violence enters the equation, we are no longer in the realm of Millian private conduct. The overall result of the discussion of "applications" is to show quite explicitly precisely what McGlynn and Ward wish to reject. Harm, whether through an act of commission or omission, must have some verifiable impact on individuals. It must "violate a distinct and assignable obligation to any other person or persons" (Mill 1977a, 281) in order for it no longer to be self-regarding (Ten 1968, 32). It cannot relate to an unverifiable social interest.

Free Expression
The harm principle protects some activities associated with consuming pornography, namely, those that take place exclusively in private spaces between consenting adults. The question of publication is more complex. When criticizing libel, for example, it is not strictly the existence of regulation that Mill opposes. Instead, it is the broad discretion given to judges and the criminal rather than civil approach that attract his criticisms (Mill 1984, 22-25). When defending freedom of the press, Mill excludes speech that creates an imminent risk of violence (Riley 2010, 80).
The right to freedom of speech has often been taken to be derivative of the harm principle. Reflecting this common view, Murray's (2009) analysis of the case for outlawing extreme pornography includes a demand for evidence of physical harm associated with pornography use. He takes the evidential bar to be proof that viewing pornography can be excluded from a right to freedom of expression that is otherwise assumed to be nonharmful. However, we know from the account of private conduct that speech is not intrinsically nonharmful. Public expression is a social act (Kateb 1996a, 227;Riley 2005a, 176). A Millian instead defends free expression on a utilitarian basis rather than referring back to the harm principle. Exercising free speech can disturb people, but "the benefits of free discussion and of the open struggle between competing conceptions of the good life outweigh the costs of such confrontation" (Waldron 1987, 423).
We can account for the common elision of the right to free speech and the harm principle by the fact that there is significant overlap both in theory and practice. For Mill, freedom of thought and discussion is "a single branch" of a "general thesis" (1977a, 227), and some elements of expression are "practically inseparable" (1977a, 226) from freedom of thought (Guest 2008). Thought is, of course, self-regarding. Furthermore, it is not the interest of those expressing ideas that count but the benefit of those able to hear different opinions (Mill 1977a, 242-45). In contemporary terms, this right of listeners represents a sort of freedom of information, amounting to the notion that the law should not normally come between a willing speaker and a willing audience, or indeed, in the case of thought itself, interfere with someone expressing ideas absent an audience, a case of testimonial freedom (Cohen 2006).
On the other hand, if it is primarily the right of the audience, then the right to expression does not extend to a limitless and absolute right to be heard. This indicates a distinction between expression made in closed settingswhether churches, assemblies, or safe spaces with their own rules of order-and expression broadcast or made in a public space. The presence of disturbing ideas in civil society is both inevitable and valuable for a Millian, but people are not obligated to witness them, and this seems particularly pertinent for violent or explicit sexual content: "To prevent serious forms of perceptible damage that nobody should be forced to suffer, laws of justice should include time, place and manner restrictions that distribute equal rights not to be confronted or bothered by speakers" (Riley 2010, 80). While a Millian framework cannot be read to support a free speech "absolutism," a pro-censorship reading of On Liberty stretches it beyond recognition.
Against this position, Dyzenhaus contends that "liberals who regard Mill as the founder of their tradition should reevaluate their position on pornography in light of Mill's curiously neglected essay The Subjection of Women" (1992b, 534). For Dyzenhaus, Mill's critique of Victorian society runs much deeper than legal institutions: "Mill does not see legally prescribed inequalities between men and women as much more than de jure recognition of de facto social relationships based ultimately on . . . the superior physical power of males" (1992b, 538).That is, the subordination of women is founded on coercion, but, as it turns out, is all the more insidious and pernicious because it takes on the guise of consent. Women are trained to become the "willing slaves" of men-indeed, to find submission to men an attractive prospect. Pornography's function, in this account, is to eroticize unequal relations between men and women, thus altering the preferences of men and women toward sexuality.
However, interpreting The Subjection of Women in light of a Millian framework suggests that Dyzenhaus's argument is flawed if the intention is to justify censorship rather than the use of persuasion, and social opprobrium, to change attitudes toward women. Mill discusses the subordinate and sycophantic representation of women in literature, but not to argue that it plays a significant causal role in women's oppression. It is a reflection of the legal position that women are actually in (1984,279). He shows that many popular beliefs about women are a reflection of discrimination, not a compelling justification for that discrimination. The problem with relations between men and women taking on the guise of consent is that, in Mill's context especially, consent really is always absent because women's consent to a whole range of personal and economic choices is ultimately not required, not even formally in law (1984,292). It is not a case of women being persuaded, or having their personalities ultimately determined, to be subordinate to men. Instead, their subordinate stance and meek demeanor ameliorate positions into which they have, in fact, been forced with the threat of legalized violence.
If Mill had supported censorship, of women's literature or anything else, in The Subjection of Women, in the hope of improving women's social status, then that would be a historically puzzling departure from the Millian framework. However, Mill did not propose censorship as a solution, suggesting that a Millian approach could be both radically feminist and anticensorship. Are these two elements consistent? In the next section, I suggest they are by showing the parallels between Millian anticensorship and contemporary queer feminist defenses of sexually explicit expression.

The Millian Alternative to Prohibition
Rights justified in political, not metaphysical, terms must inevitably stand on some practical human interest. Adumbrating rights within a Millian framework and pointing out inter alia that they protect the possession and creation of adult consensual pornography, including consensual extreme pornography, is arguably insufficient to make my defense ultimately persuasive. As McGlynn and Ward argue, "Mill . . . would have been exercised in the extreme by the supposition that his essay 'on liberty' should be used to institutionalise the cultural degradation and jurisprudential inequality of women" (2009,336).
If that is all that defending extreme pornography from criminal prohibition achieves, then I risk falling into the liberal fundamentalist trap that privileges abstract legal notions over real human interests. I propose that rights do not represent abstruse constraints, as the pro-censorship case sees them, on effective action to oppose misogynist and other harmful forms of expression. When affirmed appropriately, these rights enable effective state action and, more substantially, generate the social circumstances in which voluntary action can address the harms of some pornography. These rights allow people, including theorists, activists, porn producers, porn consumers, and sex workers, to affirm and disseminate alternative positive visions of human sexuality. When the state censors, by contrast, these activities are all too easily silenced and prohibited.

The Limits and Capacities of Criminal Regulation
My anticensorship approach acknowledges the limits of the effectiveness and benevolence of coercive state intervention (Gaus 2010, 91) and endorses a "focus on how institutions perform under worst-case conditions" (Farrant and Crampton 2008, 118), including government action. A Millian criticizes theorists who rely on judging institutions only at their most attractive moments: "Whether the institution to be defended is slavery, political absolutism, or the absolutism of the head of a family, we are always expected to judge of it from its best instances" (Mill 1984, 287). It is with these "robustness considerations" (Farrant and Crampton 2008) in mind that Millians reject traditional forms of power, especially patriarchal power, while simultaneously affirming the importance of individual rights against the modern state. For a policy regime to be worthy of defense, it should work reasonably well without assuming the benevolence of those tasked with enforcing it: "The British constitution supposes that rulers always wish to abuse their power and . . . wish to remove every check which has a tendency to prevent them from abusing their power" (Mill 1984, 19;cf. Kateb 1996, 231). While these particular claims are made in classical liberal language, there is a complementary state-skeptical and critical tradition running through much contempo-rary feminism, and this applies particularly to criminal justice approaches. Howe (2013) argues it is important to consider the outcomes, rather than just the pronounced intentions, of laws in order to understand their regulatory function. Bernstein (2012;cf. Bumiller 2013, 197) criticizes the phenomenon of "carceral feminism," where political actors co-opt feminist concerns in order to extend the coercive reach of the criminal law, not with the effect of making women safer but extending state power, whereas Bumiller advises us "to be aware of both the potentialities and limitations of using state power to advance the interests of women" (2008, 2).
The pro-censorship case discounts the likelihood of the state misusing its powers in the case of the extreme pornography ban. McGlynn and Rackley argue that abuse is unlikely because "proceedings for an offence may not be instituted without the consent of the DPP [Director of Public Prosecutions]" (2009,255). 3 A Millian skepticism, by contrast, sees relying on the judgment of a single public official to be problematic, a concern that prosecutors' use of this law in practice has somewhat vindicated. McGlynn and Ward argue instead that defending extreme pornography is a sadistic application of rights claims (2009,341). In their account, skepticism of criminal intervention involves permitting women's degradation on the basis of abstract, imaginary concerns with state power. However, when applied against individuals, the law itself could unintentionally legitimize another example of sadism: the evident satisfaction that many people experience from punishing and humiliating sexual deviants, quite typically gays and lesbians (cf. Ross 2000, 305;Strossen 2000, 231). McGlynn et al. wish to address "cultural harm," but a "culture" itself cannot be convicted or incarcerated (nor can "speech" as such). Criminal legal intervention requires an individual to punish.
As a result, criminal law that presumes to address cultural problems requires individuals to stand in, potentially as scapegoats, for the harms imputed to a wider culture: "The users of extreme pornography are the legitimate targets-they must take responsibility for the market they create in materials which often glorify sexual violence against women" (McGlynn and Rackley 2010, 10-11). I suggest that the sheer variety of people who create and possess extreme pornography renders this process of selecting individuals to punish for "cultural" crimes deeply problematic. That prosecutors have ended up treating gay men recording consensual sex acts together as "legitimate targets" is only the most obvious illustration of this issue, suggesting that this approach to regulation is both harmful (it treats some users of queer pornography as collateral damage) and ineffective at identifying plausible sources of violence against women (cf. Attwood and Smith 2010).

Context, Not Content
Having noted the dangers of criminal law, I should acknowledge where a Millian approach endorses state intervention. I propose that it is not the content but the context of the creation and publication of pornography that should be decisive. The question is whether a particular pornographic expression constitutes "a positive instigation to some mischievous act" (Mill 1977a, 260). The test here is a combination of the intent and likely effects (the illocutionary and perlocutionary elements) of the expression. A Millian supports freedom of speech but not freedom of speech act (Jacobson 1995;Langton 2009b;cf. Zivi 2014). 4 With respect to possession, this excludes pornography created through coercive means for the same reason that liberals support bans on child pornography, which records individuals who are incapable of consenting to sex (Strossen 2007, 141). This justifies proscribing pornography that records real acts to which a participant does not consent or cannot consent.
On the publication side, we should not read free speech absolutism into a Millian framework. As well as a right to access ideas, being able to isolate oneself from certain forms of expression, especially from populist discussion, may be an important part of cultivating individual character and challenging "the moral coercion of public opinion" (Mill 1977a, 223). In our contemporary world of instantly accessible and aggressively promoted digital media, this principle can be extended to ensure that people can easily avoid exposure to material they do not wish to see. This renders sexually explicit content, in the context of public broadcasts, billboards, posters, and commercial cinemas, but also Internet advertising and unsolicited material (cf. Williams 1979, 118), subject to legitimate regulation. This does not infringe upon expression rights. Free expression, on my interpretation, means that individuals need both freedom to access pornography and, equally importantly, freedom from unwanted exposure to pornography.
In addition, a Millian approach supports vigorous prosecution, or powerful civil remedies, wherever pornographic expression is used to target and harm specific individuals. In other words, when pornographic material is deployed with the intent or the effect of encouraging criminal acts, including inducing fear and harassment or assault of women and others, then it is rightly restricted and very wrong to ignore. A paradigmatic example of this is the phenomenon of "revenge porn," where perpetrators shame and threaten victims with compromising images published on the Internet (Henry and Powell 2015). A Millian endorses criminal penalties for those sharing explicit images that violate the privacy of those depicted, or designed to cause distress to specific individuals. Legitimate penalties extend to individuals hosting sites or otherwise facilitating or commercially benefiting from these acts.
In this sense, from a Millian perspective, individuals are often overregulated by having their private, consensual activities subject to criminal prosecution, but also underregulated due to a lack of access to legal remedies for harmful, nonconsensual acts committed using explicit imagery.

More Speech, Better Speech, and Voluntary Action
While a Millian endorses regulation when appropriate, some social problems are not readily amenable to criminal justice. A Millian seeks out "other instruments to persuade people to their good, than whips and scourges, either of the literal or the metaphorical sort" (Mill 1977a, 277). She, along with Cornell, is "suspicious of overreliance on law" (2000,553). While some very bad, reactionary, and misogynist ideas are expressed in some pornographic texts, a Millian response to bad ideas is generally not to restrict them but to permit more and better expression to contest them. While Dyzenhaus (1992a) rejects this optimism that the best ideas will win out, in the case of extreme pornography, there are signs that confidence in more, and freer, expression is effective at challenging prevailing sexist norms rather than reinforcing them, a strategy of "unleashing the feminine imaginary, rather than constraining men" (Cornell 2000, 553). The pro-censorship approach, almost without exception, treats the variety, and deviancy, of sexual expression that the Internet has permitted as a problem that requires urgent criminal regulation (Smith and Attwood 2013, 43). Hartley notes, by contrast, that the decentralized nature of the Internet disrupts the homogenous and heteronormative products and processes of the existing pornography industry, leaving more space for personal, and radical, expression: Piracy and the Internet have severely diminished the profitability of traditional business models, while also creating access for other viewpoints. . . . [P]reviously marginalised people now have access to the means of production. Rather than being treated as some kind of freak show, pierced, tattoed, disabled, queer and trans folk can now make movies that speak to their sensibilities and create communities that support and foster them and their sexual relations. (2013,235) Of course, part of the attraction of pornography, particularly for young people, emerges from the failure of government to provide effective sex education, and the fact that sexual health information is sometimes deliberately suppressed. Unlike extending and escalating criminal penalties (Bumiller 2013, 201), access to sex education is associated with reductions in harmful sexual activity, including, critically, instances of nonvolitional sexual encounters (Macdowall et al. 2015). There is thus a significant opportunity to reduce the harms of sexual violence, yet this approach is often ignored because it lies outside the paradigm of criminal justice.
It is with respect to education that protecting sexually explicit media can be particularly important (Strossen 2000, 163-66). McGlynn and Ward are keen to dismiss the putative value of extreme pornography in stark terms, arguing simply that "some speech is 'high-value'" and some is "low-value." The screams of pain and misery that tend to accompany images of extreme pornography fall into the latter category" (McGlynn and Ward 2009, 350). With this image, McGlynn and Ward risk reducing participants in the creation of extreme pornography to essentially silenced victims, literally (in this case) incapable of intelligible speech. However, real participants and producers of extreme pornography have a life outside of the texts they produce (and frequently perform more complex roles within them). Alternative feminist and queer approaches include the perspectives of both producers and consumers of pornography in their considerations. As a result, Cornell argues for a feminism that treats "women, including porn workers, as selves individuated enough to have undertaken the project of becoming persons. To treat women in the industry as reducible to hapless victims unworthy of solidarity refuses them that basic respect" (2000,552).
Attwood, summing up recent empirical research, identifies pornography as variously "a source of knowledge, a resource for intimate practices, a site for identity construction, and an occasion for performing gender and sexuality" (2005,65). Hartley, as both a producer and researcher of pornography, affirms the value of these kinds of expressions: "As a porn performer I can express myself as both artist and scientist" (2013,231). She acknowledges, in particular, the educative role that pornography with extreme themes can have: "Some porn, particularly porn that is more focussed on mutual pleasure in whatever form-including those that challenge conventional notions of pleasure, like consensual BDSM-can be of instructional value" (2013, 233). As "essentially live-action cartoons burlesquing social conventions" (Hartley 2013, 234), pornographic texts can be politically and socially challenging commentary (cf. Brown 2002). Just because content is sexually arousing does not render it incapable of generating critical reflection. Since violent and extreme behaviors and ideas are a feature of both society and many people's imaginations, it is inevitable that some pornographic material will reflect that in the form of extreme themes.
Ward argues that her engagement and consumption of pornography is not mere enjoyment but an important opportunity for self-development and spiritual reflection: "I cultivate a private, internal space where I can honour and observe the complexity of my sexuality as it evolves. Though I remain publicly accountable, I provide myself with moments of exploratory freedom, creative license, and orgasmic surprises" (2013, 139;cf. Donner 1993, 163). In addition, Green (2000;2005, 496) notes how antipornography feminist theory can ignore the positive role it has in supporting the formation of gay identities, often in the absence of public recognition.
These experiences and analyses urge us to include pornography within the presumptive (not absolute) protection that a Millian approach extends to discussion and opinion of all things "practical or speculative, scientific, moral, or theological" (Mill 1977a, 225). Such expression is part of "framing the plan of our life to suit our own character: of doing as we like, subject to such consequences as may follow . . . even though [others] should think our conduct foolish, perverse, or wrong" (1977a, 226;cf. Zivi 2006, 57).
Of course, a Millian is not only concerned with the potential harms associated with pornography in wider society, but also with abuse and coercion within institutions. She acknowledges the legitimacy of regulation wherever pecuniary interests are involved, even in cases where violent coercion is not present: "Fornication, for example, must be tolerated, and so must gambling; but should a person be free to be a pimp, or to keep a gambling-house? ' (1977a, 296). Acknowledging that a great deal of pornography still uses paid actors rather than performance artists, or private individuals expressing their own ideas, a Millian approach certainly endorses the regulation of the commercial pornography industry, including health and safety regulations, and outright prohibition on paying for some harmful activities.
At the same time, a Millian approach acknowledges the possibility, and potential superiority, of voluntary alternatives to direct state regulation. Reeves notes that Mill remarked in one of his less well-known essays, "on centralisation" (1977b, 579), "on the need to protect voluntary organisations and local initiative-vital incubators of liberty and diversity-from the power of the central state" (Reeves 2007, 51). Similarly, Bumiller argues that the very strength of some feminist projects against domestic violence has been their ability to keep an often hostile state, and other hierarchical forms of power, at arm's length (2013,193). As Cornell finds, feminist activism within the pornography industry has found success in campaigns for unionization and self-organization, and these have done more to achieve improvements in real welfare and working conditions than state-led approaches that too often treat "'women in the industry as if they were incapable of asserting their own personhood" (2000,552). Thus, there have been successes, but there is much more for activists to do. The strength of Millian rights is that they generate spaces for individuals to combine and act together to address harms within an industry, or a sector of society, without the permission of state officials.
Against this stance, which supports spaces where varieties of consensual sexuality can be protected and explored, McGlynn and Ward argue that Mill's personal ethical position is "intrinsically aligned to his idea of the individual as a progressive, morally and intellectually improvable, being" (2014, 508) and that he was "troubled by the meaning and role of sexual activity in society" (2014, 509). Reeves suggests that this particular vision of Mill as an "ascetic, dry, humourless, sexless, lofty intellectual" (2007,47) is an unjustified caricature. Regardless, a contemporary understanding of the Millian framework can certainly encompass a vision of autonomy that recognizes sexual exploration as an important, reflective part of a reasonable life plan for a great many individuals.

Conclusion
I contribute to the growing debate about the regulation of adult pornography, particularly in light of the recent prohibition of "extreme pornography" in the United Kingdom. I have shown that a case for censorship is hard to sustain within a recognizably Millian framework, and that this framework is more informative about the protec-tions afforded to individuals than critics have sometimes claimed. McGlynn and Ward claim that "there is a strong, liberal basis for pornography regulation" (2014,521). In an important sense, this is absolutely true, acknowledged even in the Williams (1979) report that initiated this debate. Just because material is pornographic does not exclude it from the ordinary regulation of a liberal society. When pornography is used to threaten or harm individuals or groups, like any other form of expression, legal remedies are essential. What I have shown here is that there is not a plausible understanding of the Millian approach that makes sexually explicit discussion, fictional depictions, or simple records of consenting adult sexual activity specifically less deserving of protection than other forms of expression. Moreover, affirming these rights provides resources for addressing the harms of some forms of pornography.
We should affirm a Millian framework of rights not by faithfully interpreting abstract rights, but because this approach can practically frame legislation that constrains authorities from using their powers in ways that are manifestly abusive and illiberal. Without those considerations, it is likely that the law, no matter its intentions, will be, as even Rackley and McGlynn (2013) acknowledge can happen, "misunderstood and mis-used."