I. Skepticism about Free Speech

“ 'Free speech' is just the name we give to verbal behavior that serves the substantive agendas we wish to advance”—or so literary theorist and professor of law Stanley Fish has claimed.¹

This cynical dictum is one of several skeptical challenges to freedom of speech that have been extremely influential in the American academy. I will follow the skeptics' lead by distinguishing between two broad styles of critique: the progressive and the postmodern. Fish's dictum, however, like many of the bluntest charges, belongs to neither class exclusively. As an initial characterization of the distinction between these critiques, progressive skepticism claims that freedom of speech is a bad thing, while postmodernist skepticism claims it to be conceptually impossible. Both forms of skepticism hold the classical liberal endorsement of free speech and condemnation of censorship to be both naive and reactionary. Skepticism about free speech flourishes at universities in the United States and is especially well represented among professors at the country's most prestigious law schools. As legal scholar Robert Post approvingly observes: “Liberated from traditional inhibitions against official suppression of speech, the left has mobilized to pursue a rich variety of political agendas.”²

It is odd, but not inexplicable, that Post would refer to a “rich variety” of political agendas while simultaneously describing them all as leftist. The explanation of this remark, I will ultimately suggest, can be found in the ideological conformity of the American academy. However that may be, Post correctly observes that progressives (or leftists) have sought to advance several of their political aims by suppressing objectionable speech. According to this progressive critique, freedom of speech serves to maintain the status quo by bolstering a fundamentally oppressive and unjust society. Indeed, not merely objectionable opinions, but also those who seek to protect them from suppression, have come under attack by leftist academics who—as Post delicately puts it—have been “liberated from traditional inhibitions” against censorship. The architects of campus speech codes, and advocates of legislation against pornography and (so-called) hate speech, consider freedom of speech to be a tool of oppression and its advocates unwitting oppressors.³

Critical race theorists constitute one particularly strident group of progressive skeptics. In the introduction to their influential book, Words That Wound: Critical Race Theory, Assaultive Speech, and the First Amendment, legal scholars Mari J. Matsuda, Charles R. Lawrence III, Richard Delgado, and Kimberlè Williams Crenshaw write: “The first amendment arms conscious and unconscious racists—Nazis and liberals alike—with a constitutional right to be racist.”⁴

This rhetoric seems excessive, even if we do not infer from the claims that liberals are unconscious racists and that punishment is “an appropriate response to racist speech” the conclusion that liberal advocacy of freedom of speech should be suppressed.⁵ It may be unclear why liberals, who as a matter of principle defend the speech rights of racists and critical race theorists alike, should be demonized as racists.

Consider the quintessential defender of the classical liberal conception of free speech, John Stuart Mill (1806–73). Although Mill's credentials as a feminist and antiracist are unimpeachable, he defended the “absolute freedom of opinion and sentiment on all subjects,” which he understood to include not just thought but also its expression.⁶

Even if such absolutism is untenable, this statement clearly belies Fish's cynical dictum. The opinions of Fish and the critical race theorists do not in any way advance the liberal cause, yet no liberal can place them beyond the pale of free speech protection. Hence, Fish's cynical claim founders, although he may prove to be giving an accurate gloss of his own usage.

A more challenging objection to the classical liberal conception of free speech, which we will consider presently, holds that since some limits to the protection of speech are unavoidable, there can be no real freedom of expression. But this conclusion, which I have termed ‘postmodern’, is too hasty, since even Mill did not take “absolute” freedom of speech to rule out any possible interference with verbal behavior. Rather, the first tenet of the liberal conception of free speech is the doctrine known as content (or viewpoint) neutrality. This is the view that, as Mill put it: “there ought to exist the fullest liberty of professing and discussing, as a matter of ethical conviction, any doctrine, however immoral it may be considered.”⁷

Not even the harmfulness of an opinion, per se, provides good (much less sufficient) reason to suppress it. Some commentators have obscured this point by attributing to Mill an overly strong ‘harm principle’, on which any harm to nonconsenting others provides reason to interfere with an agent's liberty.⁸ Yet Mill expressly insists that the freedom to speak should be protected “[h]owever positive anyone's persuasion may be, not only of the falsity but of the pernicious consequences” of the opinion or sentiment expressed.⁹

According to the progressive critique, however, the liberal's toleration of even false, immoral, and harmful opinion is just the problem. Progressives see the overt suppression of objectionable views as aiding in the struggle against racism and sexism, and more generally against what they consider to be a grossly unjust American society. The problem with “free speech fundamentalism,” as Fish pejoratively terms the liberal position, is that it “prevents you, as a matter of principle, from inquiring into the real world of consequences of allowing certain forms of so-called speech to flourish.”¹⁰

This looks like another crude exaggeration, since, in fact, Mill holds that one can and should inquire into the consequences of various opinions, specifically in order to combat harmful opinions with criticism and to mitigate their bad effects by noncoercive means. It is true, however, that for Mill no opinion or sentiment may be prohibited or punished, “whether the means used be physical force in the form of legal penalties, or the moral coercion of public opinion.”¹¹ The second tenet of the liberal conception of free speech thus focuses on this distinction between criticism (which is not merely permissible but encouraged) and coercion (which is proscribed).

Progressive critics reject this distinction on several grounds. First, it is possible that coercive measures might be more effective at limiting the bad consequences of some opinions, yet liberals rule out such measures on principle. (Hence, a more modest argument along the lines Fish suggests would be less easily refuted, but there is good reason to doubt the efficacy of legal suppression.)¹²

Second, progressives commonly argue that legal guarantees of free speech inadequately protect the disempowered: those who lack the means to make their voices heard. Even some scholars who sincerely affirm the value of free speech claim that the principle can justify the censorship of certain opinions. For example, legal scholar Owen Fiss argues that the state “may even have to silence the voices of some in order to hear the voices of others.”¹³ Section II will examine in detail such ‘silencing’ arguments, which claim to promote freedom of expression by selectively suppressing opinion.

Yet the embrace of politically motivated censorship by academic leftists is already old news, according to Post, which has been made obsolete by more recent and radical scholarship. This postmodern critique claims that the liberal conceptions of free speech and censorship are not only politically retrograde but also philosophically incoherent. Whereas progressives defend censorship as a political tool and attack free speech as an ideology of oppression, postmodernists advance a thesis that would undermine liberal ideals even more profoundly. Post describes the current situation this way:

These implications are claimed to be radical indeed. The innovative new scholarship supposedly demonstrates that censorship is, in the words of legal scholar and philosopher Frederick Schauer, “inevitable, necessary, and desirable.”¹⁵

Similarly, Post claims that this postmodern insight about the ubiquity of censorship undermines the liberal ideal of free speech, because to oppose censorship as such “is to assume a freedom no one has” or could ever have.¹⁶ And Fish writes, “[the] truth is not that freedom of speech should be abridged but that freedom of speech is a conceptual impossibility because the condition of speech's being free in the first place is unrealizable.”¹⁷ Fish's scattershot arguments illustrate that the postmodern critique does not preclude a progressive political agenda. Yet, though many postmodernists embrace the progressives' claims, an obvious tension threatens this alliance. If the postmodern thesis is true, then the ubiquity of censorship jeopardizes the progressive assumption that the silencing of the disempowered can be remedied.

Post acknowledges this difficulty:

This question bears directly on an issue addressed in the conclusion of this paper. How can academics complain about violations of their intellectual freedom or their right to engage in political dissent, while also claiming that freedom of speech is both a faulty ideal and a conceptual impossibility? Fish proposes an answer that is, at any rate, admirably frank. “[A]s long as so-called free speech principles have been fashioned by your enemy,” he councils, “… contest their relevance to the issue at hand; but if you manage to refashion them in line with your purposes, urge them with a vengeance.”¹⁹

This proposal should be kept in mind, as it will help explain what might otherwise be mysterious about the practical circumstances in academia that I will ultimately consider.

Although I will be concerned throughout this discussion with the concepts of free speech and censorship, I will not be doing First Amendment scholarship, which I can hardly dispute with the likes of Fiss, Post, and Schauer. But the innovative new scholarship that Post hails, even when practiced by law professors, is not primarily concerned with jurisprudence. Instead, this postmodern scholarship poses a theoretical challenge to the philosophical foundations of free speech; and, like the progressive critique, it sounds a political challenge to liberalism. I propose to take up the arguments on both the theoretical and political fronts, in defense of the putatively unsophisticated traditional position. I contend that the progressive arguments against free speech are not advocated consistently, and that they would have some surprising consequences if they were. Moreover, the innovative—or at any rate fashionable—new scholarship lauded by Post does not bear up under critical scrutiny. The postmodern critique of free speech is vulnerable to a reductio ad absurdum by its own examples, which have untenable results and elide fundamental moral distinctions.

In what follows, I will examine what I take to be the most philosophically interesting and challenging objections to freedom of speech as understood by Mill. Some of these skeptical challenges are compatible with liberalism, at least insofar as they are compelling, and those that are truly incompatible do not succeed. This new scholarship does not improve upon the Millian conception of free speech it claims to supersede. Finally, in conclusion, I want to briefly consider one sort of “real-world consequence” of the political agenda shared by the progressive and postmodern free speech skeptics. It is illuminating to examine the attitudes toward freedom of speech and inquiry prevalent on American colleges and universities, where the progressive political agenda has been most thoroughly implemented. My focus on American institutions is motivated by the unique tradition of nearly absolute deference to free speech principles in the United States, a fact agreed upon by all sides. “On almost every issue of free speech theory, doctrine, and practice,” Schauer observes, “virtually every country on the face of the earth diverges from the United States, and diverges in the direction of lesser protection.”²⁰

My focus on academia, and in particular on the humanities, is motivated by the common claim that universities are and ought to be bastions of intellectual and political freedom.

The radical critiques reveal that something worth calling ‘silencing’ does pervade American campuses. This term is most commonly used to refer to the social intolerance of divergent opinion, backed by an ideology and enforced by those who wield power in a given society. Silencing thus covers some of the middle ground between criticism and censorship. I will adopt something like Schauer's characterization of an ideology as “a prevailing idea existing within an environment in which adherence to the idea is more or less required, and challenge to the idea is more or less discouraged.”²¹

This gloss is a bit too narrow, since an ideology tends to be a system of ideas and norms, but with that small caveat Schauer's definition will serve our purposes well.

The irony of free speech skepticism is that the postmodern critique, which purports to focus “with a sharp Foucaultian lens” on the social facts of power and oppression, gets those facts wrong when it looks at itself. In fact, both the progressive and postmodern forms of skepticism obscure the abuse of official power by those whose self-conception puts them in solidarity with the powerless. Despite their insistence that neither power nor censorship is monopolized by the state, free speech skeptics conceive of dissent only in opposition to the state. Therefore, the ideology predominant on American campuses is blind to its own enforcement of orthodoxy, because it conceives of itself as “dissent” rather than “conformity,” notwithstanding the social facts. This irony has deepened since the attacks of September 11, 2001, as academics who so fervently celebrated their liberation from free speech principles now claim to be silenced when their opinions are criticized as anti-American. Patriotism has long ceased to be considered a virtue in academic culture, where overt hostility to American values and interests is far too obvious now to be disguised. Indeed, there was little attempt to disguise the ideology of dissent that pervades the humanities, until it recently became expedient to do so.

II. Censorship, Ideology, and Diversity of Opinion

Fish poses the following dilemma for the liberal view, which he seems to find decisive. Either freedom of speech serves some instrumental purpose or it serves no purpose beyond itself. If free speech has no further purpose, then “there is no compelling—that is, serious—reason for adhering to it.”²²

This conclusion is much too hasty, but I am more interested in the other horn of the dilemma. Suppose that free speech is just instrumentally valuable, whether as a means to individual autonomy, collective self-determination, the discovery of truth, or simply the flourishing of society. Fish thinks this concession is sufficiently damning. “[O]nce such a good has been specified,” he contends, “… it becomes possible to argue that a particular form of speech, rather than contributing to its realization, will undermine and subvert it.”²³ While Fish calls any argument that identifies an instrumental value of free speech ‘consequentialist’, this is vulgar consequentialism. It is equally true that once such a good has been specified, it becomes possible to argue that this good is more likely to be diminished than promoted by the suppression of opinions claimed to “undermine and subvert” it.²⁴

For instance, it is possible to argue that skepticism about free speech—which aims to subvert freedom of expression by calling it unconsciously racist or conceptually impossible—threatens to undermine the liberal ideal. Yet for the liberal, this possibility does not justify the censorship of Fish and his school, for that would be a violation of their free speech. Three features specific to the free speech debate make cautionary arguments against unconstrained direct consequentialism particularly compelling. First, Fish makes no secret of his hostility to free speech, or his willingness to use even specious and intemperate arguments to undermine it; other skeptics are more circumspect but not always more sincere. Second, those with the power to censor speech have historically been prone to accept claims of harm based on ideological sympathy rather than evidence; and so it is with contemporary free speech skeptics, as we shall see. Third, when the confidence of diverse social groups in a political convention (such as punishment or freedom of speech) is tenuous, it is especially important to maintain strict procedural fairness, untainted by special pleading.

Yet Fish's argument, despite its crudeness, reveals an important point. It has become a common skeptical strategy to identify some value underlying free speech, and then to claim that other mechanisms have “silencing” effects comparable to censorship because they similarly diminish this value by in some way inhibiting speech. This silencing phenomenon demands further explication, since it cannot simply be inferred that speech rights conflict and, therefore, must be balanced against each other. To be sure, when both the intentions and the effects of some silencing acts resemble acts of censorship, the case for assimilating them as violations of free speech is strongest. When a speaker is “shouted down,” and thus deliberately prevented from being heard, there is a compelling case that this should be treated as a violation of his speech rights. Hence, certain verbal behavior can conflict with freedom of speech, but this is not in itself very significant. Not all verbal behavior falls under the Millian defense of freedom of expression, which, of course, does not recognize any right to shout down another speaker.²⁵

A subtler point may prove more challenging, though, inasmuch as this example suggests that speech rights might extend, to some degree, to a right to be heard.²⁶ This conclusion would be more problematic. It opens up the possibility that the free speech of some might conflict with the speech rights of others, albeit not as straightforwardly as in the shouting down case. Another reason to doubt how broadly this case generalizes is that proscriptions against shouting down can easily be framed neutrally, without regard to the value of the speech being silenced—or, rather, drowned out. Nevertheless, perhaps there is something worth calling ‘silencing’ which, though short of censorship, should also be treated as a violation of speech rights.

Consider, for example, an argument made by Owen Fiss, a tempered defender of free speech who nevertheless claims that certain opinions (not just verbal behavior) have silencing effects. Fiss identifies his antagonists, in a memorable phrase, as “racists, pornographers, and the rich.” ²⁷

The challenge for Fiss lies in explicating this argument, since his targets do not deliberately suppress or literally drown out others' speech. Indeed, minorities, women, and the poor can and do voice their opinions without persecution. Yet, Fiss writes:

To be credible, though, these claims cannot simply be asserted; the “silencing dynamic” must be shown to exist and to be caused by the targeted speech rather than other social forces. It is curious that Fiss does not find it necessary to adduce evidence of this silencing, beyond citing the assertions and feelings of those who claim to be its victims. As previously noted, indirect consequentialist arguments that peremptorily rule out the suppression of speech, even for the purpose of promoting debate, are most compelling when censorial conclusions are prone to be accepted on inadequate evidence. Fiss's failure to subject these claims to any real scrutiny is telling. Further doubts are raised by the fact that the targeted speech is offensive to those who seek its suppression. Critical race theorists and antipornography feminists despise the speech they seek to censor, and, by their own admission, they do not share even Fiss's attenuated commitment to free speech principles. Many advocates of silencing arguments would favor censoring hate speech and pornography whether they have any significant silencing effect or not.

Let us grant for the sake of argument that hate speech and pornography produce the silencing effects Fiss describes. It still must be argued that they constitute a violation of speech rights. Surely not everything that diminishes someone's sense of worth impedes her participation or authority in public debate, and not every such impediment is a rights violation. Furthermore, in order to advance a silencing argument—that is, an argument for using censorship to promote free speech—Fiss needs to show that censorship is the best remedy for silencing. This requires him both to identify the value of free speech and to explicate a conception of silencing on which it can further that value despite inhibiting speech. Fiss satisfies the first requirement, however contentiously, by identifying the primary value of free speech as its essential role in democratic governance. In his view, free speech principles primarily serve to advance social rather than individual freedom:

Unfortunately, Fiss does not elaborate on this lesson taught by affirmative action, which he seems to assume will be obvious. But even if we accept Fiss's narrow and controversial view of the value of free speech, he offers little argument for his assertion that bad speech must be remedied not by more speech but less whenever “private parties are skewing debate and the [proposed] state regulation promotes free and open debate.”³⁰

For now we can postpone the discussion of silencing arguments and focus on investigating claims of silencing.

One reason Fiss seems blithe about defending silencing claims is that he accepts an extremely capacious notion of them that is too broad to be identified with the violation of speech rights. For instance, he claims that in contexts of allocation, “a certain measure of silencing is inevitable” due to the scarcity of resources. Despite acknowledging that “the applicants who are denied the grant are still ‘free’ to pursue the project on their own,” Fiss characterizes this freedom as “largely formal,” because some of those who are rejected will lack the necessary resources.³¹

But this is dangerously misleading. Surely Fiss is right that some projects will never come to fruition due to lack of funding. Scores of unproduced screenplays attest to this fact and to the folly of trying to nurture every voice and idea. Yet such putative silencing-by-not-funding differs greatly from censorship in many other respects, including both its intentions and its secondary effects, even if the results (construed narrowly as the failure of some projects) are similar. Does someone whose grant proposal is not funded have a complaint that she has been silenced, her speech rights or academic freedom violated? This question seems to hang on the reasons for the negative decision, but Fiss's gloss of silencing makes no such distinction. It is one thing to compare censorship to the silencing caused by pressure to conform, backed by social rather than legal sanctions; it is quite another to compare censorship to the failure of one's projects to be funded. The problem with too capacious a notion of silencing is that it assimilates very different cases and thereby obscures crucial moral distinctions. Frustrated screenwriters waiting tables in Los Angeles have not been “silenced” in any significant sense, though screenwriters who were “blacklisted” for their political views in the 1950s were.

Although Fiss's allusion to affirmative action is terse, there are some obvious parallels to the affirmative action debate that might be used to sharpen his central claim. It is often argued that racial diversity furthers diversity of opinion. Another argument for traditional affirmative action is that there must be a “critical mass” of like-minded (or like-bodied) students and faculty in order for minority views to flourish and be heard. Yet, however that may be, Fiss's commitment to ensuring that all opinions are given a fair hearing in free and open debate seems to support a novel sort of affirmative action. If the goal is to ensure that all sides of an argument are heard, then an institution's responsibility must be to promote intellectual diversity: diversity of opinion. In circumstances where administrators have a commitment to fostering diversity of opinion or protecting intellectual freedom—either because of the pretensions of the institution or the nature of its funding—then the disappointed applicant may have grounds for complaint. In particular, when administrators do not merely fail to promote unorthodox opinion but suppress it for ideological reasons, claims of silencing are most plausible.

In an argument decrying the entrenchment of orthodoxy in academia, Schauer similarly concludes that universities should promote unorthodox views, without suggesting that this requires the suppression of popular opinions. Since “one earmark of intellectual honesty is confrontation of the best arguments for the opposing position,” he plausibly contends that we have “an obligation either to find, or if necessary to create, the strongest argument” for unorthodox views.³²

Schauer focuses on allocation rather than regulation, and he makes an overt comparison to a kind of affirmative action that aims directly at promoting diversity of opinion:

This is an extremely interesting suggestion, especially since it seems to apply equally well to hiring decisions in academia. While Schauer draws this conclusion in defense of free speech skepticism, his argument applies quite generally. Moreover, he contends that “the notion of an ideology presupposes some population within which the relevant idea is treated as ideology.”³⁴

Schauer thus implies that intellectual honesty requires the promotion of views that are unorthodox with respect to academia itself, whenever an academic ideology exists. This conclusion should be kept in mind, as I will ultimately argue that there is a far broader and more pervasively entrenched ideology in American universities than Schauer imagines. Not only does the academic establishment fail to live up to the demands of intellectual honesty as he describes them, it deliberately and effectively inhibits debate of its dogmas. If this is true, then Schauer's argument supports the conclusion that scholars whose political views are grossly underrepresented in American universities should be favored in hiring and funding decisions.

These arguments concern the allocation of resources for speech—jobs, grants, publications, etc.—rather than the overt regulation of speech. Questions of allocation pose complex issues that I cannot pursue here, except to note that it is unclear what commitments liberalism has with regard to them. Mill, of course, was expressly committed to protecting even the most unpopular opinions from censorship. “If all mankind minus one, were of one opinion, and only one person were of the contrary opinion,” he famously declared, “mankind would be no more justified in silencing that one person, than he, if he had the power, would be justified in silencing mankind.”³⁵

But he did not mean to imply, as Fiss seems to, that the state must subsidize every unorthodox view in order to ensure that all opinions are presented to the public. Down that road lies the establishment of endowed chairs in creationist cosmology and Holocaust denial studies.³⁶ The real question is how to establish and administer the boundaries of reasonable opinion and, perhaps most importantly, what constitutes corruption of this process?

Fiss offers an answer grounded in his view of the value of freedom of speech. “The choice among the unorthodox should not turn on the administrator's judgment about the goodness or badness of the ideas being advanced or whether the administrator thinks they have merit,” he claims. “This would violate the democratic aspiration of the First Amendment, which is to leave judgments as to the merits of various ideas to the people.”³⁷

Whether or not the First Amendment is motivated by democratic aspirations, Mill's rationale for protecting the expression of unpopular opinion and sentiment arises, rather, from liberal worries about the tyranny of the majority. Nevertheless, Fiss's appeal to democratic aspirations is more plausible with respect to allocation than regulation of speech, and a limited proposal focused on combating institutionally entrenched orthodoxy would be even more compelling. Schauer plausibly concludes that when it comes to questions of allocation, “we cannot avoid the most difficult questions of what functions the institutions are supposed to serve” and “whose choices we as a society wish to privilege and whose choices we wish to suspect.”³⁸ This reference to the judgment of society suggests that, for Schauer too, these issues must be solved democratically—even if society ultimately decides to cede authority to experts. But insofar as society delegates these choices, this surely does not prevent it from scrutinizing the actions of those whom it has empowered, especially to ensure that they are not abusing their authority. When an institution that is ostensibly dedicated to free inquiry is administered by an ideological elite, buffered from external criticism, this constitutes a form of corruption. Choices that ought to belong to society at large have then been taken over in a manner that is antidemocratic and violates a basic canon of intellectual honesty. Or so Schauer's argument suggests, and I agree.³⁹

These arguments for diversity of opinion and against the entrenchment of ideology can even be taken to vindicate, to a limited extent, Fish's most cynical suggestion. Recall his dictum that what one counts as free speech is whatever serves one's substantive agenda. This claim was belied by Mill's commitment to viewpoint neutrality, which protects even the arguments of free speech skeptics, despite the fact that they are intended to subvert freedom of speech. Nevertheless, Fish can be read as predicting that when experts abuse their authority by implementing ideologically driven criteria of value under the guise of merit, they will object to external criticism by declaring that their speech rights and intellectual freedom are under attack. In particular, academic ideologues faced with external criticism can be expected to defend themselves by invoking academic freedom and accusing their critics of censorship, regardless of how dubious and self-serving their claims might be. Were free speech skeptics content with the prosaic observation that popular claims of censorship are often so far-fetched as to be practically meaningless, they would be substantially correct. But they do not care much for this theoretically modest and politically uncongenial conclusion. The free speech skeptics have far more grandiose ambitions, and their commitment to intellectual diversity will prove tenuous.

Consider, for example, Post's commentary on complaints of censorship directed at attempts to remove books for ideological reasons from a library or course curriculum:

Recall that I have suggested that the term ‘censorship’ be reserved for the suppression of speech, in accordance with the liberal conception; and that we use the term ‘silencing’ more broadly, for instance, to capture the abuse of authority that Fiss and Schauer identify. Putting aside the term ‘censorship’, then, what are we to make of Post's suggestion that all decisions about allocation are ideological and necessarily involve suppression? I contend that this is a gross exaggeration, which trades on an undifferentiated notion of ideology. Post's argument implies that if one library decides to acquire only books that it deems patriotic and bans all putatively anti-American works, while another library decides to select only the most powerful arguments for and against American political institutions, then both libraries act for reasons properly termed ‘ideological’. This is the absurd result of accepting so capacious a notion of ideology that it cannot discriminate among any evaluative criteria.⁴¹

This seems a dubious theoretical insight, inadequate for supporting the claim that silencing, much less censorship, is inescapable. Since we have two different terms available, it will be helpful not to use them synonymously. Therefore, in what follows I will let ‘silencing’ be a relatively capacious term for actions that discriminate among opinions on illegitimate grounds or that stifle expressive possibilities unjustifiably, but which may or may not violate speech rights. I will reserve the term ‘censorship’ for the overt suppression of opinion and sentiment. This is merely a terminological stipulation. It does not beg any important questions, because we can still ask whether the distinction marks an important moral difference: Should some or all silencing count as censorship?

III. Silencing for the Sake of Speech

My discussion of the more radical forms of free speech skepticism will focus primarily on Schauer's work, as he seems to have espoused almost every skeptical argument and typically gives them their most careful development. For instance, Schauer provides a clear overview of the terrain initially mapped by Post, on which there are three successively “deeper” levels of analysis of free speech and censorship: the liberal, progressive, and postmodern accounts. Like Post, he is sympathetic to the progressive cause while endorsing postmodern claims about the conceptual impossibility of free speech. Thus, while Post writes of “pervasive, inescapable, and productive silencing of expression,” Schauer calls censorship “inevitable, necessary, and desirable.”⁴²

And he, too, casts doubt on the utility of the liberal conceptions of free speech and censorship. “According to the ontology on which the traditional picture is founded,” Schauer writes, “an act of censorship is an act of external interference with the internally generated communicative, expressive, artistic, or informational preferences of some agent.”⁴³ Yet, he claims, “because this type of censorship is everywhere, it is … nowhere at all.”⁴⁴

But on what grounds do Schauer and Post, among others, attribute so expansive and useless a notion of censorship to liberalism? As usual, Schauer offers the clearest explanation. He notes, first, that the liberal conception focuses on external constraint and coercion: “agents interfering with what would otherwise be our expressive, communicative, artistic, and informational preferences.”⁴⁵

Although it will prove overly broad, this characterization points in the right direction. Mill's defense of free speech denies that society or government can legitimately “determine what doctrines or what arguments [individuals] shall be allowed to hear.”⁴⁶ According to Schauer, the trouble with this account is its commitment to an untenable theory of agency, on which our preferences, and in particular our desires to speak or otherwise communicate, are “endogenous” and “antecedent.” Our desires are supposed to be endogenous in that they are formed wholly by the agent, and antecedent in that they exist prior to the act of censorship. Yet this is misguided, he concludes, because “what we can say and how we can express ourselves is significantly a function of what others do.”⁴⁷

It would certainly be a mistake to insist that desires arise wholly from within an agent, without external influence. But why attribute this view to liberalism? Schauer tacitly assumes that because the liberal conception of censorship focuses on external coercion, it must hold that speech and thought are ordinarily—in the absence of external constraint—entirely “internally generated.” Yet nothing about the liberal conception of censorship requires denying that there are various external influences shaping our desires; the liberal's rejection of coercion is, instead, a normative commitment that draws moral distinctions between different kinds of external influences. By contrast, Post acknowledges that the postmodern conception of censorship “seems to flatten distinctions among kinds of power, implicitly equating suppression of speech caused by state legal action with that caused by the market, or by the dominance of a particular discourse, or by the institution of criticism itself.”⁴⁸

This is the fundamental disagreement. Although many limitations on our capacities, including our expressive capacities, are externally generated, liberalism holds that some such limitations are far more morally significant than others. While some influences destroy an agent's ability to make future decisions or change preferences, for instance, others do not. This can admittedly be a matter of degree, but there are several reasonably clear-cut distinctions to be drawn. At any rate, this normative claim—and not some patently false descriptive claim about the genesis of our desires—is the core commitment of the liberal conception.

Indeed, the postmodernists have no difficulty glossing the liberal concepts of free speech and censorship coherently and moderately, when it suits their purposes. Not every external influence on “communicative preferences” or “discursive practices” counts as censorship on the traditional liberal model. Thus, Post recognizes that the paradigm of censorship is “the ‘direct control’ of expression by the state”; he just doubts whether there is anything “distinctive” to be said about this class of action.⁴⁹

(While this is indeed the focus of First Amendment scholarship, Mill expressly denies the legitimacy of the power to coercively suppress speech, whether exercised by the people or by the state.) But the onus must be on the postmodernists to motivate conceptual reform or revision; they cannot simply conflate all external influences on communicative desires. When postmodernists do offer such an argument, it trades on the three-level analytical scheme. They first appeal to the progressive critique to motivate stretching the application of the concept of censorship, and then use this criticism as leverage to “unravel” the concept (as Schauer puts it), thereby rendering it “so expansive as to be empty.”⁵⁰ Although conceptual reform is common enough as a philosophical strategy, a large argumentative burden must be put on any such argument that would reform a concept to death.

I will begin by considering the conceptual reforms motivated by the progressive level of analysis, which holds that the silencing of the disempowered should be seen as a form of censorship despite the absence of overt prohibition or punishment. Then I will consider the postmodern “insight” that censorship is not merely more prevalent than commonly thought, but that it is necessarily ubiquitous. In making this argument, Schauer introduces the crucial normative claim, common to the progressive and postmodern accounts, which the liberal must resist. He writes:

This claim is stated far too broadly, and at too high a level of abstraction, to be acceptable. Perhaps some forms of silencing that fall short of the overt restriction of opinion should be seen in much the same way as censorship, but not all structures that influence, or even stifle, expressive capabilities are alike. There is a glaring difference between an upbringing that involves, say, the inculcation of religious beliefs, and one that involves sexual abuse, even if both can be said to shape one's desires. Moreover, although some of these desires will be desires to express opinions or sentiments, it obscures matters to consider these as free speech issues simply because they have something to do with expressive capacities and desires. This obfuscation threatens sound moral judgment. Arguments that some type of action should be seen in much the same way as censorship must be assessed on the merits of the particular case, not because we have swallowed a dubious theory and must stomach whatever conclusions it entails.

Some progressive skeptics of free speech marshal entirely different arguments for censorship, which do not rely on an expansive view of silencing. Most commonly, they argue that freedom of speech conflicts not with the speech rights of the disempowered but with other rights or compelling interests. Charles Lawrence claims that “the alternative to regulating racist speech is infringement of the claims of Blacks to liberty and equal protection. The best way to constitutionally protect these competing interests is to balance them directly.”⁵²

I will not treat these broader balancing arguments here; instead, I want to consider the more philosophically ambitious argument made by the progressive analysis of what Schauer calls the “ontology” of censorship. This progressive critique focuses on the claim that because some speech silences other speech, peoples' speech rights directly conflict. Hence, it is argued, such forms of silencing are indeed comparable to censorship: both violate freedom of speech.

This claim is typically introduced by appeal to the “shouting down” scenario, but that case must be extended in two respects in order to support the idea that silencing constitutes a conflict of speech rights.⁵³

First, the verbal behavior that has this silencing effect must fall under the purview of the liberal conception of free speech. Second, the conflict of speech rights cannot be resolved with a content-neutral principle. These two features keep the shouting down case from threatening the liberal conception of freedom of speech. Of course, one need not drown out someone's speech in order to render it less effective. When a scientific theory is refuted by the evidence, for instance, its advocates' claims are undermined; yet we should not conclude that they have been silenced by the force of the better argument. To call this ‘silencing’ is to give up the aspiration to draw a moral comparison between silencing and censorship. In order for a silencing argument to make good on its pretensions, then, it needs to do two things. First, if its central claim is causal, it must present evidence of a specific silencing effect. Second, it needs to make a convincing normative claim that principles of free speech should prevent this form of silencing. This sets a difficult argumentative burden, which I do not think the progressive skeptics have seriously undertaken. As we saw in Section II, Fiss hardly attempted to do so before concluding that sometimes the state must silence for the sake of speech.

There is another form of silencing argument that is even more ambitious, since it seems to circumvent the need to provide empirical evidence. Catharine MacKinnon argues that pornography subordinates and silences women not as a matter of its causal effects, but because its very existence constitutes silencing. Even MacKinnon's allies admit that it is difficult to make literal sense of this claim, though some purport to do so.⁵⁴

Another way MacKinnon seeks to circumvent the burden on her argument is by casting doubt on the distinction between speech and action, which is fundamental to the liberal conception of what is and is not within the purview of free speech protection. If certain actions like “shouting down” can be classed with mere speech, then it is much easier to show that speech rights can conflict. MacKinnon offers the following examples to call into question the liberal's basic distinction. Consider someone who gives the command “kill” to a trained guard dog; he only uses words, but his words are not protected by any principle of free speech. Similarly, a professor who offers his student an A for having sex with him is beyond the pale of such protection.⁵⁵ In both cases, regardless of the consequences of such speech, these acts of speaking constitute significant wrongs. (In the latter case, even if the student declines the proposition and gets an A nevertheless, she has been wronged.)

But such examples need not worry the liberal yet. Although Mill claims that “human beings should be free to form opinions, and to express their opinions without reserve,” he adds this proviso: the freedom of opinion does not extend to acting on one's opinions, as opposed to merely expressing them.⁵⁶

Thus,

Mill insists that even opinions, by which he means the assertion of opinions, “lose their immunity” when the context in which they are spoken makes their utterance a more deeply performative act of speech. This point has several important ramifications.

First, one cannot just look at the behavior—in particular, whether it is verbal—in order to determine what kind of action has been performed. Sometimes mere words constitute speech acts such as threats, contracts, or conspiracies, and these actions are not automatically immunized simply because they are performed by speech. However, a class of less deeply performative speech acts also exists, which we can call ‘acts of expression’, including assertion, discussion, and depiction, and this class constitutes the proper objects of a free speech principle. As Mill observes, in certain contexts what would otherwise be just an assertion can become a positive incitement and thereby lose its immunity. Similarly, an action does not count as an act of expression just because it (in one sense) expressed an opinion. The assassination of President Abraham Lincoln by John Wilkes Booth expressed contempt for the Union, but that did not give immunity to the murder. Only purely symbolic nonverbal actions, such as flag burning, count as acts of expression; even then, symbolic acts are not immunized from fire codes and the like.

Stanley Fish scoffs at arguments founded on this distinction between speech and action. He writes,

This is powerful rhetoric, but Fish gives away the game with his hyperbole. In the first place, contrary to Fish's implication, there is no value judgment inherent in the identification of some speech acts as expressive and others as performative. Warnings and promises are not insignificant, much less unworthy, and many ordinary assertions are rubbish. Even so, their worthlessness does not make them something other than opinions, nor does it place them beyond the pale of free speech protection. Fish's own examples actually undermine his point: he describes a malicious “depiction” (in a parody of a public figure), degraded “portrayals” (of women), and a genocidal “message” (about Jews). Ugly, degrading, and immoral as these cases may be, his inability to describe them except in overtly representational terms—as depictions, portrayals, messages, and even ideas—undermines his cynical conclusion. There is, in fact, a commonsense pre-theoretic distinction at work here.

Nevertheless, if Fish were not so dogmatic, he might have chosen better examples. I do not deny that there are borderline cases, where it is problematic whether some behavior counts as speech or action. Nor am I claiming that this distinction will do all the work necessary to differentiate between what should and should not be assimilated with speech for the purposes of formulating a free speech principle.⁵⁹

Owen Fiss thus gets something right and something wrong when he expresses his “doubts as to the usefulness of the speech/action distinction … because it masks all the hard judgments that the First Amendment requires.”⁶⁰ The trouble with Fiss's complaint is not that it is false, but that many of the hard judgments he has in mind are hard with respect to the distinction; that is, what makes them difficult is that they are on the borderline between the categories of speech and action. The existence of borderline cases and hard judgments cannot justify wholesale conceptual revision, because hard cases are endemic to evaluative concepts. Some distinctions are more problematic than others, of course, but the costs of drawing any given distinction must be compared to those of doing without it. And we must be live to the possibility that some philosophers might have ulterior motives for attacking a distinction that may be not so much problematic as inconvenient.

It is, therefore, worth noting that MacKinnon's examples are not hard cases for the distinction between speech and action, but easy ones. Hence, her suggestion that pornography is a verbal action analogous to an order to kill is hard to take seriously. Perhaps philosopher Rae Langton comes closest to making good literal sense of the claim that pornography silences women constitutively. She does so by going more deeply into speech act theory and focusing specifically on ‘illocutionary acts’: what we do in saying something, as opposed to what we do by saying it. For instance, in saying the right words under the right conditions, we can do such things as marry, vote, command, threaten, promise, and so forth. Langton imagines a scenario in which women would count as being silenced in a specific and literal sense. They would be illocutionarily disabled, which is to be rendered unable to perform some illocutionary act.⁶¹

In Langton's scenario, a particular kind of violent pornography that depicts women saying “no,” but really meaning yes, might make it literally impossible for women to decline sex. If so, then such pornography would constitute the silencing of women by rendering them illocutionarily disabled in this respect: everything that they can say to a sexual proposition constitutes consent to it. This is not really the case, of course, since it is quite possible for women to decline a sexual advance. Note that Langton cannot be claiming merely that refusals are not always honored, and women are sometimes raped. Rather, her claim is that, in the imagined scenario, the woman cannot (and hence does not) perform the illocutionary act of declining sex. I have argued against this claim elsewhere and will not comment on it further here, except to note one especially pertinent problem with the argument.⁶²

The silencing that is illocutionary disablement, whereby someone is unable to perform certain illocutionary acts, is not wrong or even a harm, per se. For instance, children and people already married are illocutionarily disabled from marrying, but almost everyone will agree that this is a good thing. Illocutionary disablement goes on all the time, in a myriad of ways, and it can be good, bad, or indifferent. In other words, illocutionary disablement is pervasive, necessary, and productive—which is exactly the postmodern view of censorship, writ small.

IV. The Dubious Ubiquity of Censorship

Recall that, according to Schauer, the progressive conception of censorship is not the deepest level of analysis. The most sophisticated analysis, postmodernism, does not deny the phenomenon of silencing, but it criticizes the progressive view for implying that silencing can be overcome. Schauer complains that, “Langton presupposes the existence of the unsilenced or the uncensored and thus necessarily presupposes the existence of language whose meaning is uninhibited by the language of others.”⁶³

One might worry that if the uncensored does not exist, then any complaint about the suppression of speech must be misguided. This seems like a bad result, but perhaps it is the price of following the postmodernists' premises to their radical conclusions. Once we understand that “all human behavior both constitutes and restricts our communicative possibilities,” Schauer claims, we find that the very concept of censorship is problematic.⁶⁴ This is supposed to be the great postmodern insight about the ubiquity of censorship and the conceptual impossibility of free speech. Post summarizes the argument as follows:

However, this argument is logically invalid, and not just because it has a missing premise or requires some interpretive massaging. Not only is Post's argument fallacious, it exposes the fundamental error of postmodern skepticism about free speech. We can agree that censorship is one technique by which discursive practices are maintained, and if “discursive practices” just means activities like talking and writing, then surely social life does largely consist of such practices. But there are other ways of maintaining discursive practices besides censorship. Hence, it does not follow that censorship is the norm rather than the exception; indeed, nothing about the prevalence of censorship follows from Post's premises.⁶⁶

This error is not mere carelessness. It issues from the “unraveling” of concepts that are not especially problematic, and the consequent flattening of distinctions needed to avoid philosophical confusion and to differentiate between morally disparate cases.

The trouble with the vast postmodern assimilation is that it makes an enormous difference to all of us just how our discursive practices are maintained. Sometimes we influence people by persuasion and sometimes by coercion. Sometimes the line between these two methods is hard to draw, but that does not make persuasion just another form of coercion. Conflate the two and you can no longer tell the difference between inquisition and the Inquisition. Of course, Post could stipulate that he is using the word ‘censorship’ to refer to any technique for maintaining discursive practices, but this would not help his argument. To say this is merely to change the subject: Post would no longer be talking about censorship, he would merely be using the word. Schauer's argument suffers from the same peculiar malady. He claims that because our “communicative practices” are always affected by others, it makes no sense to identify as censorship only a subset of “something that is part and parcel of all human activity.”⁶⁷

The following discussion will illustrate why, pace Schauer and Post, we must not adopt so capacious and indiscriminate a conception of censorship. I propose letting the verdict rest on our assessment of Schauer's own examples, which are supposed to illustrate the ubiquity of censorship but instead lead us into absurdity. Consider the following example, which I have divided into three distinct claims:

Notice that Schauer expressly defines silencing so as to include any circumstance that deprives someone of an expressive possibility. He thus uses the terms ‘silencing’ and ‘censorship’ synonymously, and he gives them both the broadest possible scope.

The first claim presumably refers to the effort, now largely forgotten, to protest French nuclear testing in Polynesia by boycotting French wine. For the sake of argument, let us grant that Schauer's quasi-semantic claim was true in some social circles, which I will refer to as ‘Cambridge’: to consume French wine in Cambridge, Massachusetts, during the boycott was to “make a statement” about nuclear testing. The second claim states that an expressive possibility was temporarily foreclosed by this boycott. In order to explicate this claim, imagine that we are in Cambridge during the boycott. Of course, it is not merely possible but easy to consume French wine—albeit more difficult to find good value in a claret. Yet Schauer suggests that while it was once possible to express something by consuming French wine, this possibility has been foreclosed. Prior to the boycott, it was possible to make merely a statement of taste, perhaps, but now one inevitably makes a political statement as well. My wine purchase may express more than I want it to: it says that I am the kind of person who thinks it more important to buy a good bottle than to engage in the protest. I can imagine not wanting to make this statement, but, contrary to Schauer's third claim, it strains credulity to think of myself as thereby being silenced or censored. Yet this conclusion would be forced on us, were we to accept the capacious postmodern conception of censorship.

I will limit myself to two observations about this example. First, we need to be careful about such statements-in-action. While they may sometimes be more powerful than verbal expression, they are also inarticulate. It is not at all clear what, if anything, one expressed by buying French wine before the boycott. Was it something about one's taste, anxieties, budget, feelings for France as a nation, or something else entirely? One disanalogy with censorship is common to all (nonverbal) actions—or at least all those that are not wholly symbolic. It is an open question which acts should be said to “make a statement,” and just what statements they make. Schauer may think that all actions qualify, but the more acts one sees as expressive, the less contentful those expressions become. Moreover, even during the boycott one could still express disdain for France, or the Gaullist Chirac government and its unilateralist foreign policy, with words. Another disanalogy is that even if the boycotters intend to dissuade people from buying French wine, they do not intend to stifle anyone's expression; whereas that is precisely the aim of censorship (if not always of silencing). These are significant differences, which are surely relevant to any moral assessment. This brings us to my second observation: how little justification is given for the claim upon which Schauer's whole argument rests. He supposes that anything that can be said to restrict communicative possibilities, or even to affect expressive desires, should be seen as “censorious in much the same way as we see the more overt restrictive actions of the state.”⁶⁹

This is a normative claim: a claim about how we should see these practices. But why should we think so?

Schauer postulates a purely descriptive sense of the word ‘censorship’, quite different from its use in ordinary language, in order to declare various distinctions unfounded. He thus asks, why does condemnation attach “to some acts and not others, if they are equally deserving of the label ‘censorship’ in a strictly descriptive sense?”⁷⁰

But this descriptive sense—on which anything that can be said somehow to restrict communicative possibilities counts as censorship—is all-inclusive. By his own admission, this is “part and parcel of all human activity.”⁷¹ However, the claim that all human activity equally deserves to be called censorship follows only if we accept his dubious conceptual analysis. Yet his example of the French wine illustrates why the capacious conception of censorship is misguided. It seems absurd, even decadent, to compare this case to, say, the Soviet suppression of dissidents. Even without Chateau Neuf in the faculty club, Cambridge is far from Siberia. If this seems like merely a rhetorical flourish, it will in due course prove bizarrely apt.

The problem is not only with one example, but with the whole argument for the inevitability of censorship and the conceptual impossibility of free speech. On its face, the claim that we are censored whenever someone does something that limits our expressive desires or capabilities seems far-fetched—to put it mildly. Worse, the postmodern approach threatens to conflate flippant cases with the gravest infringements of liberty. Consider another example, in which Schauer suggests that great art has some deep affinity with censorship, since “the best artists succeed in transforming artistic understanding, but in doing so they have not only made possible ways of seeing that were not possible before, they have also censored or silenced ways of seeing that used to be possible.”⁷²

How so? Schauer reports that because of Marcel Duchamp's work L.H.O.O.Q., he can no longer look at Leonardo Da Vinci's Mona Lisa without chuckling and wondering where her mustache is.⁷³ Insofar as Schauer's association of censorship with such postmodern darlings as Duchamp (and elsewhere Jenny Holzer) is supposed to lend the view panache, it is worth noting that many people cannot look at Rembrandt's Syndics of the Textile Guild without thinking of Dutch Masters cigars.

Even if Schauer's examples were convincing, they would not secure the most radical postmodern claim: that censorship is inevitable because it is stitched into the fabric of language. Whereas Fish is wont to toss off such claims without explanation or argument, Schauer, to his credit, tries to defend the idea. Here is his case:

This is a rhetorical question. Schauer claims that his act deserves to be called censorship—in the strictly descriptive sense—just as much as any other. But notice two things. First, unless the postmodern analysis is correct, it does not deserve to be called censorship in any sense, and this example suggests a reductio ad absurdum of that claim. Second, no one would seriously condemn Schauer's response, whether as censorship or anything else. The deep problem with the postmodern “insight” is that it leaves its theorists without the conceptual tools to draw fundamental moral and political distinctions. To be sure, his embrace of postmodernism does not force Schauer to condemn the correction of meaningless speech or to condone the execution of political dissidents. But by encouraging its advocates to conflate serious cases with frivolous ones, the postmodern analysis threatens to induce something worth calling ‘decadence’.

Is it uncharitable to suggest that bad philosophy, under the guise of sophistication, might render the postmodern skeptic morally obtuse about censorship? Consider one final and most remarkable reflection that Schauer offers on the subject:

Only blithe pundits, Schauer suggests—presumably in contrast to sophisticated intellectuals—claim that Rushdie was censored by the Ayatollah's death threat. Rushdie might see a difference between the two sorts of “words.” He evidently took the threat and the bounty subsequently offered for his murder—both performative speech acts—rather more seriously, as they forced him into hiding for nearly a decade. What is more, Schauer here reiterates MacKinnon's categorization of all verbal behavior as “simply speech,” deliberately ignoring the obvious fact that Khomeni's words constitute a murderous edict, whereas Rushdie's words compose a controversial novel.

According to Schauer, he is censored by the boycott of French wine and by Duchamp's mustached Mona Lisa; but it is blithe, not fully serious or sophisticated, to say that Rushdie was the target of censorship when he was threatened, terrorized, and forced into seclusion by murderous religious zealots. Rushdie could have edited himself, but instead the Ayatollah did it for him. Rushdie wrote a book; Khomeni wrote a fatwa. If these acts are indeed “quite similar” from the postmodern perspective, then I am content to leave it to the reader to decide whether this perspective should be considered sophisticated or decadent, innovative or retrograde.