[*PG1]COPYRIGHT AS A MODEL FOR FREE SPEECH LAW: WHAT COPYRIGHT HAS IN COMMON WITH ANTI-PORNOGRAPHY LAWS, CAMPAIGN FINANCE REFORM, AND TELECOMMUNICATIONS REGULATION
Abstract: Copyright raises real and troubling free speech issues, and standard responses to those concerns are inadequate. This Article aims to put copyright in the context of other free speech doctrine. Acknowledging the link between copyright and free speech can help determine the proper contours of a copyright regime that both allows and limits property rights in expression, skewing the content of speech toward change.
What is “protected expression”? Suppose you write an article criticizing a public official. If the government cannot prosecute you for the article or award damages in a libel case brought by the official, your speech is protected. On the other hand, if the government can give you an injunction or award damages against someone who copies the article, your speech is protected. So your speech can be protected against the government, or by the government. These two common meanings of protected expression are each found in different areas of the law. Speech protected against the government is First Amendment speech, and speech protected by the government is intellectual property. The First Amendment declares that speech is free, while copyright means that people may be made to pay for speech. So, which is it?
[*PG2] The conventional answer is both.1 The First Amendment gets government off speakers’ backs, while the Copyright Act enables speakers to make money from speaking and thus encourages them to enter the private marketplace of ideas. But this apparently simple relationship hides some profound tensions. When one speaker wishes to use another’s words, or even words that, taken as a whole, are “substantially similar” to someone else’s words, the government may tell her that she cannot. If she has printed books with those words in them, her books may be seized and destroyed by U.S. marshals, or she may be enjoined from trying to sell them. When such situations arise, why does free speech apparently give way?2
This Article aims to put copyright in a context of other free speech doctrine. Part I considers how copyright raises real and troubling free speech issues and why the standard responses to those concerns are inadequate.3 The conventional responses do not defend copyright law because it promotes speech, but rather analyze copyright as if it furthers a generalized legitimate government goal, one like physical safety. From that perspective, the government is required to pursue its legitimate interest without using means that impermissibly trench upon free speech. The main aspects of copyright that prevent it from impermissibly restricting free speech, in this view, are the idea/expression dichotomy and the principle of fair use. This Part argues that neither principle adequately addresses the free speech concerns generally thought relevant in other areas of free speech law. Part I also sets forth potentially less restrictive alternatives to copyright as we know it and rejects the argument that the First Amendment is simply a property regime like copyright. Thus, Part I provides a cri[*PG3]tique of copyright from the perspective of standard First Amendment theory, but its aim is not to demonstrate that copyright is or might be unconstitutional. Rather, my goal is to make possible a rethinking of standard First Amendment theory in light of copyright’s constitutionality.
Part II sets forth the free speech justification for copyright.4 Copyright is “the engine of free expression,”5 providing people with property incentives to speak and disseminate speech. The argument that copyright encourages speech may allow copyright to sweep further than purely speech-suppressing regulations. But copyright is not unique. Part II shows that the free speech issues raised by copyright are related to controversial claims about free speech laws in other contexts, such as hate speech, pornography, and campaign finance, which makes the lack of controversy over copyright law even more of a puzzle. These other arguments, concerning how private parties’ speech may suppress others’ speech, have not been integrated into prevailing free speech doctrine in the same way as the analytically similar argument about copyright. Properly understood, copyright can become the engine of free expression in a second sense: Not only does it enable free speech, but copyright can drive free speech theory in unexpected but important directions.
If we believe standard First Amendment theory, then we should believe that copyright is unconstitutional because it is designed to suppress some speech to generate other speech, a result the Supreme Court condemned in the campaign finance context. But that would be silly; copyright is constitutional, in large part because it does encourage speech by the people it protects. The problem is with the standard theory: Government is already involved in shaping available speech, and that’s a good thing. Our objections to particular government regulations—and there are valid ones—must be to their bias or ineffectiveness, not to the mere fact of government action.
[*PG4] The Supreme Court recognized the government’s fundamental involvement with creating the conditions for speech in the recent Turner Broadcasting cases, which upheld a law requiring cable operators to carry local broadcast signals on some of their channels.6 The Turner cases offer a new way to evaluate government speech restrictions that are designed to promote certain kinds of speech. By requiring substantial evidence in support of a legislative conclusion that regulation will better promote speech than inaction, the Court is attempting to balance issues of institutional competence with fears that speech will be suppressed. The Turner analysis is different from most First Amendment tests because it explicitly concerns itself with the possibility that some speech will disappear if the state regulates, while other speech will disappear in the absence of regulation.
Consistent with Turner, free speech doctrine should acknowledge that the principles supporting copyright are applicable to other areas of the law. Otherwise, copyright will remain a free speech anomaly, an area of the law with a fully articulated speech-based justification that nonetheless contradicts the rest of accepted doctrine. We should not rest content with a copyright founded on special pleading.
Part III briefly applies the theory elaborated in the earlier parts to a few aspects of copyright.7 Essentially, we should recognize that copyright’s limits are as important as the rights it grants to property owners in keeping “the engine of free expression” running properly. Acknowledging the link between copyright and free speech can help us determine the proper contours of a copyright regime that both allows and limits property rights in expression, skewing the content of speech toward change.
Copyright gives the government authority to seize books and enjoin their sale, award damages against booksellers, or even send them to jail. Following preliminary ex parte proceedings requiring only a modest showing, federal marshals may seize works accused of infringement and the machines used to reproduce those works.8 The [*PG5]proceedings may be sealed so that the defendants will not get word of what is being alleged before the marshals burst in.9 If a jury finds that the accused works are infringing by a preponderance of the evidence, they may be destroyed.10 These steps in the process of suppressing copyright infringement are considered so routine and uninteresting that opinions justifying them are rarely even published.11 If the justification were anything other than copyright, these sweeping powers would be seen as a gaping hole at the heart of free speech rights.
In standard First Amendment scholarship, claims that speech belongs to no one and that willing listeners have a right to hear anything they would like to hear are common.12 The Supreme Court has held that potential audiences are generally not required to incur extra costs to get speech that someone wants to supply them.13 This holding [*PG6]should make copying a First Amendment activity, as a copier may offer an alternative source of information that the audience wants and can obtain more easily than by negotiating with the copyright owner.14
Yet courts easily reject First Amendment claims in copyright cases.15 Free speech belongs to no one, but copyrighted speech belongs to someone. Robert Denicola and Melville Nimmer have undertaken extensive defenses of copyright against First Amendment challenges; their work laid the foundations for any subsequent inquiry.16 They both recognize minor First Amendment limits on copyright in highly important news material, but in general they find that copyright itself provides the necessary limits to address any concerns about public access or free speech rights. Two internal limits have been critical to their thinking, and to all who followed: the idea/expression dichotomy and fair use.17
[*PG7] This Part addresses the claim that copyright’s internal configuration is sufficient to avoid a First Amendment challenge. While Part II takes up the First Amendment-based argument for copyright, here I argue that non-speech arguments are insufficient on their own to defend copyright against free speech criticisms. In Section I.A., I show that the idea/expression dichotomy and the fair use defense do not eliminate free speech problems; if anything, they make copyright seem even less supportable, a confusing body of law likely to deter speakers from speech that might potentially be thought to infringe. Section I.B. suggests some less restrictive alternatives to copyright as we know it. The existence of such alternatives makes copyright seem like an excessive, and thus unconstitutional, response to the problem it was designed to solve. Finally, Section I.C. explains that recent property-based visions of the First Amendment cannot solve the problem by folding free speech law into a variant of an intellectual property regime.
The idea/expression dichotomy, now embodied in § 102(b) of the Copyright Act, holds that only expression can be copyrighted, and not the idea, process, or other more general principle that underlies the particular expression.18 Because anyone who wishes can use the ideas found in any copyrighted work, there is, it is said, no free speech [*PG8]problem, as the copyright has not taken from the public domain anything of First Amendment value.19
We tend to think of copiers, unlike other speakers, as pirates or lazy people whose speech does not further free speech values. Our image of a copier is not of an actor who recites a playwright’s lines or a local politico reciting the party platform, though these people copy too. But, we assume that those people have the right to copy or need no permission to do so, and so we don’t examine their merit. We only look at the value of unauthorized copies. While visceral reaction to pirates is natural, it does not sufficiently distinguish a copier—particularly one who is not copying wholesale and for profit—from other disreputable, but protected, speakers. In general, the First Amendment protects even speech which is not original to the speaker;20 and the Supreme Court has stated that it protects individuals’ right “not only to advocate their cause but also to select what they believe to be the most effective means to advocate their cause.”21 “[A]s we know from the example of publishing houses, movie theaters, bookstores and Reader’s Digest, communication occurs in selecting which speech to copy and distribute no less than in creating the speech in the first place.”22
Speakers are allowed to choose their preferred modes of expression because altering expression could well change the meaning and the impact of the message. Famously, the Supreme Court protected [*PG9]Paul Cohen’s right to wear a jacket proclaiming “Fuck the Draft” in public.23 The Court held that the expression can often constitute the idea: “[W]e cannot indulge in the facile assumption that one can forbid particular words without also running a substantial risk of suppressing ideas in the process.”24
These words may seem inapplicable to copyright, where generally the words are not suppressed but limited to a particular class of people who pay to use them.25 But, if the owner will only authorize their use in contexts that are favorable to the author, then the state is enabling the owner to ensure that his expression will only have one meaning and will not be available to use in oppositional ways.26 This is [*PG10]troubling because an inability to use the most evocative expression possible diminishes the power of a speaker’s message.27 The Supreme Court has recognized that, if the government allows private parties exclusive control over cable systems and then allows them to screen out objectionable speech, the First Amendment may be violated.28 The same argument can be made with respect to exclusive state-backed control over expression. Moreover, a payment requirement may put certain speech beyond the reach of a large group of speakers and listeners, which is in itself troubling.
There are two related points here: First, the idea/expression dichotomy recognizes no value in preserving a “breathing space” for free speech. In other areas, the Supreme Court has announced that we must tolerate a certain amount of valueless, destructive speech, because we want to avoid self-censorship by speakers who fear that juries or judges might find them liable.29 If courts do not err on the side of finding unprotectable ideas instead of protectable expression, they run the risk of suppressing important speech.
Second, the relationship of ideas to expression explains why expression deserves strong First Amendment protection. Even if we are confident in theory that a thesaurus and some thought will produce an alternate way to say almost anything with almost as much grace, courts never actually make this inquiry and it would be hard to imagine them doing so. To decide whether it is possible to express a particular idea in a different way, we have to determine what is idea-ish about the idea and what is its expressive raiment. That is, we would have to decide what Leaves of Grass says and how to say it in another way while still communicating its exact idea. There will be nearly as many different answers to this question as there are readers, and that is what makes Leaves of Grass so very protectable. Similarly, “It’s morn[*PG11]ing in America” is important for its expressive power in making a political point. The law protects expression (in free speech and copyright) because, in fact, expression is what makes speech worthwhile. Thus, it is incorrect to say that there is no First Amendment value in “expression,” as opposed to ideas.
Yet copyright is reconciled to free speech with the claim that expression does not mean all that much to our shared artistic, intellectual, and political lives. The idea/expression dichotomy is troubling because it denigrates the value of expression while still attempting to justify the legal protection of expression as property. We protect expression from copying not because expression is unimportant to the free flow of ideas, as the idea/expression dichotomy suggests, but because it is so important that it must be encouraged by state-backed legal protections.30
The argument that piracy is not valuable speech depends on the idea that a particular speaker is low-value, regardless of the actual content of the copied speech. The same speech by different (authorized) speakers would deserve the full range of constitutional protection. Speaker-based discrimination is not unknown to First Amendment law. The Supreme Court has, for example, approved a preference for broadcasters over cable operators in certain circumstances.31 But speaker preferences usually require the government to demonstrate that it is not discriminating on the basis of content or viewpoint and that it has a good reason for its actions.32 At the least, speaker-based discrimination should put a heavier burden of justification on copyright.
Even assuming that the use of someone else’s words provides a speaker only minor convenience, avoiding copying still burdens her speech somewhat. Generally, the state cannot impose liability on a speaker simply to protect another private party’s interests. In Pacific Gas & Electric Co. v. Public Utilities Commission, the Supreme Court struck down a requirement that an electric company allow a consumer group to insert material in its billing envelopes.33 While the electric company had no “right to be free from vigorous debate. . . it [did] have the right to be free from government restrictions that abridge its own rights in order to ‘enhance the relative voice’ of its opponents.”34 Justice Marshall, concurring, wrote that: “[w]hile the interference with appellant’s speech is, concededly, very slight, the State’s justification—the subsidization of another speaker chosen by the State—is insufficient to sustain even that minor burden.”35 One might distinguish Pacific Gas & Electric Co. because the Court faced a [*PG13]situation in which a speaker was forced to subsidize an enemy of that speaker’s viewpoint. A copier, arguably, is not an enemy of a speaker’s viewpoint. Thus, copyright does not restrict speech in order to enhance the relative voice of another, but restricts speech to let a viewpoint-identical but rights-holding speaker prevail.
This is not what really happens in many significant copyright cases, however. J.D. Salinger’s biographer, for example, hardly shared Salinger’s viewpoint, and yet was found to have infringed because he quoted Salinger’s letters. Similarly, extensive quotations from L. Ron Hubbard’s published and unpublished writings justified a finding of infringement in the Second Circuit, though those quotes were used precisely to show what a fraud Mr. Hubbard was. Here we seem to have speakers whose words (including their illustrative quotations from their targets) are being suppressed to enhance the relative voices of their opponents.
Copiers also add expression, as the Nation did when it excerpted parts of Gerald Ford’s biography as part of a story on what the biography revealed about White House politics.36 The underground cartoonists of Air Pirates created twisted caricatures of innocent Disney characters that required time, thought, and creativity,37 as did a commentary on the O.J. Simpson murder trial done in the style of Dr. Seuss.38 All were found to infringe. Particularly when it comes to nonliteral copying, courts may be incapable of deciding what constitutes “opposition” to a copyright owner’s viewpoint. What was Andy Warhol saying with those Campbell’s soup cans, anyway? How many sides does an issue of artistic judgment have?
We could say that the expression taken by a copier is not valuable as speech, even if the rest of what she says is. (Of course, it is valuable as property, which is a bit embarrassing to the theory of value.) Thus, the law states that no pirate can defend against a claim of infringement by showing how much she created herself.39 Yet other areas of free speech law resist such a conclusion. The test for obscenity, for example, requires that a work as a whole must lack literary, artistic, po[*PG14]litical or scientific value in order that it may constitutionally be suppressed. Accused purveyors of obscenity can defend themselves by showing how much they created that was not obscene. In defamation and libel law, inaccurate statements of fact—even those made with knowledge or reckless disregard for their falsity—are constitutionally protected if the overall work is “substantially correct” or lacks malice.40 In other words, defendants accused of defamation can prevail by showing how much of their work was true or in good faith.
The usual justification for looking at an accused work as a whole is that courts fear a chilling effect. If a fragment of a work could be punished for violating some prohibition, publishers would have a much more difficult time determining what was allowable; they would have to scrutinize each paragraph for possible offense if taken in isolation. Publishers would also be unable to rely on the overall message of the work, even though works are normally consumed in their entirety rather than as disconnected passages.41 The reported cases in which using small amounts of another’s copyrighted expression in a larger work led to liability are disturbing, because they allow suppression of an entire work for a small taint.42
We could conceive of the low-value speech argument in this way: Free speech law recognizes a certain set of facts about the world as [*PG15]relevant when deciding whether or not certain speech is regulable; those facts are not contained in speech but determine the level of protection such speech gets. For example, whether a person is a public or a private figure will determine whether negligent misstatements of fact about that person will subject a speaker to liability.43 Whether an exhortation to kill is made in a play or by one mobster to another will determine whether the speaker is guilty of criminal conspiracy. It could be that whether a speaker has paid the requisite fee to a copyright owner is that kind of fact.44 Facts are relevant when they prove or disprove the existence in a particular case of the harms against which a speech regulation is directed.45 The fact that a person has paid a fee to a copyright owner proves that there is no risk that her speech will negatively affect the incentives of future speakers to create copyrightable expression, or the fact that her appropriation was sufficiently transformative proves that punishing her would not serve the goal of encouraging new speech. But this justification is not based on any inherent feature of expression as opposed to ideas; it is a facet of the speech-based justification for copyright, which I take up in Part II.
First Amendment theorists have suggested that copyright infringement does not serve any value that free speech is generally thought to further. Lack of originality supposedly means that copying does not serve a self-fulfillment function, in which the speaker ex[*PG16]presses that which is most herself.46 Likewise, no one is going to riot over a copyright dispute, and so there is no “safety valve” function involved.47
The self-fulfillment and stability arguments are not very persuasive. The self-fulfillment point fails to look closely at the practices that many people actually do find fulfilling: expressing their commitment to certain cultural, political, or social groups in conventional and even stylized ways.48 Indeed, the more that a member of a group adheres to that group’s script, the “better” a member she often is. Just as a personal choice protected by the First Amendment can consist of giving allegiance to an extant faith—choosing to be a Catholic or a Democrat rather than developing one’s own religion or political party—autonomy interests are also served when a person chooses to copy what someone else has said, endorsing it as her own.49 Speech is not guaranteed only to the well-educated, with thesauruses at their fingertips, or the creative.50
[*PG17] The extension of protection to every speaker, however derivative, can be justified by reference to general democratic theory, which values the contribution of each citizen to the political process.51 The more people vote the better, even though they may well be choosing between only two options. The truth is that most equal, autonomous, choosing individuals do not have much revelatory to say; they contribute by participating, not by breaking new ideological ground. A person who recites John Stuart Mill chapter and verse is doing at least as much to further political discourse as someone who composes an original ode to liberalism. The Mill disciple will not contribute much to democratic dialogue if she is not in a position to offer cogent responses to questions from the people to whom she speaks. Even her ability to marshal quotations, however, is a contribution, since persuasiveness is not the test for protected speech.52 Moreover, a speaker’s belief that Mill’s words are appropriate to a particular political situation is itself a valuable interpretation of Mill, just as a politician who quotes the Bible in debate is taking a particular religious and political stance.53
[*PG18] As for the safety valve argument, it is probably true that there has never yet been a riot over the suppression of copyright infringement. But it is equally true that there has yet to be a riot over the suppression of books from school libraries or any number of speech restrictions that nonetheless were deemed impermissible; because of its speculative nature, the safety valve argument is generally a makeweight. And if we widen the criteria for what counts as a safety valve to include speech acts that prevent alienation from government and disrespect for the law,54 copyright appears to be clogging a fair number of safety valves. Outrage at the apparent scope of copyright law and a declared intent to violate that law are reasonably prevalent on the Internet,55 where people are more likely to publicize their dissatisfaction than when they are denied the ability to copy at Kinko’s.56 That such outlaws most likely will never be sued probably does not make them respect the broad scope of the law, and the randomness of enforcement may worsen the problem. This is certainly not a reason to reject copyright, but it does suggest that safety valve concerns are not absent in the area.57
Another basic problem with using the idea/expression dichotomy to resolve free speech concerns is that the distinction between an idea and the concrete form it takes is entirely too vague.58 Indeed, the most famous and well-received explanation of the dichotomy appeals to its vagueness. Judge Learned Hand wrote:
Upon any work . . . a great number of patterns of increasing generality will fit equally well, as more and more of the incident is left out. The last may perhaps be no more than the most general statement of what the work is about, and at times might consist only of its title; but there is a point in this series of abstractions where they are no longer protected, since otherwise the playwright could prevent the use of his “ideas,” to which, apart from their expression, his property is never extended. Nobody has ever been able to fix that boundary, and nobody ever can.59
It is unsurprising, then, that judges often disagree amongst themselves about when it is necessary to use a particular fragment of expression or whether the idea could have been expressed in some other, noncopying way.60 Particularly since infringement can be found [*PG20]even without verbatim copying, in cases of “substantial similarity,” it is difficult to distinguish idea from expression.61 Worsening the uncertainty, the modern idea of a work’s “total concept and feel” allows a finding of infringement when the overall mood of two works is essentially the same, despite the fact that there might be no single element that is literally copied.62 Neil Netanel suggests that the problem of sorting idea from expression has become even less tractable now that derivative works—works based on other copyrighted works such as a film inspired by a novel—are explicitly protected.63
A vague law that restricts speech is usually thought to be unconstitutional. Confused and uncertain, speakers will “‘steer far wider of the unlawful zone’ than if the boundaries of the forbidden areas were clearly marked.”64 A standard that freely admits that one case will never provide much guidance for the next case seems about as bad a guide to safe conduct as one could imagine. Faced with a potentially devastating lawsuit, speakers will be well-advised to steer as far as possible away from any arguable copyright infringement, to spare themselves the risks of going before a judge or jury, and they should carefully limit the expression of those for whom they may be vicariously liable.65 The potential chilling effect is thus particularly great when speakers, to reach an audience, need the help of publishers or Inter[*PG21]net service providers, whose institutional interests make one particular speaker’s material not terribly important compared to a threat of legal action for infringement or contributory infringement.66
Subjective standards for distinguishing between unlawful appropriation and legitimate citation in copyright cases are also suspect on free speech grounds. The influential Ninth Circuit infringement test requires first an objective evaluation of the similarity of two works, then a subjective evaluation.67 Yet the Supreme Court has sharply limited the availability of actions for intentional infliction of emotional distress based on speech, holding that the distinction between outrageous and non-outrageous opinion “has an inherent subjectiveness about it” that would allow defendants to be held liable just because of a jury’s “tastes” or preferences.68 This concern is consistent with vagueness law’s fear of decisions made “on an ad hoc and subjective basis, with the attendant dangers of arbitrary and discriminatory application.”69 A jury’s subjective evalution of similarity may be very difficult to predict.70
One could argue that infringement cases are unlikely to be systematically biased against the opinions of out-groups, whereas judg[*PG22]ments of outrageousness or defamation are likely to be biased. We could decide that arbitrariness, in the sense of random enforcement that falls like lightning from the sky, is not constitutionally troubling in a speech regulation. But then we have substantially revised the concept of subjectivity, locating its harm in heuristic biases that subtly and routinely lead most people to judge in ways that can be predicted based on who is speaking and who is being attacked.71
Defined in this way, the problem of systematic bias is still present in copyright. Sympathetic plaintiffs are far more likely to have their rights expansively defined than unattractive plaintiffs. Thus, lovable Mickey Mouse gets lots of protection from a countercultural portrayal when Disney sues a small comic book publisher over its scandalous parody,72 while Howard Hughes has to lump it when a legitimate publisher publishes an unfavorable biography.73 Copyright losers are often artists making unconventional art that attempts to mock or satirize society, or social critics using the expression of powerful or popular people for their own purposes. Although this group would not qualify for special protection from non-speech related laws, free speech’s concern for protecting the oddball and the unpopular speaker applies here. In any event, the uncertainty and arbitrariness of the idea/expression distinction make it a poor candidate to defend copyright against a First Amendment challenge.
The 1976 Copyright Act codified previous judicial doctrine into a statutory exception for fair use of copyrighted materials as a defense to a finding of infringement.74 The statute suggests four factors for deciding fair use claims: the nature of the copyrighted work; the purpose and character of the use, including whether it is commercial or noncommercial; the amount and substantiality of the use in relation to the copyrighted work as a whole; and the effect of the use on the market for the copyrighted work. Though the law allows courts to consider other factors, in practice they usually rely on the enumerated four. Fair use preserves ground for some use of and comment on [*PG23]copyrighted works, and courts and scholars generally agree that it therefore protects First Amendment interests.75
The “nature of the copyrighted work” factor allows courts to give more protection to fanciful works than to factual ones, preserving public access to facts and opinions about the world while fencing off the content of romance novels and police dramas.76 As a free speech protector, this factor is particularly well-suited to a Meiklejohnian theory of central political speech and peripheral entertaining speech. The “nature” factor also allows courts to further First Amendment interests in remaining silent by protecting unpublished works from copying without good justification.77
The “purpose and character of the use” factor enables courts to give more weight to uses that serve some greater good than uses that are simply made for the copier’s convenience. Educational or news-reporting uses receive more favor than pure entertainment. “Transformative” uses such as parody also get more leeway. In addition, courts also favor noncommercial uses under this factor, on the theory that someone who is not making money from a use is less likely to be a venal thief.78 The “amount and substantiality” factor protects trivial and incidental uses from liability. The “effect on the market” factor, in some versions at least, protects uses that do not really hurt the copyright owner, so that speech is not restricted unless the restriction prevents an identifiable harm. All this, copyright’s defenders argue, supports First Amendment interests in the free flow of speech by limiting the scope of copyright.
One significant problem with fair use is similar to the problem with the idea/expression dichotomy: It is too vague to provide enough guidance.79 Even those who believe that fair use serves First Amendment purposes recognize its “infinite elasticity.”80 There are four named factors, but the statute suggests that the list is not exclusive, and there is little guidance for how to weigh one against another. After decades of litigation, it is still difficult to tell when and whether one can photocopy copyrighted materials, even for scientific research.81 Inconsistencies are common in copyright cases, where fact-specific analyses combined with the multifactor fair use test make cases almost impossible to categorize. Because the outcome of any particular case is uncertain, a potential infringer/fair user has to be willing to bear the substantial costs of litigation for a chance to escape liability. This seems quite likely to prompt self-censorship.82
Control of even one copyright can allow an owner to choke off democratic dialogue, if that copyright is very important for full discussion of a particular issue of public interest. Courts thus sometimes particularize the “purpose of the use” factor of the fair use test: Not just any news reporting or scholarship evokes a public interest test, but this report is important enough to justify stretching the boundaries of fair use.83 The public interest test only increases the uncertainty generated by fair use. Apparently, it is in the public interest to find out more about Howard Hughes,84 but not about Lenny Bruce85 or Rudolph Valentino;86 “who shot JFK” but not “who shot J.R.” The public interest test also requires suspect content judgments about the quality or value of the allegedly infringing work.87
Even without the public interest subfactor, one might wonder whether fair use is unconstitutional because it discriminates on the basis of content. Fair use favors copying, even pure copying, for educational and news reporting purposes. The Supreme Court, evaluating an anticounterfeiting law that prohibited certain reproductions of images of currency but made exceptions for newsworthiness or educa[*PG26]tional value, found that these exceptions were impermissibly content-based.88 There seems to be no reason that the exceptions would lose their content-based nature when applied to copyright.
Fair use also favors criticism and parody. Reviewers get leeway, as do users who humorously savage an original. These preferences are justified on the perfectly reasonable grounds that copyright owners have non-profit-based reasons to prevent uses that are critical of the original work.89 In essence, fair use contains an analogue to the “right of reply” statute struck down in Miami Herald Publishing Co. v. Tornillo.90 Tornillo invalidated a statute that allowed people who had been criticized in a newspaper a chance to respond on the same editorial pages. Anyone who had not first been criticized would have to pay to take out an ad or convince the editors to carry his or her viewpoint. Like citizens covered by a right of reply statute, fair users have a special privilege to copy a work so long as they are criticizing what came before. They may use another’s property—a copyrighted work—without the owner’s consent, just as a person criticized by the Miami Herald could use its printing press and newsprint without the owner’s consent. If they do not disagree with the work, however, their use may trigger an obligation to pay, just like any other consumer/speaker.
Such protection for uses the copyright owner finds particularly objectionable evokes the taint of compelled affirmation, having one’s property used to endorse a message with which one resolutely disagrees.91 It also seems to conflict with the Court’s pronouncement in Hurley v. Irish-American Gay, Lesbian and Bisexual Group that a speaker’s own speech cannot be appropriated by the state as a public accommodation.92 The Hurley Court held that the organizers of a St. Patrick’s Day parade could not be forced by state anti-discrimination law to allow marchers to display signs affirming their nonheterosexual Irishness because that would change the expressive message of the [*PG27]parade contrary to the organizers’ intent. Hurley might be distinguished from fair use because there is only one parade, whereas the existence of a parody does not change the meaning of the original work. But, just as the Court recognized that signs announcing gay and lesbian identities would reflect on the meaning of the rest of the parade,93 the existence of a parody may well lead consumers to reevaluate the meaning of the original.94 Moreover, the point of copyright is that it generally gives owners rights in copies, not just in physical originals, so that a parody could fall within the scope of the author’s exclusive rights were it not for the content-based fair use exception.
The transformative (including critical or parodic) uses escape court-backed prohibition because otherwise private owners would prohibit expression they disliked. Against a background of generally neutral copyright law, the government’s hand appears to come between the speaker and the censor—but only if we accept that censorship can be carried on by private parties with state backing. And this vision of government’s role in the speech market, I will show below, is precisely what justifies copyright as a whole, not simply the transformative use preference in fair use law.
In general, regulations that restrict speech as such are required to meet fairly stringent tests. Even when a compelling government interest supports the regulation, courts seek to assure that no more speech is suppressed than necessary.95 The appropriate inquiry, therefore, is not whether having copyright is important enough to outweigh First Amendment concerns, but whether the particular regime we have is a good way of protecting authors without unnecessarily infringing First Amendment interests.
There is a standard free speech argument that applies here: “more speech” and concerted action as a response to harmful [*PG28]speech.96 Instead of regulating hate speech, for example, the targeted group should toughen up. Free speech “absolutists” argue that targeted groups should overwhelm their opponents in the marketplace of ideas by offering competing ideas, by educating the public that racism is bad.97 Groups can refuse to deal with people whose speech offends them, and lobby others to do the same.
The argument for self-help exists in copyright, though it is not yet recognized as a free speech argument. There are numerous self-help mechanisms available for content providers who want to protect original expression. For example, publishers could use contractual mechanisms to prohibit copying and seek damages against anyone who violated the contract.98 They could also attempt to enforce anticopying norms by structuring the industry to allow authorized publishers lead time or other advantages, and punishing defectors with retributive “strike” editions.99 Content providers could deal only with those who accepted their terms of service, which would include anticopying agreements. Digital watermarking and other copy-protection technologies may allow content providers to defend their intellectual property against quick copying just as a fence around a plot of land hinders easy trespass.100
This kind of self-help is likely to be significantly more effective than standard “more speech” self-help. Contracts are more persuasive as an inducement to respect copying rights than the aspirational language of equality is as a reason to respect other people. Of course, the [*PG29]private contract regime would still depend on the state as back-up, and so might seem to raise similar First Amendment problems, but at least no one would be able to control others’ use of speech without their prior consent to the seller’s terms. Futhermore, general private property and contract law, like that which protects printing presses, computer servers, and other enabling mechanisms of speech, would not be regulation specifically targeted at speech.101
As the cable industry does, content providers could also run public education campaigns against the theft of intellectual property, encouraging people to buy only from authorized providers and educating the public about how to determine if a book is an original or a knockoff.102 Noncontractual, self-help measures based purely on persuasion are available.103 One example is shareware, software that is provided for free by the creator. Users are asked to pay a fee if they decide to continue to use the product after trying it out. Shareware thrives today, even though only an estimated ten percent of users accede to this moral suasion.104 It might decrease profits, but the First Amendment arguably imposes certain costs on speakers, like the costs of developing a persuasive counter-message to unpleasant speech.
Given that there are ways for private actors to protect original content through voluntary transactions, the government arguably does not have a compelling interest in restricting speech through copyright. Yet a regime of self-help might be bad for readers and speakers in a variety of ways, as technical and contractual remedies would not have the same leeway for de minimis uses as the copyright [*PG30]law, nor would they likely distinguish between fair and unfair use or the use of a work’s idea rather than its expression.105 Contract and technological countermeasures, therefore, might well impede the free flow of information contrary to constitutional ideals. Note, however, that the claim that copyright serves content users’ speech interests better than the self-help alternative appeals to First Amendment values as such, not to a non-speech compelling government interest.
One final way to solve the First Amendment problem is to redefine the ground rules: to say that it is property, not speech, at issue.106 Copyright, the argument goes, recognizes the natural right of the creator to control and profit from his creation. The author brings the work into the world, creating it out of nothing, or out of the raw materials of experience, and is thus entitled to dispose of that which he has made, like Lear with his children.
If copyrightable speech is property, then copyright may no longer need a free speech justification. Courts occasionally say that it would be unfair to make a defendant pay for the material it used, because that would hurt free speech interests.107 In standard First Amendment contexts, however, it is unremarkable that a person may need to pay [*PG31]to speak in a particular way—to take out an ad in the paper, to print pamphlets, and so on.108
In this vision, fair use is not a necessary part of copyright; the First Amendment has nothing to say about a requirement that a person has to pay or get an owner’s consent before she can express herself in a particular way.109 The government is simply barred from preventing willing sellers and buyers from making deals.110 The First Amendment and the Copyright Clause are in harmony because one protects information against government suppression and the other protects it against “private depredation.”111 Moreover, absolute property rules, under which any interference with rights can be enjoined and punished, are more appropriate than liability rules, under which a rights violator only has to pay for the value of what he took.112
Such a theory comes at the price of a good deal of what generally seems valuable about free speech. In fact, the state’s refusal to intervene in the distribution of material goods in aid of free speech may only be palatable because speakers can choose fairly freely from the universe of ideas and expression. Jack Balkin points out that, if the government chose to close all public fora, leaving speakers to negotiate in the private market for space in which to speak, many people would sense a First Amendment difficulty.113
[*PG32] The property rights argument depends on certain unsound assumptions about the appropriate subjects and scope of ownership.114 It takes as a foundation the idea that government is supposed to protect my property, and that such protection does not count as “intervention” into the market or the private sphere.115 But this finesses the question of how information is converted into property. Why is something less “my own” if I did not think it up, so long as I said it, or made my own copy?116 Implicit in the argument is a modified “sweat of the brow” theory—information is mine if I worked to create it, and did not copy too much in the process.
The sweat of the brow theory is highly troublesome as a justification for anything like our current copyright regime. Not only has the Supreme Court rather resoundingly rejected it,117 sweat of the brow does not explain why facts and ideas are not copyrightable. In [*PG33]theory, my idea is no less my own than my expression. Both are embodied in a creative work, and my idea may be far more valuable. One answer is that tracing the source of ideas (and facts) can be too difficult, and it is cost-unjustified for the law to allow suits for anything but copying expression.118 Even if this did not sound like a just-so story, we do have a relevant example of an intellectual property regime that does not capitulate to tracing difficulties: patent law. The first inventor of an idea or discoverer of a fact could be protected in similar fashion.
The property vision also cannot explain the peculiar rights that copyright allows authors, such as the right to control derivative works, even if those works would otherwise be independently copyrightable; the right to control public performances; and translation and abridgement rights.119 In all these cases, other people may do as much or more work to bring new expression into the world, but their work does not count. Their children are illegitimate. Moreover, the property rights theory makes the limited duration of copyright particularly hard to explain. Houses (and paper and ink) do not revert to a common pool after an owner has had control of them for a certain period of time. “If I may own Blackacre in perpetuity, why not also Black Beauty?”120 The standard defense of limited duration from this perspective is that, eventually, tracing the copyright proprietor will become impossible. But surely this is only because the duration of copyright is limited: If it were unlimited, the market would generate institutions that could find owners, just as it is possible to find out who owns any particular piece of land. Plenty of permissions organizations already exist, such as ASCAP, BMI, and the Copyright Clearance Center. There is no reason from a property perspective that anything once in their catalogs has to be set free.121
But a property rights enthusiast could agree with all these criticisms, and argue that these limits should be abolished. Greater consistency would cause greater First Amendment concerns, though, and would still not answer tough questions about the scope of owner[*PG34]ship.122 Even if transaction costs are generally low in cyberspace, it will remain just as difficult to distinguish idea from expression or determine substantial similarity in bytes as it is on the printed page.123
In addition, as Eugene Volokh and Brett McDonnell point out, any interest can be reconceptualized as a property interest to defeat a speech claim.124 Your right to swing your fist ends at my face, but the law can define how far my “face” extends. There is no particular reason the law could not give me a property interest in physical and mental integrity that could be violated by exposure to pornography; no reason, that is, but the First Amendment as it is now understood.
This part examined standard justifications for copyright against free speech challenges. The usual suspects—the idea/expression distinction and fair use—attempt to provide a justification that does not depend on copyright’s speech-enhancing role. Unfortunately, neither idea/expression nor fair use bear the necessary weight, primarily because they are too vague to provide a speech user with any real certainty about what she may say.125 There are also less restrictive alternatives to copyright as we know it that would not require nearly as much overt state intervention further weakening the conventional case for copyright. And, the proposal to assimilate speech law in its entirety to property law is ultimately incapable of avoiding the difficult questions [*PG35]of how far to extend ownership of intellectual creation.126 And yet it seems inconceivable that copyright could be unconstitutional, since it serves such an important public interest. Nor has my aim been to suggest that copyright is unconstitutional. Rather, its constitutionality depends on the fact that the government interest underlying copyright is the promotion of speech.
When the conflict between free speech and copyright was first theorized, the natural response was that the expressive and communicative interests involved in copyright protection were constitutionally cognizable. Therefore, every recent discussion of copyright and free [*PG36]speech calls attention to the Supreme Court’s pronouncement in Harper & Row, Publishers, Inc. v. Nation Enterprises: “The Framers intended copyright itself to be an engine of free expression. By establishing a marketable right to use one’s expression, copyright supplies the economic incentive to create and disseminate ideas.”127 While the Court was at the time mainly concerned with copyright’s furtherance of the First Amendment privilege not to speak,128 its language has generally been taken to have wider import, covering copyright’s speech-productive incentives as well.129 In a leading treatise on freedom of speech, Professor Nimmer takes the same position. He concludes that, though expression ordinarily deserves protection from government suppression, the idea/expression dichotomy is justified by the counter-speech value of encouraging authorship.130 That is, a speech-promoting regulation can justifiably suppress more speech than a regulation with a permissible but non-speech-promoting aim.131
When government enforces copyright, it encourages a broad, diverse array of publicly available ideas and expressions, a core interest underlying the First Amendment.132 Free speech values, then, support affirmative government action to encourage speech by harnessing the power of the market. The marketplace is not just a forum where ideas compete for dominance, but a literal (if not always liter[*PG37]ary) marketplace where ideas get traded for money. What the “engine of free expression” argument means, simply enough, is that there are First Amendment interests on both sides of a copyright case. The plaintiff complaining about copying upholds the public interest insofar as a prohibition on copying preserves creators’ incentives to put creative material in the marketplace, just as the defendant upholds the public interest insofar as copying is necessary to enable broad access to information.
The full argument that there are First Amendment interests on both sides presupposes that the extent to which a speaker is heard is a constitutionally relevant fact.133 Being heard is crucial to a speaker. The government cannot require speakers to speak in the middle of the night on an island off the mainland even if it provides speakers free transportation there. Speech values are harmed when government acts in ways that substantially impair speakers’ ability to communicate. Thus, copyright aids free speech because “[e]ffective dissemination of creative work costs money.”134
Free speech theory sometimes seems to imagine a nation of speakers each yammering into a void. Copyright, by contrast, emphasizes the communal nature of creativity and speech. People only share their ideas because there is an audience, and copyright is limited because speakers depend on what was said before. Correspondingly, listeners are entitled to speech rights because they may choose to adopt the messages that others are sending and also because their varying interpretations may enrich our shared dialogue as much or more than the original message. The speech-interests-on-both-sides argument asserts that preserving a market share for speakers is constitutionally relevant because speakers often need an incentive to speak. If the government refuses to enforce copyright, the market for ideas will end up impoverished.
Copyright is not a constitutional anomaly. There are a series of areas in which First Amendment interests may be served by restricting and channeling speech. The reasons for government regulation are different, but they are analogous. To have a healthy, dynamic system of speech, there must be certain architectural limits on the system [*PG38]that shape what occurs therein.135 Many people agree with this claim as applied to property rights, but not to speech; intellectual property, which bridges the gap between the two, shows that speech cannot escape government structuring, because speech is often capable of suppressing other speech.
One major purpose of my comparison of copyright and other market failure or speech-versus-speech theories is to suggest that First Amendment absolutists, who have busied themselves fending off radical attacks in areas such as campaign finance, sexual harassment, pornography, and hate speech, should be attending to the apparently enormous exception to standard First Amendment doctrine embodied in copyright. Copyright, after all, covers every single fixed piece of expression,136 not just isolated areas of the universe of free speech. But it is not unconstitutional, because its absence would be worse for speech. Broad indictments of regulations designed to promote some speech by controlling other kinds threaten to make copyright look unconstitutional; this is a reason to reject those broad theories, or at least to cabin them.
This section elaborates on the incentive argument for copyright, explaining how it is properly described as a market failure theory that enlists government to achieve a better balance of speech, and how copyright’s incentive structure has predictable effects on content. In many cases of speech-versus-speech conflict, the market failure identified can only be addressed by some form of regulation that can be described as content-based. The question is not whether one has to accept all such regulations if one accepts any, because there are reasonable distinctions between the various kinds of market failures. Instead, my aim is to show that a basic problem of speech-versus-speech underlies several important kinds of regulations and proposed regulations, and that the case against them must not rely on the simple claim that the government has to be kept out of the world of speech.
I want to be clear that I remain uncertain about the wisdom of these various regulations. My understanding, however, is that many people support copyright and oppose one or more of the other speech regulations discussed in this section, usually without considering the relationship between those two positions, and I am interested in whether that is a consistent stance. My sense is that one’s conclu[*PG39]sion about whether copyright is fundamentally distinguishable from other regulations depends mainly on one’s beliefs about the extent to which the First Amendment should invalidate laws that disproportionately burden the speech of disadvantaged groups that are not primarily defined by the content of their speech. The First Amendment, conventionally understood, regulates the conditions under which the law may disfavor people based on their beliefs and their speech (Communists, pornographers, copiers). But, at times, the disfavored categories precede the speech, and speech regulations only have a disparate impact on people in those categories rather than creating the categories. It is possible that the First Amendment is primarily concerned with laws that create a category of disfavored speakers, not laws that may enhance prior disadvantages by regulating speech. Yet that First Amendment seems impoverished and unrealistic to me.137
The argument for government intervention is not unique to copyright. It is made by a number of prominent scholars discussing pornography, sexual harassment, hate speech, campaign finance, and new media.138 For example, racist speech is said systematically to un[*PG40]dermine and devalue the speech of minority groups, because racist speech silences minorities and makes their speech seem less credible when it does appear.139 Unlimited campaign spending arguably distorts democratic dialogue by allowing wealthy donors and interest groups to set the public agenda, while political deliberation gets lost in the scramble for cash. Cable providers may exclude broadcasters from their former audiences by refusing to carry them as part of a cable package.
The analogy can be seen by describing copyright’s incentive mechanisms in greater detail. Copying makes original authors less attractive to publishers because there is not much point in paying for what others will then take for free. Audiences will pay less attention to the original speaker if her work can be freely reproduced by others, perhaps even without attribution.140 Ultimately, copying makes authors less willing to enter into the market in the first place. The argument for copyright explains how piratical speech can have negative effects on authors and audiences’ access to speech, just as the radical case for speech regulation explains how hate speech can distort the speech incentives of minority-group members and the receptivity of potential audiences or how political donations can crowd certain views out of the public domain. Therefore, copyright and the radical theories have a family resemblance in that they identify certain mechanisms that operate through speech and that negatively affect the functioning of the marketplace for speech.
The similarity in the arguments for regulation is apparent, for example, in Catharine MacKinnon’s argument for a civil rights remedy for pornography. Not only does MacKinnon argue that pornographic speech silences women’s speech, her analysis converges with that of infringement doctrine. Both in evaluation of substantial similarity and in application of the fair use defense, copyright refuses to look at an accused work as a whole. This is because the critical issue is the harm the defendant may have done to the plaintiff by using a copyrighted work, no matter what else the defendant may have cre[*PG41]ated in the course of so doing. MacKinnon likewise rejects the “work as a whole” standard of obscenity law: “[T]aking the work ‘as a whole’ ignores [the fact that] legitimate settings diminish the injury perceived to be done to those whose trivialization and objectification it contextualizes. . . . If a woman is subjected, why should it matter that the work has other value?”141 This agreement on what is essentially a detail of the respective regulatory schemes shows how both kinds of market-failure theories attend to what a regulable work does in the world and not to what it says in itself.142
The incentive-based or speech-on-both-sides argument also appears where relatively new media are at issue and, lacking a tradition, their structure and function are contestable.143 Government has to do something about new media, and the Supreme Court has recognized that government action will inevitably balance speech against speech in such cases. The Court has accepted the theory that the Federal Communications Commission was established because an unregulated radio spectrum led to chaos. So many people were trying to talk that they cancelled each other out.144 While other forms of regulation, including a property regime based on first-in-time capture of spectrum, would also have worked, possibly better, some form of government-backed rights holding was necessary to enable broadcast speech.145 Although broadcasters were private entities, their actions threatened to “snuff out the free speech of others.” The Court ultimately found that there was no right to do so, and that there was a [*PG42]strong government interest in preventing such silencing.146 In addition, the Court has found that there is a First Amendment interest in encouraging the dissemination of a diversity of views via broadcast media.147
More recently, the Supreme Court upheld must-carry rules requiring cable providers to carry local broadcast stations in order to preserve the profitability of broadcast so that free local television will remain widely available.148 Cable providers are required to subsidize broadcast television for the greater public good, just as fair use arguably requires some authors to subsidize others for the greater good.149 The Court found the costs of must-carry rules to cable providers could be justified in large part because of concerns that cable could [*PG43]take anticompetitive action against broadcast competitors, squeezing broadcasters out of the market.
That anticompetitive possibility, however, relies on the behavior of the television audience. Even if cable did not carry broadcast channels, television households could get the benefit of both if they used manual switches. But the average viewer is unwilling to use a switch. Because the television audience is composed of technologically inept couch potatoes, cable providers could exclude broadcasters from cable households. The Court characterized this interaction between cable providers and the market they face as a matter of providers’ power: “A cable operator, unlike speakers in other media, can thus silence the voice of competing speakers with a mere flick of the switch. The potential for abuse of this private power over a central avenue of communication cannot be overlooked.”150 Crucially, the power to silence depends not on any technological facts but on behavioral facts; not on characteristics of the speaker but of the audience.151 Insofar as they look to the appropriate conditions for the maximization of speech given the way that people actually behave, the must-carry rules have the same justifications as copyright, campaign finance reform, and regulation of hate speech and pornography.
Others have noted that copyright is relevant to more politicized free speech issues. Eugene Volokh and like-minded scholars, who think that the radical theories are a very bad idea, have also become nervous about copyright.152 Meanwhile, some of the radical theorists are arguing that courts’ unhesitating acceptance of copyright, a speech restriction that serves the interests of the wealthy and power[*PG44]ful, is unprincipled and hypocritical when the radical theories of regulation are brushed aside as incompatible with free speech.153 What the people on both sides of the issue—particularly the radical theorists—have not yet discussed, however, is that copyright is not just a run-of-the-mill speech restriction. It is a member of a family of speech restrictions unified by the claim that some government regulations improve the functioning of the market for speech by acting as the equivalent of a police force keeping order.154 Copyright is the perfect demonstration of Stanley Ingber’s point that marketplace theories readily lend themselves to arguments for government intervention. Once we decide that a market is valuable because it furthers individual choice, it becomes possible to argue that individuals should be regulated in aspects of their market behavior to increase the aggregate amount of choice.155
The fact that incentives to speak have constitutional weight and deserve First Amendment consideration is important. If we accept the speech-enhancing justification for copyright, we cannot easily dismiss other market-failure claims. If copyright serves the First Amendment, we cannot say, as the Supreme Court has, that “the concept that government may restrict the speech of some elements of our society in order to enhance the relative voice of others is wholly foreign to the First Amendment.”156
One response to my claim of structural similarity between copyright and other market-failure theories is that copyright is not about [*PG45]restricting A’s speech to enhance the relative voice of B, even if it is about restricting A’s speech to B’s advantage. Because this characterization does not note that the advantage that copyright confers on B is based on the attractiveness of B’s speech—people who want it will have to pay—this reformulation makes copyright sound even worse from a free speech perspective, like an aggressive libel law or restrictions on seditious speech or bans on comparative price advertising for alcohol, all of which benefit one group at another’s expense. More importantly, the “advantage and not speech” characterization does not describe the reasoning that courts and theorists actually use to defend copyright, and it is this reasoning that is structurally similar to other, less favored arguments. Copyright is justified because of its systematic effects on future speakers—the profits that copyright makes possible will spur many people to invest in creating speech.
Once we concede that harm to market-based incentives to speak in the future is harm to First Amendment-protected interests, however, it makes little sense to limit the cognizable class of speech-suppressing private acts to that which merely copies and sucks off profits. It is not the profit-making or even the profit-stealing nature of the infringement that is constitutionally relevant. It is the decrease in the speaker’s incentive to speak, which could also be caused by speech that derided her or by speech so pervasive that her message was lost, that triggers First Amendment interests on her side.157 Put another way, it is descriptively false to say that First Amendment law is not concerned with silencing (defined as private parties’ acts that decrease other private parties’ incentives to speak) or crowding out (defined as private parties’ speech that makes other private parties’ speech less likely to be attended to).
We regulate when we think that incentives to speak deserve protection. Thus, we regulate copiers, but not people who tell other peo[*PG46]ple that they should not scream at their children. Though both decrease others’ incentives to speak, only one group does so in a way we recognize as unjustified. Alternatively, if really good government decreased the public’s incentive to speak out on public affairs, because there was no reason to complain, we would not worry about First Amendment problems. The distinction between acceptable silencing and unacceptable, thus regulable, silencing is necessary and valid. Since we can never escape choices about whose speech to favor, we should focus on why we choose one group over another.
The free speech justification for copyright may also seem distinct from other market failure arguments because it sounds in economics; it is about cold-blooded economic calculation, not artistry and the joy of creation.158 By contrast, the explanations for why pornography, hate speech, or well-funded political campaigns can suppress others’ speech seem psychological and mushy. Economic rationality works through psychological structures, of course, as any human motivation does, but we are not accustomed to thinking of it that way. It seems easier to say that people who feel threatened and oppressed by others’ speech should just grit their teeth and fight back with better ideas than to say that people who are upset by others’ copying should take pride in the joy of creation itself and should be glad that their expression reaches so many people. Nevertheless, the other market-failure theories can equally be described as problems of economic incentives, just as authorship and creativity can readily be described in romantic ways that ignore the influence of economic incentives.
The same is true on the audience’s end. Jack Balkin has recently argued that all speech competes with other speech in an important way—audience time is limited, and someone who is watching The X-Files is not debating foreign policy at the local Republican Party headquarters.159 The real scarcity that is relevant to First Amendment analysis is of people’s time and attention, not of opportunities to speak in any particular medium. Competition exists in non-economic—or at least nonmonetized—registers as well.160
[*PG47] Once we accept that speech trades off with speech, it is simply a matter of calculation to determine how much exposure to pornography decreases a woman’s incentive to participate in public life, or how much spending by the major political parties decreases a new party’s ability to reach potential converts.161 Then, if those disincentives are unjustified, we need to figure out how to counteract them. Maybe sometimes regulation would be more justified than in particular cases of copying for fun and profit.
Because the other market-failure theories are often rejected on the grounds that they impermissibly regulate the content of speech, it is useful to look in greater detail at copyright’s effects on content. In a world without copyright, information would be distributed differently. There would probably be patrons of the arts, both governmental and private, and the content of that art would be shaped by patrons’ preferences. Without copyright, coordination difficulties and free riding problems would make it difficult for the less wealthy to aggregate their resources and fund creativity; thus, as Neil Netanel argues, wealth and power would likely have more influence on the kinds of expression that would be readily distributed.162 Conversely, copyright encourages creators (and investors) toward works that may prove popular with some market segment. The desire to give a mass audience what it will pay for, while not dispositive of content, makes a significant difference in many creative decisions.163 Copyright encour[*PG48]ages the creation and dissemination of the speech of those who seek economic rewards, decreasing the relative voices of those who create for personal satisfaction, for the glory of God, or for the respect and praise of the audience.
Perhaps the model of television and radio programming would exist in some modified form in the absence of copyright, so that interesting stories would be available for free, their content pervaded with ads so that the average consumer would sit through the ads to get the story. In that case, we might expect that creators whose work does not fit well next to an ad for Burger King fries would have a harder time reaching an audience; as evidence for this proposition, consider that the markets for books and even film are much more varied than the television and radio markets.
Authorial self-help as described in Section I.C above would likely be popular, and media that would be easiest to copy-protect would receive the most investment—computer disks that could be read once and would then erase themselves, for example, instead of traditional books. Self-help would have effects on content as well as form. Copy[*PG49]right encourages investment in entertainment over facts, as facts are not copyrightable. If authors had to use other means to protect their work, they would presumably be equally able to protect facts and fictions, and therefore there would not be a copyright-induced skew toward the fanciful.
The bias of copyright, pushing new work away from what has gone before it, also has systematic effects on content and viewpoint.164 Copyright favors expression that looks like a creative genius’; the further an author gets from what has gone before, the more protection he will get.165 It does not recognize value in folklore or other traditional art forms whose richness consists in repetition of traditional themes.166 Copyright favors high and mass culture over counterculture and subculture, since marginal groups are more likely to express themselves by unauthorized reliance on popular and well-known materials, while large corporations that have a “library” of proven characters will reuse winning formulas.167 Copiers who borrow without per[*PG50]mission and then add their own content tend to be making fun, making light, attacking the conventional; they do not have a “stable” of well-recognized character and situations.168 Such copiers do not set the agenda of public discussion. They generally lack name recognition on their own—to get people to pay attention, they may need to trade on names and situations we already know.169 The owners of popular products, by contrast, have an incentive to keep their most popular products from close association with anything unpopular or unsettling.
Despite these predictable effects, copyright can be defended as content-neutral in aim. It could be that punishing copying that destroys the economic incentive to speak by substituting for a creator’s speech and satisfying demand is like prohibiting the interruption of a public speaker. John Hart Ely persuasively argues that prohibiting interruption, even by “the most coherent and trenchant political commentary,” would be perfectly constitutional because the underlying value protected by the regulation would be the right of the original speaker to speak and the audience to listen, and those rights are not dependent on the message of the interrupter or even on the fact that the interrupter has a message. Interruption that agrees or disagrees with the speaker threatens the values sought to be protected.170 Likewise, pure copying, whatever the underlying intent, harms speech and thus can be prohibited. Yet, as Part I explained, many copyright cases do not involve pure copying. Also, questions remain regarding whether certain speakers are more likely to benefit from an anti-interruption/copying regulation and whether the regulation will prevent the interrupter/copier from speaking at all.
[*PG51] Frederick Schauer argues that mechanisms that repress speech are all-pervasive, in norms about polite or appropriate speech, in conventions that limit what will be understood as intelligible communication, and in other varied pressures that lead people to watch their words. In this view, censorship, in the sense of external forces bearing on individual communication, is everywhere; the question is what kinds the government should regulate when it participates, as it must, in shaping those forces.171 Meanwhile, Wendy Gordon argues for broad rights to use pre-existing creative works on the ground that some works essentially reach out and grab audiences. A person who has been powerfully affected by a work may feel a sense of constraint, a need to respond to the thoughts and feelings generated by exposure to the initial work.172
Schauer and Gordon are describing two aspects of the same phenomenon, as prior works are part of the environment that shapes what stories we want to tell and even what we can imagine telling. Copyright generates works that affect what will be created thereafter, not just by prohibiting pure copying and by directly encouraging variation, but also by altering the background universe of information that provides the raw material for the next generation, in the literal sense of the word. Government thus participates in encouraging some kinds of content and discouraging others.
Copyright is a regime that affects a large amount of speech, but seems relatively content-neutral, and in evaluating it we have to ask exactly how stringently the First Amendment requires us to evaluate content-neutral speech regulations. Perhaps surprisingly, the breadth of copyright’s effects becomes a factor in its favor, while the more targeted radical theories seem more suspect. The following subsections explore the family of speech-versus-speech claims by contrasting copyright to the regulation of hate speech and pornography.
What is really at stake in the evaluation of speech-versus-speech claims is a judgment about what options people should have to re[*PG52]pined to speech that in some way harms them. As Kent Greenawalt has noted, the democratic aim of promoting courageous citizens, “independent of mind and hardy emotionally,” does not mean that all kinds of hardiness are equally desirable goals for First Amendment jurisprudence.173 Greenawalt argues that fortitude in the face of serious and imminent threats of violence is not the kind of hardiness that is valuable for democratic citizenship; thus, penalizing such threats does not conflict with the goal of creating robust and vigorous citizen-communicators.174 Similarly, willingness to create in the absence of the economic incentives generated by copyright is not the kind of hardiness that we should require of speakers. The question is what other kinds of hardiness government should or should not demand of its citizens.
One obvious distinction between copyright and the regulation of racist and pornographic speech is that the mechanisms by which “silencing” works in the two cases are different. Richard Delgado recently described the speech-versus-speech justification for regulating hate speech: “[h]ate messages also make the task of the minority speaker harder, because of the toll that they take on the credibility of speakers of color. . . . The very same message from a woman will register differently from one delivered by a man.”175 He concludes that
[b]ecause the message is the same, irrespective of the speaker, the reason for the different reception cannot lie in the words themselves . . . [T]he only possible origin of this different credibility lies in the system of stories and messages that we choose to tell about, and to, minorities and women—in short, hate speech.176
The speech of A, then, deprives B of effective speech, the one value that a First Amendment absolutist cannot deny.177
The economic motivations to buy or use an infringer’s product can be unrelated to the literal message of the infringed material, [*PG53]while the connection between the cognitive biases invoked by racist/sexist speech and the oppressive messages of that speech is strong. In other words, racist speech may destroy a minority speaker’s credibility or otherwise short-circuit cognitive mechanisms for evaluating speech, but the persuasive mechanism by which such suppression works is itself arguably deserving of protection. Infringement, on the other hand, makes a copyright owner’s speech economically unattractive: Why buy a newspaper when you can get the stories for free? The copier is not the enemy of the creator’s viewpoint, only of his or her livelihood.178 The infringer is piggybacking on the first speaker, while other speakers are using their targets in an entirely different way—grinding them down. Where piggybacking is not primarily antagonistic to the first speaker, despite its negative effects on incentives, racist and pornographic speech is. Certainly, regulations of hate speech and pornography can be defended on the ground that such speech inflicts harm on the people it targets, but that rationale stands in contrast to copyright’s encouragement of disagreement.
In fact, the fair use doctrine’s preference for parody and criticism parallels the argument against regulating hate speech. The argument is as follows: There are some harms to incentives to speak that the government cannot take into account when considering whether or not to regulate. In copyright, those harms are (at least) harms to incentives that may occur when vicious reviews or parodies suppress demand for a work. This is either because the interpretation of the parodist/reviewer is more important to free expression than the diminished incentive to create new works that might be savaged, or because the effect on incentives comes not through pure economic substitution but through some other mechanism. I would argue that the former proposition is bound up with the latter.
Valuing criticism over “original” creation raises the specter of content bias. Yet it can be defended as a way to keep people believing in the marketplace of ideas, since a good review is probably more credible in a world that allows bad reviews. This is a concern for proper marketplace functioning, a regulatory concern. It also paral[*PG54]lels the traditional truth-finding rationale for free speech; criticism, even false or erroneous criticism, is valuable because it tests the value of prior works and received wisdom. Furthermore, if criticism is allowed but pure substitution is not, there are still large incentives to create. Therefore, it is reasonable to value criticism or parody over an attacked work when the two come into conflict.
Perhaps racist speech and pornography are so much like criticism and parody that, even if they affect incentives to speak, they still should not be regulated. Nonetheless, the analogy between the two speech-conflict situations is still useful. In both cases we may recognize that an “unregulated” speech market is subject to skewing in favor of some speakers; there is no natural pregovernmental level of unconstrained speech. Furthermore, infringement and racist speech may operate by different noncognitive mechanisms, but the objection in both cases is that the mechanisms are fundamentally unfair. Proponents of hate speech regulation and the like believe that silencing through the coercive power of racial epithets is wrong, just as copyright defenders believe that silencing through the limitation of economic incentives is wrong.
Copyright, unlike the radical theories, seems content- and viewpoint-neutral on its face. The decision to classify a regulation as content-based or content-neutral, however, depends upon the categorization of the mechanism by which dangerous speech does harm. That is, the neutrality argument is ultimately a characterization of the effects of the kinds of speech likely to be suppressed by regulation. To see this point, take a standard example: the regulation of inciting speech that creates a clear and present danger of violence. Unlike regulation of racially derogatory fighting words, regulation of the entire category seems viewpoint-neutral. Anyone who advocates imminent violence in a way that is likely to succeed in triggering such violence will be punished; whether the violence is in support of segregation or socialism, the advocacy is illegal. The proponent of violence could, however, characterize the regulation as viewpoint-based. People who express the belief that violence is a good idea—at least those who are likely to persuade others to agree—can be punished, but those who express the opposite viewpoint, that nonviolence is the appropriate way to effect change, will never be punished. As with a law dividing pregnant persons from nonpregnant persons, it turns out that only one side loses.
[*PG55] One response to this criticism is that the advocacy of imminent violence does what it does in a different way than the advocacy of peaceful change, by destroying the audience’s ability to reflect on what it hears, which brings us back to the different-mechanisms argument.179 The argument that speech that destroys a listener’s ability to abstain from violence is not persuasion but coercion is one plausible characterization, but there are others. In a real sense, if I say “Let’s take the damn street now” and my followers do so, I have been persuasive even if I have also been inarticulate. The persuasive power of any argument often depends on an audience’s preexisting biases and favored concepts or code words, and deciding that some argument operates outside the register of persuasion is tricky business. I will not attempt to resolve the issue; I use the example simply to demonstrate that the determination that a regulation is content- or viewpoint-neutral will ultimately depend on judgments about how different kinds of speech work.
Therefore, an infringer, particularly one who takes only parts of a copyrighted work, could well argue that infringement expresses the viewpoint that copying is good and that there is nothing new under the sun.180 Popular anti-copyright rhetoric contains many such statements.181 No matter what the content of the infringed material, only infringement can express this viewpoint in the most persuasive way, because only infringement shows the audience what infringement is good for.
Copyright protects noninfringing materials, never infringing materials; it embodies the viewpoint that infringement is bad. This is only neutral if infringement operates on the universe of available speech in a way that differs from noninfringement, a way whose harms are more extensive than the harms of other kinds of speech. The harmful incentive effects of infringement make it plausible to treat copyright [*PG56]as a neutral regulation, since copyright targets a mechanism of speech-suppression and not a viewpoint as such. Note, however, that arguments for regulating hate speech, pornography, and campaign finance similarly characterize their targets as behavior with effects on the market for speech. Under the radical theories, it is acceptable to say that women are inferior or that the rich should pay no taxes, so long as that speech does not use the (particularly powerful) mechanisms of sexually explicit subordination or saturation political advertising.182
The distinction between copyright and pornography or hate speech regulation is therefore bottomed on an evidentiary disagreement. Both theories look at audience response to regulation or a lack thereof. We are confident enough about how economic rationality works that we can predict how unredressed infringement will affect the production of speech. Without copyright, some people will create because creation is independently satisfying, and some people will still pay creators, whether as patrons of the arts or out of a sense of moral obligation, but the addition of economic incentives has a significant effect on the level of speech. By contrast, mechanisms of silencing through racist speech and pornography are far less clear. Theories of human psychology can explain the mechanisms, but they seem less intuitively obvious to many people today than the idea that money motivates action.183 Pornography and hate speech promote inequality, and one effect of that is to harm the victims as speakers, but the process seems more diffusely connect to silencing than infringement is, even if the amount of silencing is as great or greater.
In addition, much discomfort with proposed regulations of pornography (and hate speech) comes from the perceived impossibility of tailoring regulation to that which silences women. A variety of non-sexually-explicit demeaning images and stories affect women’s ability to speak and be heard. Against that background, it is hard to imagine that eliminating pornography would materially affect women’s speech incentives or credibility. Ironically, copyright’s breadth seems more acceptable as an incentive scheme because its wide coverage makes it more likely to achieve its goal. Inquiry into mechanisms of silencing suggests a version of a tailoring requirement. A speech-promoting speech restriction should be targeted to cover bad, speech-[*PG57]suppressing speech. If it covers substantially more or less than that, then the law is not really promoting speech. Copyright law does not currently work this way, but it should.
Copyright seems to be content-neutral because it regulates all speech.184 By contrast, racist and pornographic speech often have a predictable and politically charged content. Yet, copyright’s interest in promoting diversity and new expression is, ultimately, content-based in the sense that the term is usually used. We usually demand that mechanisms of speech regulation be content-neutral because we fear government oppression, but we have a legitimate content-based preference for a rich and diverse array of speech.185 Thus, copyright suggests that certain broad content-based preferences are acceptable justifications for government regulation.
When courts and commentators declare that protecting authors’ expression ensures a wide variety of expression rather than a flood of copies, they invoke diversity principles.186 Diversity is a preference for certain kinds of content: new expression that would be less prevalent in the absence of regulation.187 Ex ante, diversity might seem neutral, because Congress cannot be sure when it establishes a copyright law whose ox will be gored, just as legislatures cannot be sure who will use a libel law. The content discrimination comes in ex post, when a preference for variety over repetition punishes copiers and favors transformative uses.188 In the case of libel law, at least, the ex post effects [*PG58]that disadvantage defendants are enough to subject libel law to fairly intense judicial scrutiny.189
The upshot of the diversity preference is that copyright has a disparate impact on various kinds of works and speakers. Yet it is not clear that the First Amendment erects any barrier to neutral laws that, although reasonable and limited, nonetheless have a disparate impact on some group, especially if that group is not historically disadvantaged.
Radical theorists argue for a change in government regulation precisely on the grounds that the speech suppressed under the current regime should be more widely available. We need government to balance the scales so that certain disfavored people have a chance to be heard. Some of the radical theorists think they know what these unheard voices would say; others remain more agnostic. As a matter of democratic self-governance, it may be troubling that any group is silenced no matter what it might say, and particular attention to the silence of historically disadvantaged groups makes constitutional sense (for Fourteenth Amendment, Carolene Products-type reasons).
Perhaps copyright is less troubling than other market-failure theories because copyright promotes diversity of expression, not diversity of ideas, and only the latter is a troubling kind of diversity preference. But recall the argument of Part I.A that expression and ideas are intertwined in ways that are difficult to sort out. Ideas have no form without expression. Furthermore, without copyright, no one would have property rights in ideas in any event; the incentive to have a “new” idea and clothe it in expression could well be diminished without copyright. Therefore, copyright’s protection for ideas works in tandem with its protection of expression.190
The bias of copyright may differ in another way from the biases in other speech-promoting regulations. Copyright’s goal is not just [*PG59]diversity but amount of speech—let a thousand novels bloom. I do not believe that the distinction between diversity and amount of speech is meaningful, however. One could make the same claim about campaign finance reform, hate speech regulation, antipornography ordinances, and so on. Such restrictions would enable many voices heretofore silent to begin speaking, regardless of what those voices might say. More fundamentally, the spatial metaphor for measuring speech becomes fairly useless at this point. Assume that, without copyright, people would spend less time communicating their thoughts to one another, since there would be less profit in it. We would have to make up our own stories to entertain ourselves individually. Would there really be “fewer” stories? Or would we have the “same” number of thoughts, only less advanced because they would not be enriched by others’? While I believe that two people may well improve the logic and persuasiveness of their beliefs by exchanging ideas, I am not sure it is appropriate to say that the “number” of ideas changes through communication. Copyright lends itself to counting more easily than other forms of speech regulation, perhaps, because it is easier to measure the number of magazines on a shelf than it is to measure diversity of viewpoints. But that still does not tell us whether a billion copies of a biography of Leonardo DiCaprio is a better free speech goal than ten thousand copies each of ten thousand different biographies.
“More” speech is not just about having more alternative viewpoints or novels from which to choose, but about having more tools with which to make new speech. Diversity in the marketplace is usually conceived of at static slices of time: more choices for consumers means more diversity. Diversity of speech is a different animal, as the theory behind copyright demonstrates; the variety available at one time affects what will be available later on. The preference for a dynamic diversity, one that allows speakers to generate new speech, is a content preference, but it is a justified one.
Whatever one thinks of the regulation of hate speech and pornography, comparison with copyright theory is useful to identify grounds on which one could promote or condemn particular government actions that penalize some speakers to encourage others. In the end, whether a regulation is “content-based” may not be as important as whether we can define and defend its predictable effects on various groups of speakers.
Campaign finance reform raises similar questions of whether a speech regulation designed to improve one group’s access to speech, which therefore has a predictable disparate impact on a different group, is legitimate.
The strongest justifications for campaign finance reform rest on some theory that lack of regulation has poisoned the system by which information about candidates gets to voters.191 The general idea is that no one really wants the situation we have, but that most participants are forced to play the big-spending game because of a collective action problem, a Prisoner’s Dilemma. Some proponents of reform argue that unlimited spending by candidates leads to a system in which challengers (who may have difficulty getting contributions when challenging a proven candidate) or less wealthy candidates are drowned out.192 Large modern campaigns require huge “war chests,” which in turn drives politicians to solicit wealthy donors or interest groups, creating a system in which money buys influence.193 Justice Breyer has thus endorsed the proposition that campaign finance regulation is justified as a speech-promoting speech restriction, preventing the few from drowning out the many.194
Still another argument for reform is that modern campaigning is an “arms race” in which, for defensive reasons, politicians have to spend so much time on fund-raising that they have no time for governance. Some speech and deliberation—fund-raising speech— crowds out speech and deliberation about what to do in office. Campaign finance reform, then, will not necessarily change the range of views available to the public, but it will improve the quality of public service. Regulation of campaign finance will, it is argued, produce [*PG61]more democratic deliberation.195 Just as copyright can be attacked for promoting the interests of wealthy corporations, campaign finance reform is often criticized as an incumbent-protection measure. Incumbents have greater access to non-monetary assets such as an ability to get free media exposure and name recognition. Therefore, the argument goes, campaign finance reform, by restricting campaign expenditures, will increase the relative importance of these incumbent-favoring assets.
In the campaign finance context, the Supreme Court held that it is a content preference to fear that some speech, because it is backed by deep pockets, will drown out other speech. To the Court, campaign finance regulations evinced a content-based concern with communicative impact.196 Campaign finance reformers, by contrast, consider regulation content-neutral, a concern with the volume of speech rather than its ideas. But, because the “volume” of political ads bears a fairly clear relationship to their ability to persuade, and the volume of a loudspeaker does not, the Court saw content discrimination. Like the different-mechanisms argument discussed in the preceding subsection, this judgment depends on the characterization of money as something that enables speech rather than as a mechanism by which certain speech makes its mark on the world.
In tandem with its characterization of limits on campaign spending as content-based, the Supreme Court rejected the democratic, speech-equalization rationales for campaign finance reform, on the theory that the speech of some should not be suppressed to enhance the relative voice of others.197 Like copyright, there is no ex ante restriction, but once the political season begins, the restrictions affect who can speak or what can be said.198 Also like copyright, campaign finance reform challenges us to recognize the relation between money and speech—money generates speech, in copyright by financing the production and distribution of speech, and similarly in [*PG62]campaign finance. The desire for money affects the content of individual speakers’ speech, as publishers seek to produce popular material or candidates solicit the support of wealthy donors.
Copyright and campaign finance reform are linked not only by their market-failure theories but also by a concern for democracy, in the sense that both theories postulate that citizens should have access to many speakers saying many different things. Some people criticize the radical theories because they seem to make the state responsible for deciding what is good for people, deciding which stories have not been sufficiently successful in reaching a sympathetic audience. Robert Post, for example, finds Owen Fiss’s emphasis on getting information out into the public sphere so that people can decide how to vote unappealing because it “offers a strikingly passive image of the democratic citizen, who can be brought to identify with collective self-determination merely by being provided with . . . full and accurate information.”199
The radical theorists, however, disagree with this characterization because to them there should be no easy line between speakers and audiences. Currently, some people talk too much when they should be listening, and vice versa, but that is not inevitable. Integrating copyright into other theories about how the state constructs the conditions for speech helps illuminate how no citizen, on either the creative/speaking or the copying/listening side, is passive. Copyright’s understanding of how audiences can rework expression to suit their own purposes suggests that even what we think of as passive listening may be more active than sharp distinctions between listening and speaking admit. Access to multiple viewpoints is important not just so citizens can choose, but so they can create their own viewpoints. Campaign finance reform has similarly democratic aspirations, as it attempts to enhance the political voice of groups that may currently lack the means to be heard, both by directly decreasing the importance of money and by eliminating the fund-raising pressures that may lead politicians to devote insufficient time to the issues.
A distinction between campaign finance reform and copyright may therefore rest on predictions about the groups affected by the two regulations. Campaign finance reform affects rich people, a group that seems smaller and more stable—thus more politically vul[*PG63]nerable—than the authors protected by copyright (never mind that the real beneficiaries of copyright are often from the same group). Various aspects of First Amendment law are structured to minimize disparate effects on identifiable groups, as with libel law, where we predict that unconstrained juries will be too sympathetic to the powerful and unsympathetic to their challengers to preserve vigorous reporting and editorializing. Campaign finance reform raises the same concerns. Similarly, we ought to see copyright, and its exceptions, as a law with predictable content-based effects, which should therefore be subject to some heightened review. Because of copyright’s breadth, however, the standard tests for constitutionality of speech regulations may be too stringent; copyright, and perhaps other regulations, may deserve scrutiny limited to the reasonableness of Congress’s line-drawing. But when speech is directly regulated, is the inquiry ever limited to mere rationality as with a standard economic regulation, or must courts demand something more from Congress? The next section addresses that question.
Must-carry regulations are the only speech-promoting regulations upheld in their entirety in recent years. The Court’s articulation of a theory that allowed these regulations to persist, despite their substantial and direct effects on cable providers’ speech, provides valuable guidance for what serious First Amendment analysis of copyright would look like.
Must-carry regulations ensure that cable systems carry local broadcasters at no charge, if the broadcasters so desire. The fear that prompted enactment of the must-carry law was that cable providers would shut local broadcasters out of their systems, thus destroying the local stations that had served regional populations for decades before the development of cable. Local stations, in theory, carry local news, as well as educational and informational programming that might not otherwise be found on cable.200 The crowding-out of local stations was linked to cable technology, which made it inconvenient for a viewer to switch back and forth between cable and local broadcast.
[*PG64] The Turner I Court held that “the mere assertion of dysfunction or failure in a speech market, without more, is not sufficient to shield a speech regulation from the First Amendment standards applicable to nonbroadcast media.”201 The Court emphasized that laws that single out the press are always subject to some heightened First Amendment scrutiny.202 The Court then invoked United States v. O’Brien203 as the basis for its analysis, despite very different situations. O’Brien concerned a regulation banning the destruction of draft cards that was used to prosecute an antiwar protester. It set forth a test for conduct regulations that have an incidental impact on expression, whereas the Cable Act directly regulated expression.
The Turner I Court used O’Brien because it found that none of Congress’s interests in must-carry were related to the suppression of free expression.204 In fact, the multiplication of information sources is “a governmental purpose of the highest order.”205 The Court, however, demanded a showing that the threatened harms to free television, diversity of information sources, and fair competition in the programming market were real and that regulation would alleviate those harms in a direct and material way.206 In addition, the government had the burden of showing that its regulation did not burden substantially more speech than necessary.207 Thus, the Court deferred to Congress as a fact-finder; once it determined that Congress had carried out its fact-finding responsibilities, it accorded great weight to the problems Congress identified and the remedies Congress chose. At the same time, the Court scrutinized the Cable Act carefully in order to determine whether, if the facts were as Congress found them to be, the Act regulated only as much speech as necessary to achieve Congress’s aims.
The Turner I Court justified its somewhat relaxed test for direct regulation of speech on the ground that the must-carry law was con[*PG65]tent-neutral. The test for content neutrality, it stated, was whether the government adopted a regulation because of agreement or disagreement with a message.208 Also, laws that “by their terms distinguish favored speech from disfavored speech on the basis of the ideas or views expressed are content based.”209 This definition seems like a better description of viewpoint regulation, since we usually think of content-based regulations as covering obscenity, libel, or other classes of speech that may have a broad range of “messages.” Indeed, Justice O’Connor declared in dissent that Congress’s preference for the topics covered by broadcast stations—local news, public affairs, educational programs, etc.—constituted a content preference.210
[*PG66] Despite the conceptual difficulties, however, the Court found that the must-carry provisions were content-neutral because they required carriage of broadcast stations regardless of the views those stations expressed.211 The Court also found that the congressional purpose of maintaining access to free television for all Americans was content-neutral. According to the Court, Congress was not saying that broadcast was more valuable than cable, just that it had value.212 This seems disingenuous, since Congress fairly clearly was expressing a preference for local programming over the alternatives that would otherwise appear on the cable channels reserved for must-carry. It might be more accurate to say that must-carry is reasonably content-neutral, and that the categories of speech it prefers are broad enough to be acceptable, especially given the inevitable clash between cable and broadcast speech created by the characteristics of the television-viewing audience.
Turner I and Turner II, which upheld Congress’s balancing of interests after a full examination of the record, leave a very uncertain impression of what kind of congressional findings will suffice to justify a speech regulation.213 O’Brien itself did not distinguish between situations in which Congress was attentive to fact-finding and those in which it was not. It applied a very deferential test in a case where Congress did not have much evidence before it. But the Turner cases appear to modify that test, applying it to direct regulation of expression and holding that deference is appropriate, while requiring that [*PG67]Congress be attentive to fact-finding.214 The Court stated that its “sole obligation” was to assure that Congress drew reasonable inferences based on substantial evidence, because Congress is better equipped than the courts to amass and evaluate the vast amounts of data involved in complex regulation.215
One implication of this reasoning is that, without data, Congress is not in a better position to draw reasonable inferences than the Court.216 Unaided speculation is not enough; Congress at least needs some help speculating. The Court may also have been influenced by the fact that economics and technology played large roles in Turner, whereas the justification for regulation in O’Brien was essentially based on the psychology of draft dodging. In cable regulation, and in copyright, there are clear economic principles that explain the justification for the regulation, although the application of those principles may be hotly contested.
Possibly, as in Turner II, a speech-protective justification for regulation will make the Court’s scrutiny less exacting than it would have been had the law been enacted to protect children from corruption. When speech interests exist on both sides of an issue, the courts must tread carefully. They cannot just analyze the reasonableness of the restriction, and they cannot assume that they know better than Congress even if there are content-based elements to a regulation. If speech is opposed to speech, a decision not to regulate, or to regulate in some other way, will also have content-based results. If the standard for legislation is set too high, speech will actually suffer, as it probably would if the Court struck down the Copyright Act in its entirety; if the standard is too low, interest groups may capture the legislature and overprotect some speech at the expense of other speech. (This is what has occurred for years with copyright term extension, expanded rights of various sorts, and legal protection for anti-copying measures that prevent even fair uses.)
Just as copyright’s free speech justifications have implications for the evaluation of other speech restrictions, free speech has implications for copyright. In the next few pages, I hope to offer a few examples of those implications, though many other things could be said. Reconceiving copyright as speech-promotion law helps us understand not just how to make copyright efficient at what it does, but what it should do.
For example, I began this Article with a discussion of copyright’s vagueness. I conclude, perhaps surprisingly, that vagueness is the necessary price of the benefits of copyright. (And the same might be true of other speech-promoting speech regulations, though the evidence is much less clear.)
Similarly, rethinking copyright as a speech-promotion device has several implications for fair use; I will only discuss one. The fair use preference for “noncommercial” uses should take account of what general First Amendment law recognizes, which is that speech for profit is not necessarily robustly “commercial.” Much profit-seeking speech is nonetheless easily suppressed or deformed, and commerciality as it has been understood in fair use doctrine should be narrowed in a manner more consistent with general free speech law.
Finally, I suggest a framework for evaluating copyright’s effects on speech that takes account of Congress’s ability to find relevant facts, an endeavor that Turner puts at the center of free speech analysis and that will be vital for any other speech-promoting speech regulation.
The problems of vagueness discussed in Part I seem particularly problematic given that there are speech interests on both sides of any copyright dispute. Assuming that people are generally risk averse,217 vagueness chills speech on both sides, although vagueness is almost [*PG69]universally discussed as detrimental to the interests of a copier.218 We get less original production because authors (and their publishers) cannot be certain of capturing enough of the gains of creativity, and we get less copying because legitimate users cannot be sure they will be able to fend off infringement claims. This seems like a lose-lose situation.
Eugene Volokh and Brett McDonnell have offered one way for courts to decrease First Amendment-copyright tensions. Appellate courts could review de novo findings of infringement where the case rests on substantial similarity. They argue that this practice would allow the circuits to build a body of case law that would enhance predictability.219 They take as models to be emulated the appellate-court-supervised development of the law of fair use, libel and defamation, obscenity, and the Fourth Amendment.
This list, however, does not really recommend itself as a set of models of adjudicative deliberation and clarity. In fact, every one of these areas is pretty much a mess, the Fourth Amendment most of all, despite the fact that appellate courts spend appalling amounts of time and paper sorting out Fourth Amendment cases.220 Even if libel and the like formed coherent bodies of law, I am not sure how well the lesson would apply. Libel, defamation, and obscenity lack the infinite variety of copyright. In libel and defamation, the actual malice requirement does most of the speech-protective work, and the doctrine is further limited to the subset of speech that is widely understood to be damaging to the target. Obscenity is confined to the graphically sexual; the potential variations between challenged publications are on the order of “Insert Tab A into Slot B.” In copyright, by contrast, it is difficult to understand how a finding in one case will aid others in any but the vaguest of ways. Substantial similarity rests on comparing the plaintiff’s work to the defendant’s, not to any social consensus or paradigm work. Volokh and McDonnell do not actually offer any vision of what concrete, explicit principles of substantial similarity might look like, and their silence highlights the difficulty involved when trying to analogize from an infringement case over 12 Mon[*PG70]keys221 to one over The Devil’s Advocate: Are four points of similarity enough? How does similar color stack up against similar shape?222 Whatever predictability results from the Volokh and McDonnell proposal would largely stem from the numbers—there are fewer potential three-judge panels than juries.223
The attempt to decrease uncertainty substantially is futile, because vagueness in defining the scope of copyright is the price we pay for speech.224 Arguably, most vagueness law serves to contract the number of situations in which the government can punish speech, as it is harder to define punishable speech than to identify it in practice. In copyright, though, it would be very difficult to live without an idea/expression distinction or a fair use exception. More certain regimes (no copyright at all, for example) would be even worse for speech.225
The Supreme Court’s recent decision in Reno v. American Civil Liberties Union gives some guidance about what other aspects of a First [*PG71]Amendment-influenced copyright would look like.226 The Court struck down portions of the Communications Decency Act which exposed people and entities using the Internet to liability if minors could access indecent speech. In the process, the Court’s opinion repeatedly emphasized the dangers posed by the Act to noncommercial speakers, for whom the profit motive did not operate as a counterbalance to the threat of liability and who would therefore be more easily deterred from speaking than commercial speakers.227 This reasoning suggests a free speech justification for narrowing the commerciality prong of fair use.
The current explanation of the commercial/noncommercial part of the fair use test states that commercial uses are more likely than noncommercial uses to capture the copyright owner’s market. Not only is this highly debatable—certainly repeated and widespread noncommercial use can eliminate a potential market, say for videotapes of popular shows or sound files of popular recordings228—it also faces substantial baseline problems defining what exactly the copyright owner’s “market” should be. This explanation invites claims that, if liability is imposed, a market authorized by the copyright owner will develop; these claims then produce the conclusion that economic harm is caused by the challenged use because the authorized market never materializes.229
The circularity of the market-based argument creates a need for a better justification, and free speech has it in standard explanations for regulating commercial speech. Commercial speech is robust enough, because of the profit motive, to generate a broad range of works.230 A defendant who believes that she has made a commercially successful product may be more willing to litigate a potential infringement, whereas a defendant making a noncommercial use will likely have [*PG72]neither the resources nor the inclination to risk a large judgment against her. But this would also invite courts to look more carefully at what a “commercial” use is. Not every part of a publication is “commercial” in the same way, even when the publisher wants to make money; advertisements are commercial speech in First Amendment law but the news stories right above them in the newspaper are not.231 A publisher is probably more willing to suppress the content of any particular story for fear of liability, whether for copyright infringement or another reason, than to suppress an ad. Therefore, unless near-verbatim copying is at issue—suggesting that the publisher is getting commercial advantage from copying and has not done anything else to attract consumers—courts should not let profit-seeking weigh very heavily in a non-advertising commercial use. And courts should be leery of imposing any liability for nonprofit uses, because they are more fragile and easily suppressed.
This interpretation would bring the meaning of “commercial” in copyright closer to its meaning in free speech law. In copyright, “commercial” use is defined broadly, as any speech disseminated for profit. Although the Supreme Court in Campbell rejected the proposition that a profit-seeking use is presumptively unfair when the use is also transformative, it did not reject the idea that anything that people pay for is commercial use as far as copyright is concerned. By contrast, in the First Amendment context commercial speech is determined by three factors: whether the speech is an advertisement; whether it refers to a specific product or service; and whether the speaker has an economic motive for the speech.232
There is an underlying relationship between commerciality in free speech and in copyright. The first two factors of the free speech test have less to do with the justification for lessened protection for commercial speech—its robustness—than the third. The first two factors instead cabin the principle of commercial robustness against the expansion of speech regulation. If the speech is an advertisement that refers to a specific product or service, it may be easy for the speaker to communicate its core message even in the presence of government regulation. In other cases, the core of the message may not have much intrinsic relationship to the commercial motive, as when a publisher chooses to publish books it believes will be best-sellers, regardless of whether the topic is Chicken Soup for the Teenage Soul or 101 Uses [*PG73]for a Dead Cat. Where the message and the motive have a looser relationship, the content of speech is vulnerable to government-induced deformity even though the speaker intends to keep saying something despite regulation. Thus, such speech is not “commercial” in general free speech law; it is noncommercial speech for profit.
Noncommercial speech for profit is the speech on which copyright’s incentive function operates. Advertisements do not need the inducement of copyright; the profit to be gained from selling the underlying goods would support Madison Avenue in any event. Copyright is designed to encourage precisely those creators (or, more accurately, those investors in creative work) who want to make money and whose profit motive is not as strongly tied to the message of the copyrighted work.
We can therefore identify three kinds of message-motive connections. For ads, the message is “buy X,” and the motive is profit from selling X. For general speech sold in the market, the message varies and the motive is profit from selling the speech, and maybe proselytizing, too. Finally, for nonprofit speech, the message varies and the motive is something other than profit.233 The first class of speech is particularly robust,234 although the Supreme Court has recently cautioned that the government still needs substantial justification to regulate it.
The second class is susceptible to deformation and needs greater protection from regulation.235 Government regulation of such speech may be particularly disturbing for the very reason that a profit-motivated speaker may keep speaking, only with different content, if the government regulates speech. The market will appear robust and free, but it will be pervaded by government-induced distortion. Profit-[*PG74]seeking should therefore not inherently weigh against the defendant in a fair use analysis.
The third class lacks even a generalized profit motive and is likely to be particularly fragile and deserving of heightened scrutiny when regulated. The absence of profit suggests that the motive has something to do with the specific message being communicated, which deserves special consideration in a free speech analysis.236 Noncommercial copyright uses may have market effects, but they still deserve special favor because they represent communication that could easily be suppressed.237
Who will decide where to draw the line, Congress or the courts? Though courts will defer to congressional judgments about many factual situations, the Supreme Court has repeatedly emphasized that the judiciary must ultimately determine whether laws are consistent with the First Amendment. Where First Amendment interests compete, however, the difficulties are compounded. Assuming that copyright contains some speech-enhancing elements, a range of possible regimes could work, depending on an assessment of the empirical [*PG75]validity of claims about encouraging rewards for creativity versus allowing creators to draw on what has come before.238
The Court has rejected suggestions that it should evaluate the extent of a patent monopoly to determine whether it was the best way to promote the useful arts.239 The Court emphasized the explicit constitutional grant of power to Congress: “When as here the Constitution is permissive, the sign of how far Congress has chosen to go can come only from Congress.”240 Yet in Graham v. John Deere Co., the Court held that the Copyright and Trademark Clause is
both a grant of power and a limitation. . . . [Congress may not] enlarge the patent monopoly without regard to the innovation, advancement or social benefit gained thereby. Moreover, Congress may not authorize the issuance of patents whose effects are to remove existent knowledge from the public domain, or to restrict free access to materials already available.241
Most recently, in Feist Publications, Inc. v. Rural Telephone Service Co., the Court held that the constitutional scope of copyright contained a requirement of originality; Congress was not free to allow copyright in facts or in non-original works.242 It is difficult to imagine that a copyright of infinite term would be constitutional, and Congress could probably not enact a copyright law in the old English censorial form, giving exclusive rights (and ensuring profitable production) only to [*PG76]state-approved works.243 If a court were to determine that Congress had failed to identify a speech-based justification for some aspect of copyright and the law suppressed more speech—maybe a lot more—than it promoted, it would be obligated to tell Congress to try again.
With limited empirical evidence at hand, Congress would need at least a persuasive economic theory to explain why its preferred copyright regime did not limit substantially more speech than necessary.244 The justification would not, however, require that each work protected increased the incentive to speak, since the marginal contribution of any one work is minimal. Instead, the effect of a decision to grant rights in the copyright owner or in the user should be generalized, to see what the effects on speech would be if a right or a use became widespread.245 As Justice Souter recently suggested in the campaign finance context, “the quantum of empirical evidence needed to satisfy heightened judicial scrutiny of legislative judgments will vary up or down with the novelty and plausibility of the justification raised.”246
Jessica Litman has examined the legislative history of the 1976 Copyright Act, and concludes that Congress adopted compromises between industry groups. Producers and large consumers of information such as libraries were represented, and ordinary viewers and readers were not. The result was expansive definitions of copyright holders’ rights coupled with narrow exceptions to protect the few in[*PG77]formation users at the bargaining table.247 Congress, and individual members, did not understand or even agree with the particulars of the law adopted. Rather, the legislature brokered a series of deals between industries and then wrote them into law.248 This is not our ideal of policymaking, and it does not fit the Turner vision of serious congressional consideration of the values at stake. The process was probably a good way of allocating copyright ownership as between the various contenders (authors, publishers, etc.) who were all represented, but it was a bad way of defining the scope of copyright against other parties.249 This history, and the similar genesis of industry-sponsored legislation to increase rights in information in years since, provides another reason for courts to scrutinize specific assertions of rights against information-users with greater care.250
The Turner cases suggest that Congress needs credible evidence that its copyright law enhances speech. A speech-sensitive analysis would make expansions of copyright owners’ rights such as the addition of moral rights to copyright, protection for derivative works, and the recent retroactive extension of the copyright term251 look highly suspect.
Copyright poses a serious First Amendment problem. It restricts speech pervasively and powerfully, and its contours are ill-defined. Its saving grace is that it is better for free speech than its absence would be. This article made the First Amendment case against and for copyright, concluding that copyright is justified as a way for government to promote a wide range of speech. Nevertheless, copyright’s wide-ranging effects on speech require careful balancing so that the needs of future creators are not lost in the name of protecting the property rights of those who have already spoken.
The implications of taking market-based and incentive theories seriously can justify the Supreme Court’s new approach to evaluating speech-generating regulations in the Turner cases. Although the Court treated must-carry as a free speech issue, not a property ownership issue, the Court clearly saw a market opposed to a government regulator rather than a soapbox-pounding speaker fighting Big Brother. Cable operators are not very much like orators or authors in the Romantic sense. They are shopkeepers who price and deliver a product. As such, treating must-carry as a problem of potential market failure and monopoly made sense. But, because the problem was also a First Amendment problem—having appeared after the First Amendment became a significant constraint on government action rather than before, like copyright—the Court applied a higher standard to this market regulation than it does when non-speech markets are at issue.
Like must-carry, copyright is about economics and speech. The challenge of reconciling modern constitutional doctrine on economic and social regulation with modern free speech doctrine may be the most serious constitutional difficulty of our time. Copyright forces us to recognize that government has an essential role to play in creating the conditions for speech. Furthermore, that government role is predicated on specific judgments about the value of broad classes of speech, and has systematic effects on content and expression. If government intervention and value judgments are inevitable, free speech inquiry should not focus on the necessity of government intervention, a useless debate, but rather on the kinds of value judgments that are acceptable in distinguishing speech that may be prohibited—in copyright, infringing speech—from speech that will be protected against and by government intervention.
[*PG79] The conceptual separation between copyright and free speech doctrine stems from a general assumption that speech as free speech is not about profit but about politics or self-expression, while people using speech as a profitable commodity have no real investment in its actual content. Although the reality is that the same words often play both roles, a speech claim is created by characterizing words as matters of private choice, while a copyright claim is created by characterizing them as salable property. The challenge of modern copyright law is to explain how words can be both meaningful and profitable, protected speech and protected property.